NOTCH1 activation compensates BRCA1 deficiency and promotes triple-negative breast cancer formation

Abstract

BRCA1 mutation carriers have a higher risk of developing triple-negative breast cancer (TNBC), which is a refractory disease due to its non-responsiveness to current clinical targeted therapies. Using the Sleeping Beauty transposon system in Brca1-deficient mice, we identified 169 putative cancer drivers, among which Notch1 is a top candidate for accelerating TNBC by promoting the epithelial-mesenchymal transition (EMT) and regulating the cell cycle. Activation of NOTCH1 suppresses mitotic catastrophe caused by BRCA1 deficiency by restoring S/G2 and G2/M cell cycle checkpoints, which may through activation of ATR-CHK1 signalling pathway. Consistently, analysis of human breast cancer tissue demonstrates NOTCH1 is highly expressed in TNBCs, and the activated form of NOTCH1 correlates positively with increased phosphorylation of ATR. Additionally, we demonstrate that inhibition of the NOTCH1-ATR-CHK1 cascade together with cisplatin synergistically kills TNBC by targeting the cell cycle checkpoint, DNA damage and EMT, providing a potent clinical option for this fatal disease.

Fig. 1. SB mutagenesis promotes tumourigenesis in Brca1 mammary-specific knockout mice.

a Overview of the engineered alleles in mutant mice. b Kaplan–Meier curve showing the mammary tumour-free rate for the indicated genotypes. BrWSB40 (n = 188) and BrWSB75 mice (n = 129) showed increased tumourigenesis compared to BrW control mice (n = 62). BrWSB40 versus BrW (p < 0.0001); BrWSB75 versus BrW (p < 0.0001); BrWSB40 versus BrWSB75 (p = 0.9984) by log-rank tests with GraphPad Prism. c Numbers of tumours per mouse in different groups. d IHC staining with antibodies against ER, PR and Her2. e TNBC incidence among different groups.

Fig. 2. SB-driven candidate gene identification and functional annotations.

a Venn diagram indicating CIS genes for the BrWSB and BrMSB groups. There were 40 genes shared between the two groups. b Oncoplot of the 40 genes showing their frequency in all tumour samples. c Protein–protein interaction analysis of SB-driven candidate genes using STRING online tools.

Fig. 3. Effect of Notch1 activation on cell death and tumourigenesis.

a Predicted effect of candidate genes, as indicated by their sense fraction of insertions based on the direction of the CAG promoter and the transcriptional direction of the inserted gene. Genes with a strong bias towards sense insertions are expected to be activated and might serve as oncogenic drivers. Conversely, those biased towards antisense insertions are predicted to be inactivated or yield truncated products and serve as tumour suppressors. The dashed line in the middle indicates the equal ratio of sense and antisense insertions. b Structure of Notch1 and transposon insertion sites within the Notch1 gene. Blue arrows indicate that the promoter in the transposon is in the same orientation as the host gene, and red arrows indicate different directions. Insertion frequencies are indicated by numbers. c Expression analysis of all exons of the Notch1 gene based on RNA-seq of Notch1-driven SB tumours. d Activated Notch1 protein expression analysis based on western blotting compared with non-Notch1-driven tumours. e Western blot analysis of Notch1-driven tumours (MK1370-3R) and non-Notch1-driven tumours (MK1097-5R) after shRNA knockdown. f, g Tumour volume measured at day 30 of MK1370-3RMT (f) and MK1097-5RMT (g) after shRNA-mediated knockdown of endogenous Notch1 or pan-Notch1; n = 4 biologically independent animals. The t-test was used to determine the significance of the difference between the different sets of data. h Kaplan–Meier curve showing the mammary tumours-free rate for SB mice with Notch1-driven tumours (n = 93) and non-Notch1-driven tumours (n = 152), as well as BrW (n = 62) and BrM (n = 56) control mice. Notch1-driven tumours tended to show earlier onset compared with non-Notch1-driven tumours (p < 0.0001) by the log-rank test. i Percentage of different types of ES colonies with wild-type Brca1 and knockout after treatment with 4-HT to induce Brca1 knockout in p53 mutant, ICN1-overexpressing and parental ES cell lines. Cells were inoculated with 4-HT for 3 days (to delete Brca1), followed by disassociation and replating on feeder cells with regular medium. Single ES colonies were selected 7 days later for genotyping of their Brca1 status. Data are presented as mean values ± s.d.

Fig. 4. Notch1 activation suppresses the lethality caused by BRCA1 deficiency via cell-cycle regulation.

a Overexpression of ICN1 suppressed the cell death caused by BRCA1 acute knockdown in MCF10A naïve cells. b Quantification analysis by MTT assays regarding the rescue effect of ICN1 on BRCA1 deficiency. Analysis on the 3rd day after Dox and/or lentivirus with shRNA-BRCA1 induction; n = 3 biologically independent experiments. c IF staining and foci counting d and intensity quantification e of γ-H2AX to indicate DNA damage at 48 h after BRCA1 acute knockdown with or without ICN1 overexpression; n = 17–30 independent cell measurements. f Mitotic index analysis of MCF10A cells at 1 h after 0.5 Gy gamma-irradiation treatment. A total of 20,000 cells were counted, and three repeats were used to determine the SD. g Quantification analysis of the relative M-phase portion in the different treatment groups in e; values were normalised to the no-irradiation treatment value for the individual groups; n = 3 biologically independent experiments. The p-value was determined by the t-test with one-tailed distribution. h Dynamic changes in the mitotic index at different time points after 0.5 Gy gamma-irradiation treatment; n = 3 biologically independent experiments. * indicates p < 0.05. i Pulse-chase assay to measure S-G2 and S-M progression. The yellow arrowhead indicates representative cells that progressed to G2 phase at a particular time point. j Fraction of EdU-positive cells entering G2 phase (BrdU-negative) during the chase (n = 1800). k S-M progression indicated by EdU and p-H3 double-positive cells (n = 1100). Data are presented as mean values ± s.d.

Fig. 5. Notch1 activation rescues the lethal effect caused by BRCA1 deficiency through the ATR–CHK1 axis.

a Western blotting analysis of cell-cycle checkpoint proteins at different time points after Dox administration. b Western blotting analysis of CHK1 phosphorylation after ICN1 induction after knockdown of CHK1. c MI assay of MCF10A cells when CHK1 was knocked down with/without ICN1 overexpression. Cells were collected at 1 h after 0.5 Gy irradiation treatment; n = 3 biologically independent experiments. d MTT assay evaluating the rescue effect of Brca1 acute knockdown on the 3rd day when CHK1 was knocked down with shRNA; n = 3 biologically independent experiments. e Western blotting analysis of CHK1 phosphorylation after ICN1 induction following ATR knockdown. f MI assay of MCF10A cells when ATR was knocked down with/without ICN1 overexpression. Cells were collected at 1 h after 0.5 Gy irradiation treatment; n = 3 biologically independent experiments. g MTT assay evaluating the rescue effect of Brca1 acute knockdown on the 3rd day after ATR knockdown with shRNA; n = 3 biologically independent experiments. h IP analysis indicated that ICN1 can directly bind with ATR. i, j Immunohistochemistry staining of a human TNBC patient tissue microarray for target proteins. The scatter plots indicate a positive correlation between ICN1 and p-ATR (n = 90 for i, n = 72 for j); R and p-values were obtained based on Spearman’s rank correlation coefficient R’s and probability (p) value calculator. Data are presented as mean values ± s.d.

Fig. 6. Correlation of NOTCH1 expression with human TNBC and basal-type breast cancer.

a Flowchart illustrating the strategy of TNBC correlation analysis for candidate genes. b Venn diagram showing TNBC-correlated genes obtained from TCGA and METABRIC databases. Genes labelled with red correlated positively with TNBC incidence, and genes highlighted with blue correlated negatively. c Distribution of NOTCH1 expression levels in TNBC vs. non-TNBC and basal vs. non-basal human patients. The t-test was used to determine the significance of differences between the different sets of data. d Correlation between NOTCH1 expression level and tumour subtype. Tumours were ranked based on NOTCH1 expression level, and tumour subtypes are labelled on the top. Data were obtained from TCGA (n = 403) and METABRIC (n = 1898) databases.

Fig. 7. Notch1 stimulates BRCA1-related TNBC progression by promoting EMT.

a Gene expression analysis of Notch1-driven TNBC SB tumours (n = 24) compared with non-Notch1-driven TNBC SB tumours (n = 4). b GSEA analysis reveals that upregulated genes in Notch1-driven tumours are enriched in the EMT_breast_tumour_up term. c GSEA analysis indicates that downregulated genes in Notch1-driven tumours are enriched in the EMT_breast_tumour_dn term. d Western blotting analysis of EMT markers after induction of ICN1 in MCF10A and MCF10A-BRCA1-knockdown cells. e Western blotting analysis of EMT markers after induction of ICN1 in T47D and T47D-BRCA1-knockdown cells. f Meta-analysis of BRCA1-mutant human breast tumours for EMT markers, basal markers, luminal markers and TNBC markers.

Fig. 8. Combined drug treatment of TNBC tumours.

a Effect of combined treatment of cisplatin with ATR, ATM or CHK1 inhibitors on HCC1937 cells; n = 4 biologically independent experiments. b Quantification analysis of combined drug treatment in HCC1937 cells. Cisplatin (2.5 µM), ATR, ATM or CHK1 inhibitor (0.5 µM); n = 3 biologically independent experiments. c Killing effect on TNBC cells treated with cisplatin combined with different inhibitors. The colour in the heatmap represents the relative values of the areas under the curves. d In vivo effect of cisplatin combined with ATR inhibitor (AZD6738) on tumours in the PDX model (TM00091), which carries the BRCA1 mutation and highly expresses NOTCH1. Six mice were used for each treatment group. e IHC staining of Ki67 and cleaved caspase-3 in PDX tumours after drug treatment. Five tumours from different mice were examined for each treatment group. f Quantification analysis of dead cells (arrowhead), including cleaved caspase-3-positive cells (arrow) in each treatment group. Six to ten areas from different tumours were counted. g Illustration of the ICN1 function for TNBC formation in BRCA1-defective conditions. The t-test was used to determine the significance of differences between the different sets of data. Data are presented as mean values ± s.d. * indicates p < 0.05.