Research Article The Wnt signaling receptor Fzd9 is essential for Myc-driven tumorigenesis in pancreatic islets Mariano F Zacarı́as-Fluck1,* , Toni Jauset1,2,*, Sandra Martı́nez-Martı́n1 , Jastrinjan Kaur1 , Sı́lvia Casacuberta-Serra2 , Daniel Massó-Vallés1 , Erika Serrano del Pozo1, Génesis Martı́n-Fernández1, Íñigo González-Larreategui1 , Sergio López-Estévez2 , Lamorna Brown-Swigart3 , Marie-Eve Beaulieu1 , Jonathan R Whitfield1 , Babita Madan4, David M Virshup4 , Gerard I Evan5 , Laura Soucek1,2,6,7 The huge cadre of genes regulated by Myc has obstructed the induced within 4 h of acute Myc activation in vivo, its expression was identification of critical effectors that are essential for Myc-driven maintained as long as Myc activity was sustained, and rapidly de- tumorigenesis. Here, we describe how only the lack of the receptor creased following Myc deactivation and tumor regression (Lawlor Fzd9, previously identified as a Myc transcriptional target, impairs et al, 2006). In addition, Fzd9 has been identified as a direct Myc target sustained tumor expansion and β-cell dedifferentiation in a mouse gene through expression and ChIP analysis (Lawlor et al, 2006; Sabò model of Myc-driven insulinoma, allows pancreatic islets to maintain et al, 2014). their physiological structure and affects Myc-related global gene Fzd9 is up-regulated in several types of human cancers including expression. Importantly, Wnt signaling inhibition in Fzd9-competent human gastric cancer (Kirikoshi et al, 2001), osteosarcoma (Wang mice largely recapitulates the suppression of proliferation caused by et al, 2017) and astrocytoma (Zhang et al, 2006). Knockdown of Fzd9 Fzd9 deficiency upon Myc activation. Together, our results indicate has also been shown to inhibit cell proliferation and motility in that the Wnt signaling receptor Fzd9 is essential for Myc-induced hepatocellular carcinoma cell lines (Fujimoto et al, 2009). However, tumorigenesis in pancreatic islets. as Fzd9 has also shown some tumor suppressor activity in acute myeloid leukemia (Zhang et al, 2016) and non-small cell lung cancer, DOI 10.26508/lsa.201900490 | Received 18 July 2019 | Revised 22 February its role in tumorigenesis is still controversial (Winn et al, 2005, 2006). 2021 | Accepted 22 February 2021 | Published online 2 March 2021 In this study, we explored the role of Fzd9 in mediating and maintaining Myc oncogenic function in pIns-MycERTAM;RIP7-Bcl-xL β−cell tumors. We show that Fzd9 does, indeed, play a critical Introduction function in the development of Myc-driven insulinomas. We, thus, identify a novel link connecting Myc and the Wnt pathway, which Myc is a highly pleiotropic transcription factor that governs cell appears to be a required effector of Myc oncogenic activity in β-cell expansion by coordinating diverse cellular processes, including tumorigenesis. This implies the existence of a positive feedback proliferation, dedifferentiation, biosynthetic metabolism, cell growth, loop, where Wnt/β-catenin signaling activates Myc (He et al, 1998), and angiogenesis (Dang, 2012). Using the Myc-switchable pIns- and Myc in turn activates Wnt/β-catenin signaling through Fzd9. MycERTAM;RIP7-Bcl-xL β-cell tumor model in an unbiased reversible kinetic expression analysis, we previously identified genes required for maintenance of Myc-driven β-cell tumors. We found a relatively Results small subset of Myc target genes whose change in expression (whether induction or repression) was dependent upon sustained Fzd9 is required for Myc-induced β-cell neoplasia Myc activity, and that such changes were reversed by deactivation of Myc and consequent tumor regression. One of these candidate To determine whether Fzd9 plays a significant role in Myc-induced tu- tumor maintenance genes was Fzd9, a member of the “frizzled” gene morigenesis in vivo, pIns-MycERTAM;RIP7-Bcl-xL (MycER;BclXL hereinafter) family of Wnt receptors (Lawlor et al, 2006). Fzd9 was significantly mice were crossed into a Fzd9-deficient background (Ranheim et al, 1 Mouse Models of Cancer Therapy Group, Vall d’Hebron Institute of Oncology (VHIO), Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain 2Peptomyc SL, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain 3Department of Pathology and Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, San Francisco, CA, USA 4Program in Cancer and Stem Cell Biology, Duke-National University of Singapore (NUS) Medical School, Singapore, Singapore 5 Department of Biochemistry, University of Cambridge, Cambridge, UK 6Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain 7Department of Biochemistry and Molecular Biology, Universitat Autònoma de Barcelona, Bellaterra, Spain Correspondence: lsoucek@vhio.net Daniel Massó-Vallés and Marie-Eve Beaulieu’s present address is Peptomyc SL, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain *Mariano F Zacarı́as-Fluck and Toni Jauset contributed equally to this work © 2021 Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 1 of 10 2005). The size, distribution, number and histological disposition of β-cells for 3 wk and BrdU was administered systemically 3 h before pancreatic islets in Fzd9KO/KO mice appeared identical to that in Fzd9WT/WT euthanasia. Pancreata were harvested and BrdU incorporation detected mice (Fig S1). When tamoxifen was systemically administered to Fzd9WT/WT; by immunofluorescence (Fig 2D). High levels of BrdU-positive β-cells were MycER;BclXL and Fzd9KO/KO;MycER;BclXL animals for 3-wk to activate detected in Fzd9WT/WT;MycER;BclXL mice (16.49% ± 5.4%), but, strikingly, MycER, control Fzd9WT/WT;MycER;BclXL transgenic mice rapidly developed low levels of BrdU-positive cells (2.25% ± 1.49%) were observed in the grossly hyperplastic insulinomas (Fig 1A and C), whereas Fzd9KO/KO; Fzd9KO/KO;MycER;BclXL β-cells. These levels were similar to those observed MycER;BclXL mice exhibited no detectable β-cell tumor hyperplasia in MycER transgene-negative islets (Fzd9KO/KO;BclXL; 0.52% ± 0.37%) (Fig 2E) and pancreatic islets preserved their normal size (Fig 1B and C). and in stark contrast to the higher levels observed after 3 d (Fig 2B). To exclude the possibility that Fzd9 deficiency inhibits MycER ex- Immunohistochemical staining for the proliferation marker Ki67 pression or stimulates its degradation, we monitored the presence of demonstrated that the proliferative arrest observed in Fzd9KO/KO;MycER; MycER immunohistochemically in β-cells, with an anti-ER antibody. As BclXL islets is in fact a progressive phenomenon already evident at 1 wk expected, MycER expression was undetectable in control MycER of sustained Myc activation and essentially complete by 3 wk (Fig 2F). transgene-negative islets of Fzd9KO/KO;BclXL mice, but was confirmed These results suggest that even though proliferation is initially to be consistently expressed in tamoxifen-treated Fzd9WT/WT;MycER; triggered upon Myc activation in Fzd9KO/KO;MycER;BclXL cells, Myc’s BclXL islets and was even higher in Fzd9KO/KO;MycER;BclXL islets (Fig 1D). tumorigenic potential is lost over time due to the lack of Fzd9. Hence, MycER is readily detectable in Fzd9-deficient cells, and therefore lack of Fzd9 impairs Myc-induced pancreatic β-cell expansion by mech- Fzd9 is necessary for MycER–dependent dedifferentiation and anisms unrelated to suppression of MycER expression. transformation of β-cells MycER is functionally active in the absence of Fzd9 In MycER;BclXL mice, partial Myc-induced dedifferentiation of β-cells has been previously noted as one of the pleiotropic effects of oncogenic Since Fzd9 deficiency blocks the appearance of Myc-driven hyperplastic Myc, responsible for tumorigenesis and expansion of the islets of insulinomas, we next investigated the possibility that MycER is simply no Langerhans (Pelengaris et al, 2002). To verify if such an effect was Fzd9- longer functional in tamoxifen-treated Fzd9KO/KO;MycER;BclXL islets. To dependent, we looked at the expression of both insulin and glucagon do this, we first ascertained whether activated MycER retains its ability to (markers of β-cells and α-cells, respectively). As expected, we detected drive β-cell proliferation in the absence of Fzd9. After 3 d of acute high insulin expression in control MycER transgene-negative islets tamoxifen treatment, BrdU was systemically administered to Fzd9KO/KO; (Fzd9KO/KO;BclXL) (Fig 3A). In contrast, insulin levels were somewhat BclXL, Fzd9WT/WT;MycER;BclXL and Fzd9KO/KO;MycER;BclXL mice 3 h before repressed upon 3-wk tamoxifen treatment of Fzd9WT/WT;MycER;BclXL euthanasia. Then, pancreata were collected and BrdU incorporation into mice because of Myc-induced partial dedifferentiation of β-cells (Fig β-cells analyzed by immunofluorescence (Fig 2A). Multiple BrdU-positive 3B) (Pelengaris et al, 2002). Intriguingly, however, high levels of insulin β-cells were detected in both Fzd9 wild-type (Fzd9WT/WT;MycER;BclXL) and expression were maintained in islets from Fzd9KO/KO;MycER;BclXL mice knockout (Fzd9KO/KO;MycER;BclXL) mice (21.6% ± 7.7% and 24.1% ± 16.4%, after tamoxifen treatment (Fig 3C and D), indicating blunting of Myc- respectively), whereas BrdU incorporation was almost absent from induced β-cell dedifferentiation in the absence of Fzd9. Indeed, pockets control MycER transgene-negative islets (Fzd9KO/KO;BclXL; 0.4% ± 0.5%) of fully differentiated cells can also be appreciated by hematoxylin and (Fig 2B). Thus, at least initially, Myc retains its capacity to drive β-cell eosin (H&E) staining, already after 2 wk of MycER activation in Fzd9KO/KO; proliferation even in the absence of Fzd9. MycER;BclXL but not in Fzd9WT/WT;MycER;BclXL (Fig S3). Next, to ascertain whether Myc retains its capacity to induce Glucagon positive α-cells displayed their typical proportion (~10% of apoptosis in pancreatic β-cells in the absence of Fzd9, MycER was the total islet cellular content) and distribution at the islet periphery in continuously activated for 3 d in the β-cells of Fzd9WT/WT;MycER and Fzd9-deficient mice without the MycER transgene (Fzd9KO/KO;BclXL) (Fig Fzd9KO/KO;MycER mice, without co-expression of BclXL. As seen for 3A). In contrast, in tamoxifen-treated Fzd9WT/WT;MycER;BclXL mice, Myc proliferation, shrinkage of islets resulting from β-cell apoptosis caused the appearance of α-cells scattered throughout the whole islet (previously shown in Pelengaris et al [2002]) was observed in both (Fig 3B). Notably, Myc did not trigger this relocalization in the absence of Fzd9-proficient and Fzd9-deficient islets (Fig 2C). Hence, Fzd9 is not Fzd9 (Fzd9KO/KO;MycER;BclXL mice treated with tamoxifen), where the required for Myc-induced β-cell apoptosis. α-cells displayed a more physiological distribution, although in a higher Finally, additional Myc-dependent phenotypes previously de- proportion than in control islets (Fig 3C and E). scribed (Soucek et al, 2007), such as recruitment of mast cells (Fig Overall, these results suggest that Fzd9 is required for Myc to S2A) and induction of angiogenesis (Fig S2B) were also observed induce profound and long-lasting dedifferentiation in β-cells, as after 3 d of tamoxifen treatment in Fzd9KO/KO;MycER;BclXL. part of its tumorigenic program. Together, these observations indicate that, in the absence of Fzd9, MycER is still functionally active upon tamoxifen treatment. Gene expression analysis shows down-regulation of genes involved in β-cell survival and stress response in the absence of Fzd9 absence impairs tumor expansion Fzd9 If Myc retains its ability, at least initially, to drive β-cell proliferation, why To identify potential transcriptional changes that could explain the do islets fail to expand in tamoxifen-treated Fzd9KO/KO;MycER;BclXL mice? anti-tumorigenic effect observed in the absence of Fzd9, Fzd9WT/WT; To investigate this, we asked whether Myc-induced β-cell proliferation is MycER;BclXL and Fzd9 KO/KO ;MycER;BclXL mice were treated with maintained long term in the absence of Fzd9. MycER was activated in tamoxifen for 3 d and the islet RNA was used for microarray Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 2 of 10 Figure 1. The absence of Fzd9 impairs the development of Myc-driven pancreatic insulinomas. (A, B) H&E staining (4×) of pancreas sections from Fzd9-proficient and (B) Fzd9-deficient mice treated with tamoxifen for 3 wk. (C) Quantification of the islet size from Fzd9KO/KO;BclXL, Fzd9WT/WT;MycER;BclXL and Fzd9KO/KO;MycER;BclXL mice. (D) Presence of MycER detected by immunofluorescence against ER (red) in these 3-wk treated islets. Yellow dotted lines define the periphery of the pancreatic islets. Representative images for each of the groups are shown (10×). Data information: in (C), data are represented as mean ± SD. * and ** indicate P-values below 0.05 and 0.01, respectively (Kruskal–Wallis followed by Dunn’s test). (A, D) Scale bars: 200 μm in (A), 100 μm in (D). Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 3 of 10 Figure 2. Sustained proliferation of pancreatic islets upon MycER activation requires Fzd9. (A, B) Incorporated BrdU (green) detected by immunofluorescence (10×) after 3 d of tamoxifen treatment and (B) its quantification. Percentages of BrdU-positive cells per islet from at least three mice per group are represented. (C) Representative H&E staining of pancreas sections from mice treated with tamoxifen for 3 d. Yellow dotted lines define the periphery of the pancreatic islets (4×). (D, E) Incorporated BrdU detected by immunofluorescence (10×) after 3 wk of tamoxifen treatment and (E) its quantification. Percentages of BrdU-positive cells per islet from at least three mice per group are represented. (F) Quantification of the proliferation marker Ki67 in islets from mice treated for 1 and 3 wk. Percentages of Ki67-positive cells per islet from three mice per group are represented. Data information: in (B, E, F), data are represented as mean ± SD. (B, E, F) Statistical significance of differences was examined using one-way, in (B, E), or two-way (in F) ANOVA followed by Tukey’s test. ** and *** indicate P-values below 0.01 and 0.001, respectively. (A, C, D) Scale bars: 100 μm in (A, D), 200 μm in (C). Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 4 of 10 Figure 3. Fzd9 mediates Myc-induced dedifferentiation of insulin-expressing β-cells. (A, B, C) Insulin (green) and glucagon (red) detected by immunofluorescence (10×) in pancreatic tissue sections of Fzd9KO/KO;BclXL, (B) Fzd9WT/WT;MycER;BclXL, and (C) Fzd9KO/KO;MycER;BclXL treated with tamoxifen for 3 wk. (A, B, C, D) Quantification of total insulin intensity in the islets from (A, B, C). (E) Quantification of glucagon positive cells in the periphery of the islets. Percentages of glucagon positive cells are shown. Data information: in (D, E), data are represented as mean ± SD. (D, E) Statistical significance of differences was examined using Kruskal–Wallis followed by Dunn’s test (D) and one-way ANOVA followed by Tukey’s test (E). * and ** indicate P-values below 0.05 and 0.01, respectively. Scale bar: 100 μm. analysis. This early 3-d time-point was selected to reveal those identified 933 differentially expressed genes out of 27,747 probes differences in gene expression that could be the cause, and not the (Fig 4A and Table S1). Among the genes whose expression is most consequence, of the subsequent phenotypic changes (such as decreased in Fzd9KO/KO pancreatic islets compared with Fzd9WT/WT decrease proliferation rate or dedifferentiation of β-cells). Statis- expression (fold-change < −2.5), we found several early response tical analysis of the microarray (P < 0.05 and fold-change ± 1.2) genes, comprising transcription factors and other cellular mediators Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 5 of 10 Figure 4. The absence of Fzd9 alters the expression of Myc-related, differentiation and Wnt signaling Gene Sets upon MycER activation. (A) Heat map of differentially expressed genes determined by microarray analysis performed on a pool of multiple isolated islets (~20–40) from each of four Fzd9WT/WT; MycER;BclXL mice and four Fzd9KO/KO;MycER;BclXL mice. (B, C) Gene sets related to pancreatic differentiation and (C) Myc targets. (D) qRT-PCR analysis of genes related to Gene Sets shown in (B) (GPC1) and in (C) (POLD2, SRSF7, and RPL34). (E) Immunofluorescence stainings (20×) and quantification of MYC targets MCM5 and PCNA. Total MCM5 intensity and PCNA positive nuclei per islet are shown. (F) Gene sets related to Wnt signaling pathway identified in the Gene Set Enrichment Analysis as differentially expressed when comparing expression profiles. (G) β-catenin detected by immunofluorescence (10×) in pancreatic tissue sections treated with tamoxifen for 3 d and its Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 6 of 10 of β-cell survival and stress response. JUNB and ATF3, for instance, Wnt signaling is engaged by Fzd9 and acts as a downstream are known to coordinate a β-cell survival pathway during inflam- effector of Myc-induced tumorigenesis mation (Gurzov et al, 2012), whereas NPAS4 is an important early mediator of β-cell stress response (Sabatini et al, 2013). Moreover, Given the role of Fzd9 in Wnt signaling, this pathway seemed a priori a RGS2 (Dong et al, 2017) and IRS2 (Blandino-Rosano et al, 2016) good candidate that could contribute to the multiple Myc-dependent regulate β-cell survival and mass. phenotypes observed when MycER is activated. In fact, Wnt signaling has In addition, Gene Set Enrichment Analysis (GSEA) revealed significantly been shown to regulate pancreatic β-cell proliferation (Rulifson et al, reduced mTORC1 signaling in Fzd9KO/KO mice compared with Fzd9WT/WT 2007). However, as previously noted, there was no detectable difference in controls. Importantly, mTORC1 signaling is necessary for β-cell prolifer- the histology of pancreata from Fzd9KO/KO and Fzd9WT/WT mice (Fig S1), ation, as Raptor knockdown has a direct impact on β-cell size, mass, and indicating that the absence of this receptor is not rate limiting at least survival (Blandino-Rosano et al, 2017) (NES: −2.18, q < 0.001; Table S2). during normal development of the organ. Nevertheless, in MycER- activated islets, there are significant differences in Wnt-related gene sets between Fzd9 knockout versus wild-type cells (Fzd9KO/KO;MycER;BclXL Differentiation and Myc target gene sets are differentially versus Fzd9WT/WT;MycER;BclXL) (Fig 4F). Moreover, qRT-PCR analysis of expressed in the absence of Fzd9 AXIN1 and JUNB, genes belonging to WNT and β-Catenin Gene Sets, re- spectively, confirmed their down-regulation, although they did not reach GSEA showed that several gene sets associated to β-cell differentiation statistical significance (P = 0.11 and 0.099, respectively; Fig S4E). In addition, were differentially expressed in MycER expressing Fzd9KO/KO;MycER;BclXL immunofluorescence staining for β-Actin, part of Labbe WNT3A Targets up compared with Fzd9WT/WT (Table S2). For instance, the expression of Gene Set (Fig 4F), showed that in the absence of Fzd9 expression, β-Actin genes involved in developmental processes, including transcriptional levels are significantly lower in pancreatic islets (Fig S4D). regulation of pancreatic β-cell differentiation (https://reactome.org/ Similarly, immunofluorescence against β-catenin revealed sig- content/detail/R-HSA-1266738) is significantly impaired (NES: −2.223; q: nificantly increased protein levels of this Wnt signaling downstream 0.0126). The Hedgehog/Gli pathway, whose up-regulation is responsible effector in Fzd9WT/WT;MycER;BclXL islets compared with their MycER- for β-cell dedifferentiation (Landsman et al, 2011), is significantly re- deficient counterpart Fzd9WT/WT;BclXL (Fig 4G), whereas β-catenin duced (NES: −2.221, q: 0.0125), as is TGFβR signaling, also involved in β-cell staining remained low in Fzd9KO/KO;MycER;BclXL cells. Hence, the dedifferentiation (Blum et al, 2014) (NES: −1.982, q: 0.0322) (Fig 4B). To- Wnt signaling pathway, clearly engaged when Myc is active, appears gether, these results show that several transcriptional programs en- dramatically affected by the absence of Fzd9. gaged by Myc in β-cells, normally leading to their dedifferentiation, are Given these results and the well-established role of Frizzled receptors in significantly impaired in the absence of Fzd9. Wnt signaling, we wondered whether pharmacological Wnt inhibition would Notably, GSEA also showed that several Myc-related gene sets are be capable of recapitulating the Fzd9 tumor-resistant phenotype. To verify significantly down-regulated (Fig 4C and Table S3) along with Myc-related this hypothesis, we made use of C59, a potent inhibitor of porcupine transcriptional programs like cell cycle, metabolism, and apoptosis (Fig (PORCN) that blocks Wnt palmitoylation, Wnt secretion and biological ac- S4A–C). The fact that the absence of only one Myc target gene is able to tivity (Proffitt et al, 2013). Fzd9WT/WT;MycER;BclXL mice were pretreated with affect the expression of a number of Myc-related signatures reveals Fzd9 C59 for 2 d, followed by 1 wk of C59/tamoxifen co-treatment. Then, pancreata as a key effector of Myc-driven reprogramming of pancreatic β-cells. from these C59-treated Fzd9WT/WT;MycER;BclXL mice, as well as mice treated We validated the transcriptional regulation in pancreatic islets with tamoxifen alone (Fzd9KO/KO;BclXL, Fzd9WT/WT;MycER;BclXL and Fzd9KO/KO; observed in the GSEA by performing qRT-PCR of some of the MycER;BclXL) were collected and stained to assess Ki67 positivity. Analysis of identified differentially regulated genes after 3 d of tamoxifen- individual islets shows that, as described above (Fig 3D), when MycER is induced MycER activation in Fzd9KO/KO and Fzd9WT/WT mice (Fig 4D). activated, cell proliferation in Fzd9KO/KO;MycER;BclXL islets was significantly We observed that, consistent with the microarray data, GPC1 (be- lower when compared with their Fzd9WT/WT;MycER;BclXL counterparts longing to Developmental Biology Gene Set), RPL34, SRSF7 and (18.28% ± 3.5% versus 35.99% ± 3.9%). Notably, when the Fzd9-proficient mice POLD2 (MYC Targets Gene Set), were significantly down-regulated. were treated with C59, the proliferation rate was also significantly decreased This confirms—using a distinct quantitative method—the Fzd9- (26.05% ± 3.2%) (Fig 4H), indicating that Wnt inhibition by C59 is able to largely mediated transcriptional regulation of several MycER targets pre- mimic the phenotype of the Fzd9-deficient mice. viously observed by microarray analysis. We complemented this validation approach by performing immunofluorescence stainings for other down-regulated MYC target genes (namely, MCM5 and proliferating cell nuclear antigen Discussion [PCNA]) on pancreatic islets after 1 wk of MycER activation and confirmed their reduced expression in the islets of Fzd9 knockout A vast amount of data in the literature points at Wnt signaling as one of mice (Fig 4E). the major culprits in solid and liquid tumors (Zhan et al, 2017) and Wnt quantification in the islets. Total β-catenin intensity per islet area is shown. (H) Percentage of Ki67-positive cells in individual islets from Fzd9WT/WT;MycER;BclXL and Fzd9KO/KO;MycER;BclXL after 1 wk of tamoxifen treatment. An additional group of the Fzd9WT/WT;MycER;BclXL mice was pretreated for 2 d with the Wnt inhibitor C59 and then received both C59 + tamoxifen during 1 wk. Data information: in (D, G, H), data are represented as mean ± SD, whereas in (E), as median and interquartile range. (D, E, G) Statistical significance of differences was examined using t test (D), Mann-Whitney U test (E) and Tukey’s test (G). *, ** and *** indicate P-values below 0.05, 0.01 and 0.0001, respectively. (E, G) Scale bars: 50 μm in (E), 100 μm in (G). FDR, false discovery rate; NES, normalized enrichment score. Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 7 of 10 pathway inhibition via the targeting of Frizzled receptors has been frozen at −20°C. This injectable solution was administered to mice suggested as a potential strategy to decrease growth and tumorigenicity by intraperitoneal injection (6 μl/g) every 24 h. 1 ml-syringes and of human tumors (Gurney et al, 2012). Our data indicate that simply 27 G needles were used for injection. inhibiting one receptor at a time (namely Fzd9) could be sufficient to The Wnt inhibitor C59 was formulated in PEG400 (Sigma-Aldrich) achieve a significant therapeutic impact at least in some tumorigenic at 2 g/l. C59 in PEG400 was aliquoted and stored at 4°C. Right before contexts, with the advantage of reducing potential side effects associated each treatment, an equal amount of water (1:1) was added to make a with the simultaneous inhibition of multiple Frizzled family members. final concentration of 1 g/l of C59 in 50% PEG400. Mice were then Whereas Myc expression has been traditionally placed downstream of treated with a daily dose of 10 mg/kg by oral gavage for seven the Wnt signaling pathway, as previously discussed, others have indi- consecutive days. A mixture 1:1 dilution of water and PEG400 was cated otherwise (Cowling & Cole, 2007). Here, we demonstrate for the used as vehicle for C59-untreated animals. first time the key role of the Wnt receptor Fzd9 and Wnt signaling in Myc- induced insulinomas. In this context, whereas previous results suggest Immunostaining of pancreas sections that targeting Fzd9 may not be the best strategy for cancer therapy because of its dual pro- and anti-tumorigenic character (Zeng et al, 18 h after the last administration of tamoxifen, mice were eutha- 2018), our data indicate that it might be at least a valuable target when nized with CO2 and pancreata collected. For BrdU staining, 150 μl of up-regulated in Myc-driven malignancies. In addition, our results sug- BrdU (Sigma-Aldrich) at 5 mg/ml were intraperitoneally injected 2 h gest the existence of a novel positive feedback loop between Myc and before euthanasia. Wnt signaling: deregulation of Myc might enhance Wnt signaling by up- For histological analysis, mouse pancreata were fixed with regulation of Fzd9, which would, at the same time, promote Myc paraformaldehyde through systemic cardiac perfusion, col- overexpression/deregulation—because Myc is a bona fide downstream lected in cassettes and further incubated in paraformaldehyde target of Wnt—thus feeding again Wnt signaling through Fzd9. It should for 24 h. Tissues were then paraffin-embedded, and 5-μm be noted that both our genetic and pharmacological approach to delete sections cut and stained by H&E. Additional sections were used Fzd9 and interfere with Wnt signaling, respectively, are systemic and not to perform immunostaining. In short, sections were deparaffi- specific for the islets of Langerhans only. Hence, we cannot exclude that nized, rehydrated, and subjected to high-temperature antigen such a positive feedback loop between Myc and Wnt can also have cell retrieval in 10 mM citrate buffer (pH 6.0). Primary antibodies autonomous implications. were as follows: anti-BrdU (Clone BU1/75; Bio-Rad), anti-Ki67 The identification of Fzd9 as a key Myc-driven tumorigenic (SP6; Abcam), anti-Meca32 (Meca32; BD Biosciences), anti-insulin effector is particularly relevant in view of the fact that, even if (EPR17359; Abcam), anti-glucagon (K79bB10; Abcam), anti-β-catenin Myc is one of the “most wanted” cancer targets, several direct, (D10A8; Cell Signaling), anti-MCM5 (ab75975; Abcam), anti-PCNA and indirect Myc inhibition strategies have failed because of lack (307904; BioLegend), and anti-β-Actin (A-5441; Sigma-Aldrich). of efficacy and high toxicity derived from low specificity (Whitfield Samples were incubated with primary antibodies overnight in et al, 2017). Thus, identification of key tumorigenic effectors and blocking buffer (2.5% BSA, 0.3% Triton X-100 in PBS), sections their selective targeting represents an alternative strategy to were washed, and species-appropriate secondary applied, ei- achieve high therapeutic impact in Myc-deregulated malignan- ther Alexa Fluor 488 dye–conjugated antibodies (Thermo Fisher cies. Here, we have described a good example of this alternative Scientific) or Vectastain ABC kit and DAB reagents (Vector Lab- approach. oratories). Fluorescence antibody-labeled slides were mounted in DAKO fluorescent mounting medium containing 1 μg/ml DAPI counterstain. Materials and Methods Microarray analysis of pancreatic islets Generation and maintenance of genetically engineered mice Genome-wide expression analysis was performed in isolated pan- pIns-MycERTAM;RIP7-Bcl-xL and Fzd9KO/KO mice have been previ- creatic islets. Briefly, tamoxifen-treated Fzd9WT/WT;MycER;BclXL and ously described (Pelengaris et al, 2002; Zhao et al, 2005). All the Fzd9KO/KO;MycER;BclXL (n = 4) mice were euthanized by cervical animal studies were performed in accordance with the ARRIVE dislocation. Pancreata were inflated with Collagenase P (6 ml/mouse guidelines and the 3 Rs rule of Replacement, Reduction and at 0.7 mg/ml) (Roche) in HBSS injected through the bile duct. Tissues Refinement principles. Animals were maintained and treated in were transferred to vials containing 5 ml of Collagenase P and in- accordance with protocols approved by the CEEA (Ethical Com- cubated for 20 min at 37°C with gentle shaking. Digested pancreata mittee for the Use of Experimental Animals) at the Vall d’Hebron were poured into 50 ml tubes and washed with cold HBSS by filling Institute of Oncology. Mice (both males and females) between 8 the tube, performing a short spin up to 652g and removing the and 12 wk old were used. supernatant. Pellets were resuspended in 5 ml and exocrine tissue further removed by filtering the suspension through 100 μm re- Preparation and administration of tamoxifen and C59 strainers. Then, tissue remaining in the filter was placed in a 6 cm plate in 4 ml of cold HBSS. Pancreatic islets were visualized with the Tamoxifen (Sigma-Aldrich) was dissolved in peanut oil (Thermo help of a magnifier by addition of dithizone at 0.02 g/l, hand-picked Fisher Scientific) at 10 mg/ml. Aliquots of 1 ml were prepared and and transferred into a clean 1.5 ml tube. Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 8 of 10 RNA from islets was isolated, DNAse-treated and quality Acknowledgements assessed through Agilent 2100 Bioanalyzer. RNA was reverse- transcribed to generate cDNA. Microarray was performed using a MF Zacarı́as-Fluck was supported by the Juan de la Cierva Programme of the Mouse Gene Array 2.1 ST (Affimetryx). GSEA was performed using Spanish Ministry of Economy and Competitiveness (IJCI-2014-22403) and L publicly available software provided by the Broad Institute Soucek by the Generalitat de Catalunya AGAUR 2017 grant SGR-3193, S Martı́nez- (version 3.0) with the Hallmarks, Curated, Motif, gene ontology, Martı́n by the Generalitat de Catalunya “Contractació de personal investigador novell (FI-DGR)” 2016 fellowship (2016FI_B 00592). We thank the rest of the Oncogenic Signatures and Immunological Signatures gene sets Soucek Lab for critical reading of the manuscript, Dr. Oriol Arqués for his advice from the MsigDB (Subramanian et al, 2005). The number of on the microarray analysis and Dr. Águeda Martı́nez-Barriocanal for providing permutations was set to 1,000 and the genes were ranked the proliferating cell nuclear antigen antibody. We also thank Dr. Jose Fernández- according to Signal2Noise. Heat map was generated using Heat Navarro, Group Leader of Bioinformatics Unit at VHIO, for his advice on the mapper (Babicki et al, 2016). Hierarchical clustering was per- statistical analysis of qRT-PCR. Finally, we thank the Cellex Foundation for providing research facilities and equipment and the FERO Foundation for their formed applying complete linkage method and based on Eu- support to the laboratory of L Soucek. clidean distance. qRT-PCR validation of microarray Author Contributions qRT-PCR validation was performed in isolated pancreatic islets. Briefly, MF Zacarias-Fluck: conceptualization, formal analysis, investigation, Fzd9WT/WT;MycER;BclXL (n = 3) and Fzd9KO/KO;MycER;BclXL (n = 3) were project administration, visualization, and writing—original draft, treated with tamoxifen for 3 d and pancreatic islets were then isolated from review, and editing. them following the protocol described in the previous section. RNA was T Jauset: conceptualization, formal analysis, investigation, project then extracted from the samples using RNeasy kit (QIAGEN) and quantified administration, visualization, and writing—original draft, review, using NanoDrop. Equal amounts of RNA were then DNAse-treated (NEB) and editing. and reverse transcribed to generate cDNA using iScript Reverse Tran- S Martı́nez-Martin: investigation. scription Supermix for RT-qPCR (Bio-Rad). SYBR green qRT-qPCR analysis J Kaur: investigation, formal analysis, and writing—review. was then performed on these cDNA samples with PerfeCTa SYBR Green S Casacuberta-Serra: investigation. FastMix, Low Rox (Quantabio) using QuantStudio 6 FLEX system (Applied D Massó-Vallés: investigation. Biosystems). The data thus obtained were analyzed following the com- E Serrano del Pozo: investigation. parative (ΔΔCT) method described in Livak and Schmittgen (2001). B2M G Martı́n-Fernandez: investigation. (Beta-2-microglobulin) was used as the housekeeping gene. Based on our Í González-Larreategui: investigation. hypothesis and the data from the microarray, statistical analysis was S López-Estévez: investigation. performed using one-tailed t test. Sequences of primers used are listed in L Brown-Swigart: project administration. Table S4. M-E Beaulieu: formal analysis. JR Whitfield: formal analysis and writing—original draft, review, and Image and statistical analysis editing. B Madan: resources. Islet size was quantified using ImageJ. Immunofluorescence (BrdU, In- DM Virshup: resources. sulin, Glucagon, β-catenin, MCM5, PCNA, β-Actin) and Immunohisto- GI Evan: conceptualization and formal analysis. chemistry (Ki67) stainings were quantified using ImageJ and Qupath L Soucek: conceptualization, formal analysis, project administra- (Bankhead et al, 2017), respectively. All data were represented and tion, supervision, investigation, and writing—original draft, review, analyzed using GraphPad Prism 6. One-way ANOVA and Tukey’s test, or and editing. Kruskal–Wallis followed by Dunn’s test were used to assess statistical significance when analyzing three groups of parametric or non- Conflict of Interest Statement parametric distributions. Analysis of two groups were performed with t test or Mann–Whitney. For each experiment, at least three animals per Drs B Madan and DM Virshup have a financial interest in an unrelated PORCN group were used and multiple islets quantified. Results are shown as inhibitor, ETC-159. The other authors declare no conflict of interest. mean ± SD or median and interquartile range, accordingly. Statistical significance was considered when P < 0.05. References Data Availability Babicki S, Arndt D, Marcu A, Liang Y, Grant JR, Maciejewski A, Wishart DS (2016) Heatmapper: Web-enabled heat mapping for all. Nucleic Acids Res 44: W147–W153. doi:10.1093/nar/gkw419 Microarray data were deposited under accession number GSE167073. Bankhead P, Loughrey MB, Fernández JA, Dombrowski Y, McArt DG, Dunne PD, McQuaid S, Gray RT, Murray LJ, Coleman HG, et al (2017) QuPath: Open source software for digital pathology image analysis. Sci Rep 7: 16878. Supplementary Information doi:10.1038/s41598-017-17204-5 Blandino-Rosano M, Barbaresso R, Jimenez-Palomares M, Bozadjieva N, Supplementary Information is available at https://doi.org/10.26508/lsa.201900490. Werneck-de-Castro JP, Hatanaka M, Mirmira RG, Sonenberg N, Liu M, Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 9 of 10 Rüegg MA, et al (2017) Loss of mTORC1 signalling impairs β-cell Rulifson IC, Karnik SK, Heiser PW, ten Berge D, Chen H, Gu X, Taketo MM, Nusse homeostasis and insulin processing. Nat Commun 8: 16014. R, Hebrok M, Kim SK (2007) Wnt signaling regulates pancreatic beta doi:10.1038/ncomms16014 cell proliferation. Proc Natl Acad Sci U S A 104: 6247–6252. doi:10.1073/ pnas.0701509104 Blandino-Rosano M, Scheys JO, Jimenez-Palomares M, Barbaresso R, Bender AS, Yanagiya A, Liu M, Rui L, Sonenberg N, Bernal-Mizrachi E (2016) 4E- Sabatini PV, Krentz NA, Zarrouki B, Westwell-Roper CY, Nian C, Uy RA, Shapiro BP2/SH2B1/IRS2 are part of a novel feedback loop that controls beta- AM, Poitout V, Lynn FC (2013) Npas4 is a novel activity-regulated cell mass. Diabetes 65: 2235–2248. doi:10.2337/db15-1443 cytoprotective factor in pancreatic beta-cells. Diabetes 62: 2808–2820. doi:10.2337/db12-1527 Blum B, Roose AN, Barrandon O, Maehr R, Arvanites AC, Davidow LS, Davis JC, Peterson QP, Rubin LL, Melton DA (2014) Reversal of beta cell de- Sabò A, Kress TR, Pelizzola M, de Pretis S, Gorski MM, Tesi A, Morelli MJ, Bora P, differentiation by a small molecule inhibitor of the TGFbeta pathway. Doni M, Verrecchia A, et al (2014) Selective transcriptional regulation Elife 3: e02809. doi:10.7554/elife.02809 by Myc in cellular growth control and lymphomagenesis. Nature 511: 488–492. doi:10.1038/nature13537 Cowling VH, Cole MD (2007) Turning the tables: Myc activates Wnt in breast cancer. Cell Cycle 6: 2625–2627. doi:10.4161/cc.6.21.4880 Soucek L, Lawlor ER, Soto D, Shchors K, Swigart LB, Evan GI (2007) Mast cells are required for angiogenesis and macroscopic expansion of Myc-induced Dang CV (2012) MYC on the path to cancer. Cell 149: 22–35. doi:10.1016/ pancreatic islet tumors. Nat Med 13: 1211–1218. doi:10.1038/nm1649 j.cell.2012.03.003 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Dong H, Zhang Y, Wang J, Kim DS, Wu H, Sjogren B, Gao W, Luttrell L, Wang H (2017) Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al (2005) Gene set Regulator of G protein signaling 2 is a key regulator of pancreatic beta-cell enrichment analysis: A knowledge-based approach for interpreting mass and function. Cell Death Dis 8: e2821. doi:10.1038/cddis.2016.216 genome-wide expression profiles. Proc Natl Acad Sci U S A 102: Fujimoto T, Tomizawa M, Yokosuka O (2009) SiRNA of frizzled-9 suppresses 15545–15550. doi:10.1073/pnas.0506580102 proliferation and motility of hepatoma cells. Int J Oncol 35: 861–866. Wang Q, Liu H, Wang Q, Zhou F, Liu Y, Zhang Y, Ding H, Yuan M, Li F, Chen Y (2017) doi:10.3892/ijo_00000400 Involvement of c-Fos in cell proliferation, migration, and invasion in Gurney A, Axelrod F, Bond CJ, Cain J, Chartier C, Donigan L, Fischer M, osteosarcoma cells accompanied by altered expression of Wnt2 and Chaudhari A, Ji M, Kapoun AM, et al (2012) Wnt pathway inhibition via Fzd9. PLoS One 12: e0180558. doi:10.1371/journal.pone.0180558 the targeting of Frizzled receptors results in decreased growth and Whitfield JR, Beaulieu ME, Soucek L (2017) Strategies to inhibit Myc and their tumorigenicity of human tumors. Proc Natl Acad Sci U S A 109: clinical applicability. Front Cell Dev Biol 5: 10. doi:10.3389/ 11717–11722. doi:10.1073/pnas.1120068109 fcell.2017.00010 Gurzov EN, Barthson J, Marhfour I, Ortis F, Naamane N, Igoillo-Esteve M, Winn RA, Marek L, Han SY, Rodriguez K, Rodriguez N, Hammond M, Van Scoyk M, Gysemans C, Mathieu C, Kitajima S, Marchetti P, et al (2012) Pancreatic Acosta H, Mirus J, Barry N, et al (2005) Restoration of Wnt-7a expression beta-cells activate a JunB/ATF3-dependent survival pathway during reverses non-small cell lung cancer cellular transformation through frizzled- inflammation. Oncogene 31: 1723–1732. doi:10.1038/onc.2011.353 9-mediated growth inhibition and promotion of cell differentiation. J Biol He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Chem 280: 19625–19634. doi:10.1074/jbc.m409392200 Vogelstein B, Kinzler KW (1998) Identification of c-MYC as a target of the Winn RA, Van Scoyk M, Hammond M, Rodriguez K, Crossno JT Jr., Heasley LE, APC pathway. Science 281: 1509–1512. doi:10.1126/science.281.5382.1509 Nemenoff RA (2006) Antitumorigenic effect of Wnt 7a and Fzd 9 in non- Kirikoshi H, Sekihara H, Katoh M (2001) Expression profiles of 10 members of small cell lung cancer cells is mediated through ERK-5-dependent Frizzled gene family in human gastric cancer. Int J Oncol 19: 767–771. activation of peroxisome proliferator-activated receptor gamma. J doi:10.3892/ijo.19.4.767 Biol Chem 281: 26943–26950. doi:10.1074/jbc.m604145200 Landsman L, Parent A, Hebrok M (2011) Elevated Hedgehog/Gli signaling Zeng C-M, Chen Z, Fu L (2018) Frizzled receptors as potential therapeutic targets causes beta-cell dedifferentiation in mice. Proc Natl Acad Sci U S A in human cancers. Int J Mol Sci 19: 1543. doi:10.3390/ijms19051543 108: 17010–17015. doi:10.1073/pnas.1105404108 Zhan T, Rindtorff N, Boutros M (2017) Wnt signaling in cancer. Oncogene 36: Lawlor ER, Soucek L, Brown-Swigart L, Shchors K, Bialucha CU, Evan GI (2006) 1461–1473. doi:10.1038/onc.2016.304 Reversible kinetic analysis of Myc targets in vivo provides novel Zhang Y, Jiang Q, Kong X, Yang L, Hu W, Lv C, Li Y (2016) Methylation status of insights into Myc-mediated tumorigenesis. Cancer Res 66: 4591–4601. the promoter region of the human frizzled 9 gene in acute myeloid doi:10.1158/0008-5472.can-05-3826 leukemia. Mol Med Rep 14: 1339–1344. doi:10.3892/mmr.2016.5387 Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using Zhang Z, Schittenhelm J, Guo K, Buhring HJ, Trautmann K, Meyermann R, Schluesener real-time quantitative PCR and the 2(-Delta Delta C(T)) method. HJ (2006) Upregulation of frizzled 9 in astrocytomas. Neuropathol Appl Methods 25: 402–408. doi:10.1006/meth.2001.1262 Neurobiol 32: 615–624. doi:10.1111/j.1365-2990.2006.00770.x Pelengaris S, Khan M, Evan GI (2002) Suppression of Myc-induced apoptosis in beta Zhao C, Aviles C, Abel RA, Almli CR, McQuillen P, Pleasure SJ (2005) cells exposes multiple oncogenic properties of Myc and triggers carcinogenic Hippocampal and visuospatial learning defects in mice with a progression. Cell 109: 321–334. doi:10.1016/s0092-8674(02)00738-9 deletion of frizzled 9, a gene in the Williams syndrome deletion Proffitt KD, Madan B, Ke Z, Pendharkar V, Ding L, Lee MA, Hannoush RN, interval. Development 132: 2917–2927. doi:10.1242/dev.01871 Virshup DM (2013) Pharmacological inhibition of the Wnt acyltransferase PORCN prevents growth of WNT-driven mammary cancer. Cancer Res 73: 502–507. doi:10.1158/0008-5472.can-12-2258 License: This article is available under a Creative Ranheim EA, Kwan HC, Reya T, Wang YK, Weissman IL, Francke U (2005) Commons License (Attribution 4.0 International, as Frizzled 9 knock-out mice have abnormal B-cell development. Blood described at https://creativecommons.org/ 105: 2487–2494. doi:10.1182/blood-2004-06-2334 licenses/by/4.0/). Fzd9 is essential for Myc-driven insulinomas Zacarı́as-Fluck et al. https://doi.org/10.26508/lsa.201900490 vol 4 | no 5 | e201900490 10 of 10