KDM4B is a Master Regulator of the Estrogen Receptor Signalling Cascade

The importance of the estrogen receptor (ER) in breast cancer (BCa) development makes it a prominent target for therapy. Current treatments, however, have limited effectiveness, and hence the definition of new therapeutic targets is vital. The ER is a member of the nuclear hormone receptor superfamily of transcription factors that requires co-regulator proteins for complete regulation. Emerging evidence has implicated a small number of histone methyltransferase (HMT) and histone demethylase (HDM) enzymes as regulators of ER signalling, including the histone H3 lysine 9 tri-/di-methyl HDM enzyme KDM4B. Two recent independent reports have demonstrated that KDM4B is required for ER-mediated transcription and depletion of the enzyme attenuates BCa growth in vitro and in vivo. Here we show that KDM4B has an overarching regulatory role in the ER signalling cascade by controlling expression of the ER and FOXA1 genes, two critical components for maintenance of the estrogen-dependent phenotype. KDM4B interacts with the transcription factor GATA-3 in BCa cell lines and directly co-activates GATA-3 activity in reporter-based experiments. Moreover, we reveal that KDM4B recruitment and demethylation of repressive H3K9me3 marks within upstream regulatory regions of the ER gene permits binding of GATA-3 to drive receptor expression. Ultimately, our findings confirm the importance of KDM4B within the ER signalling cascade and as a potential therapeutic target for BCa treatment.


Supplementary Figure S9. KDM4B depletion reduces ER expression in MCF-7 and BT-474 BCa cell
lines.  and  cells were transiently transfected with either scrambled (Scr) control or KDM4B (A)-(C) siRNAs (siKDM4B (A-C) for 48 hours prior to RNA extraction and quantitative PCR analysis to assess pS2 and ER expression levels, or Western analysis using antibodies to KDM4B, ER and α-Tubulin. Quantitative PCR data is a mean of three independent experiments +/-standard error.
Supplementary Figure S10. Additional KDM4B siRNAs reduce ER expression in BCa cells. A. MCF-7 cells were transiently transfected with either scrambled (Scr) control or two KDM4B siRNAs (siKDM4B (C) or (D)) in steroid-depleted media for 40 hours prior to treatment with 10 nM E 2 for 8 hours. Cells were harvested for quantitative PCR analysis to assess pS2 and GREB1 mRNA expression, or Western analysis using antibodies to ER, KDM4B and α-Tubulin. Quantitative PCR data is a mean of three independent experiments +/-standard error. B. MCF-7 and T47D cells grown in serumcontaining media were transiently transfected as in (A) and ER, KDM4B and α-Tubulin levels analysed by immunoblotting. Duplicate samples (labelled (1) and (2)) were analysed for MCF-7 cells.
Supplementary Figure S11. KDM4B depletion affects BCa cell proliferation and apoptosis. MCF-7 cells grown in steroid-depleted media supplemented with and without 10 nM E 2 and T47D cells grown in serum-containing media were transiently transfected with either scrambled control (Scr) or two KDM4B siRNAs (siKDM4B (C) or (D)) and after 96 hours WST1 proliferation (A) or BrdU incorporation assays (B) were performed. Data is the mean of three independent experiments +/standard error (* p<0.05, Student T-test).

MCF-7 and HEK293T cells. MCF-7 cells (A) or HEK293T cells (B) grown in steroid-depleted media
were transiently transfected with or without GATA-3 (50 ng/well) and increasing amounts of KDM4B (10-250ng/well) mammalian expression vectors together with either Enh1 (left panels) or Enh2 (right panels) and β-galactosidase reporter constructs. After 48 hours, cells were harvested for luciferase and β-galactosidase activities and data is representative of N=3 experiments +/-standard error (*denotes p<0.05). Figure S14. KDM4B controls ER gene expression independently of the ER. MCF-7 (A) and T47D (B) cells were transiently transfected with either scrambled (Scr) control, two KDM4B siRNAs (siKDM4B (C) and siKDM4B (D)), or an ER oligonucleotide (siER) as a positive control, in steroid-depleted media for 40 hours prior to treatment with 10 nM E 2 , supplemented with and without 1 µM 4-hydroxy-Tamoxifen (Tam), for 8 hours followed by RNA extraction. Resultant cDNA was incorporated into quantitative PCR analysis to examine pS2 and ER expression. Data is the mean of three independent experiments +/-standard error.

Supplementary Figure S15. Fulvestrant treatment does not affect KDM4B recruitment to ER gene
cis-regulatory elements. A. MCF-7 cells were grown in steroid-depleted media for 48 hours prior to vehicle or 1 µM fulvestrant pre-treatment for 195 minutes and then treated with or without 10 nM E 2 for an additional 45 minutes before ChIP using an anti-ER antibody or isotype control IgG.
Recruitment to ER target genes pS2 and GREB1 was analysed by quantitative PCR. Data is an average of three independent experiments +/-standard error. B. Cells treated as above were subject to Western analysis using ER, KDM4B and α-Tubulin antibodies. C. MCF-7 cells grown in steroiddepleted media for 48 hours were treated as in (A) with vehicle or 1µM fulvestrant for 195 minutes and then stimulated with 10 nM E 2 for 45 minutes prior to ChIP using an anti-KDM4B antibody or isotype control. KDM4B enrichment is shown at Pro A, Enh1 and Enh2 regions of the ER gene. Data is the mean of three independent experiments +/-standard error.

Supplementary Figure S16. EI-MCF-7 cells show altered requirement for KDM4B at ER gene cis-
regulatory elements. EI-MCF-7 cells grown in steroid-depleted media were transiently transfected with either scrambled (Scr) or KDM4B siRNAs for 72 hours prior to treatment with 10 nM E 2 for 45 minutes and subsequent ChIP analysis using an antiH3K9me3 antibody or isotype control. H3K9me3 enrichment is shown at Pro A, Enh1 and Enh2 regions of the ER gene. Data is the mean of three independent experiments +/-standard error.