The Dual Function of KDM5C in Both Gene Transcriptional Activation and Repression Promotes Breast Cancer Cell Growth and Tumorigenesis

Abstract Emerging evidence suggested that epigenetic regulators can exhibit both activator and repressor activities in gene transcriptional regulation and disease development, such as cancer. However, how these dual activities are regulated and coordinated in specific cellular contexts remains elusive. Here, it is reported that KDM5C, a repressive histone demethylase, unexpectedly activates estrogen receptor alpha (ERα)‐target genes, and meanwhile suppresses type I interferons (IFNs) and IFN‐stimulated genes (ISGs) to promote ERα‐positive breast cancer cell growth and tumorigenesis. KDM5C‐interacting protein, ZMYND8, is found to be involved in both processes. Mechanistically, KDM5C binds to active enhancers and recruits the P‐TEFb complex to activate ERα‐target genes, while inhibits TBK1 phosphorylation in the cytosol to repress type I IFNs and ISGs. Pharmacological inhibition of both ERα and KDM5C is effective in inhibiting cell growth and tumorigenesis. Taken together, it is revealed that the dual activator and repressor nature of an epigenetic regulator together contributes to cancer development.

(B) Kaplan Meier survival analyses for OS (overall survival) of ER-negative breast cancer patients using KDM5C as input (n=173).
(F) MCF7 cells were transfected with siCTL or a second independent siRNA specific against KDM5C (siKDM5C-2) in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M) for 24 hrs followed by FACS analysis.
(I) MCF7 cells infected with shCTL or shKDM5C-2 were subjected to immunoblotting (IB) using antibodies as indicated.

Figure S2. KDM5C activates estrogen/ER-target genes.
(A) Correlation of KDM5C's effects on estrogen response between two RNA-seq biological repeats.
(B, C) UCSC Genome browser views of RNA-seq as described in Fig. 2A for two estrogen/ERtarget genes, PGR (B) and NRIP1 (C), were shown.
(E) MCF7 cells were transfected with siCTL or a second independent siKDM5C (siKDM5C-2) in stripping medium for three days, and treated with or without estrogen (E2, 10 -7 M, 6 hrs), followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-induced genes as indicated.
(H) MCF7 cells were infected with shCTL or a second independent shKDM5C (shKDM5C-2) lenti-virus in stripping medium for three days, and treated with or without estrogen (E2, 10 -7 M, 6 hrs), followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-induced genes as indicated.
(K) WT and KDM5C (sgRNA-2) MCF7 cells were maintained in stripping medium for three days, and treated with or without estrogen (E2, 10 -7 M, 6 hrs), followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-induced genes as indicated.

Figure S3. KDM5C represses type I IFNs and ISGs expression.
(A) Correlation of KDM5C's effects on gene expression between two RNA-seq biological repeats.
(E) MCF7 cells were transfected with siCTL or a second independent siKDM5C (siKDM5C-2) for three days, followed by RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated.
(G) MCF7 cells were infected with shCTL or a second independent shKDM5C (shKDM5C-2) lenti-virus for three days, followed by RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated.
(I) Wild type (WT) and KDM5C knockdown (KDM5C (sgRNA)) MCF7 cells were subjected to RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated.
(K, M) T47D (K) and BT474 (M) cells were transfected with siCTL or siKDM5C for three days, followed by RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated.
(B, C) UCSC Genome browser views of KDM5C, ER, H3K4me1, H3K4me2, H3K4me3, H3K27Ac, P300, H3K9me3 and H3K27me3 ChIP-seq in the presence and absence of estrogen on selected active enhancer regions in the vicinity of estrogen-induced target genes, SIAH2 (B) and P2RY2 (C). Boxed regions indicated active enhancers. (A) MCF7 cells were subjected to cellular fractionation. Both nuclear and cytosolic fractions were subjected to immunoblotting (IB) analysis using antibodies as indicated. PARP1 and HSP60 were served as purity control for nuclear and cytosolic fractions, respectively.
ZMYND8 binding was examined on enhancer (E) regions nearby estrogen-target genes as indicated. Data presented was fold change of ChIP signals over siCTL after normalized to input (± SD, **P < 0.01, ***P < 0.001).
(F) MCF7 cells were subjected to immunoprecipitation (IP) with anti-P300 antibody followed by immunoblotting (IB) analysis as indicated.
(I) MCF7 cells transfected with siCTL or siKDM5C were treated with estrogen (E2, 10 -7 M, 1 hr) followed by ChIP with anti-P300 specific antibody. P300 binding was examined on enhancer (E) regions nearby estrogen-target genes as indicated. Data presented was fold change of ChIP signals over siCTL after normalized to input (± SD, **P < 0.01, ***P < 0.001).  (D) MCF7 cells transfected with siCTL or siZYMND8 were subjected to immunoblotting (IB) analysis. GAPDH was served as a loading control.
(E) MCF7 cells transfected with control siRNA (siCTL) or a second independent siRNA targeting ZMYND8 (siZYMND8-2) were treated with or without estrogen (E2, 10 -7 M, 6 hrs) followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-target genes as indicated.
(G) MCF7 cells transfected with siCTL or a second independent siRNA targeting ZMYND8 (siZYMND8-2) were subjected to immunoblotting (IB) analysis. GAPDH was served as a loading control.
(H, J) T47D (H) and BT474 (J) cells were transfected with siCTL or siZMYND8 in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M, 6 hrs) followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-target genes as indicated.
(L) MCF7 cells were transfected with siCTL, siKDM5C and siZMYND8 alone or in combination in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M, 6 hrs) followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-target genes as indicated. (± SD, *P < 0.05, **P < 0.01, ***P < 0.001).
(D) MCF7 cells were infected with siCTL or a second independent siZMYND8 (siZMYND8-2) for three days, followed by RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated.
(F) T47D cells were transfected with siCTL or siZMYND8 for three days, followed by RNA extraction and RT-qPCR analysis to examine the expression of selected type I IFNs and ISGs as indicated. Figure S8. ZMYND8 is required for the growth of ER-positive breast cancer.
(A) MCF7 cells were transfected with siCTL or a second independent siZYMND8 (siZYMND8-2) in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M) for different duration as indicated followed by cell proliferation assay (± SD, ***P < 0.001).
(B, C) T47D (B) and BT474 (C) cells were transfected with siCTL or siZYMND8 in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M) for different duration as indicated followed by cell proliferation assay (± SD, *P < 0.05, **P < 0.01, ***P < 0.001). (B) MCF7 cells were transfected with control siRNA (siCTL) or siRNA specific against KDM5C (siKDM5C) in stripping medium for three days, and treated with or without estrogen (E2, 10 -7 M, 1 hr) followed by H3K4me3 ChIP analysis. The occupancy of H3K4me3 was examined for selected active enhancer regions in the vicinity of estrogen-induced genes as indicated.
(C) Corresponding values (± SD) for heat map shown in (B).
(E) Cells as described in (B) were subjected to immunoblotting (IB) analysis to examine the levels of H3K4me3 and histone H3.
(F) Corresponding values (± SD) for heat map shown in Fig. 6B.
(G) MCF7 cells were transfected with siCTL or siKDM5C in the presence or absence of control vector, KDM5C (WT)-R or KDM5C (H514A)-R in stripping medium for three days, and then treated with or without estrogen (E2, 10 -7 M, 6 hrs) followed by RNA extraction and RT-qPCR analysis to examine the expression of selected estrogen-target genes as indicated. R, siRNAresistant.
(H) Corresponding values (± SD) for heat map shown in (G) (± SD, ns, non-significant, *P < 0.05, **P < 0.01, ***P < 0.001).    KDM5C activates estrogen/ER-target genes and represses type I IFNs and ISGs via distinct mechanisms, with KDM5C binds to ER-bound active enhancers and activates estrogen/ERtarget genes directly in an enzymatic-independent manner, while represses type I IFNs and ISGs indirectly in an enzymatic-dependent manner through inhibiting TBK1 phosphorylation.
KDM5C's dual activities in gene transcriptional regulation together promote breast cancer cell growth and tumorigenesis. Table Legend   Table S1. Proteins associated with KDM5C in both cytosol and nucleus identified by mass spectrometry. Table S2. Sequence information for all qPCR primers used in the current study. Sequence information of qPCR primers designed to detect gene expression (mRNA) or factor binding on chromatin (ChIP) were shown. F: forward; R: reverse.             Figure S12