Elevated APOBEC3B expression drives a kataegic-like mutation signature and replication stress-related therapeutic vulnerabilities in p53-defective cells

Abstract

Background: Elevated APOBEC3B expression in tumours correlates with a kataegic pattern of localised hypermutation. We assessed the cellular phenotypes associated with high-level APOBEC3B expression and the influence of p53 status on these phenotypes using an isogenic system.

Methods: We used RNA interference of p53 in cells with inducible APOBEC3B and assessed DNA damage response (DDR) biomarkers. The mutational effects of APOBEC3B were assessed using whole-genome sequencing. In vitro small-molecule inhibitor sensitivity profiling was used to identify candidate therapeutic vulnerabilities.

Results: Although APOBEC3B expression increased the incorporation of genomic uracil, invoked DDR biomarkers and caused cell cycle arrest, inactivation of p53 circumvented APOBEC3B-induced cell cycle arrest without reversing the increase in genomic uracil or DDR biomarkers. The continued expression of APOBEC3B in p53-defective cells not only caused a kataegic mutational signature but also caused hypersensitivity to small-molecule DDR inhibitors (ATR, CHEK1, CHEK2, PARP, WEE1 inhibitors) as well as cisplatin/ATR inhibitor and ATR/PARP inhibitor combinations.

Conclusions: Although loss of p53 might allow tumour cells to tolerate elevated APOBEC3B expression, continued expression of this enzyme might impart a number of therapeutic vulnerabilities upon tumour cells.

Figure 1

TP53 silencing abrogates APOBEC3B-induced cell cycle arrest and cell death. (A) Induced APOBEC3B expression results in elevated phosphorylation of γH2AX(Ser139) and RPA32(S4/8) in 293-A3B cells. 293-GFP cells were included as a negative control and express GFP but not APOBEC3B upon dox exposure. (B) Induced APOBEC3B expression results in replication stress and arrests cells in G₂/M phase. FACS profiles in cells labelled with EdU (left panels) or anti-phospho histone H3 antibody (‘histone H3’, right panels) are shown. Fraction (%) of cells in non-active S phase are shown in the left-hand panels. Fraction (%) of anti-phospho histone H3 +ve and −ve cells are shown in the right-hand panels. (C) APOBEC3B induction drives cell inhibition in vitro, which can be rescued by TP53 silencing. Growth curves of 293-A3B and 293-A3B-p53 cells treated with DMSO and 100 or 1000 ng ml⁻¹ of dox, was measured using CellTiter-Glo reagent are shown. Cell growth was analysed on the indicated day after inducing APOBEC3B-GFP expression. Each data point represents eight replicates. Two-way ANOVA was used to calculate P-values. Error bars indicate s.d. for each measured group. (D) Induced APOBEC3B expression activates the apoptotic pathway. APOBEC3B induction results in an increase in cleaved PARP1 and phosphorylated p53 on Ser15 levels 48–72 h after dox induction in 293-A3B cells as shown by western blotting. (E) TP53 silencing abolishes p53 expression while maintaining inducible APOBEC3B and GFP expression, leading to elevated phosphorylation of H2AX (Ser139, γH2AX) and RPA32 (S4/8) in 293-A3B-p53 cells 48–72 h and also 5 days after inducing APOBEC3B with 0.1–1000 ng ml⁻¹ of dox. (F) TP53 silencing alleviates APOBEC3B-driven replication stress and G₂/M arrest in 293-A3B-p53 cells as shown by FACS profiling. Legend as per panel (B). (G) APOBEC3B-GFP induction increases genomic uracil (P=0.0009) in 293-A3B-p53 cells when compared with non-induced 293-A3B-p53 cells. Each data point represents the mean of at least six replicates. A two-sided Student’s t-test was used to estimate statistical significance. Error bars indicate s.d. for each measured groups. Method described in (Rona et al, 2016). In panels (A, B and D), APOBEC3B was induced with 1 and 100 ng ml⁻¹ of dox for 48 or 72 h, DMSO was used as control (0 ng ml⁻¹).

Figure 2

Induced APOBEC3B expression drives a kataegis-like mutation signature and leads to an increase in overall C-to-T mutation load in the TCW sequence context. (A) APOBEC3B induction causes genome-wide mutation clustering that was present only in dox-treated 293-A3B-p53 cells. Mutation clusters were measured by their IMD (upper panel). Each dot represents one mutation. y Axis represents intermutation distance, that is, the genomic distance of each mutation from the previous one. x Axis represents position in human genome. The horizontal line (1000 bp) shows the cutoff for mutations with IMD≤1000 bp. (B) Cytosine mutations present only in dox-treated 293-A3B-p53 cells (clone 1) were mainly C-to-T transitions (red). (C) Cytosine mutations in kataegic clusters (mutations with IMD≤1000 bp) were mostly C-to-T transitions (red). (D) Cytosine (C-to-G and C-to-A) mutations in dox-treated 293-A3B-p53 cells were enriched in TCW trinucleotide contexts (marked with asterisk). Panels (A–C) show clone 1 as an example; other clones shown in Supplementary Figure S5.

Figure 3

APOBEC3B expression sensitizes cells to inhibitors targeting DDR-related kinases. APOBEC3B induction in 293-A3B-p53 cells causes sensitivity to: (A) 50 μM HU (Student’s t-test, P<0.0001), (B) the ATR inhibitor VX-970 (ANOVA, P<0.0001), (C) the ATR inhibitor AZD6738 (ANOVA, P<0.0001), (D) the CHEK1 inhibitor SAR020106 (ANOVA, P<0.0001), (E) the WEE1 inhibitor AZD1775 (ANOVA, P<0.0001), and (F and G) the CHEK2 inhibitor CCT241533 (ANOVA, P<0.0001). In panels (A–F), 14-day inhibitor exposures were used with or without 1000 ng ml⁻¹ dox to induce APOBEC3B. In panel (G), the indicated concentrations were used. Cell viability was assessed using CellTiter-Glo reagent at the end of each treatment course. Each data point represents the mean of three replicates. Error bars represent s.d.

Figure 4

APOBEC3B expression sensitizes cells to PARP inhibitors. APOBEC3B induction in 293-A3B-p53 cells causes sensitivity to the clinical PARP inhibitors, olaparib (A, ANOVA, P<0.0001) and (B) BMN-673 (ANOVA, P<0.0001). Fourteen-day inhibitor exposures were used with or without 1000 ng ml⁻¹ dox to induced APOBEC3B. Cell viability was assessed using CellTiter-Glo reagent at the end of each treatment course. Each data point represents the mean of three replicates. Error bars represent s.d.

Figure 5

APOBEC3B induction in 293-A3B-p53 cells causes sensitivity to DDR inhibitor combinations. (A and B) ATR inhibitors (VX-970, 10 nM and AZ20, 100 nM) potentiate the sensitivity of cisplatin in the presence of APOBEC3B induction (ANOVA, P<0.0001 as shown). (C) ATR inhibitor (10 nM) potentiates the sensitivity of olaparib (0–1000 nM) in the presence of APOBEC3B induction (ANOVA, P<0.0001 as shown). (D) Olaparib (100 nM) potentiates the sensitivity of WEE1 inhibitor in the presence of APOBEC3B induction (ANOVA, P<0.0001 as shown). Fourteen-day inhibitor exposures were used with or without 1000 ng ml⁻¹ dox to induce APOBEC3B. Cell viability was assessed using CellTiter-Glo reagent at the end of each treatment course. Each data point represents the mean of three replicates. Error bars represent s.d.