Targeting Wnts at the Source—New Mechanisms, New Biomarkers, New Drugs Babita Madan and David M

Published OnlineFirst April 21, 2015; DOI: 10.1158/1535-7163.MCT-14-1038

Review Molecular Cancer Therapeutics Targeting Wnts at the Source—New Mechanisms, New Biomarkers, New Drugs Babita Madan and David M. Virshup

Abstract

Wnt signaling is dysregulated in many cancers and is therefore ciﬁc upregulation of Wnt receptors at the cell surface regulates an attractive therapeutic target. The focus of drug development has cellular sensitivity to Wnts. Loss-of-function mutations in RNF43 recently shifted away from downstream inhibitors of b-catenin. or ZNRF3 and gain-of-function chromosome translocations Active inhibitors of Wnt secretion and Wnt/receptor interactions involving RSPO2 and RSPO3 are surprisingly common and have been developed that are now entering clinical trials. Such markedly increase Wnt/b-catenin signaling in response to secreted agents include inhibitors of Wnt secretion, as well as recombinant Wnts. These mutations may be predictive biomarkers to select proteins that minimize Wnt–Frizzled interactions. These new patients responsive to newly developed upstream Wnt inhibitors. therapies arrive together with the recent insight that cancer-spe- Mol Cancer Ther; 14(5); 1–8. 2015 AACR.

Introduction standing the mechanism underlying increased Wnt signaling can guide the therapeutic approach. Wnts are secreted proteins that are critical for normal embry- onic development and homeostasis of select adult tissues. Wnt b signaling regulates diverse developmental and homeostatic func- Downstream Wnt/ -Catenin Pathway tions, including proliferation, differentiation, cell polarity, motil- Inhibitors ity, and migration. In humans, there are 19 WNT ligands that Many downstream components of the Wnt/b-catenin pathway couple to multiple receptors and coreceptors to activate various can be mutated in cancers to stabilize b-catenin protein. Well- downstream pathways. The most extensively studied Wnt signal- recognized activating mutations include truncating mutations of ing pathway follows from the binding of Wnts to a Frizzled (FZD) APC, found in the majority of colorectal cancers, stabilizing and a Lipoprotein receptor-related protein coreceptor, LRP5 or mutations of CTNNB1 (b-catenin) in diverse tumors, and loss- LRP6. A well-studied series of intracellular events cause inhibition of-function mutations of AXIN1/2 in biliary tract and liver cancers of GSK3 and the stabilization of b-catenin (1, 2). This leads to (6). The correlation of b-catenin stabilization with cancer has led cytoplasmic accumulation and then translocation of b-catenin to to the conclusion that stabilized b-catenin can cause cancer, the nucleus where it interacts with members of the TCF/LEF family although in most cases, this has not been rigorously demonstrated of transcription factors to regulate the cell-type specific expression (8). Multiple attempts to target elevated downstream Wnt sig- of target genes (1). Wnts also have important functions indepen- naling have used small molecules that inhibit b-catenin abun- dent of b-catenin, for example, by triggering downstream path- dance or its transcriptional activating activity (Fig. 1; refs. 9–12). ways that regulate tissue polarity and cell movement through the Screens for downstream inhibitors have provided an important activation of RhoA, JNK, and nemo-like kinase signaling cascades insight that b-catenin stability can be regulated by tankyrase- (3–5). mediated degradation of AXIN (11, 13). Tankyrase inhibitors Dysregulated Wnt signaling is associated with various patho- increase AXIN abundance and therefore increase b-catenin deg- logic states, including osteoporosis, cancer, vascular diseases, and radation. Such drugs have been independently developed by fibrosis (6). Elevated Wnt signaling promotes tumorigenesis by several groups, although their oral use may be limited by mech- supporting cellular proliferation, blocking differentiation, driving anism-based gut toxicity (14). A more selective approach is use of epithelial mesenchymal transitions, and promoting both stem small molecules to modulate b-catenin's interaction with tran- cell renewal and metastasis (5, 7). Pathologic Wnt signaling can be scriptional coactivators such as CREB-binding protein (CREBBP/ driven by upstream or downstream events. Targeting the Wnt CBP). One such small molecule, PRI-724, is in phase I trials for signaling pathway can be beneficial for diseases with elevated or acute myeloid leukemia (clinicaltrials.gov NCT01606579) and diminished Wnt activity. Recent findings demonstrate that under- advanced solid tumors (clinicaltrials.gov NCT01302405; ref. 15). Inhibitors directed at b-catenin accumulation will, however, not block signaling pathways triggered by b-catenin–independent Program in Cancer and Stem Cell Biology, Duke-NUS Graduate Med- actions of Wnts. ical School, Singapore. Corresponding Author: David M. Virshup, Duke-NUS Graduate Medical School, 07-20, 8 College Road, Singapore 169857. Phone: 65-6516-7881; Fax: 65-6221- Upstream Inhibitors of Wnt Production and 2402; E-mail: [email protected] Function doi: 10.1158/1535-7163.MCT-14-1038 To inhibit both b-catenin–dependent and –independent sig- 2015 American Association for Cancer Research. naling, studies have focused on upstream elements of Wnt

www.aacrjournals.org OF1

Madan and Virshup

Figure 1. Inhibitors of canonical Wnt/ b-catenin signaling. The binding of Wnts to their receptors, Frizzleds, and coreceptors, LRP5/6, initiates a cascade of signaling events mediated by phosphorylation of Dishevelled and LRP6. The recruitment of the scaffolding protein AXIN1 to the receptor complex leads to inhibition of GSK3 and the b-catenin destruction complex. This is followed by the translocation of b-catenin to the nucleus where its association with TCF family members and coactivators, including CBP and p300, activates transcription of multiple Wnt/b-catenin target genes. Numerous inhibitors that target various nodes of the Wnt/ b-catenin pathway are indicated.

signaling, including Wnt production and Wnt ligand–receptor clinical trials (21, 22). It is likely that Ipafricept may interact interactions. Wnts secreted from sending cells interact with Friz- with a subset of the 19 Wnt ligands. Finally, Wnt signaling is zleds, LRP5/6 coreceptors, and various other receptors on the markedly stimulated by R-Spondins (RSPO) both physiologi- surface of receiving cells. Figure 2A illustrates our current under- cally, and pathologically when they are overexpressed due to standing of Wnt secretion. Figure 2B illustrates the regulation of specific chromosomal translocations (23, 24). Neutralizing cell surface Wnt receptors and the targets of upstream Wnt antibodies against RSPOs are also being developed (ref. 25; inhibitors (recently reviewed in ref. 4). and others). These recombinant proteins and monoclonal One disorder where it is clinically valuable to increase Wnt antibodies have high specificity, low off-target effects, and signaling is osteoporosis. Sclerostin (SOST) is an osteocyte-spe- infrequent dosing, but share issues of parenteral administra- cific Wnt inhibitor that interacts with LRP4 as well as LRP5/6 to tion, long half-lives, and slow off-rates. inhibit excessive bone formation (Fig. 2). SOST and LRP4 loss-of- An alternative approach to blocking Wnt signaling upstream function mutations cause increased bone mass, a finding that can is the use of orally available small molecules that prevent Wnt be mimicked therapeutically for the treatment of osteoporosis by ligand biogenesis and secretion. All mammalian Wnts require SOST or LRP4 neutralizing antibodies (16, 17). posttranslational acylation. A single essential, nonredundant Monoclonal antibodies and decoy receptors can also block enzyme, PORCN (the human ortholog of the Drosophila gene the interaction of Wnts with its receptors. Although early porcupine) adds a mono-unsaturated palmitoleate moiety to a studies examined antibodies directed against Wnts, recent work serine residue conserved in all mammalian Wnts (26–28). This has focused on inhibition of Wnt receptors. Examples include palmitoleation is essential for both Wnt secretion, and the anti-LRP6 antibodies (18), and the anti-Frizzled antibody van- binding of Wnts to their receptors (Fig. 2A; refs. 13, 29–31). tictumab (OMP-18R5; ref. 19). Vantictumab binds to 5 of the Thus, inhibition of PORCN blocks the secretion and activity of 10 Fzd receptors recognizing an epitope spanning the "cleft" all human Wnts (32, 33). Small-molecule PORCN inhibitors region of Fzd that contains highly conserved residues (19, 20). circumvent some of the limitations of the recombinant pro- Vantictumab may effectively target a subset of cancers that are tein/monoclonal antibody approaches, and will inhibit both dependent on the function of the appropriate Fzds. A related canonical and noncanonical signaling pathways. Several approach uses a soluble Fzd-based decoy receptor related to the PORCN inhibitors are being developed and have been shown soluble Frizzled-related proteins (sFRP; Fig. 2) that prevents the to be effective in cell- and mouse-based models of Wnt-depen- interaction of Wnts with their primary receptors. One such dent cancers (refs. 30, 34, 35; Madan and colleagues; unpub- recombinant protein, Ipafricept (OMP-54F28), has reached lished results). For all of the upstream Wnt inhibitors, one

OF2 Mol Cancer Ther; 14(5) May 2015 Molecular Cancer Therapeutics

Targeting Wnts at the Source

Figure 2. Upstream Wnt pathway inhibitors. A, Wnts are posttranslationally palmitoleated (orange line) by Porcupine (PORCN) in the ER, which is crucial for their interaction with Wntless (WLS) that transports them to the plasma membrane. Inhibitors of PORCN enzymatic activity prevent the palmitoleation of all Wnts, thus preventing their interaction with Wntless and subsequently their transport. B, SOST and DKK proteins inhibit Wnt signaling by interaction with LRP5/6. ZNRF3/RNF43 ubiquitinates the Frizzleds, targeting them for degradation. RSPOs function as Wnt coactivators by interacting with Leucine-rich repeat containing G protein-coupled receptors (LGR5 and others) resulting in complex formation with ZNRF3/RNF43, inhibiting their activity and leading to stabilization of Frizzleds at the cell surface. Decoy receptors and antibodies have been developed to prevent the interaction of Wnts with Frizzleds or the interaction of RSPOs with LGR5 to regulate Wnt signaling.

important challenge is to select patients who will beneﬁtfrom markers and indicators of toxicity in animal and human studies of their use. PORCN inhibitors.

Potential Toxicities of Wnt Inhibitors General Considerations for Predictive Apart from its role in development, Wnt signaling regulates Biomarkers for Upstream Inhibitor Therapy homeostasis of several adult organs, most notably the intes- in Cancer tinal epithelium, hair, and bone (36). Consistent with this, Upstream inhibition of Wnt signaling by targeting either tankyrase inhibitors demonstrated severe, mechanism-based PORCN or the interaction of Wnts with their receptors is most GI toxicity when given orally (14). In contrast with tankyrase likely to be clinically efficacious when the cancer requires con- inhibitors, PORCN inhibitors and the recombinant Frizzled tinued Wnt ligand activity for proliferation. Conversely, Wnt fi inhibitory antibodies have not shown signi cant toxic effects ligands may not be required when the downstream Wnt pathway on the intestine or skin, at least in mice, at doses that effec- is activated due to APC, AXIN, GSK3,orb-catenin mutations. In tively inhibit cancer proliferation (30, 34, 37). Doses of the these cancers, the upstream inhibitors are less likely to be PORCN inhibitor Wnt C-59 10-fold higher than the therapeu- efficacious. tic dose cause extensive loss of intestinal proliferation (37). Recent advances in understanding how the Wnt receptors The reason for this therapeutic window and the differences in Frizzled and LRP5/6 proteins are regulated, combined with toxicity of these different drugs is unclear, but may be due to next-generation sequencing of cancer genomes, have identified redundant pathways, differences in tissue penetration, or dos- genetic markers that may predict sensitivity to these upstream Wnt ing schedules. pathway inhibitors. Wnt signaling plays a critical role in bone metabolism, driving the differentiation and function of osteoblasts (38). Decreased Wnt signaling, both experimentally in mice, and in humans with Genetic Markers genetic variants, leads to defects in mineralization, osteoporosis, Mutations that sensitize cells to secreted Wnt ligands and fracture risk (38, 39). Not surprisingly, then, vantictumab and Cancers dependent on Wnt signaling are likely to have high ipafricept trials were temporarily halted apparently due to levels of Wnt ligands or Fzd receptors. Genetic mutations that increased bone turnover in the treated patients but restarted with make cancers more sensitive to Wnts by increasing Fzd abundance bone loss mitigation strategies (40, 41). Markers of bone turnover have recently been identified, and may be excellent predictive such as increase in bCTX (C-terminal telopeptide X), BSAP (bone- biomarkers. These genes, listed in Table 1, can constitute the first- specific alkaline phosphatase), osteocalcin, and P1NP (serum line tests for enrichment of patients in clinical trials of upstream type I collagen) may be useful both as pharmacodynamic bio- Wnt inhibitors.

www.aacrjournals.org Mol Cancer Ther; 14(5) May 2015 OF3

Madan and Virshup

Table 1. Genomic alterations predictive of sensitivity to upstream Wnt pathway inhibitors Gene Type of mutations Primary tissues % Mutated Citation RNF43 Loss-of-function: frameshift, indels, and nonsense mutations Biliary tract 38.5% (48) (esp. at R117 and G659 in MSI tumors) cause increased Endometrium 18% (46) levels of Frizzleds on the cell surface. Large intestine 18% (46) Esophagus 4% a Pancreas 6.1% (77) Stomach 16.3% (51) Mucinous ovarian 21% (47)

ZNRF3 Loss-of- function: frameshift, indels and nonsense mutations Endometrium 3.8% a may lead to increased levels of Frizzleds on the cell surface. Colorectal 9.7% (23, 24) Adrenocortical 19% b(TCGA)

RSPOs E1F3E(e1)–RSPO2 fusion, PTPRK(e1)–RSPO3(e2) PTPRK(e7)– Colon cancer 10% (23, 24) RSPO3(e2) fusion leading to increased levels of RSPOs, consequently preventing clearance of Frizzleds from the cell surface.

Notch 1 Loss-of- function mutations in N-terminal EGF repeats Hematopoietic and lymphoid 16.6% a responsible for receptor ligand interaction. Loss of function Colon cancer 8.3% (23, 24) of Notch may lead to elevated Wnt signaling and Wnt Esophagus 9.1% (60) ligand expression. Head and neck 18.9% b(TCGA) Cervix 8.3% b(TCGA) Low grade glioma 10.0% b(TCGA) Stomach 7.0% (51) Lung squamous cell carcinoma 8.4% b(TCGA)

Notch 2 Loss-of- function mutations in N-terminal EGF repeats Endometrium 6.7% a responsible for receptor ligand interaction. Loss of function Colon cancer 5.6% (24, 24) of Notch may lead to elevated Wnt signaling and Wnt Lung squamous cell carcinoma 7.3% b(TCGA) ligand expression. Urinary tract 6.9% a Stomach 8.0% (51)

Notch 3 Loss-of-function mutations in N-terminal EGF repeats Endometrium 7.7% a responsible for receptor ligand interaction. Loss-of- Large intestine 6.2% a function of Notch may lead to elevated Wnt signaling and Esophagus 7.8% a Wnt ligand expression. Stomach 6.6% (51) Colorectal 13.9 (23, 24) aCOSMIC (http://cancer.sanger.ac.uk/cancergenome/projects/cosmic/). bCbioportal (http://www.cbioportal.org/public-portal/).

RNF43 and ZNRF3 loss-of-function mutations extent at R117. RNF43 loss-of-function mutations are found in þ Inactivating mutations in the structurally related integral mem- other MSI cancers, including endometrial and gastric cancers þ brane E3 ubiquitin ligases RNF43 and ZNRF3 cause increased cell (46, 49–51). Whether PORCN inhibitors are effective in MSI surface Wnt receptors and hence sensitize cells to Wnts (Fig. cancers remains to be tested. 1; Table 1). RNF43 and ZNRF3 negatively regulate Wnt signaling by promoting ubiquitination, internalization, and degradation of R-Spondin translocations and gene amplification the Wnt receptors Frizzled and LRP5/6 (42, 43). Recent studies RSPOs are a family of four genes, RSPO1-4, that enhance Wnt (ref. 44; Madan and colleagues; unpublished data) have demon- signaling (ref. 52, reviewed in ref. 53). Their mechanism of action strated that the proliferation of pancreatic cancer cell lines and has recently been elucidated, and importantly, cancer-associated patient-derived xenografts with inactivating RNF43 mutations is gain-of-function mutations have been identified. RSPO proteins inhibited by PORCN inhibitors, including LGK974, Wnt-C59, can be thought of as an independent volume control on Wnt and ETC-159. Hence, RNF43 and ZNRF3 mutations, by sensitiz- signaling. RSPOs normally inhibit the activity of RNF43 and ing cells to Wnts, may also be predictive of response to both ZNRF3 by sequestering them into a heterotrimeric complex with PORCN inhibitors and Frizzled antibodies. Mutations in RNF43 LGR4/5/6 (Leucine-rich repeat containing G-protein–coupled are surprisingly common, occurring in 18% to 27% of endome- receptors 4, 5, or 6). This complex results in the inhibition and trial cancers, 3% to 5% of pancreatic cancers, 21% ovarian membrane clearance of these E3 ubiquitin ligases, leading to mucinous carcinomas, and 38.5% liver fluke associated cholan- increased Frizzled and LRP5/6 proteins on the cell surface giocarcinomas (45–48). Notably, up to 18% of colorectal cancers (43, 54). The LGR RSPO complex may also enhances b-cate- have RNF43 mutations, often associated with mismatch repair nin–dependent signaling by promoting MEK1/2-mediated phos- þ deficient, microsatellite instability (MSI ) cancers (46). These phorylation of LRP5/6 and b-catenin–independent signaling mutations are mutually exclusive with APC mutations, consistent through regulation of actin dynamics (55). with their ability to activate Wnt/b-catenin signaling. ZNRF3 An important insight into howWntscausecancercamewith mutations are found in adrenocortical and endometrial cancers the recent discovery of gain-of-function chromosomal translo- (Table 1). The MSI-associated mutations show a strong predilec- cations resulting in gene fusions with increased expression of tion for a G*C tract resulting in a frameshift at G659, and to a lesser RSPO2 and RSPO3 (23). Three recurrent translocations were

OF4 Mol Cancer Ther; 14(5) May 2015 Molecular Cancer Therapeutics

Targeting Wnts at the Source

found that fuse upstream signal sequences onto the amino- found in 5% of diffuse large B-cell lymphomas. As of yet, there terminus of the RSPO genes in about 10% of colon cancers and is no evidence that mutations in these genes will sensitize cells to to varying frequencies in head and neck, esophageal, and PORCN or other upstream Wnt inhibitors. ovarian cancers (23, 24, 56). These fusion products markedly stimulate Wnt/b-catenin signaling. As with the RNF43 muta- Upstream inhibitors may be less effective in cancers with tions, these fusions are mutually exclusive with APC mutations. downstream Wnt/b-catenin pathway mutations There are also subsets of ovarian (25%), prostate (10%), and Mutations in pathway components downstream of the plasma breast (10%) cancers with R-spondin2 gene amplifications. The membrane can activate Wnt/b-catenin signaling. Such down- tumors with elevated expression of RSPO either due to trans- stream mutations may render cells insensitive to upstream locations or due to amplifications could also potentially PORCN and Wnt inhibition. These mutations increase b-catenin respond to PORCN inhibitors, anti-Fzd antibodies, or anti- signaling primarily by preventing its phosphorylation or degra- RSPO antibodies (19, 25). Thus, RSPO2/3 gain-of-function dation. These include loss-of-function mutations in AXIN and mutations, together with RNF43/ZNRF3 loss-of-function muta- APC that impair their ability to serve as scaffolding proteins for the tions, molecularly define a population of patients with path- b-catenin degradation complex. Somatic and inherited mutations ologically increased cellular sensitivity to Wnts, who may in APC leading to prematurely truncated protein are most fre- benefit from new upstream Wnt inhibitors. quently found in colorectal cancers (49). Mutations in b-cate- The mutually exclusive nature of RNF43 or RSPO transloca- nin altering or deleting the GSK3 phosphorylation sites (Ser33, tions on one hand, and APC mutations on the other, could in a Ser37, and Thr41) are reported in various cancers, including simple world be interpreted as the need to activate b-catenin one pancreatic,gastric,liver,andendometrialcancers(64).These way or another. However, increased Wnt activity will also activate mutations prevent b-catenin degradation, leading to enhanced b-catenin–independent pathways, and in fact, some Wnts can LEF/TCF-mediated transcription. Mutations in WTX, another actually block b-catenin signaling. Conversely, APC mutations do scaffolding protein and part of the b-catenin degradation com- much more than simply elevate b-catenin protein abundance. plex, are also reported in colon cancers and prevent b-catenin Hence, it remains to be tested whether upstream activating muta- degradation (65). tions in RNF43 and RSPO in fact work in part or wholly through Although mutations in APC and b-catenin are common in b-catenin or whether they are oncogenic via b-catenin–indepen- colon cancer, these cancers still produce Wnt ligands that incre- dent mechanisms. mentally activate b-catenin (66, 67). Thus, it will be interesting to test whether upstream Wnt inhibitors regulate the growth of Loss-of-function mutations in Notch ligands cancers with downstream b-catenin activating mutations. Notch expression suppresses Wnt signaling in keratinocytes and skin tissues (57, 58). Inhibition of Notch signaling increases the expression of Wnt ligands, leading to increased Wnt signaling Gene Expression Markers (59). Consistent with this, loss-of-function mutations in Notch1 Wnt/b-catenin pathway activation turns on diverse target genes in head and neck squamous cell carcinoma cell lines sensitize that have been extensively investigated in multiple gene expres- them to PORCN inhibitors (34). Loss-of-function mutations in sion profiling studies. If gene expression data are available from Notch1 are found in diverse epithelial cancers. Notch mutations individual cancers, it may be potentially useful in predicting Wnt may predict Wnt dependency and sensitivity to upstream inhi- pathway activation. However, despite years of study, there is no bitors, but confirmatory studies are needed (60). robust, clinically useful gene expression signature that predicts Wnt pathway activation across tissues. AXIN2 is the most broadly Mutations that may sensitize cancer cells to secreted Wnt used Wnt/b-catenin target gene and high expression of AXIN2 is ligands routinely used as a measure of elevated canonical Wnt signaling in Liu and colleagues noted that approximately one of three of the research laboratory. Assessment of AXIN2 expression in clin- head and neck cancer cell lines responded to PORCN inhibition in ical samples would require mRNA analysis, as reliable monoclo- vitro, and performed exome sequencing on a subset of those lines nal antibodies are not available. Importantly, these assays will (34). They found a number of mutations that correlated with have only limited ability to guide the decision of whether to use an PORCN inhibitor sensitivity. These genes are candidate predictive upstream Wnt inhibitor, because activation of Wnt target genes biomarkers, although the associations are not yet validated and will not differentiate between upstream or downstream muta- the mechanistic connection is obscure. It is also possible that these tions in the Wnt/b-catenin pathway. tumors harbor unidentified RSPO translocations (24). Increase in the expression of various genes such as NKD1, Two genes associated with PORCN inhibitor sensitivity in head RNF43, ZNRF3, TNFRSF19, SP5, BMP7, LGR4/5, APCCD1,and and neck cancer cell lines, were FAT1 and FAM58A. FAT1 is a NOTUM correlates with AXIN2 expression, consistent with protocadherin that can interact with b-catenin and sequester it to other evidence that these are Wnt/b-catenin target genes (ref. 43; the cell membrane. Inactivation of FAT1 leads to aberrant Wnt/ A.Alok,Z.Lei,DMV;manuscriptinpreparation).Increasein b-catenin signaling in multiple types of cancer (61). In addition to the expression of RSPOs, either due to amplification or trans- head and neck squamous cell carcinomas, FAT1 mutations are locations, as noted above, may also predict enhanced Wnt frequently observed in central nervous system, colorectal, and signaling. With the development of sensitive monoclonal anti- ovarian cancers. FAM58A (family with sequence similarity 58, bodies and recent advances in techniques such as RNAscope member A) encodes cyclin M, which regulates CDK10 and inter- (68) that allow for reliable quantification of gene expression on acts with SALL1 (62, 63). Mutations in FAM58A cause an X-linked formalin-fixed tissues, profiling of expression of these genes dominant disorder characterized by syndactyly, telecanthus, ano- might be useful for patient selection. Thus, it may be possible to genital, and renal malformations. Mutations in FAM58A are develop panels that assess expression of a combination of these

www.aacrjournals.org Mol Cancer Ther; 14(5) May 2015 OF5

Madan and Virshup

genes and gene products to predict sensitivity to upstream Wnt Other pathway proteins, such as PORCN and WLS, may be useful inhibitors. markers and are the subject of active research (76). Finally, there are several phosphorylation events coupled Genes Whose Decreased Expression to Wnt-receptor engagement, including the phosphorylation Suggests Increased Wnt Signaling at the of defined residues on LRP6 and Dishevelled. Phosphoepi- tope-specific antibodies are useful tools in the research Membrane laboratory, primarily in immunoblotting, but their utility There are multiple secreted Wnt antagonists that may normally in routine IHC is neither established, nor likely to be dampen Wnt signaling but are epigenetically silenced in various clinically feasible. cancers (66, 69). Most of the reported antagonists function by preventing the interaction of Wnts with their receptors LRP5/6 or Conclusions Frizzled. Although epigenetic silencing may be a challenge to Recent studies have revealed further complexity in the Wnt/ monitor clinically, it provides an additional mechanism to sen- b-catenin pathway. Posttranslational modifications of WNTs sitize cancers to Wnts, and hence to upstream Wnt inhibitors. and AXIN have been identified and can be targeted to benefit The best-known Wnt antagonists are the Dickkopf (DKK) and patients with elevated Wnt signaling. Accumulating data dem- the sFRP families. DKKs are secreted glycoproteins that bind to onstrating the b-catenin–independent role of Wnts combined LRP5/6. DKK1/2 interacts with LRP5/6, whereas DKK4 interacts with the limitations of the current downstream pathway inhi- with LRP6. Epigenetic silencing of DKK1 and other DKK family bitors have led to development of drugs that inhibit PORCN members is described in colon cancers (reviewed in ref. 70). The and target Wnt secretion and antibodies and decoy receptors sFRPs interact with Wnts via the Wnt-binding cysteine rich that prevent the interaction of Wnts with their receptors. A domain of the Frizzleds and can activate or antagonize Wnt variety of tools may assist in selecting patients who can benefit signaling depending on their concentration (71, 72). Epigenetic from these new Wnt pathway inhibitors. Most notably, recent silencing of sFRP promoters by hypermethylation is seen in studies show that Wnt receptor abundance is regulated by colorectal and breast cancers (66, 69). Wnt inhibitory factor-1, ubiquitination and internalization via an RSPO/LGR5/RNF43 another secreted Wnt inhibitor, is silenced in several tumor types axis. Large-scale sequencing efforts have identified common (73–75). Profiling the expression of Wnt antagonists or epigenetic loss-of-function mutations in RNF43/ZNRF3 and gain-of-func- marks associated with their silencing, while challenging in clinical tion translocations in RSPO2/3 that can be clinically utilized to practice, could be useful to predict sensitivity of tumors to guide the selection of the patients who may respond to upstream Wnt inhibitors. upstream Wnt pathway inhibitors. IHC tools to assess the mutational activation of Wnt signaling at the membrane may Protein Markers also be useful to identify the patients with cancer who may fi Increased abundance of the Wnt cell surface receptors Fzds is bene t from Wnt pathway inhibitors. likely to be seen in cancers with mutant RNF43 or overexpressed fl RSPO2/3. It may be possible to develop this as a predictor of Disclosure of Potential Con icts of Interest responsiveness to PORCN and other upstream pathway inhibi- B. Madan and D.M. Virshup are co-inventors on patents on PORCN inhibitors. D.M. Virshup is a consultant/advisory board member for Experimental tors. Ideally, one could screen for increased Fzds by IHC on FFPE Therapeutics Center Singapore. samples. Broad specificity anti-Fzd antibodies have been developed and may soon be widely available for this purpose. Increased Acknowledgments levels of RSPO2 or RSPO3 protein secondary to gene amplifica- The authors thank their many colleagues in Singapore and around the world tion or translocation may also be detected by specific IHC for insights, shared results, and exciting science. We apologize for any omissions staining. in citations due to limitations in space. Cytosolic and nuclear b-catenin is a commonly used research marker of Wnt/b-catenin pathway activation but does not differ- Grant Support entiate between upstream or downstream Wnt pathway activation, D.M. Virshup is supported in part by the National Research Foundation and may be increased due to activation of alternative pathways, Singapore through Singapore Translational Research (STaR) Investigatorship AXIN2 award NMRC/STaR/0017/2013 administered by the Singapore Ministry of such as EGFR signaling. Similarly, expression is indicative of Health's National Medical Research Council. pathway activation but does not differentiate between upstream and downstream causes. In addition, AXIN2 protein levels are low Received December 9, 2014; revised February 3, 2015; accepted February 22, and not readily detectable with currently available antibodies. 2015; published OnlineFirst April 21, 2015.

References 1. Nusse R, Varmus H. Three decades of Wnts: a personal perspective on how a 5. Anastas JN, Moon RT. WNT signalling pathways as therapeutic targets in scientiﬁc ﬁeld developed. EMBO J 2012;31:2670–84. cancer. Nat Rev Cancer 2013;13:11–26. 2. Niehrs C. The complex world of WNT receptor signalling. Nat Rev Mol Cell 6. Clevers H, Nusse R. Wnt/b-catenin signaling and disease. Cell 2012;149: Biol 2012;13:767–79. 1192–205. 3. Acebron SP, Karaulanov E, Berger BS, Huang Y-L, Niehrs C. Mitotic wnt 7. Kahn M. Can we safely target the WNT pathway? Nat Rev Drug Discov signaling promotes protein stabilization and regulates cell size. Mol Cell 2014;13:513–32. 2014;54:663–74. 8. Grigoryan T, Wend P, Klaus A, Birchmeier W. Deciphering the function of 4. Yu J, Virshup DM. Updating the Wnt pathways. Biosci Rep 2014;34: canonical Wnt signals in development and disease: conditional loss- and gain- 593–607. of-function mutations of b-catenin in mice. Genes Dev 2008;22:2308–41.

OF6 Mol Cancer Ther; 14(5) May 2015 Molecular Cancer Therapeutics

Targeting Wnts at the Source

9. Takahashi-Yanaga F, Kahn M. Targeting Wnt signaling: can we safely 32. Proffitt KD, Virshup DM. Precise regulation of porcupine activity eradicate cancer stem cells? Clin Cancer Res 2010;16:3153–62. is required for physiological Wnt signaling. J Biol Chem 2012;287: 10. Lee SB, Gong Y-D, Park YI, Dong M-S. 2,3,6-Trisubstituted quinoxaline 34167–78. derivative, a small molecule inhibitor of the Wnt/beta-catenin signal- 33. Najdi R, Proffitt K, Sprowl S, Kaur S, Yu J, Covey TM, et al. A uniform human ing pathway, suppresses cell proliferation and enhances radiosensitiv- Wnt expression library reveals a shared secretory pathway and unique ity in A549/Wnt2 cells. Biochem Biophys Res Commun 2013;431: signaling activities. Differentiation 2012;84:203–13. 746–52. 34. Liu J, Pan S, Hsieh MH, Ng N, Sun F, Wang T, et al. Targeting Wnt-driven 11. Huang S-MA, Mishina YM, Liu S, Cheung A, Stegmeier F, Michaud GA, et al. cancer through the inhibition of Porcupine by LGK974. Proc Natl Acad Sci Tankyrase inhibition stabilizes axin and antagonizes Wnt signalling. U S A 2013;110:20224–9. Nature 2009;461:614–20. 35. Wang X, Moon J, Dodge ME, Pan X, Zhang L, Hanson JM, et al. The 12. Emami KH, Nguyen C, Ma H, Kim DH, Jeong KW, Eguchi M, et al. A small development of highly potent inhibitors for porcupine. J Med Chem molecule inhibitor of beta-catenin/CREB-binding protein transcription. 2013;56:2700–4. Proc Natl Acad Sci U S A 2004;101:12682–7. 36. Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 13. Chen B, Dodge ME, Tang W, Ma Z, Fan C-W, Wei S, et al. Small molecule- 2006;127:469–80. mediated disruption of Wnt-dependent signaling in tissue regeneration 37. Kabiri Z, Kabiri Z, Greicius G, Greicius G, Madan B, Madan B, et al. Stroma and cancer. Nat Chem Biol 2009;5:100–7. provides an intestinal stem cell niche in the absence of epithelial Wnts. 14. Lau T, Chan E, Callow M, Waaler J, Boggs J, Blake RA, et al. A novel tankyrase Development 2014;141:2206–15. small-molecule inhibitor suppresses APC mutation-driven colorectal 38. Zhong Z, Zylstra-Diegel CR, Schumacher CA, Baker JJ, Carpenter AC, Rao S, tumor growth. Cancer Res 2013;73:3132–44. et al. Wntless functions in mature osteoblasts to regulate bone mass. Proc 15. Lenz H-J, Kahn M. Safely targeting cancer stem cells via selective catenin Natl Acad Sci U S A 2012;109:E2197–204. coactivator antagonism. Cancer Sci 2014;105:1087–92. 39. Laine CM, Joeng KS, Campeau PM, Kiviranta R, Tarkkonen K, Grover M, 16. Chang M-K, Kramer I, Huber T, Kinzel B, Guth-Gundel S, Leupin O, et al. et al. WNT1 mutations in early-onset osteoporosis and osteogenesis Disruption of Lrp4 function by genetic deletion or pharmacological imperfecta. N Engl J Med 2013;368:1809–16. blockade increases bone mass and serum sclerostin levels. Proc Natl Acad 40. Oncomed. FDA Removes Partial Clinical Hold on OncoMed's Ipafri- Sci U S A 2014;111:E5187–95. cept [Internet]. investor.shareholder.com. 2014 [cited 2014 Sep 10]. 17. McClung MR, Grauer A, Boonen S, Bolognese MA, Brown JP, Diez-Perez A, Available from: http://investor.shareholder.com/oncomed/releasede- et al. Romosozumab in postmenopausal women with low bone mineral tail.cfm?ReleaseID¼869239 density. N Engl J Med 2014;370:412–20. 41. Oncomed. OncoMed Pharmaceuticals Announces Removal of Partial 18. Ettenberg SA, Charlat O, Daley MP, Liu S, Vincent KJ, Stuart DD, et al. Clinical Hold by the FDA for Vantictumab [Internet]. investor.sharehold- Inhibition of tumorigenesis driven by different Wnt proteins requires er.com. 2014 [cited 2014 Sep 10]. Available from: http://investor.share- blockade of distinct ligand-binding regions by LRP6 antibodies. Proc Natl holder.com/oncomed/releasedetail.cfm?ReleaseID¼868147 Acad Sci U S A 2010;107:15473–8. 42. Koo B-K, Spit M, Jordens I, Low TY, Stange DE, van de Wetering M, et al. 19. Gurney A, Axelrod F, Bond CJ, Cain J, Chartier C, Donigan L, et al. Wnt Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis pathway inhibition via the targeting of Frizzled receptors results in of Wnt receptors. Nature 2012;488:665–9. decreased growth and tumorigenicity of human tumors. Proc Natl Acad 43. Hao H-X, Xie Y, Zhang Y, Charlat O, Oster E, Avello M, et al. ZNRF3 Sci U S A 2012;109:11717–22. promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 20. Dann CE, Hsieh JC, Rattner A, Sharma D, Nathans J, Leahy DJ. Insights into 2012;485:195–200. Wnt binding and signalling from the structures of two Frizzled cysteine-rich 44. Jiang X, Hao H-X, Growney JD, Woolfenden S, Bottiglio C, Ng N, et al. domains. Nature 2001;412:86–90. Inactivating mutations of RNF43 confer Wnt dependency in pancreatic 21. Sebio A, Kahn M, Lenz H-J. The potential of targeting Wnt/b-catenin in ductal adenocarcinoma. Proc Natl Acad Sci U S A 2013;110:12649–54. colon cancer. Expert Opin Ther Targets 2014;18:611–5. 45. Kinde I, Bettegowda C, Wang Y, Wu J, Agrawal N, Shih I-M, et al. Evaluation 22. Smith DC, Gordon M, Messersmith W, Chugh R, Mendelson D, Dupont J, of DNA from the Papanicolaou test to detect ovarian and endometrial et al. Abstract B79: a first-in-human Phase 1 study of anti-cancer stem cell cancers. Sci Transl Med 2013;5:167ra4. (CSC) agent OMP-54F28 (FZD8-Fc) targeting the WNT pathway in patients 46. Giannakis M, Hodis E, Jasmine Mu X, Yamauchi M, Rosenbluh J, Cibulskis with advanced solid tumors. Mol Cancer Ther 2014;12:B79–9. K, et al. RNF43 is frequently mutated in colorectal and endometrial cancers. 23. Seshagiri S, Stawiski EW, Durinck S, Modrusan Z, Storm EE, Conboy CB, Nat Genet 2014;46:1264–6. et al. Recurrent R-spondin fusions in colon cancer. Nature 2012;488: 47. Ryland GL, Hunter SM, Doyle MA, Rowley SM, Christie M, Allan PE, et al. 660–4. RNF43 is a tumour suppressor gene mutated in mucinous tumours of the 24. Cardona GM, Bell K, Portale J, Gaffney D, Moy C, Platero S, et al. ovary. J Pathol 2013;229:469–76. Identification of R-Spondin fusions in various types of human cancer 48. Ong CK, Subimerb C, Pairojkul C, Wongkham S, Cutcutache I, Yu W, et al. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Exome sequencing of liver fluke-associated cholangiocarcinoma. Nat Gen- Association for Cancer Research; 2014 Apr 5–9; San Diego, CA. Philadel- et 2012;44:690–3. phia (PA): AACR; Abstract nr 2408. 49. Muzny DM, Bainbridge MN, Chang K, Dinh HH, Drummond JA, Fowler G, 25. Gurney A, Bond CJ. RSPO3 BINDING AGENTS AND USES THEREOF. et al. Comprehensive molecular characterization of human colon and Pharmaceuticals O, editor. United States Patent Office; 2014. pp. rectal cancer. Nature 2012;487:330–7. 1–161. 50. Wang K, Yuen ST, Xu J, Lee SP, Yan HHN, Shi ST, et al. Whole-genome 26. Lum L, Clevers H. The unusual case of porcupine. Science 2012;337: sequencing and comprehensive molecular profiling identify new driver 922–3. mutations in gastric cancer. Nat Genet 2014;46:573–82. 27. Rios-Esteves J, Resh MD. Stearoyl CoA desaturase is required to produce 51. Cancer Genome Atlas Research Network. Comprehensive molecular char- active, lipid-modified Wnt proteins. Cell Rep 2013;4:1072–81. acterization of gastric adenocarcinoma. Nature 2014;513:202–9. 28. Takada R, Satomi Y, Kurata T, Ueno N, Norioka S, Kondoh H, et al. 52. Kazanskaya O, Glinka A, del Barco Barrantes I, Stannek P, Niehrs C, Wu W. Monounsaturated fatty acid modification of Wnt protein: its role in Wnt R-Spondin2 is a secreted activator of Wnt/beta-catenin signaling and is secretion. Dev Cell 2006;11:791–801. required for Xenopus myogenesis. Dev Cell 2004;7:525–34. 29. Coombs GS, Yu J, Canning CA, Veltri CA, Covey TM, Cheong JK, et al. WLS- 53. de Lau W, Peng WC, Gros P, Clevers H. The R-spondin/Lgr5/Rnf43 module: dependent secretion of WNT3A requires Ser209 acylation and vacuolar regulator of Wnt signal strength. Genes Dev 2014;28:305–16. acidification. J Cell Sci 2010;123:3357–67. 54. Xie Y, Zamponi R, Charlat O, Ramones M, Swalley S, Jiang X, et al. 30. Proffitt KD, Madan B, Ke Z, Pendharkar V, Ding L, Lee MA, et al. Pharma- Interaction with both ZNRF3 and LGR4 is required for the signalling cological inhibition of the Wnt acyltransferase PORCN prevents growth of activity of R-spondin. EMBO Rep 2013;14:1120–6. WNT-driven mammary cancer. Cancer Res 2013;73:502–7. 55. Carmon KS, Gong X, Yi J, Thomas A, Liu Q. RSPO-LGR4 functions via 31. Janda CY, Waghray D, Levin AM, Thomas C, Garcia KC. Structural basis of IQGAP1 to potentiate Wnt signaling. Proc Natl Acad Sci U S A 2014;111: Wnt recognition by Frizzled. Science 2012;337:59–64. E1221–9.

www.aacrjournals.org Mol Cancer Ther; 14(5) May 2015 OF7

Madan and Virshup

56. Shinmura K, Kahyo T, Kato H, Igarashi H, Matsuura S, Nakamura S, et al. 67. Voloshanenko O, Erdmann G, Dubash TD, Augustin I, Metzig M, Moffa G, RSPO fusion transcripts in colorectal cancer in Japanese population. Mol et al. Wnt secretion is required to maintain high levels of Wnt activity in Biol Rep 2014;41:5375–84. colon cancer cells. Nat Commun 2013;4:2610. 57. Nicolas M, Wolfer A, Raj K, Kummer JA, Mill P, van Noort M, et al. Notch1 68. Wang F, Flanagan J, Su N, Wang L-C, Bui S, Nielson A, et al. RNAscope: a functions as a tumor suppressor in mouse skin. Nat Genet 2003;33: novel in situ RNA analysis platform for formalin-fixed, paraffin-embedded 416–21. tissues. J Mol Diagn 2012;14:22–9. 58. Proweller A, Tu L, Lepore JJ, Cheng L, Lu MM, Seykora J, et al. Impaired 69. Suzuki H, Toyota M, Carraway H, Caraway H, Gabrielson E, Ohmura T, notch signaling promotes de novo squamous cell carcinoma formation. et al. Frequent epigenetic inactivation of Wnt antagonist genes in breast Cancer Res 2006;66:7438–44. cancer. Br J Cancer 2008;98:1147–56. 59. Devgan V, Mammucari C, Millar SE, Brisken C, Dotto GP. p21WAF1/Cip1 70. Cruciat C-M, Niehrs C. Secreted and transmembrane wnt inhibitors and is a negative transcriptional regulator of Wnt4 expression downstream of activators. Cold Spring Harb Perspect Biol 2013;5:a015081–1. Notch1 activation. Genes Dev 2005;19:1485–95. 71. Uren A, Reichsman F, Anest V, Taylor WG, Muraiso K, Bottaro DP, et al. 60. Song Y, Li L, Ou Y, Gao Z, Li E, Li X, et al. Identification of genomic Secreted Frizzled-related protein-1 binds directly to wingless and is a alterations in oesophageal squamous cell cancer. Nature 2014;509:91–5. biphasic modulator of Wnt signaling. J Biol Chem 2000;275:4374–82. 61. Morris LGT, Kaufman AM, Gong Y, Ramaswami D, Walsh LA, Turcan S,¸ 72. Xavier CP, Melikova M, Chuman Y, Uren A, Baljinnyam B, Rubin JS. et al. Recurrent somatic mutation of FAT1 in multiple human cancers leads Secreted Frizzled-related protein potentiation versus inhibition of to aberrant Wnt activation. Nat Genet 2013;45:253–61. Wnt3a/b-catenin signaling. Cell Signal 2014;26:94–101. 62.GuenVJ,GambleC,FlajoletM,UngerS,TholletA,FerandinY,etal. 73. Kansara M, Tsang M, Kodjabachian L, Sims NA, Trivett MK, Ehrich M, et al. CDK10/cyclin M is a protein kinase that controls ETS2 degradation and Wnt inhibitory factor 1 is epigenetically silenced in human osteosarcoma, is deficient in STAR syndrome. Proc Natl Acad Sci U S A 2013;110: and targeted disruption accelerates osteosarcomagenesis in mice. J Clin 19525–30. Invest 2009;119:837–51. 63. Unger S, Bohm€ D, Kaiser FJ, Kaulfuss S, Borozdin W, Buiting K, et al. 74. Taniguchi H, Yamamoto H, Hirata T, Miyamoto N, Oki M, Nosho K, et al. Mutations in the cyclin family member FAM58A cause an X-linked dom- Frequent epigenetic inactivation of Wnt inhibitory factor-1 in human inant disorder characterized by syndactyly, telecanthus and anogenital and gastrointestinal cancers. Oncogene 2005;24:7946–52. renal malformations. Nat Genet 2008;40:287–9. 75. Elston MS, Clifton-Bligh RJ. Identification of Wnt family inhibitors: a 64. Tornesello ML, Buonaguro L, Tatangelo F, Botti G, Izzo F, Buonaguro FM. pituitary tumor directed whole genome approach. Mol Cell Endocrinol Mutations in TP53, CTNNB1 and PIK3CA genes in hepatocellular carci- 2010;326:48–54. noma associated with hepatitis B and hepatitis C virus infections. Geno- 76. Stewart J, James J, McCluggage GW, McQuaid S, Arthur K, Boyle D, et al. mics 2013;102:74–83. Analysis of wntless (WLS) expression in gastric, ovarian, and breast cancers 65. Major MB, Camp ND, Berndt JD, Yi X, Goldenberg SJ, Hubbert C, et al. reveals a strong association with HER2 overexpression. Mod Pathol 2015; Wilms tumor suppressor WTX negatively regulates WNT/beta-catenin 28:428–36. signaling. Science 2007;316:1043–6. 77. Wu J, Jiao Y, Dal Molin M, Maitra A, de Wilde RF, Wood LD, et al. Whole- 66. Suzuki H, Watkins DN, Jair K-W, Schuebel KE, Markowitz SD, Chen WD, exome sequencing of neoplastic cysts of the pancreas reveals recurrent et al. Epigenetic inactivation of SFRP genes allows constitutive WNT mutations in components of ubiquitin-dependent pathways. Proc Natl signaling in colorectal cancer. Nat Genet 2004;36:417–22. Acad Sci U S A 2011;108:21188–93.

OF8 Mol Cancer Ther; 14(5) May 2015 Molecular Cancer Therapeutics

Targeting Wnts at the Source−−New Mechanisms, New Biomarkers, New Drugs

Babita Madan and David M. Virshup

Mol Cancer Ther Published OnlineFirst April 21, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-1038

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/early/2015/04/20/1535-7163.MCT-14-1038. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.