Steroid hormone receptors silence genes by a chromatin-targeted mechanism similar to those used for gene activation

ABSTRACT

How genes are repressed by steroid hormones remains a matter of debate, and several indirect mechanisms have been proposed. We found that the ligand-activated progesterone receptor recruits to the promoter of downregulated genes a repressor complex composed of HP1γ, the lysine demethylase LSD1, histone deacetylases, coREST, the RNA SRA, and the ATPase BRG1. BRG1 is needed for chromatin remodeling and facilitates the deposition of linker histone variant H1.2, which compacts chromatin and hinders RNA polymerase loading and transcription. Thus, steroid hormone receptors can repress genes in ways reminiscent of those used for gene induction, namely by directly targeting factors that remodel chromatin. But while PR-dependent gene induction in T47D cells is mainly achieved by potentiating enhancer activity, repression acts at the level of gene promoters.

KEYWORDS:

• breast cancer
• chromatin remodelling
• gene repression
• histone H1
• Progesterone receptor

Eukaryotic cells adapt to environmental changes by modifying their gene expression programs via modulation of the rate of transcription of subset of genes. In eukaryotes, regulation of gene transcription is a complex process as genomic DNA is packaged within the nucleus into highly organized and dynamic chromatin, the structure of which determines the rate of RNA polymerase initiation and elongation, and thus provides an opportunity for gene regulation. In fact, many external signals impinging on eukaryotic cells activate signaling pathways converging on enzymes that modulate chromatin structural dynamics. In the case of steroid hormones, their intracellular hormone receptors play a fundamental role acting both as signaling molecules that activated multiple kinase pathways and as hormone activated transcription factors that target chromatin remodeling enzymes to specific binding sites in the genome. We have studied the action of the progesterone analog R5020 on the breast cancer cell line T47D, which express the progesterone receptor (PR). When these cells are incubated with R5020 for 6 hours, thousands of genes exhibit increased binding of PR to their enhancers and consequent increase in transcription.² However, this is only one side of the coin, as 649 genes are found repressed after hormone exposure.¹³

Till present, more attention and study has been devoted to gene activation compared to gene repression mediated by hormones. In fact, gene repression is poorly understood, with some authors even considering repression as a secondary effect of activation. It is postulated that downregulation of genes is rather a negligible process and its physiological importance has been questioned. As for the molecular mechanism, it has been proposed that downregulation occurs without interaction between the hormone receptors and the promoter or enhancer sequences of repressed genes, and also that some repressive effects could be explained by squelching of certain components of the transcriptional machinery.²¹ In the case of progestins (Pg), though the PR_B isoform is assumed to function predominantly as an activator of Pg-responsive genes, the PR_A isoform has been linked to the repressive actions of Pg.⁷ In addition, gene expression analysis performed in cell lines that express either PR_B or PR_A showed that the two PR isoforms largely regulate different subsets of genes with 6 genes uniquely downregulated by PR_B, 0 uniquely downregulated by PR_A, and 6 downregulated by both receptors.¹⁹ Thus, all these findings are controversial and the molecular mechanism of repression is not clear.

Hormone-dependent downregulation requires binding of PR and the kinases ERK and MSK1 to target genes

As mentioned above, 649 genes were downregulated after Pg treatment of T47D breast cancer cells. Gene ontology analysis revealed that these genes are associated with relevant cell functions including intracellular signaling cascades, cell proliferation, cell adhesion, and cell fate commitment. By ChIP-seq experiments, we found that PR binding sites (PRBs) are enriched close to the transcription start site (TSS) of downregulated genes in regions containing sequence motifs similar to those found in PRBs of upregulated genes, discarding the existence of specific negative responsive elements as previously proposed for estrogen receptor (ER).²²

The kinetics of PR binding analyzed by ChIP-seq after exposure to hormone for different times² showed a more rapid and transient PR binding to downregulated genes compared to upregulated genes. These findings support a direct effect of the hormone on the repression of these genes requiring a transient binding of PR to their promoters.

In the context of gene activation, PR forms a ternary complex with the active form of the kinases ERK and MSK1, which is selectively recruited to the target PRBs.²⁴ We wondered whether these kinases are also recruited to the PRBs associated to downregulated genes. ChIP experiments confirmed the recruitment of ERK1/2 as well as MSK1 to the BCAS1, KRT23, and IGFBP5, three transcriptionally repressed genes exhibiting PRBs in their promoter region. In fact, it is the “active” form of PR phosphorylated in S294 by ERK1/2 associated with the kinases ERK1/2 and MSK1, which is recruited both to activated and repressed genes. These new findings explain our previous results that inhibition of ERK or MSK signaling pathways significantly affected both up and downregulation of hormone-responsive genes.²⁴ During gene activation, MSK1 phosphorylates H3S10phos and promotes the displacement of a repressive complex,^24,25 but still the targets of MSK1 during hormonal gene repression remain to be identified.

BRG1 as part of the HP1γ–LSD1.com complex is recruited to repressed genes and needed for hormonal downregulation

We found that the ligand-activated PR recruits to downregulated genes a repressor complex composed of HP1γ, the lysine demethylase LSD1, histone deacetylases, CoREST, and the RNA SRA.¹³ The same complex is also used by the unliganded PR to keep genes silent in the absence of hormone,²⁵ in a striking case of functional convergence.

Several ATP-dependent remodeling complexes are involved in hormonal gene regulation, including the homologues of SWI/SNF complex, BAF and PBAF, which contain BRG1 and/or BRM ATPase subunits. These complexes have been shown to alter nucleosome structure and to facilitate transcription factor binding and gene activation. In the case of Pg acting on breast cancer cells, BAF has been shown to be needed for gene activation and is recruited to PRBs via an interaction with PR.²⁶ We now found that the ATPase BRG1 but not the whole BAF complex is required during Pg-dependent gene repression. Interestingly, BRG1 is recruited to repressed genes as part of the HP1γ−LSD1 complex and its remodeling activity is critical during this process. The presence of chromatin-remodelers and histone deacetylases was previously described in other repressive complexes, including SMRT, NuRD, and N-CoR, though its functional implication is still uncertain.¹⁵

We confirmed the implication of BRG1 in hormone-dependent gene repression in breast cancer cells by RNA-seq experiments. We found that 44% of the down and 26% of the upregulated genes were affected by BRG1 knockdown.¹³ The repressed genes were significantly more dependent on BRG1 than upregulated genes. Interestingly, knockdown of BRG1 affected hormone-dependent cell proliferation as well as apoptosis, highlighting its key role in these processes. In T47D cells, BRG1 is found at least in two large complexes, one associated with BAF and the other associated with HP1γ, but how the distribution of BRG1 between BAF and the HP1γ complex is controlled is still unknown. As BRG1 is phosphorylated in response to hormone and specific acetylation of BRM, the other ATPase of the BAF complex has been reported,³¹ one possibility is that post-translational modifications could influence BRG1 interactions.

Chromatin crosstalk and the role of linker histone H1.2 during hormonal gene repression

Post-translational modifications of specific histone tail residues affect transcription by altering DNA–histone interactions, and/or providing specific binding sites for proteins that affect chromatin structure. To repress gene transcription, the LSD1–coREST complex coordinates chromatin changes that are brought about by combinations of distinct chromatin-modifying enzymes. In particular, the H3K9me3 signal is important for the localization of HP1γ protein in the genome.⁸ In T47D cells, we have found an increase in the H3K9me3 signal in hormone repressed genes and we have identified SUV39H2 as the responsible methyltransferase. Thus, PR recruits the HP1γ–LSD1.com repressor complex with its associated BRG1, but what is the function of the ATPase during repression? It has been proposed that by virtue of its remodeling activity, BRG1 as part of LSD1–coREST complex facilitates access of the complex to chromatin targets.¹⁵ We studied the molecular mechanism and found that BRG1 enhanced nucleosome occupancy around the TSS of repressed genes and exhibited hormone-dependent interaction with the linker histone variant H1.2, but not with other H1 variants such as H1.3 and H1.5. BRG1, along with the HP1γ–LSD1 complex and H1.2, co-localizes at the promoter region of repressed genes in ChIP-seq experiments performed in cells exposed to hormone and BRG1 is required for H1.2 loading.¹³ Histone H1.2 deposition leads to chromatin compaction reducing access of the transcriptional machinery.

Linker histone variants are considered to be widely redundant,⁴ and we wondered about the reasons for the selective role of H1.2 in hormonal gene repression. The major differences between histone H1.2 and other somatic H1 variants are found in the C-terminal tail, where unique post-translational modifications—including acetylation—has been reported.²⁷ These differences may account for the selective binding of histone H1.2 to the bromodomain of BRG1, but mutational studies will be required to address this possibility. Depletion of H1.2 in T47D cells caused a general decrease in nucleosome spacing and cell cycle arrest in G1.²⁰ Thus, in addition to its selective binding to BRG1 H1.2 exhibits specific properties in T47D cells.

We recently found that nuclear ATP synthesis via the PARP1-PARG-NUDIX5 pathway is required for chromatin remodeling after hormone exposure.²⁸ Blocking nuclear ATP synthesis by depletion of NUDIX5 compromised both hormonal gene activation and gene repression, indicating the deposition of H1.2 by BRG1 may also depend on nuclear ATP generation.

The pioneer factor FOXA1 marks repressed PRBs

A question that remains open is how the PR along with the repressor complex discriminates PRBs associated with the repressed genes. We speculate that these sites are marked by another factor that determines the nature of the PR-associated complex by interacting with one of its subunits. Given that FOXA2 has been reported to interact with BRG1,¹¹ and that FOXA1 is important for steroid receptor dynamics,⁹ we explored its possible function in repression. We found that prior to hormone exposure, FOXA1 is significantly enriched near the PRBs that mediate hormonal repression as compared with PRBs associated to upregulated genes or to all PRBs. Knockdown of FOXA1 affected a larger proportion of downregulated than of upregulated genes (68.5% vs. 48%, respectively). Moreover, depletion of FOXA1 compromised BRG1 recruitment and prevented hormone-dependent gene repression.¹³ Thus, one possibility is that FOXA1 contributes to BRG1 targeting, though we cannot exclude additional factors. Interestingly, the activated and repressed genes sharing the HP1γ−LSD1.com repressor complex for hormonal regulation are “marked” at uninduced conditions by the unliganded receptor and FOXA1, respectively.^13,25

In conclusion, hormonal gene repression requires binding of activated PR to sequences similar to those involved in gene activation, with the distinction that repressive PRBs are near the TSS, whereas activating PRBs are preferentially in distant enhancer regions. In addition, ERK and MSK1 kinases are recruited by PR to both activated and repressed genes, though the phosphorylation target involved in repression has not been identified. The activated PR recruits to downregulated gene promoters a repressive complex composed of HP1γ, histone demethylases, histone deacetylases, the SRA RNA, and the ATPase BRG1. The BRG1 in the repressive complex increases linker histone H1.2 deposition and nucleosome occupancy, leading to chromatin compaction around the TSS that hinders RNA-pol-II loading and maintenance of PR binding (Fig. 1).

Figure 1. Model of gene repression induced by progestins in breast cancer cells. At uninduced conditions, genes are active with RNA pol II and an open chromatin configuration characterized by the presence of the pioneer factor FOXA1 and acetylated histones. Upon hormone exposure, activated PR along with the HP1γ–LSD1.com complex interacts with the ATPase BRG1 and promotes histone deacetylation, demethylation, and chromatin remodeling via BRG1, which increase nucleosome positioning and occupancy. This arrangement of nucleosomes constitutes a suitable platform for histone H1.2 binding, and thus close the target chromatin decreasing RNApol II loading and transcription.

Open questions

But we are still halfway in the study of the mechanism of repression induced by hormones. We were able to understand how a subset of 150 out of 649 genes are repressed but still around 500 genes are downregulated by other mechanisms that remain to be elucidated. This subset of bona fide repressed genes are affected by knockdown of BRG1, HP1γ, and FOXA1, and showed increase in nucleosome positioning along with H1.2 deposition after hormone induction.¹³ Our studies indicate that the deacetylases HDAC1 and HDAC2 present in other repressor complexes as NURD and Sin3, and acting in concert with Polycomb repressive complexes (PRC1/PRC2), are potential candidates to complete the list of machineries responsible for hormonal repression.¹⁶ In fact, phosphorylation of HDAC1, HDAC2, and HDAC3 stimulates enzyme activity and is also required for their incorporation into the Sin3, NuRD, and CoREST corepressor complexes.^17,23 This may be one of the functions of the kinases during repression.

What is the scope of this mechanism of gene repression? We can expect that a similar mechanism could operate with the receptors of other steroid hormones. It has been shown that ER, androgen receptor (AR), and glucocorticoid receptor (GR) interact with LSD1, BRG1, and HDACs, respectively, and that depletion of these cofactors affect receptor-dependent transcription.^1,6,12,18,29 Moreover, FOXA1 has a relevant role in hormone-receptor binding to chromatin.³⁰ However, whether other components of the repressing mechanism, in particular H1.2 deposition, are used for gene repression by other steroid hormones remains to be established.

Most of our understanding of the molecular mechanisms of ER and PR signaling comes from in vitro studies with hormone-receptor positive cell lines. Estrogens and Pg act in vivo on a subset of mammary epithelial cells and relegate biological functions to paracrine factors as Receptor Activator of Nuclear factor Kappa B Ligand (RANKL) and members of the Wnt family.¹⁰ One point certainly worth to analyze is the dynamics of PR binding in normal human mammary cells. However, several factors hinder this study: (i) normal cells in culture tend to rapidly loose hormone receptors, (ii) the amount of cells is limiting, and (iii) normal cells require the three-dimensional (3D) structure of the mammary gland to maintain their identity.

Several groups reported the recruitment of PR to progesterone-target genes as RANKL, β-casein, and ID4 in HC-11 mouse mammary epithelial cells,^3,5,14 but the genomic distribution of PR in normal human mammary epithelial cells is still unknown.

Finally, given the relevance of the downregulated genes for breast cancer cell proliferation and apoptosis, the factors involved in the repression mechanism could be the potential targets for the management of hormone-dependent cancers.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

The experimental work was supported by grants from the Departament d’Innovació Universitat i Empresa (DIUiE). We acknowledge support of the Spanish Ministry of Economy and Competitiveness (SAF2013-42497-P), “Centro de Excelencia Severo Ochoa 2013–2017,” SEV-2012-0208, and ERC Synergy Grant “4DGenome” nr 609989.