Aberrant Wnt signaling drives a number of cancers through regulation of diverse downstream pathways. Wnt/β-catenin signaling achieves this in part by increasing the expression of proto-oncogenes such as MYC and cyclins. However, global assessment of the Wnt-regulated transcriptome in vivo in genetically distinct cancers demonstrates that Wnt signaling suppresses the expression of as many genes as it activates. In this study, we examined the set of genes that are upregulated upon inhibition of Wnt signaling in Wnt-addicted pancreatic and colorectal cancer models. Decreasing Wnt signaling led to a marked increase in gene expression by activating ERK and JNK; these changes in gene expression could be mitigated in part by concurrent inhibition of MEK. These findings demonstrate that increased Wnt signaling in cancer represses MAPK activity, preventing RAS-mediated senescence while allowing cancer cells to proliferate. These results shift the paradigm from Wnt/β-catenin primarily as an activator of transcription to a more nuanced view where Wnt/β-catenin signaling drives both widespread gene repression and activation.

Significance:

These findings show that Wnt/β-catenin signaling causes widespread gene repression via inhibition of MAPK signaling, thus fine tuning the RAS-MAPK pathway to optimize proliferation in cancer.

Wnt signaling is an important regulator of development, growth, and differentiation, and is often dysregulated in cancer. In the best understood paradigm, Wnt binding to Frizzled and LRP5/6 inhibits glycogen synthase kinase 3 (GSK3) and stabilizes β-catenin, allowing it to accumulate and move to the nucleus (1, 2). Nuclear β-catenin activates gene expression by binding to members of the T-cell factor/lymphoid enhancer factor family, causing displacement of corepressors such as Groucho (3, 4). These Wnt/β-catenin activated genes such as MYC and cyclins regulate cellular proliferation and differentiation in a tissue-specific manner (5, 6). This general model explains how Wnt signaling can activate gene expression, and how excessive Wnt signaling contributes to cancer development. However, another major consequence of Wnt/β-catenin signaling, the repression of gene expression, is less well understood and its role in cancer is unknown. Based primarily on studies in Drosophila and mice, a handful of genes are known to be repressed by Wnt signaling (7–10). There are, however, no studies systematically examining the genes that are repressed by Wnt signaling.

A subset of pancreatic, biliary, adrenocortical, and colorectal cancers are Wnt addicted, and provide a powerful system to study genes and pathways regulated by Wnt signaling. Wnt-addicted cancers are driven by mutations in the E3 ubiquitin ligases RNF43/ZNRF3 or fusions in R-spondins (RSPO) that lead to the increased abundance of the Wnt receptors Frizzled and LRP5/6 at the cell surface (11–13). These cancers are highly dependent on continued Wnt production. Fortunately, Wnt secretion is now druggable. This is because all Wnts require the addition of a palmitoleate to a conserved serine residue. This palmitoleation, catalyzed by the O-acyltransferase porcupine (PORCN), mediates Wnt interaction with both the transporter Wntless (WLS) and with the cysteine-rich domain of Frizzled receptors, and is hence essential for Wnt secretion and its activity (14, 15). Pharmacologic inhibition of PORCN has proven effective in the treatment of multiple preclinical models of Wnt-addicted cancers (6, 11, 16–20).

To better understand Wnt-regulated genes, we pharmacologically blocked Wnt signaling in three different Wnt-addicted cancers, two RNF43 mutant orthotopic pancreatic cancer models and in a RSPO3-mutant colorectal cancer patient-derived xenograft (PDX), using the PORCN inhibitor ETC-159, a well-tolerated drug currently in clinical trials (21). Examining the time-dependent transcriptional changes in these models showed that 24% of expressed gene transcripts were decreased within 2–3 days of Wnt inhibition. These included well-recognized Wnt/β-catenin target genes such as AXIN2, NOTUM and important mediators of cell cycle and ribosomal biogenesis including MYC and cyclins (6, 17). Surprisingly, 28% of the transcripts were upregulated upon Wnt inhibition, that is, they were repressed by high Wnt signaling. How Wnt signaling represses these genes is currently unknown.

Here, we find that a large fraction of genes upregulated upon Wnt inhibition have ETS- and AP1-binding sites in their promoters. Investigating the pathways that are upregulated upon Wnt inhibition, we find that both tankyrase and PORCN inhibitors lead to enhanced activation of downstream MAPK signaling in multiple Wnt-driven preclinical cancer models. Concurrent targeting of both MEK and PORCN in these cancers confirmed that a large fraction of genes that are upregulated upon Wnt inhibition are regulated via a Wnt/β-catenin/MAPK pathway. We propose that one role of Wnt signaling in cancer is to dampen RAS-MAPK signaling, thereby preventing oncogene-induced senescence to maximize cancer proliferation.

Tissue culture

HEK293 were obtained from ATCC and grown in DMEM, Eagle Minimum Essential Medium (EMEM), or RPMI1640 supplemented with 10% FBS and 1% GlutaMAX (Life Technologies) in a 37°C humidified incubator with 5% CO2. AsPC-1 cells from ATCC were grown in RPMI supplemented with 10% FBS in a 37°C humidified incubator with 5% CO2. HPAF-II cells from ATCC were grown in EMEM supplemented with 10% FBS in a 37°C humidified incubator with 5% CO2. E[beta]P was a gift from Roel Nusse (Addgene plasmid, catalog no. 24313; ref. 22). This plasmid was used for generating HPAF-II cells stably overexpressing stabilized β-catenin. Selection and passage of HPAF-II cell lines harboring the stabilized β-catenin was performed with 1 μg/mL puromycin. All cell lines were regularly tested for Mycoplasma contamination and confirmed to be Mycoplasma free. The cell lines were used for experiments at a passage less than 20 after thawing.

Animal care

NOD scid gamma (NSG) mice were purchased from InVivos or Jackson Laboratories. The animal studies were approved by the Duke-NUS Institutional Animal Care and Use Committee and complied with applicable regulations. Animals were housed in standard cages and were allowed access ad libitum to food and water.

Tumor growth and mice treatment

Mouse xenograft models were established by injection of HPAF-II cells in NSG mice. For xenograft studies, HPAF-II cells or HPAF-II cells with stabilized β-catenin were suspended in 50% matrigel and injected subcutaneously into the flanks of NSG mice. Mice were treated with ETC-159 or trametinib after establishment of tumors. Trametinib or ETC-159 were formulated in 50% PEG 400 (vol/vol) in water and administered by oral gavage at a dosing volume of 10 μL/g body weight as described previously (6). The tumor dimensions were measured with a caliper routinely, and the tumor volumes were calculated as 0.5 × length × width2 as described previously (20). All mice were sacrificed 8 hours after the last dosing. At sacrifice, tumors were resected, weighed, and snap frozen in liquid nitrogen or fixed in 10% neutral buffered formalin.

Western blot analysis

For immunoblot analysis, tumors were homogenized in 4% SDS buffer using a polytron homogenizer. Equal amount of proteins were resolved on 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. Western blots were performed according to standard methods. The blots were probed with phospho-p44/42 MAPK (Erk1/2; Thr202/Tyr204) antibody (9106), total ERK (9102), actin, and p-JNK (9251) antibodies from Cell Signaling Technology, GAPDH from Abcam (ab8245) and total JNK (BD51-1570GR). The blots were developed using SuperSignal West femto or SuperSignal West Dura substrate (Thermo Fisher Scientific). The images were captured digitally using the LAS-3000 Life Science Imager (Fujifilm). For stripping and reprobing the blots, “restore Western blot stripping buffer” from Thermo Fisher Scientific was used.

Luciferase reporter assays

For reporter assays in HEK293 cells, cells were plated in 24-well plates and were transiently transfected with either an AP1 or c-Fos reporter plasmid, and WNT1 expression plasmid. After 24 hours, the cells were washed with PBS and lysed in reporter lysis buffer (E4030, Promega) containing protease inhibitors. Luciferase reporter activity was measured using firefly luciferase substrate (Promega). Cell viability as determined using lactate dehydrogenase assay was used for normalization of the reporter activity.

RNA isolation and qRT-PCR

Tumors were homogenized in RLT buffer using a polytron homogenizer. Total RNA was isolated using RNAeasy kit according to manufacturer's protocol. The RNA sequencing (RNA-seq) libraries were prepared using the Illumina TruSeq Stranded Total RNA protocol with subsequent PolyA enrichment. qRT-PCR was performed using SsoFast Evagreen Supermix (Bio-Rad).

IHC

Tissue sections of formalin-fixed paraffin-embedded tumors were deparaffinized in xylene and rehydrated in a series of ethanol gradients. For phospho-ERK (p-ERK) staining, after antigen retrieval with sodium citrate buffer pH 6.0 for 10 minutes, the endogenous peroxidase activity was blocked by incubation with H2O2. The sections were then incubated overnight with 1:200 diluted phospho-p44/42 MAPK antibody (4376, Cell Signaling Technology) followed by incubation with horseradish peroxidase–conjugated anti-rabbit secondary antibody (Dako) for 1 hour. Incubation with 3,3′-diaminobenzidine chromogen substrate resulted in brown staining of phospho-ERK positive cells and the nuclei were counterstained with Mayer's hematoxylin. Bright-field images were acquired on a Nikon Eclipse Ni-E microscope.

Gene expression analysis

RNA-seq datasets were analyzed and clustered as described previously (6). Sequences were assessed for quality and reads originating from mouse (mm10) were removed using Xenome (23). The remaining reads were aligned against hg38 (Ensembl version 79) using STAR v2.5.2 (24) and quantified using RSEM v1.2.31 (25). Reads mapping to chrM or annotated as rRNA, snoRNA, and snRNA were removed. Genes which had less than 10 reads mapping on average over all samples were removed. Differential expression analysis was performed using DESeq2 (26). Independent filtering was not used in this analysis. Pairwise comparisons were performed using a Wald test. To control for false positives due to multiple comparisons in the genome-wide differential expression analysis, we used the FDR that was computed using the Benjamini–Hochberg procedure. Gene-level counts were transformed using a variance-stabilizing transformation and converted to z-scores. Time was transformed using a square root transformation. All genes differentially expressed over time (DESeq2, FDR < 10%) were clustered using GPClust (27) using the Matern32 kernel with a concentration (alpha) parameter of 0.001 and a length scale of 6.5. For the APC-mutant colorectal PDX models (PDX1 and PDX2), microarray data normalization and selection of genes with changes in expression were performed using Partek analysis software. The genes whose expression differed more than 1.5-fold between experimental groups, P ≤ 0.05 were considered significant. P values were controlled for multiple testing. The RNA-seq data were deposited in the NCBI's Sequence Read Archive under accession number PRJNA632995.

Functional enrichment analysis

For the analysis of the Wnt-repressed genes Gene Ontology (GO) enrichments were performed with GOStats (28) using all genes differentially expressed (FDR < 10%) as background. ReactomePA (29) or Enrichr (30) was used for investigating pathway enrichments using the same background. Terms with an FDR < 10% were defined as significantly enriched. GO:BP and Reactome pathway enrichments for the models of interest are reported in Supplementary Table S1.

Transcription factor binding site analysis

Transcription factor binding site (TFBS) motifs were obtained from the JASPAR2018 database (31). Promoters were defined as 250 bp upstream and downstream from the ENSEMBL annotated transcription start site. Analysis of motif enrichment (AME) was used to search for enriched motifs in these regions using all expressed genes not in a specific cluster as background and a hit-lo-fraction of 0.5. P values reported by AME were corrected for multiple testing using FDR (32). Motifs with an FDR < 10% were defined as significantly enriched. A complete list of the motif enrichments for the Wnt-repressed clusters C2, C4, C6, and C8 are reported in Supplementary Table S2. Analysis of transcription factor binding events was performed using CHEA3 (33).

Data analysis

Data were analyzed using Prism v5.0 (GraphPad) and R/Bioconductor. Significance for all tests was set at P ≤ 0.05 unless otherwise stated. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001 in all instances.

Wnt signaling represses as many genes as it activates

We have previously reported that inhibition of Wnt signaling in HPAF-II orthotopic xenografts of human pancreatic adenocarcinoma cells with an inactivating RNF43 mutation led to changes in the expression of 11,673 genes over time (FDR < 10%; ref. 6). These differentially expressed genes fell into 64 clusters based on similar temporal dynamics of their transcriptional response to PORCN inhibition (Supplementary Fig. S1A; ref. 6). Analysis of these clusters led us to identify two global patterns of transcriptional response (termed as superclusters), with one supercluster comprised of genes consistently downregulated (Wnt-activated genes) and the other containing genes that were upregulated (Wnt-repressed genes) following inhibition of Wnt secretion (Supplementary Fig. S1B).

While Wnt signaling is mainly associated with activation of gene expression via Wnt/β-catenin pathway, there are no systematic studies examining the repression of signaling by Wnts. We focused our attention here on the genes that were upregulated upon Wnt inhibition. Notably, 28% of all expressed genes in this Wnt-driven HPAF-II pancreatic cancer were repressed directly or indirectly by high Wnt signaling. The supercluster of these Wnt-repressed genes contained 4,350 genes distributed across 17 clusters, with each cluster consisting of genes with a similar temporal response to ETC-159 (Supplementary Fig. S1B; ref. 6). We focused specifically on four clusters of Wnt-repressed genes (C2, C4, C6, C8) as these were consistently upregulated upon Wnt inhibition (Fig. 1A). Genes with representative time courses of upregulation from these four clusters after Wnt inhibition are shown in Fig. 1B.

Figure 1.

Wnt signaling represses as many genes as it activates. A, Schematic of the experimental design. HPAF-II cells were injected into the tail of mouse pancreas. Tumors were established over a period of 28 days and then the mice were treated twice a day with 37.5 mg/kg/dose of ETC-159. RNA was isolated from the tumors at the indicated time points. Four clusters of Wnt-repressed target genes were consistently upregulated following PORCN inhibition in HPAF-II orthotopic tumors. B, Representative Wnt-repressed genes from each of the four clusters are shown. TPM, transcripts per million reads. C, Treatment of a Wnt-addicted colorectal cancer PDX model with ETC-159 identified a large set of upregulated genes, many of which were also upregulated in the clusters of interest (C2, C4, C6, and C8) as identified in HPAF-II orthotopic model. The proportion of genes that were significantly upregulated (FDR < 10%) in colorectal cancer PDX at 56 hours and expressed in both models for each of the clusters is indicated. D, GO Biological Process and Reactome enrichments of Wnt-repressed genes identified in HPAF-II orthotopic tumors (C2, C4, C6, and C8), colorectal cancer patient-derived xenograft (CRC PDX), and AsPC-1 orthotopic tumors highlight processes including cell morphogenesis and cell adhesion (hypergeometric test, FDR < 10%).

Figure 1.

Wnt signaling represses as many genes as it activates. A, Schematic of the experimental design. HPAF-II cells were injected into the tail of mouse pancreas. Tumors were established over a period of 28 days and then the mice were treated twice a day with 37.5 mg/kg/dose of ETC-159. RNA was isolated from the tumors at the indicated time points. Four clusters of Wnt-repressed target genes were consistently upregulated following PORCN inhibition in HPAF-II orthotopic tumors. B, Representative Wnt-repressed genes from each of the four clusters are shown. TPM, transcripts per million reads. C, Treatment of a Wnt-addicted colorectal cancer PDX model with ETC-159 identified a large set of upregulated genes, many of which were also upregulated in the clusters of interest (C2, C4, C6, and C8) as identified in HPAF-II orthotopic model. The proportion of genes that were significantly upregulated (FDR < 10%) in colorectal cancer PDX at 56 hours and expressed in both models for each of the clusters is indicated. D, GO Biological Process and Reactome enrichments of Wnt-repressed genes identified in HPAF-II orthotopic tumors (C2, C4, C6, and C8), colorectal cancer patient-derived xenograft (CRC PDX), and AsPC-1 orthotopic tumors highlight processes including cell morphogenesis and cell adhesion (hypergeometric test, FDR < 10%).

Close modal

To test whether Wnt signaling repressed similar genes in other models, we compared this set of genes with the genes we previously identified as being upregulated in an ETC-159–treated colorectal cancer PDX driven by an RSPO3 translocation (Fig. 1C; ref. 17). Notably, comparing tumors with different tissue origins (pancreatic vs. colorectal) and with different driver mutations (RNF43 loss of function vs. RSPO3 translocation) identified that 46% (1,985/4,350) of the upregulated genes following 56 hours of treatment with ETC-159 (FDR < 10%) were shared between the two models (Supplementary Fig. S1C). Similar gene expression changes upon Wnt inhibition were also observed in another preclinical model of pancreatic cancer, AsPC-1 orthotopic xenografts (Supplementary Fig. S1D). This comparison across three Wnt-dependent models with distinct driver mutations highlights the consistency of the transcriptional response and confirms that Wnt signaling represses a large set of genes.

Next, we investigated whether specific pathways and processes were enriched in these temporal clusters of genes that are upregulated following Wnt inhibition. This analysis revealed that these clusters were enriched for cell morphogenesis and adhesion pathways, suggesting widespread changes in the signaling landscape of these tumors following ETC-159 treatment (Fig. 1D; Supplementary Table S1). Similar enrichments were observed in the upregulated genes in both the colorectal cancer PDX and AsPC-1 cancer models (Fig. 1D). This is concordant with previous observations that Wnt inhibition leads to a dramatic remodeling of tumor architecture, and that tumors treated with distinct Wnt inhibitors (PORCN inhibitors as well as anti-Frizzled and anti-RSPO3 antibodies) appear more differentiated following Wnt inhibition (6, 11, 17, 19, 34). Taken together, our systematic analysis of three independent in vivo models of Wnt-addicted cancers identified a large shared subset of genes, including regulators of differentiation, that are repressed by Wnt signaling.

Inhibition of Wnt signaling activates the MAPK signaling pathway

To investigate how Wnt signaling might repress the identified subset of genes, we performed a TFBS motif analysis. This revealed that the promoters of the genes consistently upregulated upon Wnt inhibition were significantly enriched for motifs bound by members of the ETS family (e.g., ELF3, ELF4, ELF5, and ETV6) and AP1 (Jun/Fos) related transcription factors (FDR < 10%; Fig. 2A; Supplementary Table S2). A similar enrichment for AP1 (Jun/Fos) related transcription factors was also observed in the genes upregulated upon Wnt inhibition in both the AsPC-1 pancreatic cancer xenografts as well as colorectal cancer PDX models (Fig. 2A). ETS and AP1 family members are downstream effectors of the MAPK-ERK and JNK (35–37). These results suggested that Wnt inhibition activated MAPK signaling in the tumors, and that MAPK and Wnt signaling interact to regulate the expression of a major subset of the Wnt-regulated genes.

Figure 2.

Inhibition of Wnt signaling activates the MAPK signaling pathway. A, The clusters of consistently upregulated Wnt-repressed genes were enriched for distinct sets of TFBSs including AP1 (c-Fos and Jun) and ETS family members (ELF3, ELF4, ELF5, and ETV6; FDR < 10%) in all three models of Wnt-addicted cancers, HPAF-II, colorectal cancer PDX (CRC PDX), and AsPC-1 xenografts. B and C, Wnt inhibition activates p-ERK (B) and p-JNK (C) in the HPAF-II tumors. Protein lysates from tumors from control and treated mice were resolved on SDS-PAGE and probed for p-ERK and p-RSK90. Each lane represents tumor lysate from an individual mouse. Actin or GAPDH were used as controls. D, Wnt inhibition leads to sustained increase in p-ERK levels in HPAF-II tumors. Representative image of HPAF-II tumor sections from mice treated with ETC-159 for 21 days and stained for p-ERK. E, MAPK signaling is repressed by canonical Wnt signaling. Subcutaneous xenografts from HPAF-II cells expressing stabilized β-catenin were generated. Tumor lysates from the control mice or mice treated with ETC-159 for 8 or 56 hours were resolved on SDS-PAGE and probed for p-ERK. Each lane represents tumor lysate from an individual mouse. Stabilized β-catenin prevents the ETC-159–induced increase in p-ERK. F, Wnt inhibition induces senescence, which is prevented by stabilized β-catenin. Representative images of tumor sections from the four treatment groups (treated for 7 days) stained for SA-β-galactosidase, a senescence marker, and counterstained with nuclear fast red. Blue, positive staining for senescent cells. G, The entire tumor sections were scanned and the percentage of SA-β-galactosidase positively stained area (blue) in a section from each of the groups is shown. Each dot represents quantitation of the tumor section from an individual mouse. n = 5–7 mice/group. P values calculated using Mann–Whitney U test are shown. H, Wnt inhibition regulates expression of senescence-associated genes. Expression of senescence-associated genes was analyzed in the tumors from all four treatment groups (treated for 56 hours). Each data point represents an individual tumor. **, P ≤ 0.01; ns, nonsignificant.

Figure 2.

Inhibition of Wnt signaling activates the MAPK signaling pathway. A, The clusters of consistently upregulated Wnt-repressed genes were enriched for distinct sets of TFBSs including AP1 (c-Fos and Jun) and ETS family members (ELF3, ELF4, ELF5, and ETV6; FDR < 10%) in all three models of Wnt-addicted cancers, HPAF-II, colorectal cancer PDX (CRC PDX), and AsPC-1 xenografts. B and C, Wnt inhibition activates p-ERK (B) and p-JNK (C) in the HPAF-II tumors. Protein lysates from tumors from control and treated mice were resolved on SDS-PAGE and probed for p-ERK and p-RSK90. Each lane represents tumor lysate from an individual mouse. Actin or GAPDH were used as controls. D, Wnt inhibition leads to sustained increase in p-ERK levels in HPAF-II tumors. Representative image of HPAF-II tumor sections from mice treated with ETC-159 for 21 days and stained for p-ERK. E, MAPK signaling is repressed by canonical Wnt signaling. Subcutaneous xenografts from HPAF-II cells expressing stabilized β-catenin were generated. Tumor lysates from the control mice or mice treated with ETC-159 for 8 or 56 hours were resolved on SDS-PAGE and probed for p-ERK. Each lane represents tumor lysate from an individual mouse. Stabilized β-catenin prevents the ETC-159–induced increase in p-ERK. F, Wnt inhibition induces senescence, which is prevented by stabilized β-catenin. Representative images of tumor sections from the four treatment groups (treated for 7 days) stained for SA-β-galactosidase, a senescence marker, and counterstained with nuclear fast red. Blue, positive staining for senescent cells. G, The entire tumor sections were scanned and the percentage of SA-β-galactosidase positively stained area (blue) in a section from each of the groups is shown. Each dot represents quantitation of the tumor section from an individual mouse. n = 5–7 mice/group. P values calculated using Mann–Whitney U test are shown. H, Wnt inhibition regulates expression of senescence-associated genes. Expression of senescence-associated genes was analyzed in the tumors from all four treatment groups (treated for 56 hours). Each data point represents an individual tumor. **, P ≤ 0.01; ns, nonsignificant.

Close modal

We explored this hypothesis by analyzing the MAPK pathway in HPAF-II tumors. Activating KRAS mutations occur in approximately 95% of pancreatic cancers including RNF43-mutant pancreatic cancers (38). At baseline, the Wnt-addicted HPAF-II pancreatic tumors only have a moderate amount of p-ERK despite the presence of mutant RAS. However, following Wnt inhibition upon ETC-159 treatment, we observed a marked increase in phospho-ERK1/2 (p-ERK) and its downstream substrate p-90RSK (Fig. 2B). We also observed an increase in phospho-JNK (p-JNK) beginning at 8 hours, before the increased expression of Wnt-repressed genes, with a further increase at 56 hours (Fig. 2C). This activation of p-ERK was sustained and prolonged, as it was also detected after 21 days of treatment with ETC-159 (Fig. 2D). p38 MAPK was not activated by PORCN inhibition (39). These results demonstrate that high Wnt signaling suppresses MAPK signaling pathways in vivo, even in the setting of mutant RAS, and that inhibition of Wnt signaling leads to enhanced activation of MAPK pathways.

Wnts can signal via both β-catenin–dependent and –independent pathways. To test whether the effect of PORCN inhibition on p-ERK was via degradation of β-catenin rather than an off-target or β-catenin–independent effect, we generated HPAF-II cells that expressed a stabilized β-catenin where the CK1α/GSK3 phosphorylation sites (Ser33, Ser37, Thr41, and Ser45) were mutated to Ala, thereby rendering β-catenin insensitive to proteasomal degradation upon Wnt inhibition. In these cells, β-catenin–dependent pathways are no longer affected by ETC-159 treatment. We observed that, unlike the parental HPAF-II tumors, ETC-159 treatment of HPAF-II tumors expressing stabilized β-catenin did not lead to enhanced MAPK activation (Fig. 2E). Taken together, these data show that MAPK activation is downstream of β-catenin and requires its degradation.

While activated MAPK can drive proliferation, hyperactivation of the RAS/MAPK can cause senescence (40). Indeed, the activity of the senescence marker SA-β-galactosidase was markedly increased in HPAF-II xenografts following Wnt inhibition (Fig. 2F and G). Consistent with an increase in senescent cells, the expression of LMNB1 was also markedly decreased while CDKN2B was increased (Fig. 2H). Confirming that the enhanced senescence required inhibition of the Wnt/β-catenin pathway, these changes in senescence markers were absent or blunted in ETC-159–treated HPAF-II xenografts expressing stabilized β-catenin (Fig. 2FH). Taken together, these findings support the model that Wnt inhibition drives senescence via hyperactivation of MAPK signaling.

Wnt inhibition activates MAPK in multiple models including MMTV-WNT1 mammary tumors

We next examined the AsPC-1 pancreatic cancer orthotopic xenografts and observed that, similar to the HPAF-II orthotopic tumors, there was a sustained increase in both p-ERK and p-JNK levels in the ETC-159–treated AsPC-1 tumors (Fig. 3A and B). To investigate whether Wnts suppress MAPK in other cancers, we also examined the original model for Wnt-driven cancers, mouse mammary tumor virus (MMTV)-WNT1 mammary tumors (41). In the mice carrying a MMTV LTR-Wnt1 transgene, overexpression of murine WNT1 causes a high incidence of mammary adenocarcinomas (42). Tumors arising in these mice remain Wnt dependent. We have previously reported that a second, structurally distinct (but similarly named) PORCN inhibitor, WNT-C59, is efficacious in arresting or reducing the growth of these tumors (20). We analyzed protein extracts from control or WNT-C59–treated orthotopic MMTV-WNT1 allograft tumors. As seen in Fig. 3C and D, the levels of p-ERK in the tumors from mice treated with the WNT-C59 were also increased compared with tumors from the control mice. Thus, Wnt pathway inhibition by two distinct drugs (ETC-159 and WNT-C59) increased MAPK activity in multiple in vivo models of both pancreatic and mammary cancer, demonstrating that repression of MAPK signaling is a general consequence of high Wnt signaling in cancer and not limited to cancers with mutant KRAS.

Figure 3.

Wnt inhibition activates the MAPK signaling pathway in multiple models. A and B, Wnt inhibition with ETC-159 activates p-ERK and p-JNK in the AsPC-1 xenografts. Protein lysates from tumors from control and treated mice were resolved on SDS-PAGE and probed for p-ERK or p-JNK. Each lane represents tumor lysate from an individual mouse. GAPDH were used as control. C, Wnt inhibition with WNT-C59 activates p-ERK in the MMTV-Wnt1 breast tumors. Protein lysates from tumors were analyzed as in A. Each lane represents tumor lysate from an individual mouse. D, Wnt inhibition leads to sustained increase in p-ERK levels in WNT-C59–treated MMTV-Wnt1 tumors. Representative image of MMTV-Wnt1 tumor sections from mice treated with WNT-C59 for 21 days and stained for p-ERK. E, Wnt inhibition with TNKS656 activates p-ERK in the APC-mutant PDXs. Protein lysates from control and treated mice from two independent PDXs (i) PDX1 (ii) PDX2 were analyzed as in A. Each lane represents tumor lysate from an individual mouse. GAPDH was used as control. F, Activation of Wnt signaling inhibits AP1 and c-Fos reporter activity. HEK293 cells were cotransfected with either AP1 or Fos reporters or indicated amounts of WNT1 expression plasmids and the luciferase reporter activity was measured. G, Knockout of Rnf43 and Znrf3 in pancreas leads to an increase in the expression of Axin2 as a measure of Wnt activation. H, Wnt activation leads to reduction in p-ERK levels in pancreas of mice. Representative image of pancreas from WT and Ptf1aCreRnf43fl/flZnrf3fl/fl mice stained for p-ERK.

Figure 3.

Wnt inhibition activates the MAPK signaling pathway in multiple models. A and B, Wnt inhibition with ETC-159 activates p-ERK and p-JNK in the AsPC-1 xenografts. Protein lysates from tumors from control and treated mice were resolved on SDS-PAGE and probed for p-ERK or p-JNK. Each lane represents tumor lysate from an individual mouse. GAPDH were used as control. C, Wnt inhibition with WNT-C59 activates p-ERK in the MMTV-Wnt1 breast tumors. Protein lysates from tumors were analyzed as in A. Each lane represents tumor lysate from an individual mouse. D, Wnt inhibition leads to sustained increase in p-ERK levels in WNT-C59–treated MMTV-Wnt1 tumors. Representative image of MMTV-Wnt1 tumor sections from mice treated with WNT-C59 for 21 days and stained for p-ERK. E, Wnt inhibition with TNKS656 activates p-ERK in the APC-mutant PDXs. Protein lysates from control and treated mice from two independent PDXs (i) PDX1 (ii) PDX2 were analyzed as in A. Each lane represents tumor lysate from an individual mouse. GAPDH was used as control. F, Activation of Wnt signaling inhibits AP1 and c-Fos reporter activity. HEK293 cells were cotransfected with either AP1 or Fos reporters or indicated amounts of WNT1 expression plasmids and the luciferase reporter activity was measured. G, Knockout of Rnf43 and Znrf3 in pancreas leads to an increase in the expression of Axin2 as a measure of Wnt activation. H, Wnt activation leads to reduction in p-ERK levels in pancreas of mice. Representative image of pancreas from WT and Ptf1aCreRnf43fl/flZnrf3fl/fl mice stained for p-ERK.

Close modal

The most common cause of pathologic Wnt pathway activation is mutation of APC in colorectal cancer. In this setting, tankyrase inhibitors that stabilize AXIN can inhibit β-catenin signaling (43). To ask whether inhibition of the Wnt/β-catenin pathway in APC-mutant colorectal cancer also activated MAPK and MAPK-regulated genes, we examined the effect of the tankyrase inhibitor TNKS656 on p-ERK in APC-mutant colorectal cancers (44). While tankyrase inhibitors have a smaller tumor growth inhibition in this setting, treatment of mice with TNKS656 caused an increase in p-ERK in two distinct PDXs, PDX1 and PDX2 (Fig. 3E). Consistent with this, TFBS analysis of genes upregulated after TNKS656 treatment in both the colorectal cancer PDX models showed significant enrichment for MAPK-regulated transcription factors including CEBPB and AP1 (Supplementary Fig. S2A; ref. 45). Thus, the inhibition of MAPK and activation of downstream gene expression by hyperactive Wnt/β-catenin signaling appears widespread in cancer.

We also examined whether, conversely, increased Wnt signaling could inhibit the expression of MAPK-regulated AP1 and c-Fos transcriptional reporters in vitro. Indeed, transient overexpression of WNT1 in HEK293 cells led to a reduction of both AP1 and Fos reporter activity (Fig. 3F).

As a further test of the impact of high Wnt signaling on the MAPK pathway, we utilized an additional genetically engineered mouse model of tissue-specific high Wnt signaling. Wnt signaling was activated by deletion of both Rnf43 and its paralog Znrf3 in the mouse pancreas (46) using the pancreas specific Cre driver Ptf1a-Cre. This led to constitutive enhanced Wnt signaling, as shown by increased expression of the Wnt target gene Axin2 in the Ptf1aCreRnf43fl/flZnrf3fl/fl mice compared with WT mice (Fig. 3G). The mice developed benign hyperplasia of the pancreas. Consistent with our data in pancreatic cancers, we observed that while the pancreas of WT mice showed p-ERK staining in the pancreatic lobes, the increase in Wnt signaling in Ptf1aCreRnf43fl/flZnrf3fl/fl mice pancreas led to the abrogation of p-ERK staining (Fig. 3H).

Taken together, our data from multiple cancer and noncancer models clearly demonstrate that Wnt signaling can repress MAPK signaling pathways both in vitro and in vivo.

Inhibition of MEK–ERK pathway reverses the expression of genes upregulated upon Wnt inhibition

Our findings demonstrate that Wnt inhibition leads to both (i) MAPK activation, and (ii) derepression of a large set of genes with MAPK-responsive TFBSs, suggesting that Wnt signaling is repressing MAPK-regulated genes. One prediction of this hypothesis is that inhibition of MAPK signaling should block the upregulation of some or all of the genes that are otherwise upregulated upon Wnt inhibition. We therefore treated HPAF-II pancreatic xenografts with ETC-159, the MEK inhibitor trametinib or a combination of ETC-159 and trametinib. Treatment of HPAF-II pancreatic cancer xenografts even with a low dose of ETC-159 (10 mg/kg) induced MAPK activation as shown by increased p-ERK and p-JNK, and cotreatment with ETC-159 and the MEK inhibitor trametinib (0.1 mg/kg) effectively blocked this increase (Fig. 4A and B).

Figure 4.

Inhibition of the MEK–ERK pathway reverses the Wnt signaling-mediated regulation of cellular differentiation genes. A and B, Trametinib treatment represses MAPK signaling in HPAF-II xenografts following Wnt inhibition. Representative tumor lysates from subcutaneous xenografts from control mice; mice treated with ETC-159, trametinib, or a combination of ETC-159 and trametinib were analyzed for p-ERK and p-JNK1/2. Each lane represents tumor lysate from an individual mouse. GAPDH was used as a loading control. C, MEK inhibition with trametinib blocks the upregulation of ETC-159–induced 918 Wnt-repressed genes. D, HPAF-II xenografts from four treatment groups (vehicle, ETC-159, trametinib, ETC-159, and trametinib) were harvested and gene expression analyzed by RNA-seq (Supplementary Table S3). Wnt-repressed genes including those whose expression was blunted upon trametinib treatment are shown. E, KEGG pathway enrichments of the 918 upregulated genes whose expression was blunted upon cotreatment with trametinib in tumors highlight MAPK signaling, adherens junction, actin cytoskeleton, and focal adhesion pathways (FDR < 10%). F, Promoters of 918 genes whose expression was blunted with concurrent trametinib treatment were scanned for TFBS motifs using ENCODE database using the web-based tool CHEA3. As expected, the Wnt-repressed trametinib-inhibited genes show TFBS enrichment for AP1 (Fos and Jun) and ETS family members.

Figure 4.

Inhibition of the MEK–ERK pathway reverses the Wnt signaling-mediated regulation of cellular differentiation genes. A and B, Trametinib treatment represses MAPK signaling in HPAF-II xenografts following Wnt inhibition. Representative tumor lysates from subcutaneous xenografts from control mice; mice treated with ETC-159, trametinib, or a combination of ETC-159 and trametinib were analyzed for p-ERK and p-JNK1/2. Each lane represents tumor lysate from an individual mouse. GAPDH was used as a loading control. C, MEK inhibition with trametinib blocks the upregulation of ETC-159–induced 918 Wnt-repressed genes. D, HPAF-II xenografts from four treatment groups (vehicle, ETC-159, trametinib, ETC-159, and trametinib) were harvested and gene expression analyzed by RNA-seq (Supplementary Table S3). Wnt-repressed genes including those whose expression was blunted upon trametinib treatment are shown. E, KEGG pathway enrichments of the 918 upregulated genes whose expression was blunted upon cotreatment with trametinib in tumors highlight MAPK signaling, adherens junction, actin cytoskeleton, and focal adhesion pathways (FDR < 10%). F, Promoters of 918 genes whose expression was blunted with concurrent trametinib treatment were scanned for TFBS motifs using ENCODE database using the web-based tool CHEA3. As expected, the Wnt-repressed trametinib-inhibited genes show TFBS enrichment for AP1 (Fos and Jun) and ETS family members.

Close modal

We next analyzed gene expression changes in HPAF-II tumors from four treatment groups (Vehicle, ETC-159, trametinib, or a combination of ETC-159 and trametinib) using RNA-seq. Similar to our previous studies, the number of genes that were upregulated in the tumors treated with a low dose of ETC-159 (2,683 genes) was roughly similar to the number of genes that are downregulated (3,014 genes; Fig. 4C; Supplementary Fig. S2B). Confirming our prior GO/Reactome pathway analysis (Fig. 1D), the upregulated genes were also significantly enriched for MAPK signaling and cellular differentiation (Supplementary Fig. S2C).

Importantly, the addition of trametinib blunted the upregulation of 34% (918/2,683) of these genes that are activated upon Wnt inhibition (Fig. 4C). For example, ETC-159 treatment increased the expression of multiple integrins and keratins, genes involved in cellular differentiation and reprogramming (Fig. 4D). Cotreatment with trametinib blunted the effect of Wnt inhibition on these genes (Supplementary Table S3; Fig. 4D). Further supporting this, pathway analysis indicated that genes whose expression is blunted with MEK inhibition (918 genes) represent actin cytoskeleton, focal adhesion, and leukocyte migration pathways (Fig. 4E).

Repeating the TFBS analysis, we observed that even at this lower dose of ETC-159, the 2,683 genes upregulated upon Wnt inhibition were enriched for ETS family and AP1 family transcription factors (Supplementary Fig. S2D). The 918 genes whose expression was blunted upon MEK inhibition were also highly enriched for ERK-regulated transcription factors including AP1 and ETS (Fig. 4F). Conversely, the remaining 1,765 Wnt-repressed genes whose expression was not altered by trametinib no longer had significant enrichment for binding sites of ETS family and AP1 family transcription factors nor other transcription factors in the ENCODE database. This analysis provides independent confirmation that a large subset of Wnt-repressed genes is regulated via ERK activation that contributes to cellular reprogramming. Thus, MAPK activation induced by Wnt inhibition contributes to cellular differentiation and reprogramming.

Dual Wnt and MAPK pathway inhibition slows the growth of Wnt-driven pancreatic cancers and leads to enhanced cell-cycle arrest

Our previous study in HPAF-II xenografts using a CRISPR screen demonstrated that in RNF43-mutant pancreatic cancers driven by hyperactive Wnt signaling KRAS, MAP2K1, and RAF1 remain essential genes (39). To examine this, we studied the combined effect of Wnt and MAPK inhibition on tumor growth in the HPAF-II xenograft mouse model. The combination of low doses of ETC-159 (10 mg/kg) and trametinib (0.1 mg/kg) reduced tumor growth after 21 days of treatment more effectively than either drug alone (Fig. 5A). Both ETC-159 and trametinib alone or in combination were well tolerated as the mice did not lose body weight (Supplementary Fig. S2E).

Figure 5.

Combined inhibition of Wnt and MAPK signaling reduces tumor growth. A, Trametinib and ETC-159 combination additively prevents growth of HPAF-II xenografts. NSG mice with established HPAF-II subcutaneous xenografts were randomized into four groups. Mice were gavaged daily with ETC-159 (10 mg/kg), trametinib (0.1 mg/kg), or a combination of the two for 21 days. Treatment was initiated after HPAF-II tumors were established. Terminal weights of the tumors from all mice are plotted. B, MEK inhibition has a minimal effect on the expression of Wnt-activated genes. Log2-fold change in the expression of representative Wnt-target genes from each of the treatment groups is shown. C, KEGG pathway enrichments of downregulated genes highlight processes including cell cycle, ribosome biogenesis, and DNA replication in the tumors treated with Wnt inhibitor; MAPK and Hippo signaling in the tumors treated with trametinib and proteosomal degradation pathway in the combination group (FDR < 10%). D, Wnt inhibition but not MAPK inhibition leads to reduced expression of cell-cycle genes. Change in the expression of representative cell-cycle genes from each of the treatment groups is shown. E, Combined treatment with trametinib and ETC-159 led to a significant decrease in the expression of multiple genes regulating proteasomal degradation pathway compared with treatment with either of the two compounds alone. F, ETC-159 and trametinib combined treatment increases stability of cell-cycle regulator p27 in HPAF-II xenografts. Tumor lysates from subcutaneous xenografts from control mice or mice treated with ETC-159, trametinib, or a combination of ETC-159 and trametinib were resolved on SDS gel and probed for p27. Each lane represents tumor lysate from an individual mouse.

Figure 5.

Combined inhibition of Wnt and MAPK signaling reduces tumor growth. A, Trametinib and ETC-159 combination additively prevents growth of HPAF-II xenografts. NSG mice with established HPAF-II subcutaneous xenografts were randomized into four groups. Mice were gavaged daily with ETC-159 (10 mg/kg), trametinib (0.1 mg/kg), or a combination of the two for 21 days. Treatment was initiated after HPAF-II tumors were established. Terminal weights of the tumors from all mice are plotted. B, MEK inhibition has a minimal effect on the expression of Wnt-activated genes. Log2-fold change in the expression of representative Wnt-target genes from each of the treatment groups is shown. C, KEGG pathway enrichments of downregulated genes highlight processes including cell cycle, ribosome biogenesis, and DNA replication in the tumors treated with Wnt inhibitor; MAPK and Hippo signaling in the tumors treated with trametinib and proteosomal degradation pathway in the combination group (FDR < 10%). D, Wnt inhibition but not MAPK inhibition leads to reduced expression of cell-cycle genes. Change in the expression of representative cell-cycle genes from each of the treatment groups is shown. E, Combined treatment with trametinib and ETC-159 led to a significant decrease in the expression of multiple genes regulating proteasomal degradation pathway compared with treatment with either of the two compounds alone. F, ETC-159 and trametinib combined treatment increases stability of cell-cycle regulator p27 in HPAF-II xenografts. Tumor lysates from subcutaneous xenografts from control mice or mice treated with ETC-159, trametinib, or a combination of ETC-159 and trametinib were resolved on SDS gel and probed for p27. Each lane represents tumor lysate from an individual mouse.

Close modal

We considered how the combined effects of Wnt and MEK inhibition might contribute to the additive therapeutic effects of the drugs. We examined the effect of MEK inhibition on genes that are downregulated. The lower dose ETC-159 treatment alone decreased the expression of 3,014 genes (fold change >1.5 and FDR < 5%), including well-established Wnt target genes such as AXIN2, NKD1, and NOTUM, while trametinib did not affect the expression of these genes (Fig. 5B; Supplementary Fig. S2B; Supplementary Table S3).

Somewhat surprisingly for a RAS-mutant cancer, single-agent therapy using trametinib at 0.1 mg/kg reduced the expression of far fewer (∼390) genes than did Wnt inhibition alone (3,014 genes; Supplementary Table S3; Supplementary Fig. S2B). Similar to our published data, the expression of cell-cycle regulators including cyclins, CDKs, and CDKN2A genes were altered in ETC-159–treated tumors (6). Trametinib alone did not affect cell-cycle gene expression, and no additional changes were observed in the combination group (Fig. 5D).

Interestingly, expression of 467 genes was not altered by ETC-159 or trametinib alone but was downregulated significantly only in the drug combination group (Supplementary Fig. S2B). In this group, KEGG pathway analysis demonstrated significant enrichment for the proteasome pathway (Fig. 5C). Consistent with this, we observed marked downregulation of the expression of multiple proteasome subunit genes in tumors treated with the drug combination (Fig. 5E). Proteasome inhibition in other settings induces G2–M cell-cycle arrest by stabilizing protein levels of cell-cycle regulators such as p21 and p27. Supporting the hypothesis that increased protein stability could contribute to cell-cycle arrest, we observed an increase in the abundance of CDKN1B/p27 protein in tumors from mice treated with a combination of trametinib and ETC-159 but not from the mice treated with single agents (Fig. 5F). The increased protein abundance of CDKN1B/p27 may have the effect of reinforcing the cell-cycle arrest seen with ETC-159 alone and explain in part the additive effect of inhibiting both pathways. These results suggest that combining ETC-159 with a MEK inhibitor could be an effective and safe therapeutic strategy for treating Wnt-driven pancreatic cancers.

Wnt signaling is important in both normal physiology and in cancer, and the mechanisms by which it activates gene expression are well understood. However, the fact that Wnt/β-catenin signaling can repress the expression of many genes is not widely appreciated, and the pathways by which it represses gene expression are not well understood. Here, by taking advantage of potent inhibitors of Wnt secretion, we carried out a comprehensive and time-resolved analysis of the transcriptional response to PORCN inhibition in Wnt-addicted cancer models. We identified an unexpectedly large set of genes that were repressed by Wnt signaling. Many of these Wnt-repressed genes contained ETS and AP1 transcription factor binding sites. This led to the recognition that active Wnt/β-catenin signaling represses the expression of thousands of genes in these cancers by inhibition of the mitogen-activated protein kinases, ERK and JNK. Consistent with this model, many of the genes that were turned on by Wnt inhibitors were then turned off by the MEK inhibitor trametinib. Wnt/β-catenin-driven repression of these MAPK pathways in cancer appears to be widespread, as we found evidence for it occurring in multiple Wnt-driven cancer models including RNF43-mutant pancreatic cancers, colorectal cancers with RSPO3 translocations or APC mutations, and an MMTV-WNT1 mouse mammary carcinoma. This was further confirmed in a genetically engineered mouse model where activation of Wnt signaling suppressed MAPK activation.

We and others have reported that Wnt inhibition drives differentiation of Wnt-addicted cancers (6, 11, 17, 34). Further supporting the broad impact of this Wnt-mediated repression of ERK and JNK pathway, the MAPK-regulated Wnt-repressed genes regulate actin cytoskeleton and focal adhesion. This establishes a broad mechanism how Wnt/β-catenin represses a large class of genes involved in cellular differentiation and reactivation of ERK/MAPK directional signaling following Wnt inhibition could be one potential mechanism contributing to the differentiation of tumors. The mechanism by which Wnt signaling represses MAPK is not yet understood and is the subject of ongoing investigation. Analysis of Wnt-regulated genes and our previously reported study with PORCN knockdown showed that Wnt inhibition with ETC-159 did not induce an endoplasmic reticulum stress response (47). We also investigated the possibility that signals from mouse stroma might contribute to the activation of MAPK. Indeed, after PORCN inhibition, mouse genes, presumably from tumor stroma, were modestly enriched for inflammation and associated processes, with a small number of growth factors (e.g., Ngf, Igf1, and Hgf) showing significant differences. Whether this contributes to, or is a reaction to, the induction of MAPK and senescence will require additional investigation.

Our study indicates a widespread repression of MAPK signaling by pathologically increased Wnt/β-catenin signaling in cancer. There is currently a limited amount of data that suggests this pathway is also important in normal physiology. In wound healing, WNT7A has been reported to inhibit ERK phosphorylation, thus blocking MMP9 expression (48). In the normal mouse intestine, inhibition of normal Wnt signaling transiently activated ERK and the proliferation of stem cells in intestinal crypts (49). Further studies are needed to assess the role of this pathway in development.

What is the selective advantage to cancers in repressing MAPK signaling via activation of Wnt/β-catenin signaling? Many of the cancers we studied have mutant KRAS that leads to constitutive activation of RAS/MEK/ERK signaling. We observed that Wnt-addicted KRAS-mutant pancreatic cancers had relatively modest levels of activated ERK. It is well established that isolated mutant RAS is not only poorly oncogenic, but also leads to cellular senescence (40). Consistent with this, recent studies have clarified that RAS signaling in cancer follows a Goldilocks model and has a “sweet spot” (50); either too little or too much RAS activity is not oncogenic. In this context, our data suggest that enhanced Wnt/β-catenin signaling is one mechanism that modulates excessive downstream MAPK signaling and allows RAS-mutant cancers to proliferate. Indeed, this is consistent with several mouse genetic studies showing that Wnt activation cooperates with RAS mutations in transformation. For example, in experimental models of lung carcinogenesis, Morrisey and coworkers found that stabilization of β-catenin markedly accelerated KRAS(G12D)-induced tumorigenesis (51). Similarly, Damsky and colleagues reported that β-catenin stabilization accelerated tumorigenesis and metastasis in Pten/Braf mutant melanomas (52).

Our study provides the first evidence for a Wnt-regulated repression of the MAPK pathway in cancer, which is required to orchestrate changes in RAS signaling with important consequences for cellular organization and reprogramming. Future studies dissecting the mechanism of Wnt-modulated MAPK signaling will provide further insights into RAS-mediated oncogenesis and the role of Wnt signaling in cancer.

H.G. Palmer reports grants from MERK SERONO and NOVARTIS during the conduct of the study and grants from MERK SERONO and NOVARTIS outside the submitted work. D.M. Virshup reports personal fees from Experimental Therapeutics Center Singapore outside the submitted work; in addition, D.M. Virshup has a financial interest in ETC-159. B. Madan reports grants from Singapore Ministry of Health's National Medical Research Council during the conduct of the study. B. Madan has a financial interest in the drug ETC-159 used in this study. No disclosures were reported by the other authors.

N. Harmston: Investigation, writing-review and editing. J.Y.S. Lim: Investigation. O. Arqués: Investigation. H.G. Palmer: Supervision, funding acquisition. E. Petretto: Supervision, writing-review and editing. D.M. Virshup: Conceptualization, supervision, funding acquisition, writing-review and editing. B. Madan: Conceptualization, supervision, funding acquisition, writing-original draft, writing-review and editing.

We gratefully acknowledge the assistance of members of the Petretto lab, Virshup lab, and the Experimental Drug Development Centre, A*Star, Singapore. We especially acknowledge the assistance of Hock Lee and Lowell Lin. This research is supported in part by the National Research Foundation Singapore and administered by the Singapore Ministry of Health's National Medical Research Council under the STAR Award Program NMRC/STAR/0017/2013 (to D.M. Virshup). B. Madan acknowledges the support of the Singapore Ministry of Health's National Medical Research Council Open Fund–Independent Research grant NMRC/OFIRG/0055/2017. N. Harmston is supported by MOE and Yale-NUS college and Duke-NUS Medical School start-up grants.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Niehrs
C
. 
The complex world of WNT receptor signalling
.
Nat Rev Mol Cell Biol
2012
;
13
:
767
79
.
2.
Yu
J
,
Virshup
DM
. 
Updating the Wnt pathways
.
Biosci Rep
2014
;
34
:
593
607
.
3.
Ramakrishnan
A-B
,
Sinha
A
,
Fan
VB
,
Cadigan
KM
. 
The Wnt transcriptional switch: TLE removal or inactivation?
Bioessays
2018
;
40
. DOI: 10.1002/bies.201700162.
4.
Flack
JE
,
Mieszczanek
J
,
Novcic
N
,
Bienz
M
. 
Wnt-Dependent inactivation of the Groucho/TLE co-repressor by the HECT E3 ubiquitin ligase Hyd/UBR5
.
Mol Cell
2017
;
67
:
181
93.e5
.
5.
Myant
K
,
Sansom
OJ
. 
Wnt/Myc interactions in intestinal cancer: partners in crime
.
Exp Cell Res
2011
;
317
:
2725
31
.
6.
Madan
B
,
Harmston
N
,
Nallan
G
,
Montoya
A
,
Faull
P
,
Petretto
E
, et al
Temporal dynamics of Wnt-dependent transcriptome reveal an oncogenic Wnt/MYC/ribosome axis
.
J Clin Invest
2018
;
128
:
5620
33
.
7.
Hoverter
NP
,
Waterman
ML
. 
A Wnt-fall for gene regulation: repression
.
Sci Signal
2008
;
1
:
pe43
3
.
8.
Blauwkamp
TA
,
Chang
MV
,
Cadigan
KM
. 
Novel TCF-binding sites specify transcriptional repression by Wnt signalling
.
EMBO J
2008
;
27
:
1436
46
.
9.
Grocott
T
,
Johnson
S
,
Bailey
AP
,
Streit
A
. 
Neural crest cells organize the eye via TGF-β and canonical Wnt signalling
.
Nat Commun
2011
;
2
:
265
.
10.
Gu
B
,
Watanabe
K
,
Sun
P
,
Fallahi
M
,
Dai
X
. 
Chromatin effector Pygo2 mediates Wnt-notch crosstalk to suppress luminal/alveolar potential of mammary stem and basal cells
.
Cell Stem Cell
2013
;
13
:
48
61
.
11.
Jiang
X
,
Hao
H-X
,
Growney
JD
,
Woolfenden
S
,
Bottiglio
C
,
Ng
N
, et al
Inactivating mutations of RNF43 confer Wnt dependency in pancreatic ductal adenocarcinoma
.
Proc Natl Acad Sci U S A
2013
;
110
:
12649
54
.
12.
Seshagiri
S
,
Stawiski
EW
,
Durinck
S
,
Modrusan
Z
,
Storm
EE
,
Conboy
CB
, et al
Recurrent R-spondin fusions in colon cancer
.
Nature
2012
;
488
:
660
4
.
13.
Hao
H-X
,
Jiang
X
,
Cong
F
. 
Control of Wnt receptor turnover by R-spondin-ZNRF3/RNF43 signaling module and its dysregulation in cancer
.
Cancers
2016
;
8
:
54
.
14.
Coombs
GS
,
Yu
J
,
Canning
CA
,
Veltri
CA
,
Covey
TM
,
Cheong
JK
, et al
WLS-dependent secretion of WNT3A requires Ser209 acylation and vacuolar acidification
.
J Cell Sci
2010
;
123
:
3357
67
.
15.
Janda
CY
,
Waghray
D
,
Levin
AM
,
Thomas
C
,
Garcia
KC
. 
Structural basis of Wnt recognition by frizzled
.
Science
2012
;
337
:
59
64
.
16.
Madan
B
,
Virshup
DM
. 
Targeting Wnts at the source–new mechanisms, new biomarkers, new drugs
.
Mol Can Ther
2015
;
14
:
1087
94
.
17.
Madan
B
,
Ke
Z
,
Harmston
N
,
Ho
SY
,
Frois
AO
,
Alam
J
, et al
Wnt addiction of genetically defined cancers reversed by PORCN inhibition
.
Oncogene
2016
;
35
:
2197
207
.
18.
Liu
J
,
Pan
S
,
Hsieh
MH
,
Ng
N
,
Sun
F
,
Wang
T
, et al
Targeting Wnt-driven cancer through the inhibition of Porcupine by LGK974
.
Proc Natl Acad Sci U S A
2013
;
110
:
20224
9
.
19.
Storm
EE
,
Durinck
S
,
de Sousa e Melo
F
,
Tremayne
J
,
Kljavin
N
,
Tan
C
, et al
Targeting PTPRK-RSPO3 colon tumours promotes differentiation and loss of stem-cell function
.
Nature
2016
;
529
:
97
100
.
20.
Proffitt
KD
,
Madan
B
,
Ke
Z
,
Pendharkar
V
,
Ding
L
,
Lee
MA
, et al
Pharmacological inhibition of the Wnt acyltransferase PORCN prevents growth of WNT-driven mammary cancer
.
Cancer Res
2013
;
73
:
502
7
.
21.
Chee
YC
,
Pahnke
J
,
Bunte
R
,
Adsool
VA
,
Madan
B
,
Virshup
DM
. 
Intrinsic xenobiotic resistance of the intestinal stem cell niche
.
Dev Cell
2018
;
46
:
681
5.e5
.
22.
Fuerer
C
,
Nusse
R
. 
Lentiviral vectors to probe and manipulate the Wnt signaling pathway
.
PLoS One
2010
;
5
:
e9370
.
23.
Conway
T
,
Wazny
J
,
Bromage
A
,
Tymms
M
,
Sooraj
D
,
Williams
ED
, et al
Xenome–a tool for classifying reads from xenograft samples
.
Bioinformatics
2012
;
28
:
i172
8
.
24.
Dobin
A
,
Davis
CA
,
Schlesinger
F
,
Drenkow
J
,
Zaleski
C
,
Jha
S
, et al
STAR: ultrafast universal RNA-seq aligner
.
Bioinformatics
2013
;
29
:
15
21
.
25.
Li
B
,
Dewey
CN
. 
RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome
.
BMC Bioinformatics
2011
;
12
:
323
.
26.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
27.
Hensman
J
,
Lawrence
ND
,
Rattray
M
. 
Hierarchical Bayesian modelling of gene expression time series across irregularly sampled replicates and clusters
.
BMC Bioinformatics
2013
;
14
:
252
.
28.
Falcon
S
,
Gentleman
R
. 
Using GOstats to test gene lists for GO term association
.
Bioinformatics
2007
;
23
:
257
8
.
29.
Yu
G
,
He
Q-Y
. 
ReactomePA: an R/Bioconductor package for reactome pathway analysis and visualization
.
Mol Biosyst
2016
;
12
:
477
9
.
30.
Chen
EY
,
Tan
CM
,
Kou
Y
,
Duan
Q
,
Wang
Z
,
Meirelles
GV
, et al
Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool
.
BMC Bioinformatics
2013
;
14
:
128
.
31.
Khan
A
,
Fornes
O
,
Stigliani
A
,
Gheorghe
M
,
Castro-Mondragon
JA
,
van der Lee
R
, et al
JASPAR 2018: update of the open-access database of transcription factor binding profiles and its web framework
.
Nucl Acids Res
2018
;
46
:
D260
6
.
32.
McLeay
RC
,
Bailey
TL
. 
Motif enrichment analysis: a unified framework and an evaluation on ChIP data
.
BMC Bioinformatics
2010
;
11
:
165
.
33.
Keenan
AB
,
Torre
D
,
Lachmann
A
,
Leong
AK
,
Wojciechowicz
ML
,
Utti
V
, et al
ChEA3: transcription factor enrichment analysis by orthogonal omics integration
.
Nucl Acids Res
2019
;
47
:
W212
24
.
34.
Gurney
A
,
Axelrod
F
,
Bond
CJ
,
Cain
J
,
Chartier
C
,
Donigan
L
, et al
Wnt pathway inhibition via the targeting of Frizzled receptors results in decreased growth and tumorigenicity of human tumors
.
Proc Natl Acad Sci U S A
2012
;
109
:
11717
22
.
35.
Frost
JA
,
Geppert
TD
,
Cobb
MH
,
Feramisco
JR
. 
A requirement for extracellular signal-regulated kinase (ERK) function in the activation of AP-1 by Ha-Ras, phorbol 12-myristate 13-acetate, and serum
.
Proc Natl Acad Sci U S A
1994
;
91
:
3844
8
.
36.
Treisman
R
. 
The SRF accessory protein Elk-1 contains a growth factor-regulated transcriptional activation domain
.
Cell
1993
;
73
:
381
93
.
37.
Plotnik
JP
,
Budka
JA
,
Ferris
MW
,
Hollenhorst
PC
. 
ETS1 is a genome-wide effector of RAS/ERK signaling in epithelial cells
.
Nucl Acids Res
2014
;
42
:
11928
40
.
38.
Deer
EL
,
González-Hernández
J
,
Coursen
JD
,
Shea
JE
,
Ngatia
J
,
Scaife
CL
, et al
Phenotype and genotype of pancreatic cancer cell lines
.
Pancreas
2010
;
39
:
425
35
.
39.
Zhong
Z
,
Sepramaniam
S
,
Chew
XH
,
Wood
K
,
Lee
MA
,
Madan
B
, et al
PORCN inhibition synergizes with PI3K/mTOR inhibition in Wnt-addicted cancers
.
Oncogene
2019
;
38
:
6662
77
.
40.
Serrano
M
,
Lin
AW
,
McCurrach
ME
,
Beach
D
,
Lowe
SW
. 
Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a
.
Cell
1997
;
88
:
593
602
.
41.
Nusse
R
,
Varmus
HE
. 
Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome
.
Cell
1982
;
31
:
99
109
.
42.
Tsukamoto
AS
,
Grosschedl
R
,
Guzman
RC
,
Parslow
T
,
Varmus
HE
. 
Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice
.
Cell
1988
;
55
:
619
25
.
43.
Huang
S-MA
,
Mishina
YM
,
Liu
S
,
Cheung
A
,
Stegmeier
F
,
Michaud
GA
, et al
Tankyrase inhibition stabilizes axin and antagonizes Wnt signalling
.
Nature
2009
;
461
:
614
20
.
44.
Arques
O
,
Chicote
I
,
Puig
I
,
Tenbaum
SP
,
Argiles
G
,
Dienstmann
R
, et al
Tankyrase inhibition blocks Wnt/β-catenin pathway and reverts resistance to PI3K and AKT inhibitors in the treatment of colorectal cancer
.
Clin Cancer Res
2016
;
22
:
644
56
.
45.
Nakajima
T
,
Kinoshita
S
,
Sasagawa
T
,
Sasaki
K
,
Naruto
M
,
Kishimoto
T
, et al
Phosphorylation at threonine-235 by a ras-dependent mitogen-activated protein kinase cascade is essential for transcription factor NF-IL6
.
Proc Natl Acad Sci U S A
1993
;
90
:
2207
11
.
46.
Koo
B-K
,
Spit
M
,
Jordens
I
,
Low
TY
,
Stange
DE
,
van de Wetering
M
, et al
Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors
.
Nature
2012
;
488
:
665
9
.
47.
PORCN moonlights in a Wnt-independent pathway that regulates cancer cell proliferation
.
PLoS One
2012
;
7
:
e34532
.
48.
Sundaram
GM
,
Ismail
HM
,
Bashir
M
,
Muhuri
M
,
Vaz
C
,
Nama
S
, et al
EGF hijacks miR-198/FSTL1 wound-healing switch and steers a two-pronged pathway toward metastasis
.
J Exp Med
2017
;
214
:
2889
900
.
49.
Kabiri
Z
,
Greicius
G
,
Zaribafzadeh
H
,
Hemmerich
A
,
Counter
CM
,
Virshup
DM
. 
Wnt signaling suppresses MAPK-driven proliferation of intestinal stem cells
.
J Clin Invest
2018
;
128
:
3806
12
.
50.
Li
S
,
Balmain
A
,
Counter
CM
. 
A model for RAS mutation patterns in cancers: finding the sweet spot
.
Nat Rev Cancer
2018
;
18
:
767
77
.
51.
Pacheco-Pinedo
EC
,
Durham
AC
,
Stewart
KM
,
Goss
AM
,
Lu
MM
,
Demayo
FJ
, et al
Wnt/β-catenin signaling accelerates mouse lung tumorigenesis by imposing an embryonic distal progenitor phenotype on lung epithelium
.
J Clin Invest
2011
;
121
:
1935
45
.
52.
Damsky
WE
,
Curley
DP
,
Santhanakrishnan
M
,
Rosenbaum
LE
,
Platt
JT
,
Gould Rothberg
BE
, et al
β-catenin signaling controls metastasis in Braf-activated Pten-deficient melanomas
.
Cancer Cell
2011
;
20
:
741
54
.