GATA3 and MDM2 are synthetic lethal in estrogen receptor-positive breast cancers

Abstract

Synthetic lethal interactions, where the simultaneous but not individual inactivation of two genes is lethal to the cell, have been successfully exploited to treat cancer. GATA3 is frequently mutated in estrogen receptor (ER)-positive breast cancers and its deficiency defines a subset of patients with poor response to hormonal therapy and poor prognosis. However, GATA3 is not yet targetable. Here we show that GATA3 and MDM2 are synthetically lethal in ER-positive breast cancer. Depletion and pharmacological inhibition of MDM2 significantly impaired tumor growth in GATA3-deficient models in vitro, in vivo and in patient-derived organoids/xenograft (PDOs/PDX) harboring GATA3 somatic mutations. The synthetic lethality requires p53 and acts via the PI3K/Akt/mTOR pathway. Our results present MDM2 as a therapeutic target in the substantial cohort of ER-positive, GATA3-mutant breast cancer patients. With MDM2 inhibitors widely available, our findings can be rapidly translated into clinical trials to evaluate in-patient efficacy.

Fig. 1. GATA3 and MDM2 are synthetic lethal in ER-positive breast cancer.

a Lollipop plot depicting GATA3 somatic mutations and OncoKB annotation in ER+ breast cancer derived from the TCGA PanCancer Atlas and the METABRIC datasets. b Schematic representation of the project DRIVE shRNA screen data used to identify synthetic lethal interactors of GATA3. c SLIdR-derived statistical significance (-log₁₀(P)) plotted against the difference in the mean viability scores between GATA3-mutant and GATA3-wild type breast cancer cell lines for each gene knocked down in the shRNA screen. The middle lines of the boxplots indicate medians. Box limits are first and third quartiles. The whiskers extend to the range. d Viability scores of MDM2 knock-down in GATA3-mutant and GATA3-wild type cell lines. e–g Proliferation kinetics of e GATA3-mutant MCF-7 transfected with siRNA targeting MDM2 or control (see also Supplementary Fig. 1a–c), f GATA3-wild type BT-474, g GATA3-wild type MDA-MB134 transfected with siRNA targeting GATA3, MDM2, GATA3/ MDM2, or control (see also Supplementary Fig. 1d, e). h Apoptosis assay using Annexin V and propidium iodide co-staining. From left: gating strategy to define apoptotic (blue) and live (yellow) cells; percentage of apoptotic and live cells upon MDM2 silencing in MCF-7 (see also Supplementary Fig. 1f) upon silencing of GATA3 and MDM2 alone or in combination in BT-474 and MDA-MB134. Data are mean ± s.d. n ≥ 3 biologically independent replicates. Statistical significance was determined for e–g by multiple t test and for h by two-tailed unpaired Student’s t test. b was created with BioRender.com.

Fig. 2. GATA3 status determines response to MDM2 inhibitor in vitro.

a, d, e, h, i, j Proliferation kinetics of a GATA3-mutant MCF-7 under increasing dosage of idasanutlin, d control and GATA3-WT rescued MCF-7 upon 12.5 μM idasanutlin treatment, e BT-474 upon GATA3 silencing and/or treatment with 12.5 μM idasanutlin, h BT-474 upon GATA3 p.D335Gfs overexpression and/or treatment with 12.5 μM idasanutlin, i, j GATA3-mutant MCF-7 carrying a wild-type ESR1 or mutant ESR1 (p.D538G/p.Y537S) upon treatment with 12.5 μM idasanutlin. b, g, k Apoptosis assay using Annexin V and propidium iodide co-staining b upon the increasing dosage of idasanutlin in MCF-7, g upon GATA3 silencing and/or treatment with 12.5 μM idasanutlin in BT-474, k upon treatment of 12.5 μM idasanutlin in MCF-7 carrying a wild-type ESR1 or mutant ESR1 (p.D538G/p.Y537S). c, l Immunoblot showing pro- and anti-apoptotic proteins c at 6, 12 and 24 h post-treatment with DMSO, 12.5 μM and 25 μM idasanutlin in MCF-7, l at 24 h post-treatment with DMSO or 12.5 μM idasanutlin in MCF-7 carrying wild-type or mutant ESR1 (p.D538G/p.Y537S). For all the western blots, quantification is relative to the loading control (actin) and normalized to the corresponding DMSO control. f Log-dose response curve of idasanutlin in BT-474 transfected with GATA3 siRNA or control siRNA (see also Supplementary Fig. 2d). Data are mean ± s.d. n ≥ 3 biologically independent experiments. Statistical significance was determined for a, d, e, h, i, j by multiple t test and for b, f, g, k by two-tailed unpaired Student’s t test.

Fig. 3. GATA3 expression determines response to MDM2 inhibitor in vivo.

a Schematic representation of the zebrafish xenotransplantation assay. b Barplot shows the percentages of fish that harbored or did not harbor tumors upon transplantation with GATA3-silenced or control BT-474 cells pre-treated with idasanutlin or DMSO. In total, 70–100 embryos per group were analyzed over two independent experiments. c Representative confocal images of tumor formation in zebrafish injected with fluorescent tracker-labeled BT-474 cells with GATA3 siRNA or control siRNA, pretreated with idasanutlin or DMSO. d FACS analysis showing the percentage of red-tracker labeled tumor cells extracted from the embryos. Error bars represent, in total, three replicates performed over two independent experiments. Each replicate represents the pooled lysate of 20-30 fish for each condition. e Schematic illustration of the CAM assay. f Photographs of GATA3-silenced or control BT-474 cells pre-treated with DMSO or idasanutlin implanted in CAMs and grown for 4 days post-implantation. g Volume of tumors derived from the CAM experiment (n ≥ 10 tumors over three independent experiments). Values are normalized to the mean of siCTR DMSO. h Representative micrographs of BT-474 tumors extracted 4 days post-implantation. Tumoural cells (upper) were immunostained with GATA3 (middle) and the apoptotic marker cleaved caspase 3 (lower) in the different treatment conditions (see also Supplementary Fig. 3). Data are mean ± SEM n ≥ 4 biologically independent experiments. Scale bars: c 500 μm, f 1 cm and h 50 and 100 μm. Statistical significance was determined for b by two-sided Fisher’s Exact test and for d, g by two-tailed unpaired Student’s t test. a, e were created with BioRender.com.

Fig. 4. The synthetic lethality between GATA3 and MDM2 is TP53 dependent.

a Schematic representation of the regulatory feedback loop between MDM2 and p53. b Doughnut chart showing GATA3 and TP53 mutations in ER-positive breast cancer. Mutational data were derived from the TCGA PanCancer Atlas and the METABRIC datasets. c Proliferation kinetics of TP53-mutant T-47D transfected with siRNA targeting GATA3, MDM2, GATA3/ MDM2, or control (see also Supplementary Fig. 4a). d Percentage of apoptotic cells upon silencing of GATA3 and MDM2 alone or in combination in T-47D. e Proliferation kinetics of MCF-7 transfected with siRNA targeting TP53, MDM2, TP53/MDM2, or control (see also Supplementary Fig. 4B). f Percentage of apoptotic cells upon silencing of TP53 or MDM2 alone or in combination in MCF-7. g Proliferation kinetics of MCF-7 upon TP53 silencing and/or treatment with 12.5 μM idasanutlin (see also Supplementary Fig. 4c). h Percentage of apoptotic cells upon silencing of TP53 and/or treatment with 12.5 μM idasanutlin (see also Supplementary Fig. 4d). Data are mean ± s.d. n ≥ 3 biologically independent replicates. Statistical significance was determined for b by one-sided Fisher’s Exact test, for c, e, g by multiple t test, and for d, f, h by two-tailed unpaired Student’s t test. a was created with BioRender.com.

Fig. 5. GATA3 mutations predict response to MDM2 inhibitors in ER-positive breast cancer PDOs and PDX.

a Schematic representation and representative microscopy pictures of the generation of organoids from n = 3 ER-positive breast cancers (see also Supplementary Fig. 5). b Representative micrographs of H&E, ERɑ, and GATA3 immuno-staining on the PDOs (see also Supplementary Fig. 5b). c Proliferation kinetics of GATA3-wild-type PDOs (black, patient 1) and GATA3-mutant PDOs (blue, patient 2). d Log-dose response curve of idasanutlin in GATA3-wild-type (IC50 = 5.4 μM) or GATA3-mutant (1.2 μM) PDOs (see also Supplementary Fig. 5e). e Percentage of viable cells upon treatment with different dosages of idasanutlin in GATA3-wild-type (gray) or GATA3-mutant (blue) PDOs (see also Supplementary Fig. 5f). f Representative micrographs of PDOs after five days of treatment with different dosages of idasanutlin. Scale bars are 20 and 40 μm for (b) and 200 μm for (f). g Percentage of viable cells upon treatment with different dosages of RAIN-32 in GATA3-wild-type (gray) or GATA3-mutant (blue) PDOs. h Schematic representation of the PDX model and drug treatment. i Tumor growth curve of GATA3 p.D335fs PDXs (n = 6–8 mice) treated for 29 days with vehicle, fulvestrant (200 mg/kg), RAIN-32 (50 and 100 mg/kg) alone, or RAIN-32 in combination with fulvestrant. Data are mean ± SD, n ≥ 3 biologically independent replicates. Statistical significance was determined for c, e, g, i by multiple t test and for d by two-tailed unpaired Student’s t test. a, h were created with BioRender.com.

Fig. 6. The synthetic lethality between GATA3 and MDM2 acts via the PI3K-Akt-mTOR signaling pathway.

a Schematic representation of the RNA-seq experimental setup to identify gene expression changes induced by concurrent GATA3 loss and MDM2 inhibition. Venn diagram shows the number of pathways enriched in both MCF-7 with MDM2 siRNA and MDA-MB134 with GATA3 siRNA and MDM2 siRNA. b Normalized enrichment scores of significantly up- and down-regulated pathways identified by gene set enrichment analysis in both MCF-7 and MDA-MB134. The size of the dots is proportional to the adjusted p-value as indicated in the legend. c, d Immunoblot showing markers of mTOR signaling pathway activation at 24, 48, and 72 h post-siRNA transfection in c MCF-7 cells upon MDM2 silencing and d BT-474 cells upon GATA3 and/or MDM2 silencing (see also Supplementary Fig. 6a, c, d). For all the western blots, quantification is relative to the loading control (actin) and normalized to the corresponding siCTR. e Representative immunohistochemistry micrographs of phospho-S6 stainings in BT-474 tumors extracted four days post-implantation in the CAM model (see also Supplementary Fig. 6e). f Schematic representation of the mechanistic hypothesis. Scale bars: e 50 and 100 μm. Statistical significance was determined for b by fgsea. a, f were created with BioRender.com.