
The Wnt signalling pathway is an evolutionarily highly conserved mechanism that plays crucial roles in embryonic development and tissue homeostasis across metazoan species (1). The pathway is comprised of a family of secreted glycoproteins, first discovered in Drosophila melanogaster during a mutagenesis screening (2, 3). The Wnt ligand - wingless (wg) - was identified as a key regulator of wing and haltere development, with genetic studies revealing its involvement in diverse patterning processes during development. Subsequent identification of other Wnt pathway components, such as armadillo (β-catenin), dishevelled (Dvl), and porcupine (Porc), further confirmed their essential roles in embryonic development which is showing similar phenotype upon gene ableation (4-6). In case of vertebrate, the abnormal Wnt singlaling lead to irregular digit or interdigit development, and disruption in organogenesis and limb development. Injection of Wnt1 mRNA into early Xenopus embryos confirmed their conserved roles in vertebrates and their importance in axis specification (7).
In addition to their function in developmental processes, the Wnt pathway also plays a pivotal role in adult tissue homeostasis by controlling stem cell maintenance in many organs (8). However, dysregulation of the pathway can lead to disease development such as cancer (9). Aberrant Wnt signalling has been frequently observed in diverse types of cancers, with mutations in several Wnt signalling components identified to be involved in cancer development. For instance, mutations in the adenomatous polyposis coli (APC) gene occur in about 85% of colorectal cancer (CRC) patients, as well as in hepatoblastoma, pancreatic, stomach cancer, etc. (10, 11). Mutations in CNNTB1, the gene encoding for β-catenin, are also common in various cancers, resulting in a more stable and active form (9). Furthermore, alterations in AXIN are found in a subset of colorectal cancer cases (8). These findings emphasize the crucial role of the Wnt signalling pathway in cancer development and highlight its potential as a therapeutic target.
Due to the highly conserved Wnt signalling pathway plays important roles in embryonic development and adult tissue homeostasis, dysregulation of this pathway has been linked to diseases including cancer and making it an attractive target for therapeutics (12). A number of treatments have been developed to target the Wnt pathway, including porcupine inhibitors and anti-Wnt antibodies, and drugs preventing Wnt binding to its receptors. For example, LGK974, a porcupine inhibitor that prevents Wnt secretion, is currently in phase I clinical trial for multiple solid tumors including colorectal cancer (13). Another potential treatment option targeted directly against Wnts is the usage of anti-Wnt antibodies, such as OMP-54F28 and OMP-18R5 (14). However, these treatments have so far been directed towards a limited number of well-characterized components of the pathway.
In recent years, studies have identified novel Wnt regulators, shedding light on new molecular mechanisms and offering new opportunities for targeted therapies. One pro-mising example is the E3 ubiquitin ligases RNF43 and ZNRF3, which function as tumour suppressors in stem cell homeostasis by downregulating Wnt receptors (15). However, mutations in RNF43 and ZNRF3 can transform them into oncogenes, and mutations in these genes have been found in a number of cancers, including colorectal, endometrial, and stomach cancers, as well as intraductal papillary mucinous neoplasm of the pancreas (16).
Given their importance in modulating Wnt signalling and maintaining proper stem cell activity, this review article will revisit fundamental aspects of RNF43/ZNRF3 function and discuss new findings on their regulatory mechanisms. This will provide insights into the potential of RNF43/ZNRF3 as promising candidates for targeted cancer treatments.
Wnt signalling is classified into two pathways depending on the involvement of β-catenin - canonical and non-canonical signalling. The canonical and non-canonical pathways operate using different sets of proteins to transmit downstream signals and are activated by different ligands (17). Wnt1 class Wnt ligands including Wnt2, Wnt3, Wnt3a, and Wnt8a activates canonical signalling which induces proliferation, differentiation, maturation, etc., whereas Wnt5a type Wnt ligands such as Wnt4, Wnt5a, Wnt5b, Wnt6, Wnt7a, and Wnt11 mediates non-canonical signalling which mediates cellular polarization and migration (18). These pathways are controlled by distinct downstream components, which operate in different cellular processes and environments.
In the canonical Wnt pathway, the activity of the transcriptional co-activator β-catenin is mainly regulated. In the absence of Wnt ligands, a “destruction complex” comprising of AXIN1/2, APC, glycogen synthase 3β (GSK3β), and casein kinase 1 (CK1) promotes β-catenin turnover (19-25). The scaffold proteins APC and AXIN bind β-catenin, which is then sequentially phosphorylated by CK1 and GSK3β at serine and threonine residues, respectively. The phosphorylation events result in the recruitment of the E3 ligase SCFβ-Trcp, which ubiquitylates Lys-19 and Lys-49 residues of β-catenin, thereby initiating its proteasomal degradation. On the other hand, Wnt signalling is initiated when Wnt ligands bind to a member of the Frizzled (FZD) family of seven-pass transmembrane receptors and low-density lipoprotein-related protein 5 or 6 (LRP5/6), which form a co-receptor complex in a Wnt-dependent manner (26, 27). This interaction triggers the phosphorylation of the intracellular domain of LRP5/6 and subsequently recruits and phosphorylates DVL. Phosphorylated LRP and DVL interact with AXIN in the β-catenin destruction complex, retaining the complex at the cell mem-brane. Tethering AXIN to the plasma membrane prevents β-catenin degradation, allowing β-catenin to translocate into the nucleus, where it associates with TCF/LEF DNA binding factors and activates Wnt target gene transcription (Fig 1).
Due to the importance of the Wnt signalling pathway from embryonic development to the maintenance of adult tissues, dysregulation of Wnt signalling is often associated with various diseases, including cancer (19). RNF43, an E3 ubiquitin ligase, was discovered as a gene frequently upregulated in colorectal cancer (28). Other studies reported abnormal expression of RNF43 in other types of cancers including gastric and pancreatic cancers. However, under homeostatic conditions, Rnf43 expression is very limited such as Lgr5 positive crypt stem cells (15). Fur-thermore, studies have shown that ZNRF3 mutation is also enriched in many cancer types (19).
In vitro functional analysis of the role of RNF43 by Sugiura et al. (29) has demonstrated that RNF43 has a ubiquitin ligase function since RNF43 ubiquitylates nuclear protein HAP95. Additionally, Koo et al. (15) have reported that Rnf43 and Znrf3 works as a E3 ubiquitin ligase which control Wnt signaling by performing conditional ablation of Rnf43 and its paralogue Znrf33 from the mouse small intestine induces the formation of adenomas due to hyperactive Wnt signalling. These findings indicate that Rnf43 and Znrf3 acts as a negative regulator of Fzd, reducing its cell surface concentration via ubiquitylation mediated protein degradation. In addition, Hao et al. (30) also found ZNRF3 enriched in colon cancer with hyperactive β-catenin signalling. Co-immunoprecipitation experiments have revealed that RNF43/ZNRF3 form a complex with FZD and LRP. Cell-surface biotinylation assays of Myc-FZD with cells treated with siRNA of ZNRF3 and ZNRF3ΔRING mutant resulted in an upregulation of FZD, while cells with ZNRF3 over expression showed a decreased FZD level. The mode of action of ZNRF3 occurs through FZD ubiquitylation, as a FZD KO mutant, entire intracellular Lys residues are mutated to Ala, cannot be ubiquitylated by ZNRF3 overexpression so the FZD KO expression remains at higher level than wild type FZD. These data confirm that RNF43/ZNRF3 function as negative regulators of Wnt receptors by decreasing their cell surface levels.
RNF43 and ZNRF3 are crucial genes that play a significant role in Wnt signalling activation (8). Therefore, its expression is tightly regulated at both transcriptional and post-transcriptional levels (31, 32). The activation of Wnt/β-catenin promotes RNF43/ZNRF3 expression whose, in turn, downregulate Wnt signalling by reducing receptor stability, thereby forming a negative feedback loop. Tight regulation of RNF43/ZNRF3 at the transcriptional level is essential for maintaining steady-state levels of Wnt activity.
It is noteworthy that RNF43 is a Wnt target gene that is not expressed in all cells under active β-catenin signa-lling. While Wnt stimulation in MEFs and MCF7 cell lines induces RNF43 expression, it was not detected in Wnt-stimulated ES cells (33). This observation suggests context-dependent or cell-type-specific expression of RNF43. At the protein level, RNF43 and ZNRF3 are downregula-ted through their interaction with LGR and RSPO via auto-ubiquitylation dependent degradation (15, 16, 30). This finding further demonstrated the importance of regulating RNF43/ZNRF3 expression to maintain appropriate levels of Wnt activity in various cells and tissues.
The LGRs, LGR 4/5/6, are receptors for a class of secreted Wnt agonists, known as RSPO, and a key stem cell marker in several adult tissues exhibiting active Wnt signalling, such as the crypt base columnar (CBC) cells in the small intestine (34-40). In humans, there are four different RSPO capable of binding to the extracellular domains of LGR4/5 and RNF43/ZNRF3 to form a trimeric complex that regulates the Wnt pathway (30, 35, 41). Addition of exogenous human RSPO1 to mouse colonic crypt by injection stabilizes β-catenin protein levels, leading to upregulation of the Wnt signal (42). Furthermore, Rspo2 expression positively correlates with Wnt signalling and functions as a secreted activator in Xenopus (43).
In 2011, two groups, Liu group and Clevers group, demonstrated that LGR4/5/6 directly interact with RSPO1∼4 by in vitro pull-dwon assay, leading to potentiated Wnt/β-catenin signalling and internalization of the trimeric complex (34, 35). A year later, Hao et al. (30) further elucidated this regulatory mechanism, showing that the interaction between RSPO/LGR/ZNRF3 potentiated Wnt signalling via downregulation of ZNRF3. Following studies confirmed that RSPO, LGR, and RNF43/ZNRF3 interaction is necessary for Wnt signal potentiation (44-47).
Although RSPO1 and 4 require LGR for their interaction with RNR43/ZNRF3, Lebensohn and Rohatgi (48) demonstrated that Rspo2 and 3 can downregulate RNF43/ZNRF3 even without LGRs. In mouse embryos, the expression pattern of Lgr4/5/6 only partially overlaps with Znrf3, while Rspo2, which controls limb development with Rspo3 during embryogenesis, has the same expression pattern as Znrf3 (49). In Lgr4/5/6 triple knockout cells, RSPO2 or RSPO3 still upregulate Wnt signalling, indicating that Lgr receptors are dispensable for specific R-spondins. This Lgr-independent antagonistic interaction between Rspo2/3 and Rnf43/Znrf3 is critical for limb development.
The crucial role of RNF43/ZNRF3 in downregulating FZD receptor is well-established, however, the underlying molecular mechanism remains unclear. Recent study have identified DVL as a key regulator of RNF43 activity (47). In particular, DVL1/2/3 triple knockout (TKO) in HEK293 cells resulted in a phenotype similar to RNF43/ZNRF3 depletion which is increased cell surface ZFD and LRP6 level, indicating the essential role of DVL in regulating FZD receptor levels. Further analysis revealed that DVL interacts with RNF43 and mediates RNF43-dependent FZD ubiquitylation. These findings suggest that DVL plays a critical role in regulating the RNF43/ZNRF3-FZD axis, and provide important insights into the molecular mechanism underlying this process.
The G-protein-coupled seven-transmembrane receptors FZD are integral membrane proteins that play a key role in the Wnt signaling pathway (50). The Drosophila frizzled-2 (Dfz2) was discovered as the first frizzled family gene which respond to the addition of Wg therefore inducing Wnt signaling activation (51). The FZD receptor contains a signal peptide sequence, a N-terminal cysteine-rich domain (CRD) that binds its ligand Wnt, seven hydrophobic transmembrane (TM) domains, and linker region between CRD and TM. The regions between TM I and II, III and IV, and V and VI form intracellular loops (iLoops), with two binding motifs within the third loop, iLoop3, and a highly conserved KTXXXW motif in the Fzd C-terminal region mediating the interaction with Dvl (52, 53). Recent study demonstrated that the linker region of FZD contribute to selective interaction with Wnt molecules. Ko et al. (54) identified that linker region swapped FZD4 mutant with non-canonical Wnt signaling receptor FZD6 linker region has defect on binding with WNT3A and failed to interact with LRP6 and DVL.
DVL is a family of phosphoproteins that interact with Fzd iLoop3 binding motifs I and II via its C-terminus, and with the binding motif III on the C-terminal domain of Fzd via its DEP domain (DEP-C), which is important for Wnt signaling (52). Tauriello and colleagues demonstrated that the binding of DEP-C domain of Dvl interact with Fzd to the DEP domain of Dvl has been confirmed, and a Dishevelled-interacting region on RNF43/ZNRF3 has been mapped (47, 52). However, the identification of a particular region of DVL that interacts with RNF43/ZNRF3 has not yet been identified, indicating the possibility of the existence of multiple binding sites or the presence of a mediator that facilitates the interaction. Moreover, this regulatory mechanism can be context-dependent since canonical and noncanonical Wnt pathways use different interaction and motifs. These findings provide important insights into the complex regulation of the Wnt signaling pathway and have important implications for stem cell biology.
The activity of RNF43 in regulating Wnt signaling is known to be regulated in a phospho-dependent manner through CK1 by two independent groups in 2020 (55, 56). Tsukiyama et al. has demonstrated that CK1 directly interacts with two regions in the cytoplasmic portion of RNF43, independently of other components of the β-ca-tenin destruction complex (56). Mutations in the CK1 binding site of RNF43 disrupt its function, demonstrating the requirement of CK1 for RNF43-mediated regulation of Wnt signaling. In HEK293T cells, treatment with Wnt3a leads to decreased phosphorylation of the C-terminal region of RNF43, while R-spondin1 treatment increases it. Recent studies by Tsukiyama et al. (56) have identified S474, S475, and S476 as the residues on RNF43 that are phosphorylated by CK1. Substituting these serine residues with phosphomimetic residues leads to sustained activity of RNF43, while anti-phospho mutations prevent RNF43-mediated downregulation of FZD from the cell surface. The phosphorylation status of these key residues may depend on prior phosphorylation of S478, as a S478D mutation hinders RNF43-mediated downregulation of FZD, while the phosphomimetic S478A mutation does not interfere with the function of RNF43.
Similar to RNF43, ZNRF3 also regulated by phosphorylation by CK1. Ci et al. (57) identified that ZNRF3 is targeted by the Cullin based E3 ubiquitin ligase complex SCFβTrCP for ubiquitylation mediated degradation. The authors demonstrated that CK1 is required for the SCFβTrCP mediated ZNRF3 ubiquitylation since ZNRF3 R474A phosphodegron mutant was more stable than wild-type and CK1 promotes ZNRF3 ubiquitylation followed by degradation. Moreover, two studies from the same lab identified phosphorylation mediated ZNRF3 regulation. Chang et al. (58) identified that Protein tyrosine phosphatase receptor-type kappa (PTPRK) maintains ZNRF3 ‘4Y’ endocytic signal as an unphosphorylated status to induce Wnt receptor depletion and depletion of Ptprk lead to reduction of Spemann organizer effector gene expression and promoting head and axial defects in Xenopus embryo. Additional study by Kim et al. (59) demonstrated that mesenchymal-epithelial transition factor (MET) is the kinase responsible for the phosphorylation of 4Y motif. The authors confirmed that MET bind to ZNRF3 which is induced by MET ligand HGF (hepatocyte growth factor, scatter factor) for the enhancement of Wnt signaling.
The Wnt pathway is highly conserved across metazoans, with key components of the canonical pathway ubiquitous throughout the animal kingdom (60, 61) (Table 1). While no protozoan possesses a complete Wnt signalling pathway, some of its components are present in protozoan organisms such as Gska (GSK3), CK1, Aardvark (β-cate-nin), and Wnt-related receptors found in Dictyostelium discoideum, indicating that the origin of this pathway dates back to a pre-metazoan period (62-64).
Table 1 . Conservation of Wnt pathway components in different model organisms. Except otherwise indicated, data was obtained from flybase for D. melanogaster, HGNC for H. sapiens, Xenbase for X. laevis, and wormbook for C. elegans
H. sapiens | X. laevis | D. melanogaster | C. elegans |
---|---|---|---|
DVL1,2,3 | dvl1,2,3 | dsh | mig-5; dsh-1; dsh-2 |
WNT1, 2, 2B, 3, 3A, 4, 5A, 5B, 6, 7A, 7B, 8A, 8B, 9A, 9B, 10A, 10B, 11, 16 | wnt1, 2, 2b, 3, 3a, 4, 5a, 5b, 6, 7a, 7b, 7c, 8a, 8b, 9a, 9b, 10a, 10b, 11, 11b, 16 | wg; wnt2, 4, 5, 6, 10 | mom-2; lin-44; egl-20; cwn-1; cwn-2 |
β-Catenin | β-Catenin | arm | wrm-1; hmp-1; bar-1; sys-1 |
FZD1, 2, 3, 4, 5, 6, 7, 8, 9, 10 | fzd1, 2, 3, 4, 5, 6, 7, 8, 9, 10 | fz, fz2, fz3, fz4 | mom-5; lin-17; mig-1; cfz-2 |
LEF1; TCF7 (TCF-1); TCF7L1 (TCF3); TCF7L2 (TCF4) | lef1, tcf7, tcf7l1, tcf7l2 | pan | pop-1 |
RNF43; ZNRF3 | rnf43; znrf3 | - | plr-1 |
R-spondin | r-spondin | - | - |
Lgr | lgr | - | - |
LRP5, 6 | lrp5, 6 | arr | - |
APC, APC2 | apc, apc2 | apc, apc2 | apr-1 |
AXIN1, 2 | axin1, 2 | axn | pry-1, axl-1 |
CK1α | ck1α | ck1α | kin-19 |
GSK3A, GSK3B | gsk3a, gsk3b | sgg | gsk-3 |
PP2A | pp2a | pp2a | - |
β-Trcp | btrcp | slmb | lin-23 |
In model organisms, the core elements of the Wnt pathway are well conserved, emphasizing the importance of the pathway. Regulators of the Wnt pathway, including RNF43 and ZNRF3, are highly conserved among animal species, differing primarily in their mode of protein-protein interaction via their ectodomain. Zebisch et al. (46) reported structural study of RNF43 and ZNRF3 in complex with other Wnt signaling component including Furin-like (FU) domain of Rspo2. In that study, the authors identified that non-ligand bound ZNRF3 ectodomain (ZNRF3ecto) could form dimer and ZNRF3ecto dimer become less flexible when it bound with FU domain of Rspo2 (Rspo2FU) so that 2:2 form of ZNRF3ecto-Rspo2FU could comprised. Intere-stingly, dimerization of RNF43 with Rspo was not observed. However, the binding interface between RspoFU and RNF43 or ZNRF3 was essentially the same. Although ZNRF3 ectodomain share high structural similarity with RNF43 ectodomain, Peng et al. (45) demonstrated that RNF43 and ZNRF3 also have structural difference. While ZNRF3 N-terminal strands β1 and β2 form an extended β-hairpin flap, RNF43 has a flexible loop region in its N-terminal region. However, both regulators are structurally similar and belong to the PA-TM-RING family of ubiquitin ligases. This family is characterized by three domains: a N-terminal protease-associated (PA) domain, really interesting new gene (RING) domain, and C-terminal TM domain (65).
The human PA-TM-RING family comprises 12 proteins, all of which possess these domains (65) (Fig 2). The PA domain itself is a conserved sequence motif among proteases and predicted to function as a protein-protein interaction domain. The RING domain is a catalytic E3 ligase domains, and all PA-TM-RING proteins have the ability to ubiquitylate their substrates. The RING, PA, and TM domains are highly conserved among homologous proteins. The PA-TM-RING family gene GRAIL (gene related to anergy in lymphocytes) was initially identified by Anandasabapathy et al. (66) as a key regulator of T cell anergy. Following studies have demonstrated the role of PA-TM-RING family genes as a E3 ubiquitin ligase including plant homologue RMR (receptor homology-transmembranedomain-RING-H2 motif protein) in addition to RNF43 related PA-TM-RING family genes (67, 68).
In mice and humans, Wnt pathway components share a high degree of homology, including RNF43 and ZNRF3, and human PA-TM-RING family proteins or their orthologues are also conserved in mice. Xenopus Rnf128 and Drosophila melanogaster’s Godzilla and Goliath belong to the same protein family and are known to be required for the trafficking of recycling endosomes. Godzilla is also implicated in the Wnt pathway, particularly in the regulation of Wingless transcytosis. Similarly, Caenorhabditis elegans has an orthologue of RNF43/ZNRF3, named PLR-1, which downregulates Wnt signalling by ubiquitylating C. elegans Fzd Wnt receptors Cfz-2, Lin-17, Mig-1, and Mom-5, thereby reducing the receptors’ cell-surface levels. As neither R-spondin nor LGR4/5 are conserved in C. elegans, PLR-1 may have alternative regulatory mechanisms, potentially at the transcriptional level. The highly conserved nature of Wnt pathway components across species highlights its importance in animal development and evolution.
The Wnt signalling pathway plays a crucial role in development and the maintenance of homeostasis in adult tissues and RNF43 and its paralogue ZNRF3 plays critical role on the regulation of Wnt pathway. However, alterations in its components have been associated with various malignancies. Recent research has led to a better understanding of the Wnt signalling pathway, but there is still a need for a more profound insight into the regulators of this pathway for the development of targeted therapies for Wnt dysregulation-related malignancies.
RNF43 and ZNRF3 are members of the PA-TM-RING protein family and known negative regulators of Wnt signalling. They play a significant role in the regulation of cell-surface levels of Fzd, which is a shared component of both the canonical and noncanonical pathways. Despite the recent studies, their functions in noncanonical signalling are not yet fully understood and further investigation into the impact of mutations in RNF43/ZNRF3 on malignancies will also be crucial for developing specialized treatments and predicting cancer outcomes.
Overall, a more profound understanding of the Wnt signalling pathway and its regulators will provide an opportunity for the development of more targeted therapies for malignancies caused by Wnt dysregulation. Further research on the role of RNF43/ZNRF3 in noncanonical signalling and their regulatory mechanisms is necessary to develop specialized treatments and to predict cancer outcomes.
The authors have no conflicting financial interest.
![]() |
![]() |
CrossRef (7) |