The roles of chromatin remodelers and their underlying mechanisms of action in cancer remain unclear. In this study, SMARCB1, known initially as a bona fide tumor suppressor gene, was investigated in liver cancer. SMARCB1 was highly upregulated in patients with liver cancer and was associated with poor prognosis. Loss- and gain-of-function studies in liver cells revealed that SMARCB1 loss led to reduced cell proliferation, wound healing capacity, and tumor growth in vivo. Although upregulated SMARCB1 appeared to contribute to switch/sucrose nonfermentable (SWI/SNF) complex stability and integrity, it did not act using its known pathways antagonism with EZH2 or association between TP53 or AMPK. SMARCB1 knockdown induced a mild reduction in global H3K27 acetylation, and chromatin immunoprecipitation sequencing of SMARCB1 and acetylated histone H3K27 antibodies before and after SMARCB1 loss identified Nucleoporin210 (NUP210) as a critical target of SMARCB1, which bound its enhancer and changed H3K27Ac enrichment and downstream gene expression, particularly cholesterol homeostasis and xenobiotic metabolism. Notably, NUP210 was not only a putative tumor supporter involved in liver cancer but also acted as a key scaffold for SMARCB1 and P300 to chromatin. Furthermore, SMARCB1 deficiency conferred sensitivity to doxorubicin and P300 inhibitor in liver cancer cells. These findings provide insights into mechanisms underlying dysregulation of chromatin remodelers and show novel associations between nucleoporins and chromatin remodelers in cancer.

Significance:

This study reveals a novel protumorigenic role for SMARCB1 and describes valuable links between nucleoporins and chromatin remodelers in cancer by identifying NUP210 as a critical coregulator of SMARCB1 chromatin remodeling activity.

Chromatin remodelers are responsible for nucleosome occupancy and positioning, the key mechanisms underlying the maintenance of chromatin structure, and ultimately gene expression control. Chromatin remodelers can be divided into four subfamilies of ATP-dependent nucleosome-remodeling complexes: imitation switch (ISWI), chromodomain helicase DNA-binding (CHD), switch/sucrose nonfermentable (SWI/SNF), and INO80 (1). SWI/SNF chromatin-remodeling complex, also known as the BAF (BRG1/BRM associated factor) complex, is one of the most studied complexes and usually composed of about 15 proteins, and has subunit changes depending on cell conditions (2, 3). The SWI/SNF complex remodels nucleosome occupancy and modulates transcriptional dependency in response to cellular signals, signifying the critical roles of the SWI/SNF complex during cell development and cell state transition, such as differentiation (4–7). SWI/SNF complexes can possess various combinations such as the mutually exclusive catalytic subunit SMARCA4 (BRG1) or SMARCA2 (BRM) and core subunits SMARCC1 (BAF155) or SMARCC2 (BAF170), whereas SMARCB1 (SNF5, BAF47, and INI1) is a common subunit of SWI/SNF complexes (2). Over the past two decades, genetic mutations (including germline/somatic mutations, translocations, and copy number variations) in the subunits of the SWI/SNF complex have been reported in several human cancers (8, 9). Since then, tremendous efforts have been made to understand the mechanisms underlying mutations and/or alterations in tumorigenesis (8, 10). The identification of SMARCB1 loss in rhabdoid tumors provided the first evidence on the involvement of the SWI/SNF complex in cancers (11, 12). Because SMARCB1 is generally reported to inactivate mutations in cancers, it has been considered a key tumor suppressor (11–17). Few studies have investigated alternative roles of SMARCB1 in cancers. More recently, the cryo-electron microscopy structure of Saccharomyces cerevisiae SWI/SNF bound to a nucleosome suggested that SMARCB1 plays a critical role in anchoring SWI/SF complex to DNA (18). Furthermore, intra-complex synthetic lethality is emerged as a promising therapeutic approach due to frequent SWI/SNF complex mutation rate (19), understanding of the role of SMARCB1 is dispensable.

Eukaryotic cells are characterized by the presence of a nuclear membrane that allows compartmentalization and exchange of substances between the cytoplasm and nucleus to maintain cell function (20). The majority of nucleocytoplasmic exchanges occur through nuclear pore complexes, which comprise 30 distinct nucleoporins (NUP; ref. 21). Recent evidence indicates that NUPs regulate gene expression and chromatin structure in addition to their classical transport functions, and the roles of specific NUPs are dependent on cellular conditions (22, 23). Pioneering studies have also shown that NUPs are deeply associated with the onset and progression of various diseases, such as neurologic disorders, viral infections, and cancers (24, 25). However, their potential role and underlying up/downstream mechanisms in disease progression are only beginning to be understood.

In this study, we investigated the roles of the core subunit of the SWI/SNF complex, namely, SMARCB1, in liver cancers. This study hypothesized was that SMARCB1 is an oncogene in liver cancers, and nucleoporin 210 (NUP210) was identified as a key molecular target and partner of SMARCB1.

Cell culture

The human liver cancer cell lines SK-Hep1, Huh7, SNU398 were obtained from KCLB (Korean Cell Line Bank). All cell lines were tested for mycoplasma contamination within 4 months and were no more than 25 passages. Each cell line was maintained in DMEM, RPMI1640 (Welgene) supplemented with 10% FBS (Welgene) and 100 units/mL of penicillin–streptomycin (Invitrogen). All cells were cultured at 37°C in a humidified incubator with 5% CO2.

shRNA infection

shSMARCB1 constructs were purchased from Sigma-Aldrich. For lentivirus production, MISSION lentiviral packaging mix was used. Infected derivative cells stably expressing shRNA were selected in the presence of 1.25 μg/mL puromycin.

Transient transfections

SK-Hep1, Huh7 cells in six-well plates were transiently transfected with 100 nmol/L siRNA (Bioneer) using Lipofectamin2000 Reagent from Thermo Fisher Scientific. Sixteen hours after transfection, Opti-MEM was changed with growth media.

RNA extraction and reverse transcription PCR

Total RNA was extracted using TRIzol reagent, digested with DNase I, and reverse transcribed using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Amplification of cDNA was performed on a LightCycler 480II (Roche) using the LightCycler 480 SYBR Green I Master (Roche), according to the recommended conditions. cDNAs were amplified using the following gene-specific primers. Primers were listed as Supplementary Table S1.

Chromatin immunoprecipitation assay (ChIP) and Re-ChIP

ChIP assays were performed according to instructions from Upstate Biotechnology. For each assay, 50 μg DNA, sheared by a sonication (the DNA fragment size was 200 to 500 bp), was precleared with protein A magnetic beads (Upstate Biotechnology, #16-661) and then 50 μg DNA was precipitated by SMARCB1 (abcam, #12167), H3K27Ac (abcam, #ab4729), P300 (abcam, #ab54984), SMARCA4 (abcam, #ab4081), and NUP210 (Bethyl, #A301–7955). In re-ChIP assays, the immunocomplexes with the first antibody (SMARCB1; abcam, #12167) were eluted by incubation with 10 mmol/L dithiothreitol at 37°C for 30 minutes. The eluates were diluted 50 times with immunoprecipitation (IP) dilution buffer and re-immunoprecipitated with a second antibody (P300; abcam, #ab54984). After IP, recovered chromatin fragments were subjected to real-time PCR. IgG control experiments were performed for all ChIPs and incorporated into the IP/Input (1%) by presenting the results as (IP-IgG)/(Input-IgG). ChIP primers were listed as Supplementary Table S2.

ChIP-sequencing

Library preparation and sequencing

The construction of library was performed using NEBNext Ultra DNA Library Prep Kit. Briefly, the chipped DNA was ligated with adaptors. After purification, PCR reaction was done with adaptor-ligated DNA and index primer for multiplexing sequencing. Library was purified by using magnetic beads to remove all reaction components. The size of library was assessed by Agilent 2100 bioanalyzer. High-throughput sequencing was performed as paired-end 100 sequencing using HigSeq 2500. This data set was obtained from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database (accession no. GSE122727).

ChIP-seq data analysis

The sequenced reads were trimmed using BBmap (BBDuk) and aligned to the human reference genome (hg38 assembly) using Bowtie2 (26). HOMER (findPeaks; ref. 27) was used to identify SMARCB1 binding sites (peaks) or H3K27Ac enriched sites compared with corresponding input samples in shCon and shSMARCB1 cells with an FDR-adjusted P value cutoff of 0.001. The identified peaks were annotated according to the known gene database (RefSeq). The annotated peaks were categorized into two groups (promoter or enhancer). Peaks located between −1 kb and +0.1 kb from transcription start site (TSS) were defined as promoter peaks, whereas the rest of peaks were defined as enhancer peaks. Super-enhancer regions were also identified using HOMER (findPeaks with “super” option). The read coverage tracks for visualization were constructed using HOMER (make UCSC file) with default options. H3K27Ac enriched regions in both shCon and shSMARCB1 cells were merged. Then, fragments per kilobase per million mapped reads (FPKM) of H3K27Ac were calculated and log2-transformed.

Genomic coordinates (hg19 assembly) of bivalent domains were collected from the ENCODE project database (http://rohsdb.cmb.usc.edu/GBshape/cgi-bin/hgFileUi?db=hg19&g=wgEncodeAwgSegmentation). The genomic coordinates (hg19) were converted to the hg38 genomic coordinates using the LiftOver tool in the UCSC genome browser (https://genome.ucsc.edu/cgi-bin/hgLiftOver). The genomic coordinates of CpG islands were obtained from the UCSC genome browser (http://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/).

Motif analysis

Motif analysis of SMARCB1 bound sequences depending on genomic locations (promoter and enhancer) was performed using HOMER (findMotifsGenome.pl) with the default option.

Gene ontology analysis

DAVID (28), Metascape (29), and EnrichR (30) were used to infer the biological functions of genes associated with peaks. Default parameters were used.

Transcriptome analysis

Total RNA was amplified and purified using Target Amp-Nano Labeling Kit for Illumina Expression BeadChip. Detection of array signal was carried out using Amersham fluorolink streptavidin-Cy3 according to the bead array manual. Arrays were scanned with a bead array reader confocal scanner. The quality of hybridization and overall chip performance were monitored manually by visual inspection of both internal quality control checks and the raw scanned data. Raw data were extracted using the software provided by Illumina Genome Studio v2011.1 and Gene Expression Module v1.9.0. Array probes were logarithm-transformed and normalized by the quantile method. This data set was obtained from the NCBI database (accession no. GSE107208).

Processing of RNA-seq data

RNA-seq was done at a length of 150 bp by the paired-end method in all samples. To remove adaptors and low-quality reads, trimmomatic was used in its default option. Filtered reads were aligned to the human genome reference (hg38 assembly) using STAR mapper. The mapped reads were counted and converted to TPM values using RSEM. For differentially expressed gene analysis, Fold-change and statistical significance were calculated using DEseq. For gene set enrichment analysis (GSEA) was used with preranked mode. This data set was obtained from the NCBI database (accession no. GSE154486).

GEO data analysis

mRNA expression data sets were obtained from the NCBI database (accession nos. GSE89377, GSE54236, GSE14520, GSE25097, GSE36376, GSE77314, and GSE102639).

The Cancer Genome Atlas and International Cancer Genome Consortium data analysis

RNA expressions of SMARCB1 and NUP210 were analyzed in RNA-seq-based gene expression data from The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) liver hepatocellular carcinoma (HCC) project. RNA-seq data were analyzed by first replacing all RSEM values identically equal to zero with the smallest nonzero RSEM value, and then a log2 transformation was applied.

Cell cycle analysis

Cells were collected by trypsinization and performed cell-cycle assays using the CycleTEST Plus DNA Reagent Kit (BD Biosciences). The profiles of cells were analyzed using a FACS can flow cytometer (BD Biosciences).

Soft agar colony-forming assay

The assay was performed in six-well plates. A bottom layer of agar (0.5%) with enriched DMEM media (10% FBS) was poured first. After the bottom agar solidified, SK-Hep1 cells (1.0 × 104) were seeded in top agar (0.3%) with enriched DMEM supplemented with 10% FBS and incubated at 37°C for 21 days. The culture medium was changed once or twice weekly. Colonies were visualized by staining for 1 hour with 0.005% crystal violet.

Wound healing assay

Cells were grown to confluence in six-well plates. After overnight starvation in serum-free medium, cell monolayers were scraped with a sterile micropipette tip. Initial gap widths (0 hour) and residual gap widths at 24 and 48 hours after wounding were determined from photomicrographs.

Immunocytochemistry

Cells seeded in 12-well plates containing a glass coverslip were washed with 1× PBS and then fixed with 4% neutral buffered formaldehyde solution for 30 minutes at room temperature. Cells were treated with 0.1% Triton X-100 in PBS for 5 minutes at room temperature. After being blocked with 1% Goat serum/PBS for 30 minutes at room temperature, cells were incubated with primary antibody (in 1% Goat serum/PBS) 30 minutes at room temperature. After washing three times with PBS, FITC tagged second antibody (in 1% BSA/PBS) was added to the cells and incubated at room temperature for 30 minutes. Then the slides were washed in PBS and mounted on mounting medium containing DAPI. The results were visualized on an Olympus Confocal Laser Scanning Microscope.

IHC

IHC was performed on tissue microarray (TMA) blocks consisting of 2 mm cores obtained from 238 HCCs and two normal liver tissues, after approving by the institutional review board of Samsung Medical Center. IHC was performed on TMA blocks consisting of 2 mm cores obtained from 238 HCCs and two normal liver tissues. The sections were incubated with anti-SMARCB1 antibody (abcam, #ab12167; 1:400 dilution) and anti-NUP210 antibody (abcam, #ab15601; 1:500 dilution) for overnight at cold room, after antigen retrieval with TE buffer (10 mmol/L Tris-1 mmol/L EDTA, pH 9.2). Sections were then incubated with an anti-mouse/rabbit IgG antibody (Thermo Fisher Scientific, #31190) for 20 minutes at room temperature. Antigen–antibody chromogenic reactions were developed for 30 minutes and detected using the REAL EnVision detection system K5007. IHC staining was analyzed by a semiquantitative method using H-index on a continuous scale of 0 to 300 by considering both percentage of stained cells and four intensity categories: 0 for negative, 1+ for weak, 2+ for moderate, and 3+ for strong positive. High expression was defined as H-index same as or more than median value of each H-score of nuclear staining of SMARCB1 and cytoplasmic staining of NUP 210 (SMARCB1: 140, NUP210: 100). IHC of SMARCB1 and NUP210 was performed on tumor tissue isolated from the mice. These analyses were recorded in five random fields of each slide at ×400 magnification. The stainings were scored by two blinded investigators.

Nuclear and cytoplasmic extract

Nuclear and cytoplasmic fractions were prepared using the cytoplasmic extraction buffer (10 mmol/L HEPES (pH 7.9), 50 mmol/L NaCl, 0.5 M sucrose, 0.1 mmol/L EDTA, 0.5% TX100, and 1 mmol/L DTT) and RIPA buffer (CellNest, #CNR001-0100) for nuclear protein extraction.

Western blot analysis

Cells were lysed with RIPA buffer and sonicated briefly. Cell lysates were boiled in Laemmli sample buffer, and 30 μg of each protein was subjected to SDS-PAGE. The protein concentration was measured by Bradford protein assay. Antibodies were listed as Supplementary Table S3.

Immunoprecipitation

Cells were lysed in PRO-PREP (iNtRON, #17081) supplemented with protease inhibitors (Millipore, #535140) and phosphatase inhibitors (Roche, #04906845001). For immunoprecipitation, 1 mg lysates were incubated with the appropriate antibody (1–2 μg) for 3 hours at 4°C followed by 16 hours incubation with Protein-A agarose beads (Roche, #11134515001). Immunocomplexes were washed three times with PBS before resolved by SDS-PAGE for Western blotting.

Differential salt extraction

Differential salt extraction was performed as described previously (31). Cell types were grown under standard conditions. Following the collection of 5 × 107 cells, cells were suspended in elution buffer (50 mmol/L Tris-HCl at pH 7.5, 1 mmol/L EDTA, 0.1% NP40) supplemented with protease inhibitor mixture (Millipore), incubated on ice for 5 minutes, and centrifuged. The supernatant was collected, and the pellet was suspended in elution buffer with 75 mmol/L NaCl. This process was repeated sequentially with increasing concentrations of NaCl to collect 0, 150, 300, 600, and 1,000 mmol/L NaCl soluble fractions. Each fraction was prepared in SDS (final concentration of 1%), quantified by Bradford protein assay, and analyzed by Western blot analysis.

Density sedimentation analysis

Density sedimentation analysis was performed as described previously (31). Nuclear extract (500 μg) was resuspended in 0.2 mL 0% glycerol HEMG buffer and carefully overlaid onto a 4 mL 10% to 30% glycerol (in HEMG buffer) gradient prepared in a polyallomer centrifuge tube. Tubes were centrifuged in an SW40 rotor at 4°C for 16 hours at 40,000 rpm. Fractions (0.2 mL) were collected and used in analyses.

FAIRE qPCR

FAIRE-qPCR was performed as described previously (32). Briefly, Huh7 cell was crosslinked with formaldehyde, and FAIRE and Input DNA were prepared from the cell. The results are shown as the mean and range of variation of triplicate qRT-PCR performed on the same DNA sample. Results are expressed as the percentage of input chromatin (Input DNA) and are derived from a single experiment representative of at least two independent experiments. The primer sequences are the same as the ChIP primer; NUP210_Promoter and Enhancer.

Annexin V/propidium iodide staining assay

The apoptosis rate was assessed using an Annexin V-FITC Apoptosis Detection Kit (BD Biosciences, #556547). Following resuspended in 500 μL of staining buffer with 5 μL FITC-conjugated Annexin V and 5 μL propidium iodide (PI) staining solution. The cells were analyzed using a FACS Accuri flow cytometer (BD Biosciences).

In vivo tumor growth experiment

BALB/c nude mice (male, 6–7 weeks old, 25 g) were purchased from Orient Bio and maintained under pathogen-free conditions. The Huh7 cells (5 × 106 cells per 0.1 mL Hank's Balanced Salt Solution) were injected subcutaneously in the right groin. The mice were monitored daily and tumor sizes were measured every 2 to 3 days by a digital caliper, and tumor volumes were calculated using the formula volume  =  π/6 (length × width2). The animal study was reviewed and approved by the Institutional Animal Care and Use Committee of Konkuk University.

SMARCB1 is upregulated in liver cancers and is associated with poor prognosis

To investigate the possible roles of SMARCB1 in liver cancers, we first checked the SMARCB1 expression in several tissues. SMARCB1 is ubiquitously expressed in cells and tissues, but its expression is lower in normal liver than in other tissues (Supplementary Fig. S1A). However, according to TCGA, the International Cancer Genome Consortium (ICGC), and publically available cohort studies on liver cancers, SMARCB1 expression is significantly upregulated in liver tumors compared with that in normal tissues (Fig. 1A; Supplementary Fig. S1B) and normal-cancer paired samples (Supplementary Fig. S1C). High SMARCB1 expression has been significantly associated with poor prognosis (Fig. 1B). Furthermore, its overexpression is observed at an early stage in HCC, particularly in stepwise carcinogenic progression from normal lesions to HCC (Fig. 1C). In addition, SMARCB1 expression gradually increases with the cancer stage as determined by Edmondson grades (Fig. 1D). IHC for human liver TMA was divided into four intensity categories based on the H-index, and 119 of 238 cases (50%) were classified as SMARCB1 high expression group with a median of 200 or more (Fig. 1E) and about 77% positive in HCC based on 50% cutoff (P < 0.001, Fig. 1F). Therefore, SMARCB1, a well-known representative tumor suppressor gene, is overexpressed in liver cancers and may play as a putative oncogene. To determine the role of upregulated SMARCB1 in liver cancers, we first screened for the expression of SMARCB1 and other two subunits (SMARCA4 and SMARCA2) with EZH2 in liver cancer cell lines. Most liver cancer cell lines exhibited relatively high SMARCB1 expression compared with normal liver cell line MIHA and L-02, unlike in other enzymatically active subunits (Fig. 1G). We performed a loss-of-function study by generating stable SMARCB1 knockdown cells using four shRNAs in four liver cancer cell lines, that is, SK-Hep1, Huh7, SNU398, and SNU475 that show different level ATPase subunits (Fig. 1H; Supplementary Fig. S1D), and found that SMARCB1 knockdown reduced Ser10 phosphorylation of the mitosis marker histone H3 and cell proliferation rate (Fig. 1H; Supplementary Figs. S1D and S1E) and induced G1 arrest and a decrease of wound healing capacity (Supplementary Figs. S1F and S1G). Furthermore, soft agar colony formation of SKHep1 cells had markedly suppressed by SMARCB1 knockdown, whereas MIHA, a normal liver cell, obtained increased colony formation capacity by ectopic SMARCB1 overexpression (Supplementary Fig. S1H). To assess the tumorigenic effect of SMARCB1 on liver cancer in vivo, we applied a xenograft model. We found significantly reduced tumors after 40 days using stable SMARCB1 knockdown Huh7 cells compared with control cells (Fig. 1I), whereas the whole body weight was not changed. These results suggest that SMARCB1 acts as a putative oncogene in liver cancers.

Figure 1.

Overexpressed SMARCB1 acts as a putative oncogene in liver cancers. A, The relative SMARCB1 gene expression levels in nontumor and tumor in TCGA and ICGC datasets. The median expression level of each group is indicated by horizontal lines (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). B, Overall survival dependent on SMARCB1 expression by Kaplan–Meier survival curves using the TCGA and ICGC datasets. P values were obtained with the log-rank test. C, Expression changes of SMARCB1 in patients with multistage liver disease of GSE89377. CH, chronic hepatitis; CS, cirrhosis; DN, dysplasia nodule; eHCC, early hepatocellular carcinoma; avHCC, advanced hepatocellular carcinoma (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. normal). D, Expression changes of SMARCB1 in three-stage liver cancer patients of TCGA dataset (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. G1). E, Representative figures of SMARCB1 IHC staining four intensity categories were applied: 0 for negative, 1+ for weak, 2+ for moderate, and 3+ for strong positive (bar, 100 μmol/L). F, The bar chart of the positive expression of SMARCB1 in liver cancer TMA dependent on the indicated cutoff (%). G, Western blots of endogenous SWI/SNF complex subunit (SMARCB1, SMARCA4, and SMARCA2), EZH2, and β-ACTIN protein in normal liver cell lines and liver cancer cell lines. H, mRNA of SMARCB1 (top) and protein (bottom) levels of SMARCB1, SMARCA4, SMARCA2, SMARCC1, SMARCC2, pH3S10, β-ACTIN, and Histone H3 after the loss of SMARCB1. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). I,In vivo tumor growth analysis in the xenograft, nude mice injected with Huh7 shSMARCB1 or shCon (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon).

Figure 1.

Overexpressed SMARCB1 acts as a putative oncogene in liver cancers. A, The relative SMARCB1 gene expression levels in nontumor and tumor in TCGA and ICGC datasets. The median expression level of each group is indicated by horizontal lines (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). B, Overall survival dependent on SMARCB1 expression by Kaplan–Meier survival curves using the TCGA and ICGC datasets. P values were obtained with the log-rank test. C, Expression changes of SMARCB1 in patients with multistage liver disease of GSE89377. CH, chronic hepatitis; CS, cirrhosis; DN, dysplasia nodule; eHCC, early hepatocellular carcinoma; avHCC, advanced hepatocellular carcinoma (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. normal). D, Expression changes of SMARCB1 in three-stage liver cancer patients of TCGA dataset (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. G1). E, Representative figures of SMARCB1 IHC staining four intensity categories were applied: 0 for negative, 1+ for weak, 2+ for moderate, and 3+ for strong positive (bar, 100 μmol/L). F, The bar chart of the positive expression of SMARCB1 in liver cancer TMA dependent on the indicated cutoff (%). G, Western blots of endogenous SWI/SNF complex subunit (SMARCB1, SMARCA4, and SMARCA2), EZH2, and β-ACTIN protein in normal liver cell lines and liver cancer cell lines. H, mRNA of SMARCB1 (top) and protein (bottom) levels of SMARCB1, SMARCA4, SMARCA2, SMARCC1, SMARCC2, pH3S10, β-ACTIN, and Histone H3 after the loss of SMARCB1. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). I,In vivo tumor growth analysis in the xenograft, nude mice injected with Huh7 shSMARCB1 or shCon (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon).

Close modal

SMARCB1 contributes to the biochemical stability of the BAF complex and its chromatin affinity

To examine the effect of SMARCB1 on the biochemical stability of the BAF complex, the expression levels of other SWI/SNF complex subunits were evaluated after SMARCB1 knockdown. Although the mRNA expression levels of other subunits remained unchanged, knockdown of SMARCB1 causes a decrease of other core subunits (SMARCC1/C2) and existed ATPase subunits (SMARCA4 or SMARCA2), suggesting that SMARCB1 contributes to SWI/SNF complex subunits protein stability (Fig. 1H), suggesting that SMARCB1 contributes to SWI/SNF complex subunits' protein stability. Next, 10% to 30% glycerol gradient-based density sedimentation of nuclear extracts obtained from SK-Hep1, Huh7, and SNU475 cells was performed after SMARCB1 loss. The results showed that SMARCB1 loss caused an approximately one-fraction shift of mainly expressed subunits such as SMARCA4 in Huh7 and SNU475 cells (Fig. 2A, top). Although it was difficult to observe the distinct fraction shift due to a significant decrease in the total amount of expressed subunits, it was validated that recovery of SMARCB1 restored the expression and integrity of existed subunits (Fig. 2A). In addition, the effect of SMARCB1 on the chromatin affinity of the BAF complex was investigated using a NaCl-based differential salt extraction assay. The result indicated that existed other subunits almost disappeared from chromatin upon loss of SMARCB1, and again recovered chromatin affinity upon SMARCB1 ectopic expression (Fig. 2B). Taken together, we demonstrated that overexpressed SMARCB1 increases the biochemical stability of the BAF complex and its chromatin affinity in liver cancer cells.

Figure 2.

SMARCB1 has unique pathways in liver cancer. A, Density sedimentation analyses using 10% to 30% glycerol gradients (4 mL; 0.25 mL/fraction) on nuclear extracts from SK-Hep1, Huh7, and SNU475 cells under the condition of shCon, shSMARCB1, and SMARCB1 re-expressed shSMARCB1 SUZ12 is indicated as a sedimentation coefficient control. B, Western blots of BAF complex subunits in differential salt extraction experiments.

Figure 2.

SMARCB1 has unique pathways in liver cancer. A, Density sedimentation analyses using 10% to 30% glycerol gradients (4 mL; 0.25 mL/fraction) on nuclear extracts from SK-Hep1, Huh7, and SNU475 cells under the condition of shCon, shSMARCB1, and SMARCB1 re-expressed shSMARCB1 SUZ12 is indicated as a sedimentation coefficient control. B, Western blots of BAF complex subunits in differential salt extraction experiments.

Close modal

SMARCB1 possibly has unique pathways in liver cancers distinct from its known pathways

Next, a gene expression microarray analysis using SK-Hep1 and Huh7 cells were used to characterize the transcriptional landscape after SMARCB1 knockdown. Functional annotation showed that differentially up- and downregulated genes belonged to multiple cancer-related gene ontology (GO) categories (Supplementary Fig. S2A), and DAVID analysis revealed the top six clusters included defense responses, apoptosis, angiogenesis, response to extracellular stimuli, cell growth, and glycolysis (Fig. 3A), suggesting that SMARCB1 is deeply associated with cancer cell maintenance.

Figure 3.

SMARCB1 has unique pathways in liver cancers. A, Functional annotation analysis using DAVID web-based software B, Western blots of EZH2, its target H3K27me3, and histone H3 protein. C, Western blots of TP53, its target P21, and β-ACTIN protein. D, Western blots of AMPK, its active form phospho-AMPK, and β-ACTIN protein. E, A pie chart of the genomic location distribution of the transcription factor SMARCB1. F, Motif analysis of SMARCB1-bound sequences. G, GO analysis of SMARCB1-related genes.

Figure 3.

SMARCB1 has unique pathways in liver cancers. A, Functional annotation analysis using DAVID web-based software B, Western blots of EZH2, its target H3K27me3, and histone H3 protein. C, Western blots of TP53, its target P21, and β-ACTIN protein. D, Western blots of AMPK, its active form phospho-AMPK, and β-ACTIN protein. E, A pie chart of the genomic location distribution of the transcription factor SMARCB1. F, Motif analysis of SMARCB1-bound sequences. G, GO analysis of SMARCB1-related genes.

Close modal

As a solid mechanism underlying the tumor suppressor activity of SMARCB1, unbalanced antagonism with EZH2, both dependently and independently of catalytic activity, has been reported (17, 33). To determine whether SMARCB1 is also antagonistic to the polycomb complex in liver cancer cells, we measured the global levels of EZH2 and H3K27me3 as well as the expression levels of EZH2 target genes, but we found no significant differences although within mild reduction in Huh7 cells (Fig. 3B; Supplementary Fig. S2B), demonstrating that there might be no direct antagonism between SMARCB1 and EZH2 in liver cancers. Regarding other possible underlying mechanisms, there have been two notable reports on the relationship between SMARCB1 and TP53 (16, 34). Among these, one report demonstrated that SMARCB1 inactivation resulted in TP53 downregulation, which was accompanied by apoptosis and polyploidy, and concomitant TP53 mutation accelerated tumorigenesis (16). To identify a possible relationship between SMARCB1 and TP53, three liver cancer cell lines with different TP53 mutation states (i.e., wild-type in SK-Hep1 cells, mutant in Huh7 cells, and null in Hep3B cells) were subjected to Western blot analysis and transcriptome analyses. There was no change in the levels of TP53, CDKN1A (P21), and other downstream genes of TP53 after SMARCB1 knockdown, regardless of the mutation state (Fig. 3C; Supplementary Fig. S2C). More recently, SMARCB1 loss has been shown to result in endoplasmic reticulum stress, and cells are vulnerable to targeting proteostasis and autophagy by the MYC/TP53 axis (35, 36). However, we could not detect significant differences among these gene sets (Supplementary Fig. S2D).

Another report suggested that SMARCB1 loss activates AMPK, and activated AMPK suppresses TP53 translation (34). To better characterize this axis in liver cancers, AMPK activation by phosphorylated AMPK and the mRNA expression levels of AMPK signaling genes were investigated. We observed different effects on the three cell lines based on the extent of AMPK phosphorylation (i.e., inhibition in SK-Hep1 cells, activation in Huh7 cells, and no change in SNU398 cells; Fig. 3D). In addition, there were slight differences in terms of the activation of AMPK signaling genes among the cell lines, regardless of AMPK activation in SK-Hep1 and Huh7 cells (Supplementary Fig. S2E). Therefore, it is difficult to generalize the relationship between AMPK and SMARCB1, warranting further studies. Recently, a study showed that SMARCB1 knockout causes oncogene-induced senescence (OIS) by increasing the expression of an oncogene, such as EGFR (37). However, the knockdown of upregulated SMARCB1 in liver cancers did not seem to be related to the OIS pathway (Supplementary Fig. S2F). Taken together, these results suggest that SMARCB1 has unique pathways distinct from its known tumor-suppressive pathways in liver cancers.

To delineate the downstream activity of SMARCB1, ChIP sequencing (ChIP-seq) of SMARCB1 in SK-Hep1 cells was performed. The results showed that 29% of the SMARCB1 peaks were located within the promoter region (1 kb upstream and 100 bp downstream centered at the TSS, whereas other regions were defined as enhancer regions), the rest peaks were located in enhancer regions (Fig. 3E). Line plots and binding heat maps showed that SMARCB1 is well mapped to chromosomal regions (Supplementary Fig. S2G). We compared with existing publically available SMARCB1 and SMARCA4 in the other liver cancer cell line HepG2 (GSE102559 and GSE69566). SMARCB1 binds to where SMARCB1 or SMARCA4 binds in HepG2 cells, although the binding affinity is slightly weak (Supplementary Fig. S3A). Although motif analysis of the promoter and enhancer regions revealed dissimilarities in terms of characteristics (Fig. 3F), functional annotation showed shared processes such as cell morphogenesis, protein localization to the membrane, and cell proliferation/death (Fig. 3G), suggesting the cross-talk between promoters and enhancers. We further proceeded motif analysis of SMARCB1 and compared it with HepG2 enriched motif. Regardless of the location, enriched motifs overlapped well between our SMARCB1 and SMARCB1 in HepG2, whereas SMARCA4 in HepG2 did not (Supplementary Fig. S3B). It is likely that further study is needed to determine why the two subunits of the SWI/SNF complex are so different in the motif analysis. However, we could validate our SMARCB1 ChIP-seq data and motif analysis. These SMARCB1 binding data, in combination with transcriptome analysis results, strongly demonstrate that SMARCB1 contributes to the liver cancer oncogenic signature.

NUP210 is a key target of SMARCB1

While preparing this manuscript, three reports demonstrated that SMARCB1 mediates enhancer regulation (31, 38, 39). Similarly, we found that SMARCB1 loss induced a mild decrease in the global level of acetylated lysine27 of histone H3 (H3K27Ac), a known representative active enhancer modification (Fig. 4A and B). To evaluate the genome-wide effect of H3K27Ac on chromatin, ChIP-seq of H3K27Ac was performed in control and SMARCB1 knockdown cells. The results showed that although there was a global decrease in H3K27Ac level, chromatin was relatively stable (Fig. 4C, denoted as gray dots), and there was no significant effect on the super-enhancer (Supplementary Fig. S3C), similar to that observed in conditional SMARCB1 knockout cells (38). Next, we focused on decreased H3K27Ac peaks and concurrently bound SMARCB1, as denoted by green and yellow dots, respectively (Fig. 4C), and performed GO analysis (Supplementary Fig. S3D). To identify more significant SMARCB1 targets, the expression levels of genes denoted by green dots were quantified in SK-Hep1 and Huh7 cells after SMARCB1 knockdown. NUP210 was present among the top 10 most downregulated genes in both cell lines (Fig. 4D), which confirmed NUP210 downregulation upon SMARCB1 loss (Fig. 4E). This finding was further validated using conditional SMARCB1 knockdown, which preferentially reveals the direct consequence of SMARCB1 by modulating the expression level in a quantitative and temporal way. The results suggested that NUP210 expression and biological effects such as wound healing capacity and cell proliferation are highly dependent on SMARCB1 levels (Fig. 4FH). Notably, we found a novel enhancer region of NUP210, located 136 kb downstream from TSS, which had a CpG island (Fig. 4I), presumably a highly active enhancer (40). We could observe peaks of SMARCB1 and SMARCA4 in HepG2 cells in the NUP210 enhancer (Supplementary Fig. S3E). Using ChIP-PCR, we further reconfirmed a significant decrease in H3K27Ac and P300 binding to the enhancer region upon SMARCB1 loss, rather than to the promoter region (Fig. 4J and K). These data suggest that NUP210 is a substantial target of SMARCB1 in liver cancers.

Figure 4.

NUP210 is a significant downstream molecule of SMARCB1. A, Western blots enhancer markers (H3K27Ac, H3K4me1) and histone H3. B, Immunocytochemistry analysis with anti-H3K27Ac (red) and DAPI (blue). C, A scatter plot showing H3K27Ac ChIP-seq signals on promoters (yellow) and enhancers (green) in SK-Hep1 cells. D, The color bar for differential gene expression with a 1.5-fold decrease in the nearby H3K27Ac signal. E, Western blots of P300, NUP210, SMARCB1, and β-ACTIN cytosol and nuclear extracts. F, qRT-PCR of SMARCB1 and NUP210 mRNA levels in conditional knockdown cells. Conditional SMARCB1 knockdown SK-Hep1 and Huh7 cells were cultured in the presence (+) of doxycycline for 7 days, after 7 days absence (−) of doxycycline (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). G and H, Wound healing and growth rate analysis after the loss of SMARCB1. The data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. w/doxycyclin shCon). I, Representative a screenshot of ChIP-seq signals near the NUP210 gene. J, A schematic diagram of NUP210 DNA regulatory regions. K, ChIP-qPCR showing H3K27Ac and P300 binding at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). ns, nonsignificant.

Figure 4.

NUP210 is a significant downstream molecule of SMARCB1. A, Western blots enhancer markers (H3K27Ac, H3K4me1) and histone H3. B, Immunocytochemistry analysis with anti-H3K27Ac (red) and DAPI (blue). C, A scatter plot showing H3K27Ac ChIP-seq signals on promoters (yellow) and enhancers (green) in SK-Hep1 cells. D, The color bar for differential gene expression with a 1.5-fold decrease in the nearby H3K27Ac signal. E, Western blots of P300, NUP210, SMARCB1, and β-ACTIN cytosol and nuclear extracts. F, qRT-PCR of SMARCB1 and NUP210 mRNA levels in conditional knockdown cells. Conditional SMARCB1 knockdown SK-Hep1 and Huh7 cells were cultured in the presence (+) of doxycycline for 7 days, after 7 days absence (−) of doxycycline (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). G and H, Wound healing and growth rate analysis after the loss of SMARCB1. The data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. w/doxycyclin shCon). I, Representative a screenshot of ChIP-seq signals near the NUP210 gene. J, A schematic diagram of NUP210 DNA regulatory regions. K, ChIP-qPCR showing H3K27Ac and P300 binding at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). ns, nonsignificant.

Close modal

NUP210 may be a novel tumor supporter in liver cancers

NUP210 is known as a cell-specific NUP and critical regulator of myoblast and neuroprogenitor differentiation (41–43). However, to the best of our knowledge, NUP210 has not yet been studied in liver cancers. Nonetheless, we propose that NUP210 is a putative tumor supporter in liver cancers based on expression and survival data retrieved from TCGA and ICGC (Fig. 5A and B; Supplementary Fig. S4A), upregulation in normal cancer paired samples (Supplementary Fig. S4B) and gradual increases in stepwise cancer progression, as demonstrated by Edmondson grades (Fig. 5C and D). IHC staining of NUP210 showed NUP210 is highly expressed HCC than normal liver tissue, and 133 out of 238 (55.9%) were classified as NUP210 higher expression group with a median 140 and more (Fig. 5E). To determine the importance of NUP210, it was knocked down using two specific siRNAs in three live cancer cell lines (Fig. 5F). The depletion of NUP210 mimics the effects of SMARCB1 knockdowns, such as decreases in pH3S10 levels (Fig. 5F), lower proliferation rates (Fig. 5G), G1 cell-cycle arrest (Supplementary Fig. S4C), and reduced wound healing capacity (Fig. 5H). In addition, the proportion of NUP210 high expression was higher in the SMARCB1 high expression group than low expression group (63.9% vs. 47.9%, P = 0.013; Fig. 5I). There is a statistically significant correlation between IHC expression between SMARCB1 and NUP210 (r = 0.245, P < 0.001) using H-index on a continuous scale (Supplementary Fig. S4D). This correlation was further validated using IHC analysis of SMARCB1 and NUP210 in tumor tissue obtained from the xenograft model established in Fig. 1, which confirmed NUP210 downregulation upon SMARCB1 loss (Supplementary Fig. S4E). There was a positive correlation between SMARCB1 and NUP210 in TCGA data (Supplementary Fig. S4F). Upregulation of SMARCB1 and NUP210 was recently validated in large-scale liver cancer Proteomic study data (Supplementary Fig. S4G; ref. 44). Taken together, the data suggest that NUP210 mediates the putative oncogenic function of SMARCB1.

Figure 5.

NUP210 might be an important tumor supporter in liver cancers. A, The relative NUP210 levels in TCGA and ICGC datasets The median expression level of each group is indicated by horizontal lines (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). B, Overall survival data by Kaplan–Meier analysis. P values were obtained with the log-rank test. C, NUP210 levels in patients with three-stage liver cancer from the TCGA dataset (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). D, SMARCB1 levels in patients with multistage liver disease of GSE89377. CH, chronic hepatitis; CS, cirrhosis; DN, dysplasia nodule; eHCC, early hepatocellular carcinoma; avHCC, advanced hepatocellular carcinoma (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. G1). E, Representative figures of NUP210 IHC staining (bar, 100 μmol/L). F, qRT-PCR of NUP210 mRNA levels (top) and Western blot analysis (bottom) of NUP210, P300, SMARCB1, pH3S10 (proliferation marker), H3K27Ac, and histone H3. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). G, Growth rates of NUP210-depleted cells. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). H, Wound healing analysis after the loss of NUP210. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). I, Pie chart showing the ratio of NUP210 expression to SMARCB1 level.

Figure 5.

NUP210 might be an important tumor supporter in liver cancers. A, The relative NUP210 levels in TCGA and ICGC datasets The median expression level of each group is indicated by horizontal lines (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). B, Overall survival data by Kaplan–Meier analysis. P values were obtained with the log-rank test. C, NUP210 levels in patients with three-stage liver cancer from the TCGA dataset (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. nontumor). D, SMARCB1 levels in patients with multistage liver disease of GSE89377. CH, chronic hepatitis; CS, cirrhosis; DN, dysplasia nodule; eHCC, early hepatocellular carcinoma; avHCC, advanced hepatocellular carcinoma (mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. G1). E, Representative figures of NUP210 IHC staining (bar, 100 μmol/L). F, qRT-PCR of NUP210 mRNA levels (top) and Western blot analysis (bottom) of NUP210, P300, SMARCB1, pH3S10 (proliferation marker), H3K27Ac, and histone H3. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). G, Growth rates of NUP210-depleted cells. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). H, Wound healing analysis after the loss of NUP210. Data are presented as the mean ± SEM (*, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. siCon). I, Pie chart showing the ratio of NUP210 expression to SMARCB1 level.

Close modal

SMARCB1 and NUP210 may support liver cancers by activating cholesterol homeostasis and xenobiotic metabolism

Next, we performed RNA sequencing in NUP210 knockdown Huh7 cells to obtain the transcriptome signature of NUP210. We identified 131 upregulated and 256 downregulated differentially expressed genes (Fig. 6A, red and blue colored dots). Although upregulated genes are enriched in extracellular matrix organization, cellular adhesion, and morphogenesis, downregulated genes are deeply involved in intracellular energy such as amino acid metabolism, insulin-like growth factor (IGF) regulation, inflammatory response, and lipid localization (Fig. 6B). Given that NUP210 knockdown and SMARCB1 knockdown induce a mild reduction of global H3K27ac (Fig. 4), we focused on downregulated genes. GSEA of loss of NUP210 or SMARCB1 showed that both gene expression signature is negatively engaged in cholesterol homeostasis and xenobiotic metabolism (Fig. 6C and D). These two pathways are not only liver-specific but also deeply involved in liver cancers (45, 46). Liver cancer cells need a large amount of cholesterol as material for cellular membrane; therefore, repression of cholesterol homeostasis may play an essential role in antitumorigenic effects induced by the loss of SMARCB1-NUP210. Xenobiotic metabolism occurs predominantly in the liver, and targeting aberrant metabolism in liver cancers has been considered as promising therapeutic opportunities. Furthermore, we could observe that these core genes are highly upregulated in liver cancers than normal tissues in TCGA data (Fig. 6E and F). Taken together, transcriptome analysis suggested that the SMARCB1-NUP210 axis works as a critical part of the liver cancer network. However, further research remains on the mechanism of direct carcinogenesis of NUP210-mediated SMARCB1, the presence of other mediators of SMARCB1, or the independent function of NUP210 in tumorigenesis.

Figure 6.

SMARCB1–NUP210 axis works as a critical part of the liver cancer network. A, Volcano plot of RNA-seq after NUP210 knockdown. B, GO analysis of differential expression genes using METASCAPE web-based software. C and D, Common significant gene set of NUP210 and SMARCB1 using GSEA. E and F, Core genes expression patterns in liver cancer TCGA data.

Figure 6.

SMARCB1–NUP210 axis works as a critical part of the liver cancer network. A, Volcano plot of RNA-seq after NUP210 knockdown. B, GO analysis of differential expression genes using METASCAPE web-based software. C and D, Common significant gene set of NUP210 and SMARCB1 using GSEA. E and F, Core genes expression patterns in liver cancer TCGA data.

Close modal

NUP210 serves as a chromatin scaffold for SMARCB1 and P300

Of note, NUP210 knockdown induced a global decrease in H3K27Ac levels, although there was no decrease in P300 and SMARCB1 levels (Fig. 5F). Therefore, the mechanism underlying this contribution of NUP210 to the global H3K27Ac level was investigated. NUPs are composed of nuclear pore complexes, which primarily function as channels to regulate the exchange of metabolites and macromolecules between cytoplasm, and nucleoplasm and P300 is known to shuttle nuclear and cytosol. Although NUP210 is known as a transport-independent NUP (41), we hypothesized that downregulated NUP210 interferes import P300 to the nucleus. There was no significant change in P300 localization upon SMARCB1 or NUP210 loss (Supplementary Figs. S5A and S5B). Next, we investigated the chromatin accessibility of P300 and SMARCB1. The chromatin binding of P300 and NUP210 weakened without SMARCB1 (Supplementary Fig. S5C), and the chromatin affinity of P300 and SMARCB1 decreased without NUP210 (Fig. 7A). The result was confirmed that decreased chromatin affinity was rescued by SMARCB1 ectopic expression (Supplementary Fig. S5C; Fig. 7A). Remarkably, the direct binding of P300 and SMARCB1 decreased upon NUP210 loss (Fig. 7B), and P300 enrichment was reduced at NUP210 enhancer in this situation (Fig. 7C). We further validated that SMARCB1 and P300 actual recruitments at NUP210 enhancer are decreased upon NUP210 knockdown by ChIP-Re-ChIP using the P300 antibody following SMARCB1 antibody (Fig. 7D). Additional ChIP assays using SMARCA4 and NUP210 antibodies (Supplementary Fig. S5D) suggested that SMARCB1 (SWI/SNF complex) -NUP210-P300 bind simultaneously at NUP210 enhancer for NUP210 expression. Furthermore, Formaldehyde-Assisted Isolation of Regulatory Elements (FAIRE)-qPCR confirmed that SMARACB1 and NUP210 are necessary for the open chromatin structure of NUP210 DNA regulatory regions (Fig. 7E), proposing that NUP210 is a key mediator of the interaction between SMARCB1 and P300 and of chromatin binding of SMARCB1 and P300.

Figure 7.

NUP210 is an enhancer chromatin scaffold for SMARCB1 and P300. A, Western blots of differential salt extraction experiments under the condition of shCon, shNUP210, and SMARCB1 overexpressed shNUP210. B Protein immunoprecipitation analysis with anti-SMARCB1 and anti-P300 antibodies. Equal amounts of protein were loaded in each lane (1% input loaded). C, ChIP-qPCR showing P300 binding at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). D ChIP-Re-ChIP-qPCR showing binding of SMARCB1 and P300 at the NUP210 enhancer (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). E, FAIRE-qPCR analysis at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). F, Schematic summary. ns, nonsignificant.

Figure 7.

NUP210 is an enhancer chromatin scaffold for SMARCB1 and P300. A, Western blots of differential salt extraction experiments under the condition of shCon, shNUP210, and SMARCB1 overexpressed shNUP210. B Protein immunoprecipitation analysis with anti-SMARCB1 and anti-P300 antibodies. Equal amounts of protein were loaded in each lane (1% input loaded). C, ChIP-qPCR showing P300 binding at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). D ChIP-Re-ChIP-qPCR showing binding of SMARCB1 and P300 at the NUP210 enhancer (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). E, FAIRE-qPCR analysis at the NUP210 enhancer and promoter (mean ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. shCon). F, Schematic summary. ns, nonsignificant.

Close modal

SMARCB1 deficiency confers sensitivity to P300 inhibitor and doxorubicin

Next, we used the p300 inhibitor SGC-CBP30 (47) to determine if SMARCB1 functions through P300. The effect of P300 inhibition on apoptosis was more sensitive in SMARCB1 knockdown cells (Supplementary Fig. S6A). Moreover, SMARCB1 loss sensitized liver cancer cells to doxorubicin, widely used in chemotherapy for various cancers (Supplementary Fig. S6A). One paper reported that the loss SMARCB1 increases ABCB1, which confers doxorubicin resistance in residual SWI/SNF complex in HAP1 cells, a near-haploid cell line derived from chronic myeloid leukemia patient (48). To get a clue for this discrepancy, we performed compared enrichment Rank-Rank Hypergeometric Overlap (RRHO) and found that there is no correlation between our data (Supplementary Fig. S6B). This could be because of the background of cells or systems. By checking Top25 up/down-regulated genes upon SMARCB1 or SMARCA4 in HAP1 cells, we found that most gene expression patterns did not seem to follow the directions except the upregulation of ABCB1 in Huh7 cells, not SK-Hep1 (Supplementary Fig. S6C). Despite the upregulation of ABCB1, SMARCB1 knockdown Huh7 cells showed increased sensitivity upon doxorubicin treatment. We further checked mRNA levels of TOP2A and ABCB1 by real-time PCR and the suggested mediators of doxorubicin sensitivity using our transcriptome data of what mechanism offers sensitivity to our system, we did not get any clue (Supplementary Figs. S6D and S6E). Although we did not solve the underlying mechanism of what SMARCB1 lacking offers drug sensitivity, our results might provide important insight into the development of novel combination therapies for liver cancers.

SMARCB1 is overexpressed and acts as a putative oncogene in liver cancers. Comprehensive analyses of SMARCB1 and H3K27Ac based on ChIP-seq and gene expression upon SMARCB1 loss revealed that NUP210 is a crucial target of SMARCB1. Highly expressed SMARCB1 binds to a novel enhancer region of NUP210, resulting in NUP210 upregulation. Concurrently, NUP210 upregulation not only provides additional chromatin scaffold proteins of SMARCB1 and P300 but also affects the direct interaction of the two proteins, ultimately exacerbating the putative oncogenic role of SMARCB1 (Fig. 7F).

Although overexpressed SMARCB1 apparently acts as a putative oncogene in liver cancers (Fig. 1), understanding the mechanisms of SMARCB1 is not straightforward because some mechanisms are consistent with known tumor-suppressive mechanisms and others are distinct. SMARCB1 contributes to the stability of other subunits (Fig. 1H), BAF complex integrity (Fig. 2A), and chromatin affinity of the BAF complex (Fig. 2B). SMARCB1 loss reduces the level of the global enhancer mark H3K27Ac and entails a mild lapse of enhancer activities but not that of super-enhancers (Fig. 4C; Supplementary Fig. S3C; ref. 39) as like acting as a tumor suppressor. However, there was no antagonism with the polycomb protein EZH2 or dependency on P53, AMPK, MYC, and/or proteostasis pathways (Fig. 3BD). Our results provide insightful views on the oncogenic functions of SMARCB1, previously known only as a tumor suppressor gene, in the context of a multifaceted chromatin remodeler.

Integrated genome-wide analyses and biochemical studies demonstrated, for the first time, that NUP210 is a critical downstream molecule of SMARCB1, where SMARCB1 binds to a novel enhancer region of NUP210 that has a CpG island located 136 kb downstream of TSS (Fig. 4I). Although canonical CpG islands are usually located within promoter regions, an interesting study reported that orphan CpG islands located far from the promoter region are functionally active enhancers (40). Therefore, we anticipate this region could be a critical DNA regulatory region for the regulation of NUP210 expression. Furthermore, most promoter regions that bind to SMARCB1 contain CpG islands (92.7%), similar to a fairly high proportion of enhancers (38.5%; Supplementary Fig. S7A), suggesting that SMARCB1 may have special roles via CpG island-related regulation. Although a reporter assay may be essential to define functional enhancers, we could not validate in here. However, we could find a strong relationship between this region and the NUP210 promoter based on integrative reference human epigenome data (49). Both regions showed active chromatin states in liver tissue and liver cell line, whereas tissue such as skin and adipocyte exhibited repressive states (Supplementary Fig. S7B). Very recently, this region is denoted as a promoter of IQSEC1 transcript4 by GENCODE34 v34, not NCBI Refseq (Supplementary Fig. S7C), however this transcript expression is not detected in liver tissue or liver cancer cell line HepG2, suggesting that this is not an active promoter (Supplementary Fig. S7D). In addition, there was no significant expression change of IQSEC1 total transcripts upon SMARCB1 knockdown. Furthermore, we could validate that this region belongs to the NUP210 promoter domain, not IQSEC1 transcripts in liver tissue based on a solid 3D genome browser Hi-C and Virtual 4C data (http://3dgenome.fsm.northwestern.edu/; Supplementary Figs. S7E and S7F) and GTEx eQTL (expression quantitative trait loci), which studies associate genomic and transcriptomic data sets from the same individuals to identify loci that affect mRNA expression. Altogether, we demonstrated that this region is presumably a NUP210 functional enhancer rather than IQSEC1 transcript 4 in liver cell contexts. Further investigation remains to understand whether this region plays differently depending on the context.

Several pioneering studies conducted over the past few years have shown that SMARCB1 and possibly other BAF complex subunits are deeply involved in the establishment and/or maintenance of enhancers (31, 38, 39, 50). SWI/SNF subunits ChIP-seq and biochemical assays in embryonic fibroblast and cancer cell lines demonstrated SWI/SNF-dependent enhancers are essential for developmental controlling genes (35), and this close relationship between SWI/SNF complex and enhancer is distinct in SMARCB1 loss contexts via complex integrity and antagonism with polycomb complex (36, 37). More recently, NUPs have been highlighted as key regulators in many systems via the maintenance of chromatin structure and subsequent gene regulation (22, 51). Although one definitive study has suggested that one specific NUP, namely, NUP98, mediates enhancer-promoter looping in metazoans (52), no report has yet discussed the involvement of NUPs and BAF complex in enhancer-related roles simultaneously. This is the first study to report that NUP210 loss reduces the level of the global enhancer mark H3K27Ac (Fig. 5F), which weakens SMARCB1 and P300 binding and further induces closed chromatin structure (Fig. 7). We have tried NUP210 ChIP-sequencing to address how NUP210 bridges P300 and SMARCB1 at its enhancer. Unfortunately, we did not yield definite peaks that are probably due to the quality of the NUP210 antibodies. However, because the P300 and SWI/SNF complexes are already very well known as enhancer-binding molecules, the effect on the physical interaction between SMARCB1 and P300 with or without NUP210 suggested that NUP210 serves as a platform for chromatin dependent process. These results might have filled the gap between the BAF complex and enhancers via the elucidation of NUP210 as a mediator. Domain experiments of these three proteins or NUP210 ChIP-seq using competent antibodies may provide important clues on how NUP210 bridges P300 and SMARCB1 interaction in the future. Whether the NUP210 function is specified in the context of SMARCB1 overexpression and whether other NUPs and BAF complex subunits also have roles in defining enhancer warrant further investigations.

Several studies have suggested that NUPs employ various oncogenesis mechanisms beyond the classic transport-dependent functions, such as chromosomal translocation, transcription, and establishment of chromatin boundaries (24). NUP210 is known as a tissue-specific NUP that plays key roles in myogenesis, neuronal differentiation, and, more recently, adaptive immune response (41–43, 53). However, the role of NUP210 in cancers, particularly liver cancers, remains to be elucidated. This study results revealed that NUP210 could be a novel tumor supporter (Fig. 5), which promotes the binding to SMARCB1 and P300 to chromatin (Fig. 7). In addition, SMARCB1 knockout melanoma RNA-seq analysis showed that the NUP210 level significantly reduced (37), suggesting that the relationship between SMARCB1 and NUP210 is not necessarily limited to liver cancers (Supplementary Fig. S7G). Gene and nuclear pore complex interaction could sperate the heterochromatin and euchromatin boundaries (54). NUP98 is involved in establishing boundaries in HOX gene loci by fused with PHD in the AML model (55); yet to date, whether NUP210 plays this kind of role in liver cancer context remained an open question. Because NUP210 probably has other mechanisms beyond those mentioned to support cancers with SMARCB1, it would be very interesting and worthwhile to conduct a detailed study on NUP210 in cancers, particularly relevant to other perturbations BAF complex subunits.

Chromatin remodelers and NUPs have cellular context-dependent compositions, and this type of heterogeneity may offer advantages in the regulation of chromatin structure and function in various dynamic processes (2, 21, 22). Based on our results, we propose that the intimate relationship between chromatin remodelers and NUPs plays a key role in the regulation of gene expression and further cellular identity by delineating the putative tumor supportive liaison between the chromatin remodeler subunits SMARCB1 and NUP210 in liver cancers. In further studies, it would be interesting to explore the convoluted relationship between other NUPs and chromatin remodelers, which would undoubtedly shed light on the physiological significance and development of therapeutic strategies for the treatment of many diseases due to the dysregulation of chromatin remodelers.

K.H. Son reports personal fees from Dankook University during the conduct of the study. K. Kang reports being a shareholder in Deargen Inc. No disclosures were reported by the other authors.

S.H. Hong: Data curation, investigation, writing-original draft. K.H. Son: Resources, software, visualization, methodology. S.Y. Ha: Resources, data curation, formal analysis. T.I. Wi: Formal analysis, validation. S.K. Choi: Formal analysis, validation. J.E. Won: Formal analysis. H.D. Han: Formal analysis, validation. Y. Ro: Methodology. Y.-M. Park: Resources. J.W. Eun: Formal analysis. S.W. Nam: Formal analysis. J.-W. Han: Resources. K. Kang: Resources, data curation. J.S. You: Conceptualization, resources, supervision, funding acquisition, investigation, visualization, writing-original draft, project administration, writing-review and editing.

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2016R1C1B3007534 and 2016R1A5A2012284).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Clapier
CR
,
Iwasa
J
,
Cairns
BR
,
Peterson
CL
. 
Mechanisms of action and regulation of ATP-dependent chromatin-remodelling complexes
.
Nat Rev Mol Cell Biol
2017
;
18
:
407
22
.
2.
Ho
L
,
Crabtree
GR
. 
Chromatin remodelling during development
.
Nature
2010
;
463
:
474
84
.
3.
de la Serna
IL
,
Ohkawa
Y
,
Imbalzano
AN
. 
Chromatin remodelling in mammalian differentiation: lessons from ATP-dependent remodellers
.
Nat Rev Genet
2006
;
7
:
461
73
.
4.
Chi
TH
,
Wan
M
,
Lee
PP
,
Akashi
K
,
Metzger
D
,
Chambon
P
, et al
Sequential roles of Brg, the ATPase subunit of BAF chromatin remodeling complexes, in thymocyte development
.
Immunity
2003
;
19
:
169
82
.
5.
Gresh
L
,
Bourachot
B
,
Reimann
A
,
Guigas
B
,
Fiette
L
,
Garbay
S
, et al
The SWI/SNF chromatin-remodeling complex subunit SNF5 is essential for hepatocyte differentiation
.
EMBO J
2005
;
24
:
3313
24
.
6.
Wilson
BG
,
Roberts
CW
. 
SWI/SNF nucleosome remodellers and cancer
.
Nat Rev Cancer
2011
;
11
:
481
92
.
7.
You
JS
,
De Carvalho
DD
,
Dai
C
,
Liu
M
,
Pandiyan
K
,
Zhou
XJ
, et al
SNF5 is an essential executor of epigenetic regulation during differentiation
.
PLos Genet
2013
;
9
:
e1003459
.
8.
Biegel
JA
,
Busse
TM
,
Weissman
BE
. 
SWI/SNF chromatin remodeling complexes and cancer
.
Am J Med Genet C Semin Med Genet
2014
;
166C
:
350
66
.
9.
Kadoch
C
,
Crabtree
GR
. 
Mammalian SWI/SNF chromatin remodeling complexes and cancer: mechanistic insights gained from human genomics
.
Sci Adv
2015
;
1
:
e1500447
.
10.
Kadoch
C
,
Hargreaves
DC
,
Hodges
C
,
Elias
L
,
Ho
L
,
Ranish
J
, et al
Proteomic and bioinformatic analysis of mammalian SWI/SNF complexes identifies extensive roles in human malignancy
.
Nat Genet
2013
;
45
:
592
601
.
11.
Versteege
I
,
Sevenet
N
,
Lange
J
,
Rousseau-Merck
MF
,
Ambros
P
,
Handgretinger
R
, et al
Truncating mutations of hSNF5/INI1 in aggressive paediatric cancer
.
Nature
1998
;
394
:
203
6
.
12.
Roberts
CW
,
Galusha
SA
,
McMenamin
ME
,
Fletcher
CD
,
Orkin
SH
. 
Haploinsufficiency of Snf5 (integrase interactor 1) predisposes to malignant rhabdoid tumors in mice
.
Proc Natl Acad Sci
2000
;
97
:
13796
800
.
13.
Lee
D
,
Kim
JW
,
Seo
T
,
Hwang
SG
,
Choi
EJ
,
Choe
J
. 
SWI/SNF complex interacts with tumor suppressor p53 and is necessary for the activation of p53-mediated transcription
.
J Biol Chem
2002
;
277
:
22330
7
.
14.
Roberts
CW
,
Leroux
MM
,
Fleming
MD
,
Orkin
SH
. 
Highly penetrant, rapid tumorigenesis through conditional inversion of the tumor suppressor gene Snf5
.
Cancer Cell
2002
;
2
:
415
25
.
15.
Doan
DN
,
Veal
TM
,
Yan
Z
,
Wang
W
,
Jones
SN
,
Imbalzano
AN
. 
Loss of the INI1 tumor suppressor does not impair the expression of multiple BRG1-dependent genes or the assembly of SWI/SNF enzymes
.
Oncogene
2004
;
23
:
3462
73
.
16.
Isakoff
MS
,
Sansam
CG
,
Tamayo
P
,
Subramanian
A
,
Evans
JA
,
Fillmore
CM
, et al
Inactivation of the Snf5 tumor suppressor stimulates cell cycle progression and cooperates with p53 loss in oncogenic transformation
.
Proc Natl Acad Sci U S A
2005
;
102
:
17745
50
.
17.
Wilson
BG
,
Wang
X
,
Shen
X
,
McKenna
ES
,
Lemieux
ME
,
Cho
YJ
, et al
Epigenetic antagonism between polycomb and SWI/SNF complexes during oncogenic transformation
.
Cancer Cell
2010
;
18
:
316
28
.
18.
Han
Y
,
Reyes
AA
,
Malik
S
,
He
Y
. 
Cryo-EM structure of SWI/SNF complex bound to a nucleosome
.
Nature
2020
;
579
:
452
5
.
19.
Schick
S
,
Rendeiro
AF
,
Runggatscher
K
,
Ringler
A
,
Boidol
B
,
Hinkel
M
, et al
Systematic characterization of BAF mutations provides insights into intracomplex synthetic lethalities in human cancers
.
Nat Genet
2019
;
51
:
1399
410
.
20.
Alberts
B
,
Johnson
A
,
Lewis
J
,
Raff
M
,
Roberts
K
,
Walter
P
.
Molecular biology of the cell
.
New York, USA
:
Garland Science
; 
2002
.
21.
Raices
M
,
D'Angelo
MA
. 
Nuclear pore complex composition: a new regulator of tissue-specific and developmental functions
.
Nat Rev Mol Cell Biol
2012
;
13
:
687
99
.
22.
Ptak
C
,
Wozniak
RW
. 
Nucleoporins and chromatin metabolism
.
Curr Opin Cell Biol
2016
;
40
:
153
60
.
23.
Raices
M
,
D'Angelo
MA
. 
Nuclear pore complexes and regulation of gene expression
.
Curr Opin Cell Biol
2017
;
46
:
26
32
.
24.
Kohler
A
,
Hurt
E
. 
Gene regulation by nucleoporins and links to cancer
.
Mol Cell
2010
;
38
:
6
15
.
25.
Sakuma
S
,
D'Angelo
MA
. 
The roles of the nuclear pore complex in cellular dysfunction, aging and disease
.
Semin Cell Dev Biol
2017
;
68
:
72
84
.
26.
Langmead
B
,
Salzberg
SL
. 
Fast gapped-read alignment with Bowtie 2
.
Nat Methods
2012
;
9
:
357
9
.
27.
Heinz
S
,
Benner
C
,
Spann
N
,
Bertolino
E
,
Lin
YC
,
Laslo
P
, et al
Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities
.
Mol Cell
2010
;
38
:
576
89
.
28.
Huang da
W
,
Sherman
BT
,
Lempicki
RA
. 
Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources
.
Nat Protoc
2009
;
4
:
44
57
.
29.
Tripathi
S
,
Pohl
MO
,
Zhou
Y
,
Rodriguez-Frandsen
A
,
Wang
G
,
Stein
DA
, et al
Meta- and orthogonal integration of influenza “OMICs” data defines a role for UBR4 in virus budding
.
Cell Host Microbe
2015
;
18
:
723
35
.
30.
Kuleshov
MV
,
Jones
MR
,
Rouillard
AD
,
Fernandez
NF
,
Duan
Q
,
Wang
Z
, et al
Enrichr: a comprehensive gene set enrichment analysis web server 2016 update
.
Nucleic Acids Res
2016
;
44
:
W90
7
.
31.
Nakayama
RT
,
Pulice
JL
,
Valencia
AM
,
McBride
MJ
,
McKenzie
ZM
,
Gillespie
MA
, et al
SMARCB1 is required for widespread BAF complex-mediated activation of enhancers and bivalent promoters
.
Nat Genet
2017
;
49
:
1613
23
.
32.
Simon
JM
,
Giresi
PG
,
Davis
IJ
,
Lieb
JD
. 
Using formaldehyde-assisted isolation of regulatory elements (FAIRE) to isolate active regulatory DNA
.
Nat Protoc
2012
;
7
:
256
67
.
33.
Kim
KH
,
Kim
W
,
Howard
TP
,
Vazquez
F
,
Tsherniak
A
,
Wu
JN
, et al
SWI/SNF-mutant cancers depend on catalytic and non-catalytic activity of EZH2
.
Nat Med
2015
;
21
:
1491
6
.
34.
Xu
Y
,
Yan
W
,
Chen
X
. 
SNF5, a core component of the SWI/SNF complex, is necessary for p53 expression and cell survival, in part through eIF4E
.
Oncogene
2010
;
29
:
4090
100
.
35.
Howard
TP
,
Arnoff
TE
,
Song
MR
,
Giacomelli
AO
,
Wang
X
,
Hong
AL
, et al
MDM2 and MDM4 are therapeutic vulnerabilities in malignant rhabdoid tumors
.
Cancer Res
2019
;
79
:
2404
14
.
36.
Carugo
A
,
Minelli
R
,
Sapio
L
,
Soeung
M
,
Carbone
F
,
Robinson
FS
, et al
p53 is a master regulator of proteostasis in SMARCB1-deficient malignant rhabdoid tumors
.
Cancer Cell
2019
;
35
:
204
20
.
37.
Wang
L
,
Leite de Oliveira
R
,
Wang
C
,
Fernandes Neto
JM
,
Mainardi
S
,
Evers
B
, et al
High-throughput functional genetic and compound screens identify targets for senescence induction in cancer
.
Cell Rep
2017
;
21
:
773
83
.
38.
Alver
BH
,
Kim
KH
,
Lu
P
,
Wang
X
,
Manchester
HE
,
Wang
W
, et al
The SWI/SNF chromatin remodelling complex is required for maintenance of lineage specific enhancers
.
Nat Commun
2017
;
8
:
14648
.
39.
Wang
X
,
Lee
RS
,
Alver
BH
,
Haswell
JR
,
Wang
S
,
Mieczkowski
J
, et al
SMARCB1-mediated SWI/SNF complex function is essential for enhancer regulation
.
Nat Genet
2017
;
49
:
289
95
.
40.
Bell
JSK
,
Vertino
PM
. 
Orphan CpG islands define a novel class of highly active enhancers
.
Epigenetics
2017
;
12
:
449
64
.
41.
D'Angelo
MA
,
Gomez-Cavazos
JS
,
Mei
A
,
Lackner
DH
,
Hetzer
MW
. 
A change in nuclear pore complex composition regulates cell differentiation
.
Dev Cell
2012
;
22
:
446
58
.
42.
Gomez-Cavazos
JS
,
Hetzer
MW
. 
The nucleoporin gp210/Nup210 controls muscle differentiation by regulating nuclear envelope/ER homeostasis
.
J Cell Biol
2015
;
208
:
671
81
.
43.
Raices
M
,
Bukata
L
,
Sakuma
S
,
Borlido
J
,
Hernandez
LS
,
Hart
DO
, et al
Nuclear pores regulate muscle development and maintenance by assembling a localized Mef2C complex
.
Dev Cell
2017
;
41
:
540
54
.
44.
Jiang
Y
,
Sun
A
,
Zhao
Y
,
Ying
W
,
Sun
H
,
Yang
X
, et al
Proteomics identifies new therapeutic targets of early-stage hepatocellular carcinoma
.
Nature
2019
;
567
:
257
61
.
45.
De Matteis
S
,
Ragusa
A
,
Marisi
G
,
De Domenico
S
,
Casadei Gardini
A
,
Bonafe
M
, et al
Aberrant metabolism in hepatocellular carcinoma provides diagnostic and therapeutic opportunities
.
Oxid Med Cell Longev
2018
;
2018
:
7512159
.
46.
Kuzu
OF
,
Noory
MA
,
Robertson
GP
. 
The role of cholesterol in cancer
.
Cancer Res
2016
;
76
:
2063
70
.
47.
Hammitzsch
A
,
Tallant
C
,
Fedorov
O
,
O'Mahony
A
,
Brennan
PE
,
Hay
DA
, et al
CBP30, a selective CBP/p300 bromodomain inhibitor, suppresses human Th17 responses
.
Proc Natl Acad Sci U S A
2015
;
112
:
10768
73
.
48.
Dubey
R
,
Lebensohn
AM
,
Bahrami-Nejad
Z
,
Marceau
C
,
Champion
M
,
Gevaert
O
, et al
Chromatin-remodeling complex SWI/SNF controls multidrug resistance by transcriptionally regulating the drug efflux pump ABCB1
.
Cancer Res
2016
;
76
:
5810
21
.
49.
Roadmap Epigenomics
C
,
Kundaje
A
,
Meuleman
W
,
Ernst
J
,
Bilenky
M
,
Yen
A
, et al
Integrative analysis of 111 reference human epigenomes
.
Nature
2015
;
518
:
317
30
.
50.
Kadoch
C
,
Williams
RT
,
Calarco
JP
,
Miller
EL
,
Weber
CM
,
Braun
SM
, et al
Dynamics of BAF-Polycomb complex opposition on heterochromatin in normal and oncogenic states
.
Nat Genet
2017
;
49
:
213
22
.
51.
Ibarra
A
,
Hetzer
MW
. 
Nuclear pore proteins and the control of genome functions
.
Genes Dev
2015
;
29
:
337
49
.
52.
Pascual-Garcia
P
,
Debo
B
,
Aleman
JR
,
Talamas
JA
,
Lan
Y
,
Nguyen
NH
, et al
Metazoan nuclear pores provide a scaffold for poised genes and mediate induced enhancer-promoter contacts
.
Mol Cell
2017
;
66
:
63
76
.
53.
Borlido
J
,
Sakuma
S
,
Raices
M
,
Carrette
F
,
Tinoco
R
,
Bradley
LM
, et al
Nuclear pore complex-mediated modulation of TCR signaling is required for naive CD4(+) T cell homeostasis
.
Nat Immunol
2018
;
19
:
594
605
.
54.
Liang
Y
,
Hetzer
MW
. 
Functional interactions between nucleoporins and chromatin
.
Curr Opin Cell Biol
2011
;
23
:
65
70
.
55.
Wang
GG
,
Song
J
,
Wang
Z
,
Dormann
HL
,
Casadio
F
,
Li
H
, et al
Haematopoietic malignancies caused by dysregulation of a chromatin-binding PHD finger
.
Nature
2009
;
459
:
847
51
.