Identification of LncRNAs scaffolding a chromatin-associated gene regulatory complex in breast cancer cells comprising ERα, DOT1L and menin

Abstract

In Estrogen Receptor (ER)-positive Breast Cancer (BC), the development of resistance to endocrine therapies remains a significant clinical challenge, often linked to transcriptional and epigenetic deregulations. The main hallmark in this neoplasm, the nuclear receptor ERα, drives oncogenesis by recruiting multiple coregulators, including the histone methyltransferase DOT1L and the scaffold protein menin, forming a pro-tumorigenic machinery. However, the precise mechanism of this complex assembly is not fully understood. Our study investigates the molecular basis of ERα-DOT1L-menin functional association, focusing on the potential scaffolding role of long non-coding RNAs (lncRNAs).

To identify lncRNAs specifically involved in this network, we performed native nuclear immunoprecipitation coupled to RNA sequencing (RIP-Seq) for DOT1L and menin in MCF-7 BC cells. Cross comparison between these datasets and existing ERα-interacting RNA data revealed six common molecules. Among these, we focused on PVT1, a known oncogenic RNA previously characterized by our group and able to co-recruit ERα and PRC2 complex onto specific chromatin loci, whose high expression, along with ERα, DOT1L and menin, correlates with worse patients’ survival. Functional experiments confirmed that PVT1 knock down reduces ERα, menin and DOT1L association, supporting its role as a molecular scaffold in this complex. Transcriptomic profiling upon ERα, DOT1L, menin, or PVT1 blockade uncovered a common gene signature enriched in pathways covering main BC hallmarks. Many of deregulated genes were directly bound by all three proteins, suggesting a coordinated transcriptional regulation, revealing a critical axis involving ERα, DOT1L, menin and PVT1. Targeting this RNA-dependent chromatin associated regulatory complex could offer novel therapeutic strategies for BC treatment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-026-02702-9.

Main text

Among female malignancies, breast cancer (BC) represents one of the leading causes of death. As a multifactorial and heterogeneous disease, its initiation and progression are influenced by several factors, including transcriptional and epigenetic imbalances [1, 2]. About 80% of all newly diagnosed BCs are categorized as estrogen receptor (ER) positive due to the expression of ERα, a ligand-inducible transcription factor, representing the primary oncogenic driver in this neoplasm. The presence of ERα indicates tumors that are likely to respond to endocrine therapies, mainly based on estrogen synthesis blockade through aromatase inhibitors or causing its degradation using antiestrogens (selective estrogen receptor modulators SERM or selective estrogen receptor degraders SERD). Although endocrine therapies improve overall survival, approximately 30% of early-stage ER-positive BC patients develop resistance to treatment leading to metastasis formation and poor clinical outcome [3, 4]. Despite recent studies have improved our understanding in this field, further research is needed to clarify how alterations in the chromatin landscape and deregulation of epigenetic factors contribute to the occurrence of resistance. Upon estradiol (E2) stimulation, ERα undergoes conformational changes that facilitate the recruitment of hundreds of coregulators to chromatin in a highly orchestrated manner, to transduce the hormonal signal in the modulation of target gene expression. This is not a simple switch, but it happens in a highly sophisticated process involving multiple proteins and layers of regulation. The coregulators include transcriptional coactivators and corepressors that modulate chromatin structure and histone modifications, thereby enabling precise gene expression activation or repression at ERα-responsive genomic loci [5]. One of the functions mediated by this network is to dynamically modify chromatin architecture to either create or abolish a permissive environment for transcription. This is usually epigenetically regulated by “writer/reader/eraser” molecules [6–8].

ERα was described to interact, among others, with the histone methyl transferase DOT1L, and the scaffold protein MEN1 (menin), through a tightly coordinated chromatin regulatory network, sustaining estrogen signaling to induce BC progression [9, 10]. Particularly, DOT1L and menin co-localize onto specific chromatin loci regulating a broad set of genes that are central to BC cell proliferation and survival [9]. This partially explains why the pharmacological inhibition of either DOT1L or menin suppresses ERα expression and reduces BC cell proliferation [9]. In a previous work, by applying Tandem Affinity Purification (TAP) coupled with mass spectrometry (MS) following in vitro RNase digestion, it has been observed that the interaction of ERα with transcriptional coregulators is, in some cases, mediated by bridging RNAs, thus highlighting the critical role of such molecules in the nuclear assembly of ERα involving multimolecular complexes [11]. Among these, lncRNAs, known for facilitating the formation of molecular complexes and involved in several way in estrogen signaling modulation [12], may be responsible for mediating the interaction between the receptor and other molecular partners, as we have recently described through native immunoprecipitation of nuclear ERα-interacting RNAs coupled to NGS (RIP-Seq) [13]. These molecules represent a fascinating landscape as active participants in ERα functional networks, and are currently investigated as putative therapeutic targets especially in overcoming drug resistance [14].

Here, we investigated the molecular basis underlying the formation of the nuclear complex containing DOT1L, menin and ERα, particularly focusing on lncRNA molecules participating in this network, to achieve a more comprehensive view of its assembly and functional relevance.

The predicted interaction network of these three factors by using STRING tool is displayed in Fig. 1A. We noticed a reduction of DOT1L recruitment following ERα co-immunoprecipitation in the presence of RNAse treatment, (Fig. 1B). To further investigate the nature of the observed protein-protein interaction, we employed the yeast two-hybrid (Y2H) assay. Also in this case, it was evidenced that DOT1L-ERα interaction is not direct if compared with the hetero-dimerization of ERα and ERβ proteins evaluated as positive control (Fig. 1C).

Fig. 1.

ERα-DOT1L-menin network assembling. A STRING network representative image for ERα, DOT1L and menin protein-protein interaction. How the interaction has been determined is indicated into the color legend on the top of the panel. B WB showing DOT1L enrichment following ERα immunoprecipitation with or without RNase treatment (assessed by Agilent Tapestation on the left) in MCF-7 nuclear extracts. IgG was used as negative control. C Y2H assay showing protein interactions by using plasmid expressing ERα, ERβ, and DOT1L alone (all used as negative control), ERα and ERβ (used as positive control) and ERα and DOT1L in combination. Controls and interactions were tested for the growth on DO-2, DO-3 selective media with or without product. D Scatter plot (left) showing enriched RNAs (log2 + 2) identified through DOT1L and menin nuclear RIP-Seq. IgG was used as negative control. Pie charts (right) representing the enriched RNA categories (protein coding in blue and non-coding in red). E Venn diagram (up) showing overlapping lncRNAs associated with ERα, DOT1L and menin and heatmap (low) showing enrichment values of the six common lncRNAs. Each row represents one lncRNA while the columns represent the log2FC of enrichment values in menin, ERα and DOT1L RIP-Seq

Despite already described the interaction between ERα and menin is still controversial. In fact, based on biochemical evidence, it was previously observed that menin can directly bind ERα during transcription in a hormone-dependent manner [15, 16]. In the meantime, this seems to be also mediated by RNAs [11, 13]. Putting altogether, the interaction between the two proteins seems to be direct but RNA molecules may mediate or stabilize menin-ERα interaction within multicomponent nuclear complexes, such as one specifically involving DOT1L, inducing a dynamic model in which lncRNAs can act as scaffold in ERα-depending transcriptional regulation.

In order to identify possible RNA molecules involved in DOT1L-menin-ERα complex formation, we performed a native nuclear RIP-Seq in MCF-7 BC cell line. Particularly, we identified 2232 and 1189 RNAs interacting with DOT1L and menin respectively (Fig. 1D). Among others, 8.4% and 15.7% of DOT1L- and menin-interacting RNAs, respectively, were classified as non-coding molecules (Fig. 1D, Table S1). The comparisons among lncRNAs interacting with the three factors investigated allowed the identification of 6 common lncRNAs represented by AC105046.1, AL365436.2, AC027237.3, MIR4435-2HG, PVT1 and AC037198.1 (Fig. 1E).

Being already confirmed the potential of PVT1 as an oncogenic factor and its essential (fitness) role in BC cell survival [13, 17], we decided to investigate its role within this regulatory network, also taking into account that the simultaneous overexpression of all the four investigated factors is significantly associated with worse overall survival based on TCGA data (Fig. 2A). To investigate whether DOT1L-menin-ERα complex formation might be mediated by PVT1, we performed DOT1L immunoprecipitation, before and after PVT1 knockdown. This was performed in two different MCF-7 cell clones, one overexpressing both ERα and DOT1L and another overexpressing only DOT1L to have better IP efficiency due to the low expression of this factor in BC cell lines. As expected, it was observed a minor recruitment of both ERα and menin following PVT1 knock-down (Fig. 2B). Moreover, we further confirmed the enrollment of PVT1 in the investigated chromatin complex through an RNA-pulldown coupled to SDS-PAGE in MCF-7 BC cell lines (Fig. 2C). To investigate the functional significance of this network, we examined the transcriptome profiles of MCF-7 cells following either PVT1, DOT1L, menin and ERα blockade. These RNA-Seq data were previously generated and analyzed by Melone et al. [13], Nassa et al. [10], and Salvati et al. [9]. The cross-comparison displayed 50 common deregulated genes (Fig. 2D, Table S2). Of note, these genes were involved in key cancer hallmarks namely estrogen response both early and late, myogenesis, hypoxia, glycolysis, epithelial to mesenchymal transition and mTORC1/KRAS signaling (Fig. 2E).

Fig. 2.

lncRNA PVT1 is a core molecule mediating ERα-chromatin interaction network. A Kaplan–Meier curves, generated using GEPIA2 tool, depicting the probability of overall survival considering TCGA data of BRCA patients with low and high co-expression of ESR1, DOT1L, menin and PVT1. B WB showing co-immunoprecipitation of DOT1L with ERα and menin in cells (overexpressing ERα and DOT1L (left) or DOT1L (right)) after scramble or PVT1 knockdown, and the corresponding densitometric ratio on the bottom. Asterisks indicate statistically significant differences using unpaired t-test (***p < 0.001, ****p < 0.0001). C WB showing ERα, DOT1L and menin retrieved by PVT1 pulldown assay. Streptavidin beads were used as negative control. D Venn diagram displaying the uniquely and the overlapping DEgenes identified following treatment with ICI (fulvestrant), MI-2 (MEN1 inhibitor), EPZ004777 (DOT1L inhibitor), and PVT1 knockdown. E Dot plot showing enriched hallmark gene sets in genes commonly deregulated by ERα, DOT1L, menin and PVT1 blockade. Dot color indicates normalized enrichment score (k.k), and size reflects significance (–log10 FDR q-value). F Word cloud of most representative enriched transcription factors associated to the 50 common deregulated genes. Size reflects degree of enrichment. G Tile plot showing 50 common deregulated genes. Each column represents a gene; the rows indicate the genomic position of either menin, DOT1L or ERα binding site

Furthermore, upstream regulator analysis performed by using IPA tool identified potential molecules acting upstream in the regulatory network and, among them, most were known as ERα interactors or involved in estrogenic signaling modulation, such as ARID1A, HIF1A, SMARCA4 and NCOA3 (Fig. 2F) [10]. This supports the hypothesis that the observed complex can potentially cooperate with other factors to orchestrate the transcriptional regulation of key genes involved in sustaining tumor cell survival. Indeed, many of differentially expressed genes exhibit at least one ChIP-Seq peak for DOT1L, menin, and ERα, indicating direct co-occupancy of these factors on their regulatory regions (Fig. 2G).

In the intricate landscape of estrogen depending transcriptional regulation, ERα, menin, and DOT1L are key molecular players that engage in a sophisticated relationship to control target gene expression. The cooperation between these factors allows the formation of a functional complex that fine-tunes the expression of genes critical for cancer cell proliferation and survival. As predicted, ERα acts as the molecular ‘hub’, directing the menin-DOT1L complex to specific genomic locations to orchestrate transcriptional regulation [9]. This small sub-network underlines the complexity of cancer biology and presents potential avenues for therapeutic intervention by targeting the key components of the transcriptional machinery. In this context, the most likely role for lncRNAs is to act as a scaffold by physically mediating the interaction between ERα, menin, and DOT1L. In this context, the data shown here suggest that the oncogenic lncRNA PVT1 may be a critical component of this regulatory complex, ensuring that its target genes are controlled by ERα at the right time and place. Based on the known multifunctional roles of lncRNAs, other such molecules, still to be identified and characterized are likely to be enrolled in transcriptional and post-transcriptional regulation of gene expression by ERα, including its binding to chromatin and association with nascent transcripts, stabilization and translation of mRNAs [18] and multi-molecule shuttling from cytoplasm to nucleus and vice versa.

The results shown here demonstrate that targeting lncRNAs involved in chromatin regulatory complexes offers a promising strategy to reprogram the cancer epigenome, potentially re-sensitizing resistant cells to endocrine therapy or enhancing their responsiveness to novel treatments [2].

Methods

Cell culture

The human BC cell line MCF-7 (HTB-22) was purchased from the American Type Culture Collection (ATCC) and cultured according to manufacturer’s guidelines. MCF-7-FlagERα clone was generated by stably transfecting a full-length-3xFlag-ESR1 plasmid; the constructs were kindly provided by Dr. Ruggero [18]. All cell lines were authenticated by STR profiling and routinely tested for mycoplasma contamination.

Immunoprecipitation

For immunoprecipitation, 2.5 µg of anti-DOT1L or anti-ERα and Rabbit IgG Isotype Control were conjugated with 35 µl of equilibrated Dynabeads M-280 Sheep AntiRabbit IgG (Thermo Fisher, Milan, Italy), in rotation overnight at 4 °C and the experiment was performed as previously described [10]. Nuclear protein extracts were treated with 100 µg/ml RNaseA (Cat. 12091021, Invitrogen, Massachusetts, USA) and incubated 1 h at 4 °C with gentle rotation prior to binding step. For immunoprecipitation, 500 µg of nuclear protein extracts were incubated with antibody-conjugated beads for 2 h at 4 °C. Following incubation, the beads were washed sequentially with IPP150 buffer (7.14 mM HEPES pH 7.5, 8.92% glycerol, 150 mM NaCl, 0.54 mM MgCl2, 0.07 mM EDTA pH 8, 1x PIC), then with wash buffer (50 mM Tris-HCl pH 7.6, 150 mM NaCl and 1x PIC) and finally resuspended in Laemmli buffer for western blot analysis.

Yeast 2 hybrid assay

This direct 1-by-1 interaction assay was performed by Hybrigenics Services SAS, Paris, France (http://www.hybrigenics-services.com). The coding sequence of the human DOT1L fragment (aa. 1-1537, GenBank accession number gi:163965372) was PCR-amplified and cloned in frame with the LexA DNA binding domain (DBD) into plasmidnpB27 (N-LexA-DOT1L-C). The plasmid pB27 is derived from the original pBTM116 [19]. The DBD construct was checked by sequencing the entire insert. The coding sequence of the human ERα (aa 1-595, GenBank accession number gi:170295798) was PCR-amplified and cloned in frame with the LexA DNA binding domain (DBD) into plasmid pB27 (N-LexA-ER alpha-C), and with the Gal4 Activation Domain (AD) into plasmid pP13 (N-ER alpha-Gal4-C). The plasmid pP13 is derived from the original pGADGH. Both constructs were checked by sequencing. The coding sequence of the human ERβ (aa 1-530, GenBank accession number gi: 94538323) was PCR-amplified and cloned in frame with LexA DNA binding domain (DBD) into plasmid pB29 (N-ER beta-LexA -C), and with the Gal4 Activation Domain (AD) into plasmid pP13 (N-ER beta-Gal4-C). Both constructs were checked by sequencing. 1-by-1 Y2H interaction assays. Bait and prey constructs were transformed in the yeast haploid cells, YPP64T strain (TT7 strain + pdr5::loxP, snq2:: loxP, yor1:: loxP, pdr1:: loxP, pdr3:: loxP) and YHGX13 (Y187 ade2-101::loxP-kanMX-loxP, mat ) strains. The yeast strain has been optimized by deleting several genes controlling drug efflux. This allows testing the addition of products during the test and checking the effect on the interaction. The diploid yeast cells were obtained using a mating protocol with both yeast strains [20]. The assays are based on the HIS3 reporter gene (growth assay without histidine).

RNA immunoprecipitation (RIP)

For both DOT1L and menin-associated RNA immunoprecipitation (RIP), 7 µg of anti-DOT1L, anti-menin or anti-IgG Isotype Control were conjugated overnight at 4 °C with 100 µl of Dynabeads M-280 Sheep AntiRabbit IgG (Thermo Fisher). The RIP procedure was performed as previously described [13]. After the binding, beads were washed 3 times with RIP Buffer in rotation for 5 min at 4 °C, and twice quickly. Once discarded all the supernatant, 1 ml of TRIzolTM (Life Technologies, Thermo Fisher) was added directly to the beads and RNA extraction was performed according to the manufacturer’s guidelines.

RNA sequencing

Libraries preparation for transcriptome profiling was performed by using the TruSeq Stranded Total RNA Library Prep Gold (Cat. 20020599, Illumina, San Diego, California, USA) according to manufacturers’ protocol. Libraries were sequenced on Novaseq 6000 platform (Illumina) using 2 × 100 bp paired end mode.

Data analysis

RIP-Seq data analysis was performed as in Melone et al. [13]. Enriched genes where selected if they showed FC > 1. Gene Ontology was performed on MSigDB [21] using Hallmark as geneset, while the upstream regulator analysis was generated through the use of QIAGEN IPA (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA). GOPlot and expression plots were drawn with R (v. 4.4.1). STRING [22] was used to predict protein-protein interaction and GEPIA2 to perform the overall survival analysis.

Antibodies and compounds

The antibodies employed for immunoprecipitation experiments were: anti-ERα (ab3575, Abcam, Cambridge, UK), Rabbit IgG Isotype Control (31235, Thermo Fisher) anti-DOT1L (A300-954 A Bethyl Laboratories, Montgomery, Texas, USA) and anti-menin (A300-105 A, Bethyl Laboratories).

The antibodies used for western blot experiments were: anti-ERα (F-10 sc-8002, Santa Cruz Biotechnology, Dallas, Texas, USA), anti-menin (A300-105 A, Bethyl Laboratories) and anti-DOT1L (ab72454, Abcam).

Transfection experiments

MCF-7-FlagERα clone was seeded into 150 mm plates at a confluence of 80% in standard growth medium and, the day after, transiently transfected by using Lipofectamine RNAiMax (Invitrogen, Waltham, USA) in OptiMem medium (Gibco) according to the manufacturer’s instructions. For ASO-mediated knock-down, 1200 pMol/well of molecules targeting PVT1 (#231193149, IDT) and a NC (#231193151, IDT) were transfected and incubated for 72 h at 37 °C/5% CO₂. For DOT1L overexpression, plasmid containing tagged DOT1L coding sequence (Origene, Rockville, Maryland, USA) was transfected for 48 h at 37 °C/5% CO₂ using PEI reagent (Cat #23966, Polyscience, Warrington, Pennsylvania, USA) according to manufacturer protocol.

RNA pulldown

MCF-7 BC cells were UV-crosslinked with 400 mJ/cm² 254 nm UV as mentioned by Xu et al. [18]. Crosslinked cells were then harvested in ice-cold PBS and the lysed by using RIPA Buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% NP-40, 0.1% SDS, 50 mM NaF, 0.5% sodium deoxycholate, 1x protease inhibitor, 1 mM PMSF and 100 U/mL RNase inhibitor) on ice for 10 min and then centrifuged at maximum speed for 30 min at 4 °C. The supernatant, containing the protein extract, was then conjugated with 2 volumes of Hybridization Buffer (750 mM NaCl, 1% SDS, 50 mM Tris HCl pH 7, 1 mM EDTA, 15% formamide, 1x protease inhibitor, 1 mM PMSF and 100 U/mL RNase inhibitor) and incubated with the specific probes for RNA-Pulldown at 37 °C for 4 h in rotation. Five biotinylated DNA probes targeting PVT1 (Seq1 caaggaaatccactggaaggtg; Seq2 aatacttgaacgaagctccatg; Seq3 aagtaaacagagatctcaaccc; Seq4 cggatggaacgtgcatcagatg: Seq5 acaggtacaataacacacgtca) were employed by using a total of 100 pmol/mL ratio. Then, the probes plus the protein extracts were incubated with Streptavidin C-1 magnetic beads (Invitrogen) at 37 °C, 30 min with shaking to induce the binding between the streptavidin and the biotin of the probes.

Beads were collected on a magnetic rack and washed 3 times with pre-warmed wash buffer (2x NaCl and Sodium Citrate, 0,5%SDS, 1x protease inhibitor, 1 mM PMSF and 100 U/mL RNase inhibitor) to remove non-specifically bound proteins. Bound proteins were eluted by boiling beads in 2× sample buffer for 5 min at 100 °C and analyzed by SDS–PAGE followed by immunoblotting. As a control, pull-down was also performed with beads alone to assess non-specific binding.

Abbreviations

AD

Activation Domain

ARID1A

AT-Rich Interaction Domain 1A

ATCC

American Type Culture Collection

BC

Breast cancer

ChIP

Chromatin Immunoprecipitation

DBD

DNA binding domain

DOT1L

Disruptor of telomeric silencing 1-like

E2

Estradiol

ER

Estrogen Receptor

ESR1

Estrogen Receptor 1

FC

Fold Change

HIF1A

Hypoxia Inducible Factor 1 Subunit Alpha

IPA

Ingenuity Pathway Analysis

KRAS

KRAS Proto-Oncogene

MEN1

Menin 1

MS

Mass Spectrometry

mTORC1

mTOR Complex 1

NCOA3

Nuclear Receptor Coactivator 3

NGS

Next Generation Sequencing

PVT1

Plasmacytoma Variant Translocation 1

RIP

RNA immunoprecipitation

SERD

Selective Estrogen Receptor Degraders

SERM

Selective Estrogen Receptor Modulators

SMARCA

SWI/SNF Related BAF Chromatin Remodeling Complex Subunit ATPase

4

4

STR

Short Tandem Repeats

TAP

Tandem Affinity Purification

TCGA

The Cancer Genome Atlas

Y2H

Yeast two-Hybrid

Authors’ contributions

Conceptualization: R.T., G.N., A.W., V.M. and A.S.; writing and original draft preparation: V.M. and A.S.; bioinformatic and statistical analysis: D.P. and D.M.; experimental procedures: V.M., A.S., L.P., M.N., M.I., A.C., J.L.; writing, review, and editing: G.N., F.R., R.T., and A.W; funding acquisition G.N., R.T. and A.W. All the authors have approved the final version of the manuscript.

Funding

Work supported by: AIRC Foundation for Cancer Research, Grant No. IG-23068 (to A.W.), Italian Ministry of University and Research PNRR-MUR NextGenerationEU PRIN 2022, cod. 202282CMEA - CUP: D53D23007790001 (to G.N.), cod. 2022A7HJEM - CUP: D53D23005050006 (to R.T.) and cod. 2022Y79PT4– CUP: D53D23008040006 (to A.W.); PNRR-MUR NextGenerationEU PRIN-PNRR 2022 cod. P2022N28FJ - CUP: D53D23016530001 (to G.N.); Italian Ministry of Health, Young Researcher Grant cod. GR-2021-12373937 - CUP: D53D23016530001 (to G.N.) and University of Salerno, Fondi FARB (to R.T., G. N. and A.W.).

Data availability

The dataset generated and analysed during the current study is available in the ArrayExpress repository, E-MTAB-15354. RNA-Seq data were exploited from Melone et al. [13], Nassa et al. [10], and Salvati et al. [9].

Declarations

Ethics approval and consent to participate

N/A.

Consent for publication

N/A.

Competing interests

The authors declare no competing interests.