Transcriptional repression of estrogen receptor alpha by YAP reveals the Hippo pathway as therapeutic target for ER+ breast cancer

Abstract

Extensive knowledge has been gained on the transcription network controlled by ERα, however, the mechanism underlying ESR1 (encoding ERα) expression is less understood. We recently discovered that the Hippo pathway is required for the proper expression of ESR1. YAP/TAZ are transcription coactivators that are phosphorylated and inhibited by the Hippo pathway kinase LATS. Here we delineated the molecular mechanisms underlying ESR1 transcription repression by the Hippo pathway. Mechanistically, YAP binds to TEAD to increase local chromatin accessibility to stimulate transcription of nearby genes. Among the YAP target genes, Vestigial-Like Protein 3 (VGLL3) competes with YAP/TAZ for binding to TEAD transcription factor and recruits the NCOR2/SMRT repressor to the super-enhancer of ESR1 gene, leading to epigenetic alteration and transcriptional silencing. We developed a potent LATS inhibitor VT02956. Targeting the Hippo pathway by VT02956 represses ESR1 expression and inhibits the growth of ER⁺ breast cancer cells as well as patient-derived tumour organoids. Moreover, histone deacetylase inhibitors, such as Entinostat, induce VGLL3 expression to inhibit ER⁺ breast cancer cells. Our study suggests LATS as unexpected cancer therapeutic targets, especially for endocrine-resistant breast cancers.

Fig. 1. VGLL3 is essential for ESR1 regulation by the Hippo pathway.

a Venn diagram showing the overlap of YAP and TEAD2 peaks between parental and LATS1/2 dKO cells. b Differentially expressed genes (DEGs) between LATS1/2 dKO and WT parental cells with statistical significance (orange/red dots, p < 0.01). Only genes that have positive YAP-TEAD ChIP peaks were included in the analyses. c qPCR analysis of VGLL family genes in organoids derived from the mammary epithelial tissue of Lats1^+/+Lats2^+/+ mice and Lats1^fl/flLats2^fl/fl mice, which both infected with Cre-encoding adenovirus. d, e MCF-7 cells ectopically expressing VGLLs were subjected to immunoblot with indicated antibodies (d) or qRT-PCR for ESR1 mRNA (e). f, g MCF-7 cells with LATS1/2 dKO, VGLL3 KO, LATS1/LATS2&VGLL3 triple knockout (tKO), or wild-type (Parental) were subjected to immunoblot with indicated antibodies (f) or qPCR for ESR1 (g), n = 3. h MCF-7 cells with LATS1/2 dKO, YAP/TAZ dKO, LATS1/2 and YAP/TAZ quadruple knockout (qKO), or wild-type (Parental) were subjected to qPCR for VGLL3 mRNA. i Wild-type or TEAD1-4 qKO MCF-7 cells expressing a vector, Flag-YAP(5SA) or Flag-YAP(5SA/S94A) were subjected to qPCR for VGLL3. j ChIP-qPCR analysis of YAP binding in the VGLL3 promoter locus between wild-type and TEAD1-4 qKO cells. k Genome track visualisation of indicated signals at the VGLL3 locus between LATS1/2 dKO (red) and parental (black) MCF-7 cells. l Activation of VGLL3 luciferase reporter by TEAD4 and YAP. m, n Total RNA extracted from MCF-7 cells treated by the indicated chemicals for 24 h were subjected to qPCR analysis for VGLL3 (n) or ESR1 (m) mRNA. o Wild-type and VGLL3 KO MCF-7 cells treated with 1 μM Entinostat or DMSO (Ctrl) for 24 h and ESR1 expression was measured by qPCR. p Growth of wild-type and VGLL3 KO MCF-7 cells in the presence of 1 μM Entinostat or DMSO (Ctrl) for 4 days was determined by cell counting. For c, e, h–p, n = 3 with mean ±  SEM. Two-sided, unpaired t-test for c, e, j; one-way ANOVA Tukey test for g–I, o–p. n.d. not detectable, n.s. not significant; **p < 0.01, ***p < 0.001; Source data are provided in the Source Data file.

Fig. 2. VGLL3 represses ESR1 transcription by binding TEAD and recruiting NCOR2.

a List of transcriptional regulators identified by VGLL3-TurboID mass spectrometry. b VGLL3 interacts with TEAD. MCF-7 cells were transiently transfected with plasmids expressing the indicated proteins. Protein–protein interaction was examined by IP-western blot using the indicated antibodies. c TEAD is required for VGLL3-induced ERα downregulation. Wild-type and TEAD1-4 qKO MCF-7 cells expressing a control vector or Flag-VGLL3 cDNA were subjected to immunoblot analysis. d YAP/TAZ are dispensable for VGLL3-induced ERα downregulation. Wild-type or YAP/TAZ dKO MCF-7 cells transduced with control vector or Flag-VGLL3 cDNA were subjected to immunoblot analysis. e Heatmap and line graph of the ChIP-seq profiles for doxycycline (dox) inducible HA-tagged VGLL3 (iHA-VGLL3), TEAD2 (iHA-TEAD2), and control vector (iHA-Ctrl) in LATS1/2 deficient cells or iHA-VGLL3 in LATS1/2 & TEAD1-4 6KO cells at the summits of iHA-VGLL3 peaks. f Scatterplots depicting the correlation between iHA-VGLL3 and iHA-TEAD2 ChIP-seq signals in LATS1/2 deficient cells. g Enrichment of the TEAD motif in VGLL3 ChIP-seq. MEME-AME Motif enrichment analysis for VGLL3 binding peaks against JASPAR CORE database (1404 profiles) and ordered by log (p-value). h The TEAD motif (JASPAR Matrix ID: MA1121.1) was enriched at VGLL3 ChIP-seq peaks. i De novo motif analysis for the VGLL3 binding peaks by MEME-Suit. j Venn diagram showing the overlap of VGLL3 (purple), TEAD2 (yellow) or YAP (blue) peaks in LATS1/2 dKO MCF-7 cells. k NCOR2 is required for VGLL3 to repress ESR1. MCF-7 cells with CRISPR-cas9 sgRNA targeting individual putative VGLL3 binding partners were transfected with Flag-VGLL3 or control vector. Immunoblot analysis was performed with the indicated antibodies. l NCOR2 is required for ESR1 repression by LATS1/2 deletion. MCF-7 cells were infected with CRISPR sgRNA targeting individual VGLL3 binding partners or in combination with sgRNA targeting LATS1/2. m VGLL3 interacts with NCOR2. MCF-7 cells expressing Flag-VGLL3, or control vector, were immunoprecipitated with Flag antibody. Western blotting for co-precipitated endogenous NCOR2 was determined. n VGLL3 mediates TEAD-VGLL3-NCOR2 complex formation. MCF7 transfected with indicated proteins and immunoprecipitated with anti-Myc-tag or IgG control antibodies were subjected to immunoblot with indicated antibodies. Asterisk indicates non-specific band. Source data are provided in the Source Data file.

Fig. 3. The Hippo-VGLL3 targets the super enhancer locus of ESR1.

a Co-enrichment of TEAD2 and VGLL3 at the ESR1 super enhancer locus. Genome track visualisation of iHA-TEAD2, iHA-VGLL3, PolII and pPolII-S2 signals at the ESR1 distal super enhancer locus for LATS1/2 dKO (red), parental (black) cells, and iHA-VGLL3 in LATS1/2 & TEAD1-4 6KO cells (blue). b Genome track comparison of iHA-VGLL3 at the ESR1 super enhancer locus upon doxycycline (dox) induction of iHA-VGLL3 at day 2, day 7 and day 14. c LATS1/2 deficiency alters chromatin status of the ESR1 super enhancer locus. Genome track visualisation of ATAC, H3K4me1, H3K27ac, H3K27me3, H3K9me3, H3K4me3, H3K36me3, H4K20me1, CTCF and IgG signals at the ESR1 distal super enhancer locus between LATS1/2 deficient (red) and parental (black) MCF-7 cells. d VGLL3 expression phenocopies the ESR1 super enhancer histone modifications associated with LATS1/2 deletion. Genome track visualisation of H3K4me1, H3K27ac, H3K27me3 and H3K4me3 signals at the ESR1 distal super enhancer locus by ectopic expression of VGLL3 (red) or control vector (black) in MCF-7 cells. e LATS1/2 deficiency decreases the interaction between ESR1 promoter and the distal super enhancer locus. In the virtual in-situ umi-4C plots, the dotted line and arrow denotes the viewpoint drawn from ESR1 promoter locus (purple) or selected genomic intervals (cyan). Domainogram colour (log₂ fold difference) are relative to the maximum profile to the presented genomic window. SE1-5 represent TEAD-VGLL3 positive peaks within the ESR1 super enhancer locus whereas ATAC1-3 are putative distal regulatory elements outside the super enhancer. f, g LATS1/2 dKO diminishes the distal interaction between the ESR1 promoter and the super enhancer locus. Quantification of the UMI-4C contact intensities between ESR1 promoter locus and eight genomic intervals (ATAC1-3 and SE1-5) in the ESR1 regulatory region. ESR1 promoter locus (f) or eight genomic intervals (g) were used as 4C baits. Error bars, estimated binomial s.d.; n.s. not significant; **p < 0.01, ***p < 0.001; Source data are provided in the Source Data file.

Fig. 4. LATS is important to maintain ESR1 expression and growth of patient-derived breast tumour organoids.

a Establishment of ER⁺ breast tumour organoids. Representative bright field image, H&E staining and ERα immunohistochemistry of tumour organoid and the matching patient biopsy tissue. Scale bar, 50 μm. b, c LATS1/2 deletion induces YAP target genes and decreases ESR1. LATS1/2 were deleted by lentivirus mediated CRISPR in tumour organoid (BTO-02). Expression of YAP-TEAD target genes, ESR1 (b), and VGLL family genes (c) were determined by qPCR. d YAP inhibits ESR1 expression in patient-derived breast tumour organoids. qPCR analysis of YAP target genes and ESR1 in BTO-02 transduced with control vector, constitutively active YAP(5SA), or the TEAD binding defective YAP(5SA/S94A). e–g ERα mediates the growth inhibitory effect of LATS deficiency in ER⁺ breast tumour organoids. Organoids BTO-02 and BTO-04 infected with lentivirus expressing ERα encoding gene or a control vector followed by further transduction of CRISPR-cas9 targeting LATS1/2 or non-specific sequence were seeded in Matrigel. Representative images of colony growth (e), quantification of cell growth of BTO-02 (f) and BTO-04 (g). Scale bar for e, 150 μm. For b–d, f, g, n = 3. Mean ±  SEM. Two-sided, unpaired t-test for b–d; one-way ANOVA Tukey test for f, g; n.d. not detectable; n.s. not significant; **p < 0.01, ***p < 0.001; Source data are provided in the Source Data file.

Fig. 5. Pharmacological LATS inhibition suppresses ESR1 transcription.

a Chemical structure of the LATS inhibitor VT02956 and its inactive analogue VT02484. b, c Inhibition of LATS kinase activity by VT02956. Activity of the LATS1 (b) and LATS2 (c) kinases were measured in the presence of increasing concentration of VT02956 or VT02484 in vitro. d VT02956 inhibits YAP phosphorylation in cells. YAP phosphorylation was measured in the presence of increasing concentration of VT02956 in HEK293A (purple curve) or 4T1 (orange curve) cells. The HTRF phospho-YAP assay was carried out as described in the Methods. For b–d, n = 2. e Time course of VT02956-induced YAP/TAZ dephosphorylation. Immunoblot of the indicated proteins in HEK293A cells treated with VT02956 or VT02484 at 2 μM for the indicated time points. Phot-tag detects reduced YAP mobility due to phosphorylation at multiple sites. f VT02956 dose-dependently induces YAP/TAZ dephosphorylation and ERα reduction. MCF-7 cells were treated with increased dose (0.1, 0.5 and 2 μM) of the LATS inhibitor VT02956 or inactive analogue VT02484 for 2 days prior to immunoblot analysis. g, h LATS inhibitor VT02956 increases the expression of YAP target genes, reduces expression of ESR1 and its target genes in MCF-7 cells. qPCR analysis of YAP/TAZ target genes (g) or ESR1 and ERα target genes (h) in breast tumour organoids BTO-02 treated with 2 μM of VT02956 or VTVT02484 for 2 days as indicated. For g, h, n = 3. i YAP/TAZ are required for VT02956-induced ERα downregulation. Wild-type (WT) or YAP/TAZ dKO MCF-7 cells were treated with 2 μM VT02956, 2 μM VT02484, or DMSO control for 2 days, and then analysed by immunoblot for ERα expression. Mean ± SEM. Two-sided, unpaired t-test for g, h; ***p < 0.001. Source data are provided in the Source Data file.

Fig. 6. Pharmacological LATS inhibition suppresses ER⁺breast cancer cell growth.

a Outline of the 3D breast tumour organoid (BTO) drug response assay. b Representative bright-field microscopic images of BTO-02. Tumour organoid cells were infected with lentivirus expressing ERα or control vector, and cultured in the presence of DMSO or VT02956. Pre-Treatment, at day 0 of compound treatment. Post-treatment, 12 days after compound treatment. Scale bar, 150 μm. c VT02956 requires LATS1/2 and YAP/TAZ to inhibit MCF-7 growth. MCF-7 cells with LATS1/2 dKO, YAP/TAZ dKO, or wild-type (WT) were treatment with 0.5 μM or 2 μM VT02956 for 4 days. Cell proliferation was measured by cell counting and normalised to DMSO treated WT cells. d VT02956 and Palbociclib show synergistic anti-cancer effect. MCF-7 cells were treated with VT02956 (2 μM), VT02484 (2 μM) or Palbociclib (0.1 μM) alone or in combinations for 9 days. The panels show colony-formation assay stained with crystal violet from three independent experiments. e, f ERα-Y537S knock-in MCF-7 cells are resistant to 4-OHT, but still sensitive to LATS deletion. Fold change of cell proliferation is presented as comparison of ‒E₂ versus +E₂ (e) or DMSO versus 4-OHT (f). g VT02956 inhibits the growth of MCF-7 cells that harbour hormone therapy resistant ESR1 mutation. MCF-7 cells with ERa Y537S or D538G knock-in were treated with 1 μM 4OHT, 0.2 μM Fulvestrant, 2 μM VT02956 and 2 μM VT02484 or DMSO for 9 days. The panel shows colony-formation assay from three independent experiments. h VT02956 inhibits the growth of T47D cells with hormone therapy resistant ESR1 mutation. T47D cells with ERa Y537S knock-in were treated with 1 μM 4OHT, 0.2 μM Fulvestrant, 2 μM VT02956 and 2 μM VT02484 or DMSO for 14 days. The panel shows colony-formation assay from three independent experiments. For c, d–h, n = 3; mean ± SEM; One-way ANOVA Tukey test; n.s. not significant; **p < 0.01, ***p < 0.001; Source data are provided in the Source Data file.

Fig. 7. A model of the Hippo-YAP-VGLL3 axis in the regulation of ESR1 expression and ER⁺ BRCA growth.

A proposed model for ERα regulation by Hippo-YAP via VGLL3 and NCOR2. Pharmacological targeting of LATS with kinase inhibitor VT02956 or VGLL3 induction with benzamide derivative histone deacetylase (HDAC) inhibitors suppresses ERα dependent tumour growth.