Fig 2: Effects of LE or vehicle on protein expression on the MCAO and reperfusion injury after the administration of DMSO or XAV939. (a) Representative Western blots of proteins in the penumbra region of the left hemisphere; (b–d) Levels of Akt and pAkt in the experimental groups. The DMSO+MCAO+Veh group and XAV939+MCAO+LE 20% group decreased significantly in pAkt level compared to DMSO+Sham+Veh group. pAkt was significantly increased in the DMSO+MCAO+LE 20% group compared to the DMSO+MCAO+Veh group and XAV939+MCAO+LE 20% group; (e–g) Levels of GSK-3ß and pGSK-3ß in the experimental groups. GSK-3ß and pGSK-3ß levels of DMSO+MCAO+Veh group and XAV939+MCAO+LE 20% group were significantly lower than the DMSO+Sham+Veh group. The DMSO+MCAO+LE 20% had significantly increased pGSK-3ß and GSK-3ß levels compared to the DMSO+MCAO+Veh and XAV939+MCAO+LE 20% groups; (h–m) Wnt signal-related protein expressions of experimental groups. Significantly decreased levels of Wnt1, Wnt3, PORCN and ß-catenin were observed in the DMSO+MCAO+Veh group compared to the DMSO+Sham+Veh group. The DMSO+MCAO+LE 20% group had significantly increased protein levels of Wnt1 and PORCN compared to the DMSO+MCAO+Veh group. The XAV939+MCAO+LE 20% group had decreased expression levels of Wnt1 and PORCN compared to the DMSO+MCAO+LE 20% group. The ß–catenin expression level in the DMSO+MCAO+LE 20% and XAV939+MCAO+LE 20% groups did not differ significantly. There was a significant increase in pß–catenin and tankyrase 1 in the DMSO+MCAO+Veh compared to the DMSO+Sham+Veh group. Pß–catenin and tankyrase1 was significantly decreased in the DMSO+MCAO+LE 20% compared to the DMSO+MCAO+Veh group. The XAV939+MCAO+LE 20% group had significantly increased pß–catenin and tankyrase 1 expression levels compared to the DMSO+MCAO+LE 20% group; (n–q) Inflammatory protein expressions in the experimental groups. Significantly increased expressions of inflammatory markers were observed in the MCAO-injured groups compared to the DMSO+Sham+Veh group. The DMSO+MCAO+LE 20% group had significantly decreased inflammatory protein expression levels compared to the DMSO+MCAO+Veh group. Significant decrease in inflammatory protein expressions were observed in the XAV939+MCAO+LE 20% group compared to the DMSO+Sham+Veh group. TNF-a expression level was increased significantly in the XAV939+MCAO+LE 20% group compared to the DMSO+MCAO+LE 20% group. Data are presented as mean ± standard error of the mean (SEM); n = 8 for each group; * p < 0.05, ** p < 0.01 vs. DMSO+Sham+Veh, # p < 0.05, ### p < 0.001 vs. DMSO+MCAO+Veh, † p < 0.05, ††† p < 0.001 vs. DMSO+MCAO+LE 20%, one-way ANOVA followed by Tukey’s multiple comparison test.

Fig 3: Elevated WNT/ß-catenin signaling and increased expression of WNT ligands are hallmarks of HPV8-E6 driven cSCC. a Left panel: nuclear accumulation of ß-catenin in the invasive front of HPV-driven cSCCs (black arrowheads). Healthy skin shows membranous expression of ß-catenin (gray arrowhead). Scale bar = 50 µm. Right panel: in situ hybridization for Axin2 (red stain) shows increased expression of this WNT/ß-catenin signaling target in tumors in comparison with healthy skin. Nuclei are counterstained with hematoxylin (blue stain). Samples were collected 4 weeks after tumor induction. Scale bar = 100 µm. Representative samples of a total of four biological replicates are depicted. b Enhanced amount of Lgr5 transcripts in CSCs (EpCAM+ CD34+ CD49f+) compared to non-CSC tumor cells (EpCAM+ CD34-) and stromal cells (EpCAM- CD34- CD49f-) (upper panel). The Wnt target gene Cd44 shows low expression in the stroma and high expression in all epithelial tumor cells (lower panel). Cells from HPV-driven cSCC from two different mice were sorted and expression was quantified by qRT-PCR. Samples were collected 4 weeks after tumor induction. Symbols represent individual mice. This experiment was done once. c In situ hybridization of healthy skin and HPV-driven cSCC for Wnt16 (red), shows higher expression of Wnt16 in tumors compared with healthy skin. Asterisks marks unspecific chromogen accumulation in sebaceous gland cells. The section is counterstained with hematoxylin. Scale bar = 50 µm. Samples were collected 4 weeks after tumor induction. Representative samples of a total of three biological replicates are depicted. d PORCN (red) staining in healthy skin and HPV-driven cSCC. E-cadherin (green) marks epithelial cells, DAPI (blue) marks nuclei. Scale bar = 100 µm. PORCN is mostly expressed in tumor epithelial cells along the tumor–stroma interphase of the tumor, whereas it is absent from healthy epidermis. Insets show a larger magnification of the region marked by the white square. Samples were collected 4 weeks after tumor induction. Representative samples of a total of four biological replicates are depicted. e Overlapping expression of PORCN (green) and CD34 (red) in healthy skin with hair follicles and tumor tissue is indicated by arrowheads. Scale bar = 50 µm. Insets show a larger magnification of the region marked by the white square. Samples were collected 4 weeks after tumor induction. Representative samples of a total of three biological replicates are depicted. f Increased expression of PORCN (red) in human cSCC in comparison with healthy human skin. E-cadherin (green) marks epithelial tissue, nuclei are counterstained in blue. Scale bar = 100 µm. Insets show a larger magnification of the region marked by a white square. Additional pictures are shown in Supplementary Figure 3. A representative sample of a total of nine is depicted. Krt14-HPV8(E6) mice were described previously [13] and were bred to FVB mice (Harlan Laboratories, Envigo) in house. Mice were kept under specific pathogen-free conditions at the Laboratory Animal Services Center at the University of Zurich and 6–8-weeks-old, sex-matched mice were used for all experiments. To induce cSCC, ~ 4 cm2 of shaved dorsal skin was irradiated with UVA (5 J/cm2) plus UVB (1 J/cm2) using the UV 802 L Waldmann device. Murine tumor samples were collected 4 weeks after tumor induction. Experiments were performed in accordance with the Swiss federal and cantonal regulations on animal protection and were approved by the Cantonal Veterinary Office Zurich. The Swiss law on animal protection demands that groups sizes are as small as possible. The clear biological differences allowed statistical differences with small group sizes. No animals were excluded from analysis in any experiment. In case of treatment of mice with established tumors, mice were randomized in two groups before start of treatment based on tumor size; in all other cases mice were not randomized. The investigators were not blinded. Groups were compared with an unpaired Student’s t-test and show the mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. Based on comparable SD similar variance was assumed. The above information applies to all animal experiments in this report. Human SCC samples (nine samples in total; three were dedifferentiated, five were moderately differentiated, one was well-differentiated) were obtained from biobanks managed by the University Research Priority Project “Translational Cancer Research” and the research project “Skintegrity”. All material were surplus biopsies from patients who had signed an informed consent that was approved by the Cantonal Ethics Commission Zurich (EK647 and EK800). For flow cytometry and sorting, tumors were collected in PBS and digested for 2 h at 37 °C in 2.4 mg/ml Dispase (Roche), cut into pieces and digested again for one hour at 37 °C in 1 mg/ml collagenase Type IV (Worthington Biochemical Corporation) and 0.1 mg/ml DNase (Sigma-Aldrich). Antibodies against the following proteins were used: CD45.1 (clone A20, BioLegend), CD31 (clone MEC13.3, BioLegend), EpCAM (clone GoH3, BioLegend), CD34 (clone RAM34, eBioscience). Dead cells were excluded using Zombie Violet Fixable Viability Kit (BioLegend). Doublets were excluded by FSC-A versus FSC-H and SSC-A versus SSC-H gating. Immunohistochemistry on frozen sections was performed on tissue fixed in 4% PFA for 1 h, left to sink in 30% sucrose and embedded in OCT. Ten-µm-thick cryosections were blocked for 1 h at RT with 2.5% Hings and 2.5% BSA in 0.1% Tween in PBS (PBST) (blocking buffer), and then stained with biotin-conjugated anti-CD34 (clone RAM34, eBioscience) and unconjugated rabbit anti-PORCN (clone ab105543, Abcam) overnight at 4 °C in blocking buffer. Secondary antibodies (see below) were added for 1 h in blocking buffer at RT, then samples were mounted in FluorSave (CalBiochem). Standard protocols were used for embedding and cutting formalin-fixed paraffin-embedded (FFPE) tissue. After deparaffinization, on both mouse and human samples, antigen retrieval was performed in 10 mM trisodium citrate buffer pH 6. Staining was performed as described above for frozen sections. The following antibodies were used for mouse and human FFPE samples: Mouse-anti-ß-catenin (clone14, BD transduction labs), mouse-anti-E-cadherin (BD transduction labs), rabbit anti-PORCN (Abcam ab105543), rabbit anti-Ki67 (Abcam), rabbit anti-HMGA2 (SantaCruz), rabbit anti-MMP13 (clone 3H13L17, ThermoFisher), and rabbit anti-phospho-ERK (Cell Signaling). Secondary antibodies used were goat-anti-mouse AlexaFluor 488 and goat-anti-rabbit AlexaFluor 555. For the PORCN staining a biotin-labeled goat-anti-rabbit secondary antibody was used, followed by the ABC kit (VectaShield), and the Cy3 tyramide amplification kit (PerkinElmer). Staining for nuclear ß-catenin was performed with biotin-labeled secondary antibodies, followed by DAB staining (VectorLabs). When mouse primary antibodies were used on murine tissues, VectorLabs MOM kit was used to block endogenous antigens. In situ hybridization was performed using the RNAscope kit (Advanced Cell Diagnostics) according to the manufacturer’s instructions. Probes for Axin2 and Wnt16 were obtained from the same company. For qRT-PCR, RNA was isolated from sorted cells using the NucleoSpin RNA XS kit (Machery-Nagel) according to the manufacturer’s instructions. qRT-PCRs using SybrGreen were performed on cDNA synthesized with the Roche Transcriptor High Fidelity cDNA Synthesis Kit after RNA isolation by standard TRI-Reagent protocols. Reactions were performed in triplicates and monitored with the ABI Prism 7900HT system (Applied Biosystems). The following 5’–3’ primers (Microsynth) were used for qRT-PCR. Gapdh, fwd AACTTTGGCATTGTGGAAGG, rev ATCCACAGTCTTCTGGGTGG; Lgr5, fwd CTCCACACTTCGGACTCAACAG, rev AACCAAGCTAAATGCACCGAAT

Fig 4: Inhibition of WNT secretion by the PORCN inhibitor LGK974 impairs the initiation of HPV-driven cSCC. a Experimental design. Treatment with LGK974 (6 mg/kg) or vehicle was started 7 days prior to tumor induction by UV-irradiation. Mice were treated daily until the end point at day 28. Control mice were treated with vehicle. The vehicle-treated group consisted of four, the LGK974-treated group of five mice. The experiment was performed twice with similar results. b Representative macroscopic display of a vehicle-treated (left panel) and LGK974-treated (middle panel) tumor. Tumor weight at endpoint (right panel). Symbols represent individual mice. c Representative H&E staining of vehicle-treated (left panel) and LGK974-treated (right panel) tumors. Scale bar = 100 µm. d Quantification of transcripts for Pthlh, Ptprz1, and Cd44 in LGK974- and vehicle-treated tumors shows significant reduction of markers for tumor malignancy upon treatment. Symbols represent individual mice. e Representative staining showing stabilization of ß-catenin in cytoplasm and nucleus of vehicle-treated tumors, whereas it is mostly membrane associated upon LGK974 treatment (Scale bar = 50 µm). f Representative image of Axin2 in situ hybridization showing reduced transcripts (red) upon LGK974- treatment. Nuclei are counterstained with Hematoxylin. Scale bar = 50 µm. The Porcupine inhibitor LGK974 was applied as described in ref. [31]. In brief, LGK974 was dissolved in DMSO and diluted in citrate buffer pH 3 to a final concentration of 1 mg/ml. Mice were treated daily with 6 mg/kg LGK974 or vehicle (20% DMSO in citrate buffer pH 3) per os. Staining and qPCR Protocols are described in the legend of Fig. 1. The primers used were Pthlh, fwd ATCCCCGACGCCTATGTAA, rev GGGGAAAAAGCAATCAGAGA; Ptprz1, fwd GCCAGTTGTTGTCCACTGC, rev CCTTTGAGAACGAATGTGCTT; Cd44, fwd CTCCTTCTTTATCCGGAGCAC, rev TGGCTTTTTGAGTGCACAGT

Fig 5: Effects of LE or vehicle on protein expressions after MCAO and reperfusion injury. (a) Representative Western blots indicating the expression of specific proteins in the penumbra region of the left hemisphere; (b–d) Expression and phosphorylation of Akt in the experimental groups. pAkt levels in MCAO+Veh group decreased significantly compared to the Sham+Veh group. pAkt was significantly increased in the MCAO+LE 20% group compared to the MCAO+Veh group; (e–g) Expression and phosphorylation levels of GSK-3ß (pGSK-3ß) in the experimental groups. pGSK-3ß and GSK-3ß levels in the MCAO+Veh group was significantly lower than the Sham+Veh group. The MCAO+LE 20% had significantly increased pGSK-3ß and GSK-3ß levels compared to the MCAO+Veh group; (h–m) Wnt signal-related protein expressions of experimental groups. Decreased levels of Wnt1, Wnt3, PORCN and ß-catenin were observed in the MCAO+Veh group compared to the Sham+Veh group. The MCAO+LE 20% group had significantly increased protein levels of Wnt1, PORCN and ß-catenin compared to MCAO+Veh group. Increased levels of pß-catenin and tankyrase 1 were observed in the MCAO+Veh group compared to the Sham+Veh group. The MCAO+LE 20% group had significantly decreased levels of pß-catenin and tankyrase 1 compared to the MCAO+Veh group; (n–q) Inflammatory protein expressions of experimental groups. Significantly increased expressions of inflammatory markers were observed in MCAO-injured groups compared to the Sham+Veh group. The MCAO+LE 20% group had significantly decreased inflammatory protein expression levels compared to the MCAO+Veh group. Data are presented as mean ± standard error of the mean (SEM); n = 8 for each group; * p < 0.05, ** p < 0.01 vs. Sham+Veh, # p < 0.05, ### p < 0.001 vs. MCAO+Veh, one-way ANOVA followed by Tukey’s multiple comparison test.