Inducing and exploiting vulnerabilities for the treatment of liver cancer

Abstract

Liver cancer remains difficult to treat, owing to a paucity of drugs that target critical dependencies^1,2; broad-spectrum kinase inhibitors such as sorafenib provide only a modest benefit to patients with hepatocellular carcinoma³. The induction of senescence may represent a strategy for the treatment of cancer, especially when combined with a second drug that selectively eliminates senescent cancer cells (senolysis)^4,5. Here, using a kinome-focused genetic screen, we show that pharmacological inhibition of the DNA-replication kinase CDC7 induces senescence selectively in liver cancer cells with mutations in TP53. A follow-up chemical screen identified the antidepressant sertraline as an agent that kills hepatocellular carcinoma cells that have been rendered senescent by inhibition of CDC7. Sertraline suppressed mTOR signalling, and selective drugs that target this pathway were highly effective in causing the apoptotic cell death of hepatocellular carcinoma cells treated with a CDC7 inhibitor. The feedback reactivation of mTOR signalling after its inhibition⁶ is blocked in cells that have been treated with a CDC7 inhibitor, which leads to the sustained inhibition of mTOR and cell death. Using multiple in vivo mouse models of liver cancer, we show that treatment with combined inhibition of of CDC7 and mTOR results in a marked reduction of tumour growth. Our data indicate that exploiting an induced vulnerability could be an effective treatment for liver cancer.

Extended Data Figure 1. Upregulation of CDC7 mRNA correlates with poor prognosis of HCC patients and TP53 knockdown sensitizes TP53 wild-type liver cancer cells to CDC7 inhibitor.

a, 38 common hits (among top 50 most strongly depleted hits in each cell line) were identified by CRISPR screen in Hep3B and Huh7 cells. Hits in red represent factors targetable with small molecule compounds. Blue font represents non-targetable hits. b, Western blot analysis of CDC7, p-MCM2 and MCM2 levels in non-transformed cell lines and liver cancer cell lines with HSP90 serving as a loading control. c, Immunohistochemical analysis showing increased CDC7 expression in HCC tissues compared to paired adjacent non-tumour tissues. d, According to the level of CDC7 mRNA obtained from the TCGA database (n=365), HCC patients were classified into 3 groups: the top 33.3% were considered as high expression, the medium 33.3% were considered as intermediate expression and the lowest 33.3% were considered as low expression. Kaplan-Meier curves depicting that upregulation of CDC7 mRNA correlates with poor prognosis of HCC patients. Statistical significance was calculated using a two-sided Log-rank test. e, TP53 wild-type liver cancer cell lines (SK-Hep1 and Huh6) were stably transduced with control pLKO vector or with three independent shRNAs targeting p53 (shp53#1, shp53#2, shp53#3). Based on knockdown efficiency, shp53#1 and shp53#3 were selected for further experiments. f, SK-Hep1 and Huh6 expressing a control shRNA (pLKO) or knockdown of p53 (shp53) were exposed to the indicated concentrations of XL413 in colony formation assays. Cells were fixed, stained and photographed after 10-14 days of culture. g, Control or shp53 expressing SK-Hep1 and Huh6 were exposed to the indicated concentrations of XL413 for 5 days. CellTiter-Blue viability assays revealed that TP53 knockdown synergizes with XL413 in SK-Hep1 and Huh6 cells (Graphs represent mean ± s.d. from six technical replicates). For gel source images, see Supplementary Figure 1. Data in a-b, e-g are representative of three independent biological experiments. Data in c are representative images from immunohistochemical analyses using a tissue microarray (TMA) containing 80 HCC specimens.

Extended Data Figure 2. CDC7 inhibition induces senescence selectively in TP53 mutant liver cancer cells.

a, TP53 mutant liver cancer cell lines were cultured in the presence of 10 μM XL413 for 4 days, which induces senescence as detected by SA-β-gal staining. b, Growth curves measured by Incucyte live cell analyses of TP53 mutant liver cancer cell lines either untreated, continuously treated with XL413, or treated with 10 μM of XL413 for 5 or 6 days prior to treatment withdrawal (Graphs represent mean ± s.d. from five technical replicates). c, Representative images of H3K9Me3 staining in TP53 mutant liver cancer cell lines exposed to 10 μM XL413 for 7 days. d, XL413 treatment induces a senescence-associated secretory phenotype (SASP) in Hep3B and Huh7 cells treated with 10 μM XL413 for 7 days. mRNA expression of SASP genes was determined by qRT-PCR analysis (Graphs represent mean ± s.d. from four technical replicates). e, Liver cancer cells were cultured in the presence of 10 μM XL413 for 4 days and apoptotic cells were visualized by caspase-3/7 apoptosis assay. Data in a are representative of three independent biological experiments. Data in b-e are representative of two independent biological experiments.

Extended Data Figure 3. Pharmacological or genetic inhibition of CDC7 induces a senescent phenotype in TP53 mutant liver cancer cells.

a, b, TP53 mutant liver cancer cell lines (Hep3B, Huh7, SNU398, MHCC97H and HCCLM3; blue) and TP53 wild-type liver cancer cell lines (SK-Hep1 and Huh6; red) were seeded at low confluence and grown in the absence or presence of the CDC7 inhibitors LY3177833 or TAK-931 at the indicated concentrations, in long-term colony formation assays. Cells were fixed, stained, and photographed after 10-14 days of culture. c, d, Growth curves measured by Incucyte live cell analyses of TP53 mutant liver cancer cell lines (Hep3B and Huh7; blue) and TP53 wild-type liver cancer cell lines (SK-Hep1 and Huh6; red) exposed to LY3177833 or TAK-931 (Graphs represent mean ± s.d. from four technical replicates). e, f, Liver cancer cells were cultured in the presence of the CDC7 inhibitors LY3177833 or TAK-931 at the indicated concentration for 4 days. SA-β-Gal staining revealed that CDC7 inhibitors (LY3177833 or TAK-931) selectively induced senescence in TP53 mutant liver cancer cells (blue) and not in TP53 wild-type liver cancer cells (red). g, TP53 mutant liver cancer cell lines (Hep3B and Huh7) and TP53 wild-type liver cancer cell lines (SK-Hep1 and Huh6) were stably transduced with control pLKO vector or with two independent shRNAs targeting CDC7 (shCDC7#1, shCDC7#2) and the efficiency of CDC7 knockdown in liver cancer cell lines was evaluated by western blot. h, Colony formation assays of TP53 mutant (blue) and TP53 wild-type liver cancer cell lines (red) with and without CDC7 knockdown were performed. Cells were fixed, stained, and photographed after 10 days of culture. i, CDC7 knockdown induced senescence in TP53 mutant Hep3B and Huh7 cells, but not in TP53 wild-type SK-Hep1 and Huh6 cells. Senescence was detected by SA-β-Gal staining. For gel source images, see Supplementary Figure 1. Data in a-i are representative of three independent biological experiments.

Extended Data Figure 4. CDC7 inhibition leads to DNA damage accumulation specifically in TP53 mutant liver cancer cells.

a, b, Western blot analysis of liver cancer cell lines treated with CDC7 inhibitors (LY3177833 or TAK-931) for 7 days. CDC7 inhibition induces the expression of the DNA damage marker γH2AX in TP53 mutant liver cancer cells while lower γH2AX together with functional upregulation of p53 and p21^cip1 were observed in TP53 wild-type liver cancer cell lines post-XL413 treatment. c, Representative neutral comet assay images of TP53 mutant (Hep3B and Huh7) and TP53 wild-type (SK-Hep1 and Huh6) liver cancer cell lines treated with XL413 for 7 days. d, Heatmap displays log₂ fold gene expression changes in TP53 wild type (BJ and SK-Hep1) and TP53 mutated (Huh7 and Hep3B) cells upon XL413 treatment (10 μM, 4 days). e, Gene set enrichment analysis (GSEA) was performed in RNA sequencing data from Hep3B, Huh7, SK-Hep1 and BJ cells treated with 10 μM XL413 for 4 days and identified DNA repair signatures (Recombinational Repair and Fanconi Anemia Pathway) to be significantly different in TP53 mutant and TP53 wild-type cells (See methods for more detailed information). f, Neutral comet assays were performed on SK-Hep1 and Huh6 cells treated with 20 μM XL413 combined with AZD6738 (ATR inhibitor, 2.5 uM) or MK-8776 (CHK1 inhibitor, 2.5 uM) for 3 days. The value of tail moments in each treatment group were normalized based on the mean value of the control cells (n=50 per cell line/condition, graphs represent mean ± s.d., analysed with unpaired two-sided Student t-test. g, h, H2B-GFP Hep3B and Huh7 cells were cultured in absence or presence of XL413 (10 μM) and time-lapse microscopy was performed over 96 h to measure the length of mitosis (Graphs represent mean ± s.d., n=30 per cell line/condition, analysed with unpaired two-sided t-test). i, Mouse liver cancer cell lines with different genetic background (Nras^G12V;Myc^OE;Trp53^−/− and Nras^G12V;Myc^OE;Cdkn2a^ARF−/−) were exposed to the indicated concentrations of CDC7 inhibitors (XL413, LY3177833 or TAK-931) for 7 days in colony formation assays. j, Western blot analysis of mouse liver cancer cell lines treated with XL413, LY3177833 or TAK-931 for 7 days. For gel source images, see Supplementary Figure 1. Data in a-c, f, g, h are representative of two independent biological experiments. Data in i, j are representative of three independent biological experiments.

Extended Data Figure 5. CDC7 inhibition induces senescence selectively in TP53 mutant cancer cells.

a, TP53 mutant lung cancer cell lines (blue) and TP53 wild-type lung cancer cell lines (red) were seeded at low confluence and grown in the absence or presence of XL413 at the indicated concentration for 10-14 days in colony formation assays. b, Lung cancer cells were exposed to 10 μM XL413 for 4 days, which induces senescence selectively in TP53 mutant cells as detected by SA-β-gal staining. c, p53 expression was assessed in isogenic TP53^-/- and TP53^+/+ HCT116 colon cancer cell lines by western blot. d, HCT116 TP53^+/+ and HCT116 TP53^-/- cells were seeded at low confluence and grown in the absence or presence of XL413 at the indicated concentration for 7 days in a colony formation assay to assess their proliferation capacity. e, HCT116 TP53^+/+ and HCT116 TP53^-/- cells were cultured in the presence of 10 μM XL413 for 4 days, and senescence was selectively induced in TP53^-/- HCT116, as detected by SA-β-Gal staining. For gel source images, see Supplementary Figure 1. Data in a-e are representative of two independent biological experiments.

Extended Data Figure 6. Sertraline selectively induces apoptosis in XL413-induced senescent cells through suppression of mTOR signalling.

a, Schematic outline of the GPCR compound screen. Huh7 cells were treated with 10 μM XL413 for 5 days prior to seeding in 96-well plates. All compounds were tested at 4 different concentrations for 6 days and cell viability was measured using CellTiter-Blue assay. b, c, Graph depicting the effects of compounds on cell viability. Each point represents a single compound, with % activity calculated by dividing the cell viability score in the presence of 5 μM of that compound by the mean viability of the negative control. Blue dots indicate compounds inducing cell death in both control and XL413-induced senescent cells. Sertraline (red dot) induced selective cell death in XL413-induced senescent cells. Representative images of the effects of different compounds on XL413-treated and untreated cells were shown. d, Control cells and XL413-induced senescent cells were sequentially cultured with increasing concentrations of sertraline for 48 h and apoptotic cells were visualized by caspase-3/7 apoptosis assay. e, Control and XL413-treated cells were sequentially exposed to 10 μM sertraline and growth curves were measured by Incucyte live cell assay (Graphs represent mean ± s.d. from three technical replicates). f, Control and XL413-treated cells were sequentially treated with vehicle or 10 μM sertraline for 96 h in colony formation assay. g, Control and XL413-treated Huh7 and Hep3B cells were treated with sertraline (10 μM) for 24 h and western blot analyses of the indicated mTOR signalling pathway proteins were performed. h, Hep3B and Huh7 cells were treated with 10 μM XL413 for 10 days prior to sequential sertraline treatment (10 μM, 24 h) and RNA sequencing was performed. GSEA indicates that the gene set related to downregulation of mTOR signalling was negatively enriched in liver cancer cells sequentially treated with XL413 and sertraline (See methods for more detailed information). i-j, TP53 mutant liver cancer cell lines (SNU449 and PLC/PRF/5) and lung cancer cell lines (NCI-H358 and PC9) were treated with 10 μM XL413 or vehicle for 5-7 days and sequentially exposed to increasing concentrations of AZD8055. Apoptotic cells were visualized by caspase-3/7 apoptosis assay 96 h post-AZD8055 treatment. k, TP53 mutant Hep3B and Huh7 cells were treated with 10 μM XL413 or vehicle for 7-10 days. Control cells and XL413-induced senescent cells were plated and exposed to increasing concentrations of the mTORC1/2 inhibitor AZD2014 prior to caspase-3/7 apoptosis assay at 96 h. l, TP53 wild-type liver cancer cell lines (SK-Hep1 and Huh6) were treated with 10 μM XL413 or vehicle for 5-7 days prior to exposure to increasing concentrations of AZD8055. Apoptotic cells were visualized by caspase-3/7 apoptosis assay 96 h post-AZD8055 treatment. m, Control cells and XL413-induced senescent cells were treated with AZD2014 for 48 h. Western blot analysis was performed with the indicated antibodies (left panel) and the levels of p-S6RP and p-4EBP1 was normalized to total-S6RP and total-4EBP1, respectively (right panel), showing that AZD2014 treatment leads to strong mTOR signalling inhibition in XL413-induced senescent cells. For gel source images, see Supplementary Figure 1. Data in a-f are representative of three independent biological experiments. Data in h-m are representative of two independent biological experiments.

Extended Data Figure 7. The Receptor Tyrosine Kinase feedback activation induced post-AZD8055 treatment is disrupted in XL413-induced senescent cells.

a, Control cells and XL413-treated Hep3B cells were treated with AZD8055 for 48 h and extracted proteins were analysed using a Human Phospho-Receptor Tyrosine Kinase Array Kit (left panel). The levels of p-RTK proteins were normalized to positive controls (right panel). b, Activation of Receptor Tyrosine Kinases (RTKs) identified by RTK arrays and phosphorylation of SHP2 were validated by western blot analyses. c, Hep3B cells were treated with AZD8055 for 48 h prior to mRNA extraction and quantification of the indicated Receptor Tyrosine Kinase genes were performed by qRT-PCR analysis (Graph represents mean ± s.d. from three technical replicates). d, Hep3B cells were treated with AZD8055, and cell lysates were collected at the indicated time points to perform western blot analyses with the indicated antibodies. e, Hep3B cells were exposed to increasing concentrations of the AKT inhibitor (MK-2206) in combination with AZD8055 and long-term colony formation assays were performed, revealing the synergistic effects of these compounds on cell viability. f, Hep3B cells were treated with AZD8055, MK-2206 or the combined compounds at the indicated concentrations for 5 days and apoptotic cells were visualized by caspase-3/7 apoptosis assay. For gel source images, see Supplementary Figure 1. All experiments shown, except for the RTK array analyses, are representative of two independent biological experiments.

Extended Data Figure 8. AZD8055 does not induce apoptosis in Cisplatin or Alisertib-induced senescent cells

a, BJ/ET/Ras^V12 cells were treated with 100 nm 4-OHT for 21 days to induce senescence, as detected by SA-β-gal staining. b, Control or senescent BJ/ET/Ras^V12 cells were treated with either vehicle or 400 nM AZD8055 for 96 hours and apoptotic cells were visualized by caspase-3/7 apoptosis assay. c, Hep3B cells were cultured in the presence of Cisplatin or Alisertib (Aurora-A kinase inhibitor) for 4 days at the indicated concentrations and senescence induction was detected by SA-β-gal staining. d, Hep3B cells were treated with Cisplatin (1 μg/ml) or Alisertib (250 nM) for 4 days and subsequently exposed to vehicle or 400 nM AZD8055 for 96 hours. Apoptotic cells were visualized by caspase-3/7 apoptosis assay. e, Control cells, Cisplatin, Alisertib or XL413-induced senescent cells were treated with AZD8055, and cell lysates were collected at the indicated time points. Western blot analyses were performed with the indicated antibodies revealing that mTOR signalling feedback loop is functional in Cisplatin and Alisertib-induced senescent cells while it is efficiently inhibited in XL413-induced senescent cells. For gel source images, see Supplementary Figure 1. Data in a-e are representative of two independent biological experiments.

Extended Data Figure 9. Pro-senescence treatment combined with mTOR inhibitor suppresses tumour growth in liver cancer xenografts.

a, Representative images of γH2AX and SA-β-Gal staining performed on formalin-fixed, paraffin-embedded or frozen sections from subcutaneous Huh7 xenografts treated with vehicle, XL413, AZD8055 or combination for 12 days. b, Representative images of SA-β-Gal staining performed on frozen sections from subcutaneous SK-Hep1 xenografts treated either with vehicle or XL413 for 21 days. c, Tumour volume measurements in vehicle, XL413, AZD8055 or combination-treated mice bearing Huh7 and MHCC97H xenografts at endpoint (12 days and 22 days, respectively, for sample size see Figure 4a). One mouse in the vehicle group and one mouse in the XL413 group were excluded in the analysis, since the maximum permitted tumour volumes (2,000 mm³) were reached prior to trial endpoint. Graph shows mean ± s.e.m., analysed with two-sided unpaired Student t-test. d, e, Longitudinal tumour volume progression in Huh7 and MHCC97H tumour-bearing mice treated with vehicle or sorafenib for 16 or 22 days, revealing that sorafenib therapy displays limited efficacy in these two different xenograft models. Graph shows mean ± s.e.m. f, g, Representative images of H&E, PCNA, cleaved caspase-3 and p-4EBP1 staining performed on formalin-fixed, paraffin-embedded Huh7 and MHCC97H xenografts from mice sacrificed after the last dose of vehicle, XL413, AZD8055 or combination treatment. Data in a, b, f, g are representative of three independent biological experiments.

Extended Data Figure 10. Pro-senescence treatment combined with mTOR inhibitor suppresses tumour growth in p53-deficient, immunocompetent somatic murine models of HCC

a, Schematic representation of hydrodynamic tail vein gene delivery of the c-Myc proto-oncogene transposon system and a CRISPR-Cas9 vector targeting either Trp53 or Pten tumour suppressor, used to induce HCC 2-3 weeks post-HDTV injection (HDTVi). b, Quantification of SA-β-Gal staining performed on frozen sections from Myc^OE;Pten^KO or Myc^OE;Trp53^KO HCC 14 days post treatment with vehicle or XL413 monotherapy (Myc^OE;Trp53^KO also showed in Fig. 4f). For Myc^OE;Pten^KO analyses: vehicle, n=9 biologically independent nodules out of 3 mice; XL413, n=16 biologically independent nodules out of 3 mice. For Myc^OE;Trp53^KO, vehicle, n=41 biologically independent nodules out of 7 mice; XL413, n=81 biologically independent nodules out of 11 mice. Graph shows the mean ± s.e.m. of the number of SA-β-Gal positive cells per tumour nodule/mm². Statistics were calculated by two-sided unpaired Student t-test. c, Trial design to evaluate the efficacy of the pro-senescence treatment combined with mTOR inhibitor in Myc^OE;Trp53^KO HCC-bearing mice: Animals were monitored by weekly MRI post-HDTVi and enrolled into vehicle, XL413 (100mg/kg, daily gavage), AZD8055 (20mg/kg, daily gavage) or XL413+AZD8055 combination treatment groups at first signs of tumour development by MRI. Drugs were administered 6 days / week, and mice were sacrificed when symptomatic. Immunohistochemical (IHC) analyses confirmed Myc expression and p53 knockout in endpoint Myc^OE;Trp53^KO HCC. d, Longitudinal individual body weight curves from Myc^OE;Trp53^KO tumour-bearing mice treated with the combination of XL413+AZD8055 therapies. e, Individual tumour growth curves from mice treated with vehicle, XL413, AZD8055 or combination treatment were calculated based on MRI images from Myc^OE;Trp53^KO tumour-bearing mice. f, Myc^OE;Trp53^KO tumour volumes from HCC-bearing mice treated with vehicle (n=5, as shown in Fig. 4c), Sorafenib (n=4) or XL413 + AZD8055 (n=6) at day 0 and day 14. Graphs show mean ± s.e.m. analysed with two-sided unpaired Student t-test. g, h, Representative images of SA-β-Gal (g) and p16 (h) staining performed on frozen and paraffin-embedded sections, respectively, from Myc^OE;Trp53^KO tumour-bearing mice treated with the indicated drugs and sacrificed at the intermediate timepoint (14-16 days in time-matched treated cohorts. Quantification are shown in Fig. 4f, g). Scale bar, 50 µm. i, Myc^OE;Trp53^KO tumour-bearing mice treated with vehicle, XL413, AZD8055 or combination were sacrificed at the indicated time point post-treatment. Tumours were dissociated as single cell suspension and flow cytometry analyses were performed to determine the content of tumour-associated macrophages (CD45⁺ CD11b⁺Ly6C^-Ly6G^-), CD8 T cells (CD45⁺ CD3⁺CD19^- NK1.1^-CD8⁺) and CD4 T cells (CD45⁺ CD3⁺ CD19^- NK1.1^-CD4⁺) relative to total CD45⁺ leucocytes and cell proliferation (Ki67⁺) was determined within CD8 T cells and CD4 T cell populations. Graphs show mean ± s.e.m., analysed with two-sided unpaired Student t-test (for sample size, see methods). j, Myc^OE;Trp53^KO HCC-bearing mice were treated with XL413 (n=20) or XL413+AZD8055 combination (n=8) for 14 days. Among the XL413-treated mice, a subset (n=10) was sacrificed directly 14 days post treatment, concomitantly to the combination-treated group. The rest of the XL413-treated mice (n=10) underwent XL413-drug withdrawal for 4 days. The absolute number of senescent cells per tumour nodule were visualized by SA-β-Gal staining performed on frozen sections and quantified for each treatment group (XL413 n=60 biologically independent nodules out of 10 mice; XL413 withdrawn n=63 biologically independent nodules out of 10 mice; XL413+AZD8055 n=57 biologically independent nodules out of 7 mice). Graphs show mean ± s.e.m. analysed with two-sided unpaired Student t-test. Data in c are representative of three independent biological experiments.

Figure 1. A two-step screen identifies CDC7 as a target for senescence-inducing strategy in liver cancer.

a, Hep3B and Huh7 cells were transduced with a lentiviral kinome gRNA library and three independent replicates were cultured for 14 days (T14). gRNA barcodes from T0 and T14 samples were recovered by PCR and analysed by next generation sequencing. The y axis shows log2-fold change in abundance (ratio of gRNA frequency in T14 sample to that in T0 sample). The x axis depicts the average read-count in the T0 sample. b, 38 common hits (among top 50 most strongly depleted hits in each cell line) were identified by CRISPR screen in Hep3B and Huh7 cells. c, Heatmap indicates the effects of compounds (ten compounds targeting 14 hits identified by CRISPR screen, 5 μM for 4 day-treatments) on inducing senescence-associated β-galactosidase (SA-β-Gal) activity in non-transformed cell lines (BJ and RPE-1) and liver cancer cell lines (Hep3B and Huh7). d, CDC7 mRNA expression in paired tumour and non-tumour tissues from GSE14520 cohort (n=213) and TCGA database (n=50). Paired two-sided t-test. Data in graphs are mean ± s.d. Data in a-c are representative of three independent biological experiments.

Figure 2. CDC7 inhibition selectively induces senescence in TP53 mutant liver cancer cells.

a, Long-term colony formation assay was performed over 10-14 days on TP53 mutant liver cancer cell lines (blue), TP53 wild-type liver cancer cell lines (red), and non-transformed cell lines (purple) cultured with the indicated concentrations of XL413. b, Liver cancer cells and non-transformed cell lines were cultured in the presence of 10 μM XL413 for 4 days and induction of senescence was assessed by staining for SA-β-Gal activity. c, GSEA (gene set enrichment analysis) of senescence signature (FRIDMAN_SENESCENCE_UP) in sequence data from Hep3B and Huh7 cells treated with 10 μM XL413 for 4 days (See methods for more detailed information). d, Proteins were extracted from TP53 mutant and wild-type liver cancer cell lines treated with XL413 (10 μM) for 7 days and analysed by western blotting. e, Neutral comet assays were performed on Hep3B, Huh7, SK-Hep1 and Huh6 cells cultured with 10 μM XL413 for 7 days. The value of tail moments in each treatment group were normalized based on the mean value of the control cells. (n=50 per cell line/condition, data in graphs are mean ± s.d. and analysed by unpaired two-sided t-test). f, GSEA in TP53 mutant (Hep3B and Huh7) and TP53 wild-type (SK-Hep1 and BJ) cells treated with 10 μM XL413 for 4 days (FA=Fanconi Anemia, DSB=double strand break, HR=homologous recombination). For gel source images, see Supplementary Figure 1. Data in a, b are representative of three independent biological experiments. Data in d, e are representative of two independent biological experiments.

Figure 3. AZD8055 selectively triggers apoptosis in XL413-induced senescent cells.

a, Hep3B and Huh7 cells were treated with 10 μM XL413 or vehicle for 10 days prior to sequential exposure to increasing concentrations of AZD8055 for 5-7 days in colony formation assays. b, Apoptotic cells were determined by caspase-3/7 apoptosis assay 96 h after AZD8055 treatment. c, Control cells and XL413-induced senescent cells were treated with AZD8055 for 48 h prior to western blot analyses with the indicated antibodies. d, Control cells and XL413-induced senescent cells were treated with AZD8055, and cell lysates were collected at the indicated time points prior to western blot analyses with the indicated antibodies. e, f, Long-term colony formation and caspase-3/7 apoptosis assays showing the synergistic effect of mTOR and SHP2 inhibitors on Hep3B proliferation. g, Hep3B cells were treated with AZD8055, SHP2 inhibitor or the combination of both drugs at the indicated time points prior to western blot analysis with the indicated antibodies. For gel source images, see Supplementary Figure 1. Data in a-f are representative of three independent biological experiments. Data in g are representative of two independent biological experiments.

Figure 4. CDC7 inhibition induces senescence in vivo and suppresses tumour growth when combined with mTOR inhibition in multiple liver cancer models.

a, Huh7 and MHCC97H cells were grown as tumour xenografts in BALB/c nude mice. Longitudinal tumour volume progression in Huh7 and MHCC97H tumour-bearing mice treated with vehicle, XL413, AZD8055 or combined therapies for 12 or 22 days, respectively. Graph shows mean ± s.e.m. b-g, Analyses of combination therapy response in the Myc^OE;Trp53^KO HCC somatic murine model. b, Representative MRI images out of 9 independent experimental cohorts, of day 0 and day 14 of mice enrolled in vehicle, XL413, AZD8055 or combination treatment. The yellow line indicates the visible tumour area used to calculate the tumour volume. c, Tumour volumes were calculated based on MRI images from HCC-bearing mice with matched initial tumour volume. Graph shows mean ± s.e.m. from mice treated with vehicle (n=5), XL413 (n=8), AZD8055 (n=11) or combination (n=8) at the intermediate time point post-treatment initiation (day 14-16 in matched treatment groups, unpaired two-sided t-test). d, Survival curve generated from Myc^OE;Trp53^KO-tumour bearing mice treated with vehicle (n=11; median survival 17 days), XL413 (n=10; median survival 22.5 days), AZD8055 (n=11; median survival 20 days) or combination (n=11; median survival 33 days). e, Survival curve generated from independent cohorts of Myc^OE;Trp53^KO-HCC bearing mice treated with sorafenib (n=4; median survival 19.5 days) or combination of XL413 and AZD8055 (n=8; median survival 41.5 days). Vehicle group from Fig. 4d is used as a reference. d, e, Statistical significance was calculated using a two-sided Log-rank test. f, g, Graphs show mean ± s.e.m. of number of SA-β-Gal⁺ (f) or p16⁺ (g) cells per tumour nodule/mm² (unpaired two-sided t-test. For sample size, see methods).