Figure S1. FAK Silencing Effects on Cell Cycle Regulators. (A) Relative Mrna Expression

SupplementaryFigures

Figure S1. FAK silencing effects on cell cycle regulators. (a) Relative mRNA expression of FAK and CDH1 genes as measured by quantitative real-time PCR (qRT-PCR) in Control and shFAKHCC cells. Values are expressed as fold mean ± SD (***, P < 0.001; n = 3). (b)Representative western blots for FAK and E-Cadherin in Control and shFAKHCC cells. GAPDH is reported as a loading control(n = 3).(c)Growth curve inshFAK and Control HCC cells. Values are plotted as mean ± SD(*, P < 0.05; **, P < 0.01; n = 3). (d) Fold change (mean ± SD)of soft-agar colony counts for Control and shFAKHCC cells cultured for 28 days (**, P < 0.01;***, P < 0.001; n = 4). (e)Quantification of proteins reported in Figure 1d, expressed as arbitrary units. Values are mean ± SD (*, P < 0.05; **; P < 0.01; ***, P < 0.001; n = 3).(f)Quantification of proteins reported in Figure 1e, expressed as arbitrary units. Values are mean ± SD (*, P < 0.05; **; P < 0.01; ***, P < 0.001; n = 3).

Figure S2.FAK silencing reduces growth of HCC mouse xenografts. (a)Representative imaging of the liver at 54 days p.i from CTRL and shFAK NOD/SCID mice bearing intrahepatic tumours (black line marks tumours). (b) Histogram represents average weight of tumour from CTRL and shFAK nude mice at different time points (***, P < 0.001, n = 6). (c)Representative IF of caspase-3 expression in tumour samples from CTRL and shFAK NOD/SCID mice at 54 days p.i. (40 × magnification).(d)Relative mRNA expression of FAK and CDH1 genes as measured by qRT-PCR in tumours from CTRL and shFAK nude mice. Values are expressed as fold mean ± SD (*, P < 0.05; n = 5). (e) Representative IHC of E-Cadherin expression of tumour from CTRL and shFAK nude mice(40 × magnification).

Figure S3. FAK regulates EZH2 expression and activity. (a) Relative mRNA expression of FAK and EZH2 gene as measured by qRT-PCR after 24 hours of HCC cell transfection with an empty vector (Vector) or with a plasmid overexpressing FAK (PTK2). Values are expressed as fold mean ± SD (**, P < 0.01; ***, P < 0.001; n = 3). (b) Relative expression of miR-200b as measured by qRT-PCR in Control and shFAK HepG2 cells. Values are expressed as fold mean ± SD (***, P < 0.001; n = 3). (c) Representative IHC of EZH2 expression in intrahepatic tumour samples from CTRL and shFAK xenografts at 54 days p.i. (20 × magnification). (d) Representative IHC of H3K27me3 expression in intrahepatic tumour samples from CTRL and shFAK xenografts at 54 days p.i. (20 × magnification). (e)Representative IF of FAK/CK18/DRAQ5 (green/red/blue, left panels) and EZH2/H3K27me/DRAQ5 (green/red/white, right panels) expression in tumour and non-tumour liver tissue from a patient with T3N1M0 HCC (60 × magnification).(f)Average fluorescence intensity calculated for FAK, EZH2 and H3K27me3 in HCC samples according to regional lymph node metastasis(*, P < 0.05; **, P < 0.01).

Figure S4. Construction of primer sets to analyse binding p53-sites into EZH2 promoter region. (a)Relative expression of miR-101 as measured by qRT-PCR in Control and shFAK HepG2 cells. Values are expressed as fold mean ± SD of three independent experiments repeated in triplicate.(b)To get insight about the possible binding of p53 to EZH2 promoter, we first analyzed EZH2 promoter characteristic. Human promoter region for EZH2 gene were obtained by Ensemble [1]. Ensemble tools revealed, upstream from the EZH2 ATG, a region of 4179 bp that showed the strongest promoter activity. (c) Analysis wasfocused on the CpG islands of the 5′ regions of the genes selected, and more specifically in the promoter, revealed a region of 1566 bp that was used to design primers for Chip assay of p53 binding on EZH2 promoter. In this region, we designed 5 forward primers and 5 reverse primers. As showed by the virtual laboratory PROMO ( most of these primary pairs amplify regions which contain predicted binding sites for p53 and only one primer pair (F3/R3) is free of consensus sequences for p53 with dissimilarity <20% (highlighted in red). (d) ChIP for p53 binding to EZH2 promoter was amplified qPCR with the colored pairs of primers. Input, and anti-IgG antibody were used as controls. Input DNA was used to assess RLP30 transcript for normalization (n = 2).

Figure S5. Huh7.5 cells display a different behaviour under FAK silencing. (a) Relative mRNA expression of FAK gene as measured by qRT-PCR in Control and shFAK Huh7.5 cells. Values are expressed as fold mean ± SD (***, P < 0.001; n = 3). (b)Representative western blot for FAK protein in Control and shFAK Huh7.5 cells. GAPDH is reported as a loading control. (c) Relative mRNA expression of EZH2 gene as measured by qRT-PCR in Control and shFAK Huh7.5 cells. Values are expressed as fold mean ± SD (*, P < 0.05;n = 3). (d)Representative western blots for EZH2 in total lysate of Control and shFAK Huh7.5 cells. GAPDH is reported as a loading control. (e)Growth curve of Huh7.5 cells stably expressing FAK shRNAs or not (shFAK, Control). Values are plotted as mean ± SD(*, P < 0.05; **, P < 0.01; n = 3). (f)Percentage of apoptoticControl and shFAKHuh7.5 cells measured by Annexin V staining by flow cytometry. Values are plotted as mean ± SD(*, P < 0.05; n = 3).

Figure S6. FAK interacts with EZH2. (a) Relative mRNA expression of CDKN2A gene as measured by qRT-PCR in Control and shFAK HepG2 cells. Values are expressed as fold mean ± SD (*, P < 0.05; n = 3).Representative imaging of FAK-EZH2 interaction (n = 2).Total lysates from different amount of Control and shFAK HepG2 cells were subjected to immunoprecipitation with (b) anti-FAK antibody (Ab) followed by Western blot using anti-EZH2 antibody, or with (c) anti-EZH2 antibody (Ab) followed by Western blot using anti-EZH2 antibody. Total cell lysate, protein A (PA) and antibody alone were used as controls. GAPDH on not bound (inputs) was used as loading control(n = 2).

Figure S7.Rescue of EZH2 and FAK expression counteract the effect of FAK silencing in both HepG2 anf Huh7.5 cells (a) Relative mRNA expression of EZH2 gene as measured by qRT-PCR after 72 hours from plasmid transfection in shFAK+PCDNA3 and shFAK+EZH2 HepG2 cells. Values are expressed as fold mean ± SD (***, P < 0.001; n = 3). (b)Heatmap representation of the expression of cancer-related genes in shFAK HepG2 cells after 72 hours from plasmid transfection (shFAK+PCDNA3 vs. shFAK+EZH2 HepG2). (c) Venn diagram showing the overlapping between genes up-regulated upon FAK silencing and those down-regulated upon EZH2 over-expression. (d) Fold change (mean ± SD)of soft-agar colony counts for Control+PCDNA3, shFAK+PCDNA3 and shFAK+EZH2 HepG2 cells cultured for 28 days (*, P < 0.05; n = 3). (e)Percentage of apoptoticof shFAK+Vector and shFAK+PTK2 Huh7.5 cells measured by Annexin V and flow cytometry. Values are plotted as mean ± SD (n =2).(f)Relative mRNA expression of EZH2 and NOTCH2 genes as measured by qRT-PCR after 48 hours from plasmid transfection in shFAK+Vector and shFAK+PTK2 Huh7.5 cells (n = 2).

Figure S8. Rescue of EZH2 and FAK expression counteract the effect of FAK silencing in both HepG2 and Huh7.5 cells (a)Cell proliferation analysis in HepG2 and Huh7.5 cells after 48h of treatment with DMSO (Vehicle), 0.5μM or 1μM PND 1186. Values are mean of three independent experiments. (b) Cell cycle analysis in in HepG2 and Huh7.5 cells after 48h of treatment with DMSO (Vehicle), 0.5μM or 1μM PND 1186. Values are plotted as mean ± SD (*, P < 0.05; **, P < 0.01 vs Vehicle;n = 3). (c) Representative WB for FAK and pTyr397FAK (p-FAK) in HepG2 and Huh7.5 cells after 48h of treatment with DMSO (0), 0.5μM or 1μM PND 1186. GAPDH is reported as a loading control (n =2). c) Representative WB for cyclin D1 (cycD1) and cdk4 in HepG2 and Huh7.5 cells after 48h of treatment with DMSO (0), 0.5μM or 1μM PND 1186. -tubulin is reported as a loading control (n = 2).