MUC1-C regulates NEAT1 lncRNA expression and paraspeckle formation in cancer progression

The MUC1 gene evolved in mammals for adaptation of barrier tissues in response to infections and damage. Paraspeckles are nuclear bodies formed on the NEAT1 lncRNA in response to loss of homeostasis. There is no known intersection of MUC1 with NEAT1 or paraspeckles. Here, we demonstrate that the MUC1-C subunit plays an essential role in regulating NEAT1 expression. MUC1-C activates the NEAT1 gene with induction of the NEAT1_1 and NEAT1_2 isoforms by NF-κB- and MYC-mediated mechanisms. MUC1-C/MYC signaling also induces expression of the SFPQ, NONO and FUS RNA binding proteins (RBPs) that associate with NEAT1_2 and are necessary for paraspeckle formation. MUC1-C integrates activation of NEAT1 and RBP-encoding genes by recruiting the PBAF chromatin remodeling complex and increasing chromatin accessibility of their respective regulatory regions. We further demonstrate that MUC1-C and NEAT1 form an auto-inductive pathway that drives common sets of genes conferring responses to inflammation and loss of homeostasis. Of functional significance, we find that the MUC1-C/NEAT1 pathway is of importance for the cancer stem cell (CSC) state and anti-cancer drug resistance. These findings identify a previously unrecognized role for MUC1-C in the regulation of NEAT1, RBPs, and paraspeckles that has been co-opted in promoting cancer progression.


INTRODUCTION
Paraspeckles are nuclear condensates formed by the lncRNA NEAT1 and multiple associated RNA binding proteins (RBPs) [1].The NEAT1 locus encodes (i) a 3.7 kb NEAT1_1 transcript which is dispensable for paraspeckle formation, and (ii) a 23 kb NEAT1_2 isoform that functions as an essential scaffold for binding of paraspeckle proteins [2].NEAT1_1 overlaps with the 5' end of the NEAT1_2 isoform, which is extended by alternative 3' end processing [2].NEAT1_1 is polyadenylated, whereas the 3' end of NEAT1_2 contains a triple-helix structure that protects it from degradation [3,4].The proportion of NEAT1_2 to NEAT1_1 dictates the properties of paraspeckles [4], emphasizing the importance of mechanisms regulating expression of the two isoforms [5].Binding of transcription factors, such as p53, MYC, CEBPB, and others, to the NEAT1 promoter region activates transcription with isoform switching between NEAT1_1 and NEAT1_2 [6].Paraspeckle abundance is dependent on NEAT1_2 and not NEAT1_1, which plays a role in cell metabolism [6,7].NEAT1_2 includes hydrophobic A, hydrophilic B, and hydrophobic C regions that recruit the splicing factor proline-and glutaminerich (SFPQ), non-POU domain-containing octamer-binding protein (NONO), fused in sarcoma (FUS), RNA binding motif protein 14 (RBM14), DAZ-associated protein 1 (DAZAP1), HNRNPK and HNRNPH3 RBPs essential for paraspeckle biogenesis [2,[8][9][10].Among these, the Drosophila behavior human splicing family members, SFPQ and NONO, form heterodimers involved in pre-mRNA splicing, DNA repair and transcriptional regulation [11].Oligomerization of SFPQ and NONO also enhances the recruitment of other RBPs, such FUS, that are essential for paraspeckle formation and are of importance in sequestering nuclear RNAs and regulating gene expression [2,10].Retention of these RBPs and RNAs contribute to liquid-liquid phase separation (LLPS) of paraspeckles [12].
MUC1-C regulates gene activation in chromatin by interacting with transcription factors and effectors of epigenetic reprogramming [15].Paraspeckles regulate gene transcription through RNA-RNA interactions and by sequestering RNAs and proteins.MUC1-C and paraspeckles thus regulate gene expression by distinct mechanisms.There is no known involvement of MUC1-C in integrating regulation of gene expression in chromatin with that in paraspeckles or other subnuclear biomolecular condensates.The present results demonstrate that MUC1-C is necessary for NEAT1 expression in human cancer cells by NF-κB-and MYCmediated pathways.We show that MUC1-C is also required for induction of SFPQ, NONO, and FUS by MYC-dependent mechanisms.MUC1-C-induced activation of NEAT1 and RBP genes is integrated by recruitment of the BRG1 and PBRM1 components of the PBAF complex and increases in chromatin accessibility.Our findings further demonstrate that MUC1-C is necessary for paraspeckle formation in association with driving cancer progression.
MUC1-C and NEAT1 form an auto-regulatory pathway that drives inflammatory and EMT gene signatures Downregulation of NEAT1_1 and NEAT1_2 with a NEAT1siRNA or a different NEAT1shRNA (Supplementary Fig. S2A and S2B) demonstrated that NEAT1 is dispensable for the regulation of MUC1-C mRNA levels but is necessary for expression of the MUC1-C protein (Fig. 2A, B).These results were extended by demonstrating that silencing NEAT1 decreases stability of the MUC1-C protein (Supplementary Fig. S2C), in support of a MUC1-C/NEAT1 auto-inductive pathway in which (i) MUC1-C drives NEAT1 transcription and (ii) NEAT1 regulates MUC1-C by a post-transcriptional mechanism.To assess regulation of transcriptomes by NEAT1, we performed RNA-seq on BT-549 cells with NEAT1 silencing and identified 1098 downregulated and 1135 upregulated genes (Supplementary Fig. S2D).GSEA demonstrated that NEAT1 associates with regulation of the HALLMARK TNFA SIGNALING VIA NFKB and HALLMARK MYC TARGETS V2 gene signatures (Fig. 2C and D).Using these signatures, comparison of RNA-seq data from NEAT1-and MUC1-C-silenced cells identified 210 and 262 common downregulated and upregulated genes, respectively (Fig. 2E).GSEA also demonstrated that MUC1-C and NEAT1 regulate common sets of genes in the HALLMARK TNFA SIGNALING VIA NFKB (Supplementary Fig. S2F), HALLMARK MYC TARGETS V2 (Supplementary Fig. S2G), HALLMARK INFLAMMATORY RESPONSE (Fig. 2F; Supplementary Fig. S2H) and the HALLMARK MESENCHYMAL TRANSITION (Fig. 2G; Supplementary Fig. S2I) Fig. 1 MUC1-C activates NEAT1 expression by MYC-and NF-κB-mediated mechanisms.BT-549/tet-MUC1shRNA (A) and MDA-MB-468/tet-MUC1shRNA (B) cells treated with vehicle or DOX for 7 days were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR using primers listed in Supplementary Table S1.The results (mean±SD of 5 determinations) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1).C BT-549/CshRNA and BT-549/MUC1shRNA#2 cells were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of at least 3 independent biological replicates) are expressed as relative levels compared to that obtained for CshRNA-expressing cells (assigned a value of 1).D BT-549/tet-MUC1shRNA cells expressing a tet-MUC1-CD vector were treated with vehicle or DOX for 7 days and analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of at least 3 independent biological replicates) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1) (left).Lysates were immunoblotted with antibodies against the indicated proteins (right).E BT-549 cells treated with vehicle or 5 μM GO-203 for 48 h were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of 5 determinations) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1).F Schema of the NEAT1 gene with highlighting of the dELS and PLS regions.Soluble chromatin from BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-MUC1-C and anti-MYC (left) or with anti-MUC1-C and anti-NF-κB p65 (right).The DNA samples were amplified by qPCR with primers for the NEAT1 dELS region (left) and the PLS region (right).The results (mean ± SD of 3 determinations) are expressed as percent input.G BT-549/tet-MYCshRNA cells treated with vehicle or DOX for 7 days were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of 4 determinations) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1)(left).Lysates were immunoblotted with antibodies against the indicated proteins (right).H BT-549/CshRNA and BT-549/NF-κBshRNA cells were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean ± SD of 6 determinations) are expressed as relative levels compared to that obtained for CshRNA-expressing cells (assigned a value of 1)(left).Lysates were immunoblotted with antibodies against the indicated proteins (right).gene signatures.Among these genes, we confirmed as selected examples that MUC1-C and NEAT1 regulate expression of the dual phosphatase DUSP2, which is a target of NF-κB and MYC (Fig. 2H), and (ii) IL-6, which is linked to chronic inflammation and EMT (Fig. 2I).These findings indicated that MUC1-C and NEAT1 form an auto-inductive pathway that (i) regulates NF-κB and MYC target genes and (ii) associates with chronic inflammation and EMT.Fig. 2 NEAT1 regulates MUC1-C expression in an auto-inductive pathway.A BT-549/CsiRNA and BT-549/NEAT1siRNA cells were analyzed for MUC1-C transcripts by qRT-PCR.The results (mean±SD of 3 biological replicates) are expressed as relative levels compared to that obtained for CsiRNA cells (assigned a value of 1)(left).Lysates were immunoblotted with antibodies against the indicated proteins (right).B BT-549/CshRNA and BT-549/NEAT1shRNA cells were analyzed for MUC1-C transcripts by qRT-PCR.The results (mean±SD of 3 independent biological replicates) are expressed as relative levels compared to that obtained for CshRNA cells (assigned a value of 1)(left).Lysates were immunoblotted with antibodies against the indicated proteins (right).C, D GSEA of RNA-seq data from NEAT1siRNA vs CsiRNA cells using the indicated HALLMARK gene signatures.E Venn diagrams of common downregulated and upregulated genes in NEAT1-and MUC1-C-silenced cells.F, G GSEA of RNA-seq data from NEAT1siRNA vs CsiRNA cells using the indicated HALLMARK gene signatures.H, I BT-549/CshRNA, BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days, and BT-549/NEAT1shRNA cells were analyzed for DUSP2 (H) and IL-6 (I) transcripts by qRT-PCR.The results (mean ± SD of 3 independent biological replicates) are expressed as relative levels compared to that obtained for DOX-and CshRNA cells (assigned a value of 1).
MUC1-C/NEAT1 pathway confers chemoresistance MUC1-C promotes resistance of cancer cells to cytotoxic and targeted agents [14,15,42]; whereas less is known regarding involvement of NEAT1 in the drug-resistant phenotype.In this regard, studies of BT-549 cells selected for resistance to paclitaxel (PTX) [18] demonstrated that BT-549/PTX-R cells have increased NEAT1_1 and NEAT1_2 levels compared to parental BT-549 cells (Fig. 6A).Moreover, targeting MUC1-C in BT-549/PTX-R cells downregulated NEAT1_1 and NEAT1_2 expression (Fig. 6B; Supplementary Fig. S6A).Of functional significance, silencing Fig. 3 MUC1-C regulates SFPQ, NONO, and FUS expression by MYC-dependent mechanisms.A BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for nascent SFPQ and NONO transcription.The results (mean±SD of at least 3 independent biological replicates) are expressed as relative gene transcription compared to that obtained in vehicle-treated cells (assigned a value of 1).B BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for the indicated transcripts by qRT-PCR.The results (mean±SD of 3 biological replicates) are expressed as relative mRNA levels compared to that obtained for vehicle-treated cells (assigned a value of 1).C BT-549/tet-MYCshRNA cells treated with vehicle or DOX for 7 days were analyzed for SFPQ and NONO transcripts by qRT-PCR (left).The results (mean±SD of at least 3 replicates) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1).Lysates were immunoblotted with antibodies against the indicated proteins (right).D Schema of the SFPQ gene with highlighting of the dELS1 and dELS2 regions.Soluble chromatin from BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-MYC.The DNA samples were amplified by qPCR with primers for the SFPQ dELS1 and dELS2 regions.The results (mean ± SD of 3 replicates) are expressed as percent input.E Schema of the NONO gene with highlighting of the PLS region.Soluble chromatin from BT-549/tet-MUC1shRNA cells treated with vehicle of DOX for 7 days was precipitated with anti-MYC.The DNA samples were amplified by qPCR with primers for the NONO PLS region.The results (mean ± SD of 3 replicates) are expressed as percent input.F BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for the indicated transcripts by qRT-PCR.The results (mean±SD of 3 biological replicates) are expressed as relative mRNA levels compared to that obtained for vehicle-treated cells (assigned a value of 1).G BT-549/tet-MYCshRNA cells treated with vehicle or DOX for 7 days were analyzed for FUS transcripts by qRT-PCR.The results (mean±SD of at least 3 biological replicates) are expressed as relative mRNA levels compared to that obtained for vehicle-treated cells (assigned a value of 1) (left).Lysates were immunoblotted with antibodies against the indicated proteins (right).

DISCUSSION
MUC1-C governs auto-inductive circuits that activate inflammatory, proliferative and remodeling pathways associated with the wound-healing response of barrier tissues [14,15,42].These pathways become established by prolonged adaptation of barrier epithelia to chronic inflammation in the progression to cancer [14,15,33,42].The present results uncover a previously unrecognized role for MUC1-C in integrating activation of the inflammatory NF-κB pathway with driving expression of the NEAT1 gene in cancer cells.TP53 and other stress-responsive TFs, but not NF-κB to our knowledge, have been linked to NEAT1 activation [6].
NEAT1 is widely overexpressed in human cancers and is associated with poor clinical outcomes, which have been attributed in part to an essential role for forming paraspeckles in the response to loss of homeostasis [7].Our studies demonstrate that MUC1-C/NF-κB complexes occupy a NEAT1 PLS region.MUC1-C binds directly to NF-κB and promotes NF-κB occupancy on its target genes [14,15,42,44,54].Concordant with this function, silencing MUC1-C decreased occupancy of NF-κB on the NEAT1 PLS region.MUC1-C activates the BAF and PBAF chromatin remodeling complexes and induces global changes in chromatin accessibility across the entire genomes of cancer cells [25][26][27]38].We report that occupancy of MUC1-C/NF-κB complexes on the NEAT1 PLS recruits BRG1 and PBRM1 in association with increases in chromatin accessibility (Fig. 8A).As confirmation of MUC1-C dependence, targeting MUC1-C decreased chromatin accessibility of the NEAT1 promoter with downregulation of both NEAT1_1 and NEAT1_2 transcripts [6].These results identify involvement of MUC1-C in driving NEAT1_1 and NEAT1_2 expression by a mechanism involving chromatin remodeling of the NEAT1 gene.
The binding of the MUC1-C cytoplasmic domain to the MYC HLH-LZ region regulates expression of MYC target genes [24].MYC has been linked to induction of NEAT1 expression [6].Along these lines, we found that MUC1-C is necessary for (i) MYC occupancy of the NEAT1 PLS region, (ii) recruitment of BRG1 and PBRM1, (iii) PBRM1-driven increases in chromatin accessibility, and (iv) MYCmediated NEAT1 expression (Fig. 8A).Surprisingly, we found that, in addition to NEAT1, MUC1-C signaling is necessary for induction of the essential SFPQ, NONO and FUS proteins [2,9], indicating that MUC1-C plays parallel roles in integrating NEAT1 and RBPs necessary for the formation of paraspeckles (Fig. 8B).We found that MUC1-C-induced expression of SFPQ, NONO and FUS is mediated by MYC and not NF-κB.The MUC1-C cytoplasmic domain serine-rich motif confers binding to NF-κB, whereas the redox-sensitive CQC motif binds directly to the MYC HLH-LZ region [15], suggesting that MUC1-C could function in integrating NF-κB-and MYC-mediated induction of NEAT1 with MYC-driven activation of the essential RBP-encoding genes.By way of analogy with NEAT1 regulation, we found that MUC1-C is necessary for (i) MYC occupancy of the SFPQ, NONO, and FUS genes, and (ii) PBRM1-dependent increases in chromatin accessibility in association with induction of their expression (Fig. 8B).We also found by coimmunoprecipitation and localization studies that MUC1-C associates with SFPQ and FUS, but not NONO (Fig. 8C).Our results further demonstrate that MUC1-C differentially regulates the abundance of these essential RBPs in that targeting MUC1-C decreased NONO and FUS, while increasing SFPQ levels in the cytoplasm.The functional significance underpinning this differential regulation of NONO and FUS vs SFPQ is unclear; however, SFPQ activates the cytoplasmic RIG-I stress response with increases in IFNβ production [55], which is of interest in that this pathway is induced by MUC1-C in cancer cells in association with chronic inflammation, DNA damage resistance and immune evasion [38,46].These results demonstrate that MUC1-C plays a Fig. 4 MUC1-C activates NEAT1 and RBP genes by PBAF-mediated increases in chromatin accessibility.A BT-549 cells expressing a CshRNA, BRG1shRNA or PBRM1shRNA were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of three independent biological replicates) are expressed as relative RNA levels compared to that obtained for CshRNA-expressing cells (assigned a value of 1).B Soluble chromatin from BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-BRG1 and anti-PBRM1.The DNA samples were amplified by qPCR with primers for the NEAT1 PLS region.The results (mean ± SD of 3 biological replicates) are expressed as percent input.C Genome browser snapshot of ATAC-seq data from the NEAT1 PLS in BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days (left).Chromatin was analyzed for accessibility by nuclease digestion (right).The results (mean ± SD of at least 3 biological replicates) are expressed as % undigested chromatin.D Genome browser snapshot of ATAC-seq data from the NEAT1 PLS in BT-549/ CshRNA and BT-549/PBRM1shRNA cells (left).Chromatin was analyzed for accessibility by nuclease digestion (right).The results (mean ± SD of 3 replicates) are expressed as % undigested chromatin.E BT-549 cells expressing a CshRNA, BRG1shRNA or PBRM1shRNA were analyzed for SFPQ and NONO transcripts by qRT-PCR.The results (mean±SD of 3 independent replicates) are expressed as relative RNA levels compared to that obtained for CshRNA-expressing cells (assigned a value of 1).F Soluble chromatin from BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-MUC1-C, anti-BRG1, and anti-PBRM1.The DNA samples were amplified by qPCR with primers for the SFPQ dELS2 region.The results (mean ± SD of three independent biological replicates) are expressed as percent input.G. Genome browser snapshot of ATAC-seq data from the SFPQ dELS2 in BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days (left).Chromatin was analyzed for accessibility by nuclease digestion (right).The results (mean ± SD of 3 replicates) are expressed as % undigested chromatin.H Genome browser snapshot of ATAC-seq data from the SFPQ dELS2 in BT-549/CshRNA and BT-549/PBRM1shRNA cells (left).Chromatin was analyzed for accessibility by nuclease digestion (right).The results (mean ± SD of 3 replicates) are expressed as % undigested chromatin.previously unreported role in regulating the expression of SFPQ, NONO, and FUS by a mechanism of chromatin remodeling that is integrated with activation of the NEAT1 gene.
MUC1-C is activated by disruption of homeostasis and induces lineage plasticity, EMT, and epigenetic reprogramming associated with wound healing [14,15,42].Paraspeckles are increased in response to diverse stress-inducing alterations, which include inflammation, cell lineage transitions, and cancer [10,56,57].MUC1 first appeared in mammals and, along these lines, paraspeckles are mammalian-specific structures [14,15,42,58].The evolutionary concordance of MUC1-C and paraspeckles supported the contention that they may intersect in promoting the response to loss of homeostasis, which in settings of chronic inflammation could contribute to cancer (Fig. 8C).Underpinning this intersection, dysregulation of MUC1-C [14,15,42] and paraspeckle formation [50-52, 57, 59] each contribute to cancer progression.In line with their integration in promoting cancer, our results further demonstrate that MUC1-C and NEAT1 regulate Fig. 6 MUC1-C/NEAT1 pathway contributes to chemoresistance.A BT-549 and BT-549/PTX-R cells were analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of three replicates) are expressed as relative levels compared to that obtained for BT-549 cells (assigned a value of 1).B BT-549/PTX-R cells expressing tet-MUC1shRNA were treated with vehicle or DOX for 7 days and analyzed for NEAT1_1 and NEAT1_2 transcripts by qRT-PCR.The results (mean±SD of three replicates) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1).C BT-549/PTX-R cells expressing CshRNA or NEAT1shRNA were analyzed for ABCC4 and ABCC5 transcripts by qRT-PCR.The results (mean±SD of three independent replicates) are expressed as relative levels compared to that obtained for CshRNA cells (assigned a value of 1).D BT-549/PTX-R cells expressing CshRNA or NEAT1shRNA were treated with 1 nM PTX for 24, 48, and 72 h.Cell viability was assessed by Alamar blue assay.The results (mean ± SD of 3 biologic replicates each with 5 determinations) are expressed as relative levels compared to that obtained for CshRNA cells (assigned a value of 1).E BT-549/PTX-R cells expressing CshRNA or NEAT1shRNA treated with PTX were analyzed for colony formation.Shown are representative photomicrographs of stained colonies (left).The results (mean ±SD of three biologic replicates) are expressed as relative absorbance compared to that for untreated cells (assigned a value of 1) (right).F Analysis of the TCGA BRCA cohort assessing the correlation of MUC1 with NEAT1 in patient samples.Disease-free survival of patients with grade 3 TNBC tumors treated with cytotoxic chemotherapy expressing high vs low levels of MUC1 (G) and NEAT1 (H).
common sets of genes associated with chronic inflammation, induction of EMT and activation of the NOTCH1/2 stemness factors.The CSC state is of importance in conferring therapeutic resistance [60][61][62][63].We found that MUC1-C and NEAT1 contribute to cytotoxic drug resistance in different cancer models.In addition, our results consolidate the interaction between MUC1-C and NEAT1 by demonstrating in enriched CSC populations that (i) MUC1-C is necessary for regulation of NEAT1_2, and  (ii) targeting NEAT1 suppresses MUC1-C expression.To lend further credence to the MUC1-C/NEAT1 auto-regulatory pathway, we show that MUC1-C and NEAT1 are necessary for CSC selfrenewal capacity.These results collectively indicate that (i) MUC1-C is required for NEAT1 expression and thereby paraspeckle formation and (ii) MUC1-C and NEAT1 contribute to the regulation of genes that promote cancer progression.In this way, MUC1-C integrates changes in chromatin with the formation of subnuclear paraspeckles to regulate gene expression.
Our results further indicate that MUC1-C could contribute to the formation of other biomolecular condensates.Along these lines, MUC1-C (i) regulates SFPQ, NONO, and FUS expression, and (ii) forms complexes with SFPQ and FUS that promote LLPS in other settings, such as at sites of DNA damage repair [64][65][66].MUC1-C also localizes to PML bodies that, like paraspeckles, are biomolecular condensates restricted to mammals [67].Additionally, MUC1-C regulates (i) the XIST lncRNA [68], which initiates X chromosome inactivation by forming phase-separated structures [69,70], and (ii) TDP-43 [68], which regulates XIST and paraspeckle formation [71].MUC1-C is modified by galectin-3, includes an intrinsically disordered domain, forms multimers, and binds to RNPs, which are characteristics that can contribute to the formation of biomolecular condensates [15,42,72].Subsequent studies will therefore be needed to determine if MUC1-C is necessary for the formation of other nuclear or cytoplasmic stress bodies.
Short tandem repeat (STR) analysis was performed for authentication of the cells.Mycoplasma contamination was monitored using the MycoAlert Mycoplasma Detection Kit (Lonza, Rockland, MA, USA).Cells were cultured for 3 months when performing experiments.

Quantitative reverse-transcription PCR (qRT-PCR)
Total cellular RNA was isolated with Trizol reagent (Thermo Fisher Scientific).The High Capacity cDNA Reverse-Transcription Kit (Applied Biosystems, Grand Island, NY, USA) was used for cDNA synthesis) as described [31].cDNAs were amplified with the Power SYBR Green PCR Master Mix (Applied Biosystems) and the CFX96 Real-Time PCR System (BIO-RAD, Hercules, CA, USA) as described [31].

Click-iT Nascent RNA Assay
Nascent RNA labeling with EU was performed using the Click-iT Nascent RNA Capture kit (Invitrogen) according to the manufacturer's protocol.Briefly, cells were pulsed with 0.5 mM EU for 24 h.Total RNA was then isolated, and the nascent transcripts were captured on streptavidin magnetic beads.cDNA synthesis was performed using the Superscript VILO cDNA synthesis kit (Invitrogen) followed by analysis with qRT-PCR.

RNA-seq analysis
Total RNA was isolated with the RNeasy Plus Mini Kit (Qiagen) from cells cultured in triplicates.Library preparation was performed using TruSeq Stranded mRNA (Illumina, San Diego, CA, USA) as described [28].Human genome (GRCh38.74) was used to align raw sequencing reads with STAR as described [28].Gene counts were normalized and differential expression analysis was performed using DESeq2 as described [28].The fgsea (v1.8.0) package in R was used for differential expression rank order and GSEA.Queried gene sets included those available through the Molecular Signatures Database (MsigDB) as described [28].

Cell proliferation assays
BT-549/PTX-R cells and DU-145/DTX-R cells with and without NEAT1 silencing were seeded at 6000 cells per well in 96-well plates.After 24 h, cells were treated with 1 nM PTX for 24, 48, and 72 h or 1 nM Docetaxel for 24,48 and 72 h.The Alamar Blue assay was used to assess cell viability (Thermo Fisher Scientific).Fluorescence intensity (560 nm excitation/590 nm emission) was measured in quintuplicate.

Tumorsphere formation assays
Single-cell suspensions were cultured in MammoCult Human MediumKit (Stemcell Technologies) at a density of 5000 cells/well of a 6-well ultralow attachment culture plate (Corning) for 10 days as described [27].Tumorspheres with a diameter >50 microns were counted under an inverted microscope in triplicate wells.

RNAscope and protein multiplexing
Cells were grown on chambered slides, fixed with freshly made 4% PFA for 10 min, washed with 1X PBS and permeabilized with PBS-Triton (0.5%) for 15 min at room temperature.Cells were washed with PBS-Triton (0.1%) and then twice in 2× SSC-T for 5 min at room temperature at each wash step.The slides were rinsed twice in MilliQ-grade water and treated with the ACDBio Hydrogen Peroxide pre-treatment for 10 min at room temperature.The slides were then rinsed with MilliQ-grade water and the cells were incubated with RNAScope probes against (i) NEAT1 (1125421-C1; biotechne, Minneapolis, MN, USA) overnight at 40 °C.Probe signal amplification was performed using the ACDBio RNAScope 2.5 HD Detection Brown kit (322310; bio-techne) and reaction with Biotium CF647 Tyramide (96022; bio-techne) to visualize probe staining.Slides were washed with TBS-T and then incubated with primary antibodies against MUC1-C (HM-1630-P1ABX; Thermo Fisher Scientific) overnight at 4 °C.After washing again with TBS-T, slides were incubated with AF568-tagged secondary antibodies against mouse IgG (A10037; Thermo Fisher Scientific) and Armenian hamster AF488-tagged Ig (ab173003; Abcam) for 2 h at room temperature.Slides were washed with TBS-T, stained with DAPI, mounted with Prolong Gold and visualized under a confocal microscope.

ATAC-seq
ATAC-seq libraries were generated from three biologically independent replicates per condition as described [27,75].The raw ATAC-seq data were processed using the pipeline: (https://github.com/macs3-project/genomics-analysis-pipelines).To generate the signal tracks for the Integrative Genome Browser (IGV) snapshots, we used MACS2 to pileup the aligned ATAC-seq read pairs and normalized the pileup values by the million read depth of each library as described [27].

Chromatin accessibility assay
DNAse1 chromatin accessibility assays were performed on chromatin isolated as described [27].Aliquots of chromatin were left untreated or digested with 3 U/100 μl Dnase I (Promega, Madison, WI, USA) for 5 min at room temperature as described [27].DNA was purified and amplified by qPCR using primers listed in Supplementary Table S2.qPCR results were analyzed according to the formula 100/2 Ct (Dnase I) −Ct (no Dnase I) .The data were normalized to input DNA without Dnase I treatment.

Flow cytometry
Cells were washed with PBS and harvested with 0.05% TryplE.Cells were resuspended in FACS staining buffer (Invitrogen) (10 6 cells/100 μl).Conjugated anti-human CD44 (FITC, cat.#555478; BD Biosciences, San Diego, CA, USA) was added to the cell suspension at concentrations recommended by the manufacturer and then incubated at 4 °C in the dark for 30-40 min.The labeled cells were washed in the staining buffer, then fixed in PBS containing 2% paraformaldehyde, and then analyzed on MACSQuant Analyzer 10 Flow Cytometer (Miltenyi Biotec, Waltham, MA) to acquire 20,000 events for each sample.Data were analyzed with FlowJo v10.6.2 (BD Biosciences) software.

Analysis of human TNBC tumors
Raw Read Counts of the Breast Invasive Carcinoma data collection were downloaded from TCGA using TCGAbiolinks (version 2.30).Reads were normalized to TMM (Trimmed Mean of M-values) values using endeR for subsequent Spearman's Correlation Analysis.Survival curves based on MUC1 and NEAT1 expression levels were generated using the Kaplan-Meier Plotter (http://kmplot.com/analysis/).Patients with grade 3 breast cancer treated with chemotherapy were included in this analysis.The statistical difference was calculated using the log-rank test.The prognostic value of gene expression levels was assessed using the Cox proportional hazards regression model.

Statistics
Each experiment was performed at least three times.Data are expressed as the mean ± SD.The unpaired Mann-Whitney U-test was used to determine differences between means of groups.A p value of <0.05 denoted by an asterisk (*) was considered statistically significant.

Fig. 5
Fig. 5 MUC1-C-induced regulation of RBP proteins.A Lysates of BT-549 cells were immunoprecipitated with a control IgG or anti-MUC1-C.The input lysate and precipitates were immunoblotted with antibodies against the indicated proteins.B Immunofluorescence staining of MUC1-C, SFPQ, and FUS in BT-549 cells.Nuclei were stained with DAPI.The enlarged inset images and colocalization analyses are on the right.Pearson's coefficients of colocalization: MUC1-C + SFPQ = 0.625, MUC1-C + FUS = 0.594.C Cytoplasmic and nuclear lysates from BT-549/ CsiRNA and BT-549/NEAT1siRNA cells were immunoblotted with antibodies against the indicated proteins.D Total cell lysates of BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were immunoblotted with antibodies against the indicated proteins.E Cytoplasmic and nuclear lysates from BT-549/CsiRNA and BT-549/NEAT1siRNA were immunoblotted with antibodies against the indicated proteins.F Total cell lysates from BT-549/CsiRNA and BT-549/NEAT1siRNA cells were immunoblotted with antibodies against the indicated proteins.G Representative NEAT1 RNA-FISH images of BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days with highlighting of red NEAT1 foci.H Representative NEAT1 RNA-FISH images of BT-549/CshRNA and BT-549/MUC1shRNA#2 cells with highlighting of red NEAT1 foci.

Fig. 7
Fig. 7 MUC1-C/NEAT1 pathway drives the CSC state.A RNA-seq was performed in triplicate on BT-549/NEAT1siRNA and BT-549/CsiRNA cells.GSEA was performed using the BENPORATH ES 2 gene signature.B Candidate box plots showing common NEAT1 and MUC1-C driven stemness genes from RNA-seq data.C Lysates of BT-549/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were immunoblotted with antibodies against the indicated proteins.D Lysates from BT-549/CsiRNA and BT-549/NEAT1siRNA cells were immunoblotted with antibodies against the indicated proteins.E BT-549 cells grown as monolayers in 2D culture and as tumorspheres in 3D culture were analyzed for the indicated RNA levels by qRT-PCR.The results (mean±SD of three replicates) are expressed as relative levels compared to that obtained for BT-549 2D cells (assigned a value of 1).F BT-549/tet-MUC1shRNA 3D cells treated with vehicle or DOX for 7 days were analyzed for the indicated RNA levels.The results (mean±SD of three replicates) are expressed as relative levels compared to that obtained for vehicle-treated cells (assigned a value of 1).G BT-549/CshRNA 3D and BT-549/NEAT1shRNA 3D cells were analyzed for tumorsphere formation.Photomicrographs are shown for the tumorspheres (bar represents 100 μm; left).The results (mean ± SD of three determinations) are expressed as tumorsphere number (right).H BT-549/NEAT1shRNA 3D cells transfected with an empty or MUC1-C-expressing vector were analyzed for tumorsphere formation.Photomicrographs are shown for the tumor spheres (bar represents 100 μm; left).The results (mean ± SD of three determinations) are expressed as tumorsphere number (right).

Fig. 8
Fig. 8 Schema depicting MUC1-C-mediated regulation of NEAT1 and NEAT1-binding proteins.A MUC1-C binds directly to NF-κB p65 and MYC in regulating their target genes.Based on the present results, MUC1-C/NF-κB complexes occupy the NEAT1 gene PLS region and recruit BRG1 and PBRM1 with increases in chromatin accessibility and expression.As confirmation of MUC1-C dependence, silencing MUC1-C decreased occupancy of NF-κB, BRG1, and PBRM1, and chromatin accessibility of NEAT1 gene PLS region.Silencing MUC1-C similarly decreased occupancy of MYC, BRG1, and PBRM1 on the NEAT1 dELS region.In concert with these results, MUC1-C, NF-κB, MYC, BRG1, and PBRM1 were necessary for NEAT1_1 and NEAT1_2 expression.B MUC1-C/MYC complexes integrate activation of NEAT1 with regulation of the SFPQ, NONO and FUS genes.MUC1-C was shown to be necessary for occupancy of MYC, BRG1, and PBRM1 on the (i) SFPQ dELS region, and (ii) NONO and FUS PLS regions.MUC1-C, MYC, BRG1, and PBRM1 were also necessary for chromatin accessibility of the SFPQ, NONO, and FUS genes and their expression.C MUC1-C forms complexes with SFPQ and FUS, but not NONO, and MUC1-C localizes with SFPQ and FUS in paraspeckles.As found for NEAT1_2, SFPQ, NONO, and FUS expression, silencing MUC1-C decreased paraspeckle formation.In terms of functional significance, MUC1-C and NEAT1 drive gene signatures associated with intrinsic chronic inflammation and cancer progression.MUC1-C and NEAT1 were necessary for self-renewal capacity and drug resistance in concert with driving inflammatory memory and the CSC state.