Uveal melanoma is a rare and aggressive cancer that originates in the eye. Currently, there are no approved targeted therapies and very few effective treatments for this cancer. Although activating mutations in the G protein alpha subunits, GNAQ and GNA11, are key genetic drivers of the disease, few additional drug targets have been identified. Recently, studies have identified context-specific roles for the mammalian SWI/SNF chromatin remodeling complexes (also known as BAF/PBAF) in various cancer lineages. Here, we find evidence that the SWI/SNF complex is essential through analysis of functional genomics screens and further validation in a panel of uveal melanoma cell lines using both genetic tools and small-molecule inhibitors of SWI/SNF. In addition, we describe a functional relationship between the SWI/SNF complex and the melanocyte lineage–specific transcription factor Microphthalmia-associated Transcription Factor, suggesting that these two factors cooperate to drive a transcriptional program essential for uveal melanoma cell survival. These studies highlight a critical role for SWI/SNF in uveal melanoma, and demonstrate a novel path toward the treatment of this cancer.

Uveal melanoma is a rare cancer that arises from the melanocytes of the uvea. Although it originates in melanocytes, the underlying mutations and unique immune environment of the eye distinguish this cancer from the more common cutaneous melanoma. Uveal melanoma is mainly characterized by driver mutations in GNAQ or GNA11 which lead to hyperactivation of the G proteins, resulting in downstream MAPK pathway activation (1, 2). In addition, chromosome 8q amplification and BAP1 loss of heterozygosity (chromosome 3) or silencing are also commonly observed and correlate with increased aggressiveness and poor prognosis (3). To date, targeted agents, primarily against the MAPK pathway and its effectors, such as PKC, and even immunotherapy, have resulted in limited clinical responses, and surgery and radiotherapy remain the most common treatments, with overall poor prognosis upon detection of metastatic disease (4). Due to the paucity of available treatments, further studies to understand the biology of the disease, as well as elucidate novel therapeutic targets, remain critical.

We set out to uncover novel dependencies in uveal melanoma through analysis of previously performed unbiased pooled short hairpin RNA (shRNA) screens, and discovered an unanticipated role of the SWI/SNF chromatin–remodeling complex in survival of uveal melanomas. The SWI/SNF complex represents an important tumor suppressor in cancer, with approximately 20% of tumors harboring mutations in one or more of its subunits (5, 6). More recently however, BRG1/SMARCA4, the catalytic subunit of the SWI/SNF complex, has been shown to be essential for cancer survival, such as in acute myeloid leukemia where it works in concert with leukemic transcription factors to facilitate MYC expression (7, 8). In addition, it has been demonstrated that SWI/SNF drives active enhancer state maintenance in a lineage-specific manner, further suggesting its context-specific role in tumor maintenance (9–11). By this logic, the SWI/SNF complex represents a targetable node in the complex lineage-specific transcriptional machinery that may drive certain cancers.

In this study, we show that uveal melanoma models are dependent on the SWI/SNF catalytic subunits BRG1 and BRM (also known as SMARCA4 and SMARCA2, respectively) by genetic knockdown. We also demonstrate broad activity of small-molecule inhibitors of BRM/BRG1 ATPase activity (12, 13) across a panel of uveal melanoma cell lines. In an effort to dissect the mechanism of SWI/SNF dependence, our work reveals a functional link between SWI/SNF and the transcription factor Microphthalmia associated Transcription Factor (MITF). Together, our data identify SWI/SNF as a novel target in uveal melanoma and reveal the therapeutic potential of applying small-molecule inhibition of SWI/SNF for the treatment of uveal melanoma.

Cell lines and reagents

BRM011, BRM014, and BRM017 (synthesis described in refs. 12, 13) stocks were dissolved at 10 mmol/L in DMSO. Doxycycline stock solution was made at 100 μg/mL in water. Shield 1 (Clontech) was dissolved at 0.5 mmol/L in ethanol.

Cell lines were obtained from the ATCC (MP41, MM28, MP46, MP65, MP38, and SW13), Sigma (Mel202), Leiden University medical center (92.1 and OMM1), and Lonza (human epidermal melanocytes), and cultured in manufacturer-recommended media (92.1 and OMM1 were cultured in RPMI1640 + 10% FBS). All parental lines were SNP profiled using the Fluidigm assay described previously (14) and tested negative for Mycoplasma infection by qPCR (tests performed by Idexx Biosciences between March 2016 and February 2019). Cell lines were used for no more than 1 to 3 months after thawing depending on doubling time.

Cell line engineering

shRNA cloning into pLKO-based inducible vectors and cell line generation were described previously (15). Cell line details are annotated in Supplementary Table S3.

DD-BRG1 was assembled by adding Shield destabilization domain (DD) sequence (16) to the N-terminus of BRG1 open reading frame (ORF). ACTL6A cDNA (Invitrogen) and DD-BRG1 were cloned into an in-house constitutive expression vector (lentiviral with EF1alpha promoter driving ORF expression). Flag-HA-streptavidin–tagged MITF-M ORF was cloned into pLNCX-2 (Clontech).

Growth, viability, and apoptosis assays

To measure cell growth, cell lines were plated in 96-well plates (92.1 5,000 cells, OMM1 2,500, MP41 2,500), then imaged on an IncuCyte (4x objective), and analyzed using IncuCyte Zoom 2016B software.

For viability assays, cells were plated in 384-well plates (Corning 3765; 92.1 500 cells, OMM1 500, MP41 500, Mel202 1,500, MP46 1,000, MM28 4,000, MP38 3,000, MP65 1,500, SW13 1,000, and melanocytes 3,000). Plates were dosed with an 11-point, 3-fold serial dilution using the Echo550 (Labcyte). After 5 days, Cell Titer Glo (Promega) was added and luminescence measured (PHERAstar, BMG Labtech). Growth inhibition values were calculated as described previously (17). Normalized data were fit using the three-parameter nonlinear regression function in GraphPad Prism 7. Absolute AC50s (AAC50) were reported as concentrations of compound where curve fit crosses 0.5.

For caspase activity, cells were plated and treated as above (92.1 1,500 cells, OMM1 1,500, MP41 1,500, Mel202 3,000, MP46 3,000, MM28 5,000, MP38 5,000, MP65 3,500, SW13 3,000). After 48 hours, Caspase 3/7 Glo (Promega) was added and luminescence measured. Normalization was performed relative to untreated wells and plotted as fold activity.

Single-point viability and caspase activity assays were performed in indicated 92.1 shRNA lines treated with 250 nmol/L BRM011, BRM014, BRM017 or 100 ng/mL doxycycline for 3 days and then assayed as above and analyzed by normalizing relative to untreated wells.

Compound profiling in 92.1

Cells were plated in 384-well plates in duplicate at 2,000 cells per well. Cells were dosed with an 8-point, 3-fold serial dilution (compounds 8–14; Supplementary Table S1) on an Echo 550. Three days after treatment, cell viability was measured using ATPlite 1 step (Perkin Elmer) and luminescence measured. AAC50 for compounds were determined using an in-house statistics package (HELIOS).

Western blot

Cells were harvested in T-PER lysis buffer (Thermo Fisher Scientific), 50 mmol/L DTT, and HALT protease/phosphatase inhibitor cocktail (Thermo Fisher Scientific), and then diluted in Laemmli sample buffer (Bio-Rad). Proteins were separated by SDS-PAGE and transferred to 0.2 μmol/L nitrocellulose membrane (Bio-Rad). Antibody information can be found in Supplementary Table S4. Blots were visualized on a Bio-Rad Chemidoc imager by chemiluminescence (Pierce ECL, Thermo Fisher Scientific).

RNA sequencing

Sample preparation and sequencing

92.1 cells were plated in triplicate at 150,000 (72- or 48-hour treatments) or 250,000 (24-hour treatment) cells per well in a 6-well plate. After treatment, RNA was isolated using Qiagen's RNeasy Plus kit according to the manufacturer's instructions. We assessed RNA integrity using the Agilent 2100 and Agilent RNA 6000 Nano Kit.

Sample libraries were generated as per the manufacturer's specifications on the Hamilton STAR robotics platform using the TruSeq Stranded mRNA Library Prep Kit, High Throughput (Illumina), and 200 ng input RNA. The PCR-amplified RNA sequencing (RNA-seq) library products were quantified using the Advanced Analytical Fragment Analyzer Standard Sensitivity NGS Fragment Analysis Kit (Agilent). Samples were diluted to 10 nmol/L in Elution Buffer (Qiagen), denatured, and loaded between 2.5 and 4.0 pmol/L on an Illumina cBOT using the HiSeq 4000 PE Cluster Kit (Illumina). Sequencing was performed on a HiSeq 4000 at 75 base pair paired end with 8 base pair dual indexes using the HiSeq 4000 SBS Kit, 150 cycles (Illumina), and sequence intensity files were generated on instrument using the Illumina Real Time Analysis software. The resulting intensity files were demultiplexed with the bcl2fastq2 software and aligned to the human transcriptome using PISCES version 2018.04.01.

Differential expression and pathway enrichment analysis

Differential expression was determined using limma from Bioconductor (PMID: 25605792). Genes were called differentially expressed if they had an average Log2 expression ≥0 (reported as AveExpr in limma), an adjusted P value ≤ 0.01, and an absolute Log2 fold change of 0.5 relative to DMSO control. Gene set enrichment analysis for differentially expressed genes was performed using the hypergeometric test with FDR-adjusted P values for two pathway sets downloaded from MSigDB (PMID:16199517; Hallmark and KEGG).

Chromatin immunoprecipitation qPCR

Assay was designed and performed by Active Motif. Details can be found in Supplemental Methods.

ATAC-Seq

Sample preparation and sequencing

Note that 100,000 cells were processed for ATAC-Seq using OMNI-ATAC-Seq protocol (18). After tagmentation, samples were purified using Zymo DNA clean and concentration kit according to the manufacturer's protocol. Purified DNA was used for library amplification using Nextera Index kit (Illumina) and NEBNext High-Fidelity PCR master mix (New England biolabs). Number of amplification cycles (final <11) was determined individually for each sample by qPCR using SYBR green and calculated as number of PCR amplification cycles required to achieve one-third maximum fluorescent intensity. Amplified libraries were purified using AMPure XP beads (Beckman). Final product quality and concentration was determined using the D5000 DNA Tape on the Agilent Tapestation. Samples were diluted to 8 nmol/L in Elution Buffer (Qiagen), denatured, and loaded at 6.4 pmol/L on a Miseq at 75 base pair paired end with 8 base pair dual indexes using the MiSeq Reagent Kit v3, 150 cycles (Illumina). Sequencing data were used to recalculate sample concentration adjusting for sample representation. Using updated concentrations, samples were diluted to 8 nmol/L in Elution Buffer, denatured, and loaded at 6.4 pmol/L on an Illumina cBOT using the HiSeq 4000 PE Cluster Kit (Illumina). Libraries were sequenced on a HiSeq 4000 at 75 base pair paired end with 8 base pair dual indexes using the HiSeq 4000 SBS Kit, 150 cycles (Illumina), and sequence intensity files were generated on instrument using the Illumina Real Time Analysis software. Detailed data analysis methods can be found in Supplemental Methods.

RT-qPCR

Cells were plated at 15,000 cells per well in 96-well plate in biological quadruplicate per treatment condition and treated with BRM011, BRM014, or BRM017 (0, 0.1, 1, 10, 100, 1,000 nmol/L) for 24 hours. After treatment, cells were lysed using Cells-to-Ct Bulk Lysis reagents (Thermo Fisher Scientific) and cDNA synthesized using Cell-to-Ct Bulk RT reagents (Thermo Fisher Scientific) according to the manufacturer's protocol. RT-qPCR was performed with Taqman Fast Advanced Master Mix (Applied Biosystems) on a Viia 7 real-time PCR system (Applied Biosystems). Probes are annotated in Supplementary Table S4. Relative quantification for each sample was calculated using the 2−ΔΔCt method (TBP normalized and expressed as fold change relative to DMSO treated). Graphing and analyses were performed using GraphPad Prism 7 (Graphpad Software).

In vivo efficacy study

Mice were maintained and handled in accordance with the Novartis Institutes for BioMedical Research (NIBR) Institutional Animal Care and Use Committee (IACUC), and all studies were approved by the NIBR IACUC. Female athymic nude mice (Charles River) were acclimated in NIBR animal facility (12-hour light/dark cycle) with ad libitum access to food and water for at least 3 days before manipulation. Mice (6–8-week-old) were inoculated subcutaneously in the right dorsal axillary region with the 92.1 cell line (10 × 106 cells in 50% Matrigel). Tumor volumes and body weights were monitored twice per week, and the general health condition of mice was monitored daily. Tumor volume was determined by measurement with calipers and calculated using a modified ellipsoid formula, where tumor volume (TV, mm3) = [((l × w2) × 3.14159))/6], where l is the longest axis of the tumor and w is perpendicular to l. When average tumor volume reached approximately 200 mm3, animals were randomly assigned to receive daily dosing of either vehicle or BRM014 20 mg/kg. Compound treatments began 20 days after 92.1 cell implantation, and tumor samples were collected for pharmacodynamics analysis after 27 days of daily dosing.

RT-qPCR

Lysis buffer from Qiagen RNeasy plus mini kit was added to tumors samples, which were then homogenized using Lysing Matrix D beads using a Precellys 24 homogenizer. Samples were further homogenized using Qiashredder columns (Qiagen), and RNA isolated using the Qiagen RNeasy plus mini kit according to the manufacturer's protocol. cDNA synthesis was performed according to the manufacturer's protocol using ABI high-capacity cDNA synthesis kit and 1 μg input RNA. RT-qPCR was performed in technical duplicate (2 reactions per tumor) with FastStart Universal Probe master mix with Rox (MilliporeSigma) on a CFX384 Touch Real-Time PCR Detection System (Bio-Rad). Probes are annotated in Supplementary Table S4. Relative quantification for each sample was calculated using the 2−ΔΔCt method (β-actin normalized and expressed as fold change relative to vehicle control). Graphing and analyses were performed using GraphPad Prism 7 (Graphpad Software).

Data accession

RNA-seq and ATAC-seq datasets can be accessed under BioProject ID: PRJNA622863 (http://www.ncbi.nlm.nih.gov/bioproject/622863).

SWI/SNF complex is essential for uveal melanoma proliferation

Pooled screening efforts using shRNA have uncovered novel dependencies across a variety of cancer types (19, 20), but uveal melanoma is underrepresented across these studies due to the general scarcity of models in this field. Due to this, only three cell lines, OMM1 (GNA11 Q209L), 92.1 (GNAQ Q209L), and MEL285 (GNAQ/11 WT), were included in the large-scale DRIVE screening effort (19). Notably, in the two GNAQ/11-mutant cell lines, OMM1 and 92.1, various subunits of the SWI/SNF chromatin remodeling complexes, including BRG1/SMARCA4, ACTL6A/BAF53a, SMARCB1, ARID1A, and SMARCE1, were among the top sensitizers (Fig. 1A).

Figure 1.

Uveal melanoma cell lines are dependent on SWI/SNF complex members. A, OMM1 and 92.1 pooled shRNA screen data from McDonald et al (2017) showing all genes tested ranked by sensitivity score (calculated using ATARiS method). MITF, GNAQ, GNA11, and SWI/SNF subunits that scored in the screen are highlighted. B–D, Top, growth as measured by confluence for indicated cell lines is graphed. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images for indicated time point. Scale bar, 0.5 mm.

Figure 1.

Uveal melanoma cell lines are dependent on SWI/SNF complex members. A, OMM1 and 92.1 pooled shRNA screen data from McDonald et al (2017) showing all genes tested ranked by sensitivity score (calculated using ATARiS method). MITF, GNAQ, GNA11, and SWI/SNF subunits that scored in the screen are highlighted. B–D, Top, growth as measured by confluence for indicated cell lines is graphed. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images for indicated time point. Scale bar, 0.5 mm.

Close modal

To validate the SWI/SNF subunit sensitizers, we engineered doxycycline-inducible shRNAs against the genes that scored as hits. Interestingly, 92.1 showed a profound dependency on BRG1 alone (Fig. 1B and C), but not BRM (Supplementary Fig. S1D and S1E). However, in two other uveal melanoma lines, OMM1 and MP41, knockdown of neither BRG1 nor BRM alone had an effect on cell growth (Fig. 1B; Supplementary Fig. S1A–S1E). Consistent with the screening data, robust depletion of obligate SWI/SNF subunit ACTL6A (21, 22) in OMM1 led to a growth arrest (Fig. 1D; Supplementary Fig. S1G), whereas partial depletion was insufficient (Supplementary Fig. S1F and S1G). This phenotype was confirmed to be on-target as evidenced by a growth rescue with ectopic expression of shRNA-resistant ACTL6A (Fig. 1D; Supplementary Fig. S1G). These results suggest that the ACTL6A dependency potentially reflects the importance of losing the combined catalytic activity of BRG1 and BRM, because ACTL6A is present in both BRG1- and BRM-containing SWI/SNF complexes. We tested this possibility by simultaneous knockdown of BRM and BRG1 in the same cell lines, which notably led to a robust growth effect (Fig. 2, Supplementary Fig. S2A and S2B). Together, these data reveal that the SWI/SNF complex is essential in uveal melanoma.

Figure 2.

Uveal melanoma cell lines are dependent on the ATPase activity of SWI/SNF. Top, growth as measured by confluence for indicated cell lines is graphed. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images at t = 156 h. Scale bar, 0.5 mm.

Figure 2.

Uveal melanoma cell lines are dependent on the ATPase activity of SWI/SNF. Top, growth as measured by confluence for indicated cell lines is graphed. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images at t = 156 h. Scale bar, 0.5 mm.

Close modal

Next, we wished to determine the specific role of SWI/SNF ATPase activity in the observed dependency. Although the K785R mutation has been shown to abolish BRG1 catalytic activity without affecting the protein's ability to incorporate into the complex (23), it has also been described as a dominant negative allele (24). Due to this, we used a DD BRG1 construct to engineer 92.1 (Supplementary Fig. S2C), which repressed expression of the allele in the absence of the stabilizing Shield 1 ligand (Supplementary Fig. S2D; ref. 16). We observed successful rescue of proliferation upon expression of the wild-type allele, but did not see any rescue with the ATPase dead DD-BRG1 (Supplementary Fig. S2E). These data suggest the importance of the ATPase activity of BRG1; however, we cannot rule out other explanations due to the dominant negative behavior of this allele.

Uveal melanoma cell lines are exquisitely sensitive to small-molecule inhibition of SWI/SNF ATPase activity

The discovery of genetic dependency on SWI/SNF provided a clear rationale to test the activity of dual BRG1/BRM ATPase small-molecule inhibitors in this lineage. The two most potent inhibitors, BRM011 and BRM014, show equivalent inhibition of BRG1 and BRM catalytic activity in vitro, as well as robust inhibition of BRM-dependent gene expression and cancer cell growth (13). The activity of these compounds against BRM and BRG1 in biochemical characterization suggests that they would similarly inhibit both proteins in cells. This was tested and confirmed by measuring concordant antiproliferative activity with analogs of various biochemical potencies in the 92.1 cell line which shows growth arrest upon BRG1 or dual BRG1/BRM knockdown but not BRM knockdown alone (ref. 13; Supplementary Table S1). Taken together, these results suggest that these compounds are capable of inhibiting both BRG1 and BRM cellular activity. We then treated a panel of uveal melanoma cell lines all expressing key members of the SWI/SNF complex (BRG1, BRM, ARID1A, and ACTL6A; Supplementary Fig. S3A), with BRM011 and BRM014. Cell viability was measured after treatment with BRM011 or BRM014 and compared with BRM017, a structurally related cell-inactive analog (12). Across the cell line panel, there was a profound dose-dependent sensitivity to both active compounds, whereas the control analog BRM017 showed a growth effect at only the highest concentrations (>100 fold shifted from concentrations in which an equivalent phenotype is seen with BRM011; Fig. 3A). In addition, increased caspase 3/7 activity after compound treatment was measured in 92.1, Mel202, MP38, and MP41, indicating that these cells are undergoing apoptosis upon SWI/SNF inhibition (Fig. 3B), whereas OMM1, MM28, MP46, and MP65 show a stasis phenotype (Fig. 3A; Supplementary Fig. S3B). The increase in caspase activity observed upon compound treatment was similar to that observed in the 92.1 dual BRG1/BRM shRNA lines, suggesting that the apoptotic phenotype is likely the result of SWI/SNF inhibition (Supplementary Fig. S3C). In summary, 7 of the 8 cell lines in our panel were highly sensitive to the dual inhibitor BRM014 with absolute AC50s (AAC50) <100 nmol/L (Supplementary Table S2). Of note, we found no underlying mutation pattern in the SWI/SNF complex members that would explain the observed sensitivity to inhibition (Supplementary Fig. S3D).

Figure 3.

Chemical inhibition of SWI/SNF causes growth arrest in uveal melanoma cell lines. A, Viability after uveal melanoma cell lines were treated with BRM011, BRM014, and BRM017 for 5 days is plotted relative to DMSO treated and baseline (0.0) is set at day 0 viability. Error bars are shown as S.D., N = 4. Data were fit using GraphPad Prism. B, Fold caspase activity in uveal melanoma cell lines treated with BRM011, BRM014, and BRM017 for 48 hours is plotted relative to DMSO treated. Error bars are shown as SD, N = 4.

Figure 3.

Chemical inhibition of SWI/SNF causes growth arrest in uveal melanoma cell lines. A, Viability after uveal melanoma cell lines were treated with BRM011, BRM014, and BRM017 for 5 days is plotted relative to DMSO treated and baseline (0.0) is set at day 0 viability. Error bars are shown as S.D., N = 4. Data were fit using GraphPad Prism. B, Fold caspase activity in uveal melanoma cell lines treated with BRM011, BRM014, and BRM017 for 48 hours is plotted relative to DMSO treated. Error bars are shown as SD, N = 4.

Close modal

To confirm the selectivity of our compounds, we treated the BRG1/BRM-deficient cell line SW13 (25, 26). In this cell line, we only saw partial activity of BRM011 in the viability assay (Amax = 50%), and no activation of caspase activity (Supplementary Fig. S3E; Supplementary Table S2). In order to understand the potential for differential responses between tumor and normal tissue, we also treated nontransformed melanocytes and observed a partial antiproliferative effect. This activity, particularly for compound BRM014, was significantly less potent on the melanocytes (BRM011 AAC50 = 46.2 nmol/L, BRM014 AAC50 = 454.8 nmol/L; Supplementary Fig. S3F; Supplementary Table S2) than observed for many of the uveal cell lines (more than 500-fold shifted relative to 92.1), suggesting the potential for a window of SWI/SNF inhibition between the malignant uveal and normal melanocyte population.

The sensitivity observed with small-molecule inhibition of SWI/SNF ATPase activity further validates the SWI/SNF dependency observed with genetic perturbation of the complex, and indicates that the dependency may be more broadly applicable across uveal melanoma models. The activity of BRM011 and BRM014 against both BRG1 and BRM allowed successful interrogation of this dependency across a panel of uveal melanoma lines, which may have otherwise not have shown sensitivity to inhibition of only one catalytic subunit alone, such as in the case of OMM1 and MP41.

Genomic profiling reveals a functional link between SWI/SNF and MITF activity

We next interrogated the mechanisms underlying SWI/SNF-driven growth in uveal melanoma, testing the hypothesis that a functional relationship with an essential transcription factor may direct the complex to important loci where SWI/SNF-mediated chromatin remodeling could drive an essential transcriptional program. We identified MITF as another dependency common to both 92.1 and OMM1 in the pooled screening data (Fig. 1A). MITF and BRG1 have been previously described to work together in various reports showing that BRG1 is essential to drive the MITF-dependent transcriptional program (27–31). In order to determine whether this is relevant in the uveal lineage, we first knocked down MITF using an inducible shRNA in 92.1, OMM1, and MP41, and observed a growth arrest in all three cell lines (Fig. 4A; Supplementary Fig. S4A and S4B). This phenotype was confirmed to be on target by rescuing the growth arrest in 92.1 through exogenous overexpression of MITF (Fig. 4A).

Figure 4.

SWI/SNF perturbation affects an MITF-dependent transcriptional program. A, Top, growth as measured by confluence for indicated cell lines. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images for indicated time point. Scale bar, 0.5 mm. B, Venn diagram of overlap in transcriptional changes under indicated conditions. Genes were determined to be significantly up- or downregulated if absolute log fold change (logFC) ≥0.5, average expression ≥1 and P ≤ 0.01. C, Heatmap of genes for which expression changes were measured under indicated conditions (absolute logFC ≥0.5, average expression ≥ 1 and P ≤ 0.01) in at least one condition is plotted together with expression across all other conditions. The color scale bar indicates the logFC ranges for the comparisons to untreated samples. D, ChIP-qPCR for MITF and BRG1 at indicated loci. N = 3. Error bars shown as SD. E, Median ATAC-Seq read density (RPKM, reads per kilobase per million mapped reads) is plotted for peaks at TSS. Data are shown for 92.1 treated with either BRM011 (100 nmol/L) or DMSO for 24 hours. F, ATAC-Seq tracks showing DCT and TYR promoter regions in 92.1 treated with BRM011 or DMSO for 24 hours. Average density for 3 replicates is shown. G, Viability is plotted for 92.1 engineered to express exogenous MITF or empty vector, then treated with BRM011 or BRM014 for 5 days. Viability is plotted relative to DMSO treated and baseline (0.0) is set at day 0 viability. Error bars are shown as SD, N = 4.

Figure 4.

SWI/SNF perturbation affects an MITF-dependent transcriptional program. A, Top, growth as measured by confluence for indicated cell lines. Expression of shRNAs was induced by addition of doxycycline. () + doxycycline (100 ng/mL), untreated. N = 3, error bars shown as SD. Bottom, representative images for indicated time point. Scale bar, 0.5 mm. B, Venn diagram of overlap in transcriptional changes under indicated conditions. Genes were determined to be significantly up- or downregulated if absolute log fold change (logFC) ≥0.5, average expression ≥1 and P ≤ 0.01. C, Heatmap of genes for which expression changes were measured under indicated conditions (absolute logFC ≥0.5, average expression ≥ 1 and P ≤ 0.01) in at least one condition is plotted together with expression across all other conditions. The color scale bar indicates the logFC ranges for the comparisons to untreated samples. D, ChIP-qPCR for MITF and BRG1 at indicated loci. N = 3. Error bars shown as SD. E, Median ATAC-Seq read density (RPKM, reads per kilobase per million mapped reads) is plotted for peaks at TSS. Data are shown for 92.1 treated with either BRM011 (100 nmol/L) or DMSO for 24 hours. F, ATAC-Seq tracks showing DCT and TYR promoter regions in 92.1 treated with BRM011 or DMSO for 24 hours. Average density for 3 replicates is shown. G, Viability is plotted for 92.1 engineered to express exogenous MITF or empty vector, then treated with BRM011 or BRM014 for 5 days. Viability is plotted relative to DMSO treated and baseline (0.0) is set at day 0 viability. Error bars are shown as SD, N = 4.

Close modal

To identify whether MITF and SWI/SNF share transcriptional targets as previously demonstrated in cutaneous melanoma (28, 31), expression analysis by RNA-seq was performed after SWI/SNF (BRG1/BRM) or MITF knockdown or treatment with BRM011 in 92.1. As expected, robust downregulation of the shRNA target genes was observed across the knockdown conditions (Supplementary Fig. S4C). In addition, there was a significant overlap in genes affected by BRG1/BRM dual knockdown versus MITF knockdown (P < 1E-5; Fig. 4B), as well as these two genetic perturbations and chemical inhibition of SWI/SNF activity by BRM011 (Supplementary Fig. S4D). Among the genes affected by both MITF and SWI/SNF knockdown, pathways affecting hypoxia, apoptosis, cell cycle, and differentiation were enriched (P < 1E-4), consistent with the profound growth effect observed upon knockdown (Supplementary Fig. S4E). The pattern of transcriptional changes across the three different perturbations indicates that SWI/SNF and MITF control an overlapping gene set (Fig. 4C). Taken together, these data suggest that SWI/SNF catalytic activity plays an important role in the MITF transcriptional program, and support a functional relationship between SWI/SNF and MITF.

To further interrogate the relationship between SWI/SNF and MITF, we performed chromatin immunoprecipitation and quantification by PCR (ChIP-qPCR) to monitor co-occupancy of BRG1 and MITF at discrete genomic loci. Specifically, the DCT and TYR promoters were probed in 92.1 cells because both show modulation by MITF and SWI/SNF knockdown (Supplementary Fig. S4F). Both MITF and BRG1 localized to the promoters of DCT and TYR consistent with previous reports (Fig. 4D; ref. 31), suggesting that SWI/SNF and MITF colocalize to certain loci to modulate gene expression.

Next, assay for transposase accessible chromatin (ATAC-Seq) was used to investigate the changes in chromatin structure that occur upon SWI/SNF inhibition. The resulting reads (indicating open regions of chromatin) from 92.1 cells treated with BRM011 or DMSO for 24 hours were annotated relative to their closest gene as residing either downstream, overlapping, or upstream of the gene body or in the promoter (Supplementary Fig. S4F, top). Among these peaks, a large proportion were observed to change in accessibility (∼18%) when cells were treated with BRM011, with the majority decreasing in accessibility (Supplementary Fig. S4F, bottom). We chose to focus on changes in promoter-associated peaks because SWI/SNF has been previously described to play an important role in nucleosome remodeling at promoters (32, 33), and, as expected, we observed a marked decrease in chromatin accessibility near transcriptional start sites (TSS; Fig. 4E). Of note, chromatin closing was observed at the promoters of DCT and TYR (Fig. 4F; Supplementary Fig. S4F, bottom), the same loci where BRG1 occupancy was observed by ChIP-qPCR (Fig. 4D), and which showed decreased expression upon compound treatment by RNA-seq (Supplementary Fig. S4G). To further probe the relationship between SWI/SNF and MITF, we tested for enrichment of MITF-binding sites in the accessible chromatin regions identified by ATAC-Seq. We observed a significantly larger fraction of MITF-binding sites in the peaks which were depleted upon BRM011 treatment than those that either did not change or were enriched (Supplementary Fig. S4H), supporting a subset of overlapping activity for SWI/SNF and MITF in the genome.

Finally, we wished to confirm the effect of SWI/SNF inhibition on MITF and MITF target gene expression. In order to do this, we treated the panel of uveal melanoma cell lines with increasing doses of BRM011, BRM014, and BRM017 and measured expression of MITF-M, TYR, and DCT by RT-qPCR. We consistently observed across the cell lines more potent repression of the target genes by BRM011 and BRM014 as compared with BRM017 (Supplementary Fig. S5A), and this sensitivity to target gene modulation upon compound treatment was generally consistent with the BRM014 AAC50s of these cell lines.

MITF overexpression partially rescues cell death after SWI/SNF inhibition

The RNA-seq data showed modulation of MITF transcript levels upon BRM011 treatment, and a similar phenotype was observed at the protein level (Supplementary Fig. S5B). To determine the extent to which the phenotype upon inhibition of SWI/SNF was dependent on the downregulation of MITF, we overexpressed MITF in 92.1 (Supplementary Fig. S5B), and then treated these cells with BRM011, BRM014, and BRM017. The resulting viability effects were compared with empty vector alone. Upon compound treatment, MITF expression was retained in the overexpression cell line, and there was a large shift in compound sensitivity as compared with empty vector (Fig. 4G; Supplementary Fig. S5C), indicating a critical role for SWI/SNF-mediated regulation of MITF in maintaining survival of uveal melanoma cells. However, due to the incomplete rescue observed, additional growth-promoting pathways affected by SWI/SNF independently of MITF likely contribute to the phenotype.

BRM014 treatment results in the growth arrest of a uveal melanoma tumor xenograft

Finally, to characterize the phenotype of SWI/SNF inhibition in vivo, the 92.1 model was grown as a tumor xenograft in SCID mice and treated with BRM014. Once the tumors reached an average of 200 mm3, we orally dosed the animals with either vehicle or BRM014 at 20 mg/kg once daily. Similar to the in vitro sensitivity, treatment with BRM014 led to significant tumor growth inhibition (Day 41—10.32 %T/C; Fig. 5A). As observed in independent studies (13), the 20 mg/kg daily dosing regimen did not result in any significant changes in body weight (Fig. 5B). We tested the modulation of key genes identified in the RNA-seq dataset in tumors collected at the end of treatment by RT-qPCR, and saw robust inhibition across a number of these genes (Fig. 5C). In particular, genes involved in cell proliferation such as CDK2, as well as MITF target genes involved in melanocyte differentiation (MLANA, PMEL, and RAB27A), were downregulated, suggesting successful target engagement in tumors. Of note, we did not observe tumor regression as would be expected from the apoptosis seen in vitro, and in fact saw outgrowth in a subset of the tumors after 2 weeks (Fig. 5D), indicating that SWI/SNF activity may not be fully inhibited in this dosing regimen.

Figure 5.

Chemical inhibition of SWI/SNF activity leads to growth arrest of 92.1 in vivo. A, Tumor growth as measured by tumor volume for either vehicle or BRM014 treatment. B, Body weight for animals in study shown in A. C, Expression of indicated genes in endpoint samples of tumors in study shown in A. Error bars shown as SEM for all graphs. N = 6 for vehicle treatment and 10 for BRM014 treatment. P values indicated on figure were calculated using Student t test (*,P < 0.05; **, P < 0.01; and ***, P < 0.001). D, Tumor volume of BRM014-treated tumors is plotted individually. Tumors shown in purple showed stasis, and tumors shown in blue demonstrated regrowth at end of treatment.

Figure 5.

Chemical inhibition of SWI/SNF activity leads to growth arrest of 92.1 in vivo. A, Tumor growth as measured by tumor volume for either vehicle or BRM014 treatment. B, Body weight for animals in study shown in A. C, Expression of indicated genes in endpoint samples of tumors in study shown in A. Error bars shown as SEM for all graphs. N = 6 for vehicle treatment and 10 for BRM014 treatment. P values indicated on figure were calculated using Student t test (*,P < 0.05; **, P < 0.01; and ***, P < 0.001). D, Tumor volume of BRM014-treated tumors is plotted individually. Tumors shown in purple showed stasis, and tumors shown in blue demonstrated regrowth at end of treatment.

Close modal

Uveal melanoma is a rare cancer with limited treatment options, especially targeted therapeutics. Recent work has suggested that hyperactivating mutations in GNAQ antagonize PRC2-mediated repression of differentiation markers, maintaining cells in a dedifferentiated state, and that inhibition of mutant GNAQ leads to differentiation of these cells, which can be counteracted by inhibition of PRC2 (34). Similar to the epigenetic factor PRC2, SWI/SNF has also been shown to be a key player across cancers of many different genetic backgrounds, and provides an exciting potential therapeutic target (35). Here, we describe a dependency on the SWI/SNF complex in uveal melanoma using both genetic and chemical perturbation of the complex. In contrast to our findings, a recent report identified BRD9, a member of the noncanonical BAF complex (ncBAF), as a tumor suppressor in uveal melanoma (36). Such work suggests there are potential differences in penetrance of perturbing BRD9, which is specific to ncBAF versus BRG1 and BRM, which are members of all the functional SWI/SNF complexes.

Although models for uveal melanoma are limited, the significant antiproliferative activity of BRM011 and BRM014 across the various uveal melanoma models suggests that this may represent a dependency on SWI/SNF that could broadly affect approaches to disease treatment. Importantly, there are examples of cancers, such as small cell carcinoma of the ovary hypercalcemic type, in which expression of both BRG1 and BRM has been lost (37, 38), suggesting that SWI/SNF complex activity is not universally essential in cancers, and further defining the importance of the lineage-specific context when defining SWI/SNF dependence. Along those lines, we find that the dependency in this lineage is mediated, at least in part, by the melanocyte lineage–specific transcription factor MITF. Importantly, these factors colocalize to certain genetic loci and affect transcription of an overlapping gene set. Furthermore, the data shown here suggest a novel therapeutic strategy for treatment of uveal melanoma via inhibition of the SWI/SNF complex. The extensive investigation of the effects of dual inhibitors in vivo reported by Jagani and colleagues (12) suggests that a major limitation of dual BRM/BRG1 inhibition may be due to on-target toxicity. Interestingly, although 40% to 70% modulation of downstream transcriptional markers of SWI/SNF activity was sufficient to see robust growth inhibition in 92.1, similar levels of target gene/pharmacodynamic inhibition in BRG1-mutant lung cancer models did not induce a growth arrest (12, 13), suggesting that the requirements of SWI/SNF inhibition for efficacy can vary depending on the disease context. It is also important to note that although compound treatment was able to elicit apoptosis in 92.1 in vitro, this was not true in vivo. This could be due to resistance to treatment or indicate that a higher dose is required for a robust growth inhibition, and merits further investigation. Our work provides an important foundation from which future studies to investigate rational combination partners with SWI/SNF inhibition should provide important insights into design of highly efficacious treatments for uveal melanoma. It will also be interesting to determine if these discrepancies between lineages hold true using molecules with alternative mechanisms of action, such as the recently described BRG1/BRM PROTAC (39), as this will provide further mechanistic insight into the structural versus catalytic role of the SWI/SNF catalytic subunits. Together, these data provide an exciting new paradigm for treatment of SWI/SNF-dependent cancers.

A. Desplat is a current employee of Novartis, who supported the research published here. D. Abramowski is a current employee of Novartis, who supported the research published here. H. Möbitz is a shareholder of Novartis Pharma AG. C.D. Adair reports a patent for U.S. Provisional Patent Application No. 62/765,138 pending. J.P.N. Papillon reports a patent for WO2020/35779 A1 pending. D. Castelletti is a current employee of Novartis, who supported the research published here. No potential conflicts of interest were disclosed by the other authors.

F. Rago: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing. G. Elliott: Data curation, software, formal analysis, methodology. A. Li: Investigation. K. Sprouffske: Data curation, software, formal analysis, methodology. G. Kerr: Data curation, software, formal analysis, methodology. A. Desplat: Investigation. D. Abramowski: Investigation. J.T. Chen: Data curation, investigation, methodology. A. Farsidjani: Investigation. K.X. Xiang: Investigation. G. Bushold: Investigation. Y. Feng: Investigation. M.D. Shirley: Data curation, software, formal analysis, methodology. A. Bric: Supervision, investigation. A. Vattay: Investigation. H. Möbitz: Investigation. K. Nakajima: Investigation. C.D. Adair: Investigation. S. Mathieu: Investigation. R. Ntaganda: Investigation. T. Smith: Investigation. J.P.N. Papillon: Supervision. A. Kauffmann: Supervision. D.A. Ruddy: Supervision. H.-e.C. Bhang: Supervision, methodology. D. Castelletti: Supervision. Z. Jagani: Conceptualization, supervision, methodology, writing-original draft, project administration, writing-review and editing.

The authors would like to thank Lin Fan and Vera Ruda for sequencing of RNA-seq Samples, Lindsey U. Rodrigues for assistance with sample processing, Emilie Niemezyk, Franklin Chung, and Dan Rakiec for technical expertise, Matthew Crowe for sharing nontransformed melanocytes, Joshua Korn for cell line mutational data, and Eusebio Manchado, Andrew Wylie, Joshua Korn, Vesselina Cooke, Serena Silver, E. Rob McDonald, Darrin Stuart, Francesco Hofmann, and Jeffrey Engelman for discussions and feedback.

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.
Van Raamsdonk
CD
,
Bezrookove
V
,
Green
G
,
Bauer
J
,
Gaugler
L
,
O'Brien
JM
, et al
Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi
.
Nature
2008
;
457
:
599
602
.
2.
Amaro
A
,
Gangemi
R
,
Piaggio
F
,
Angelini
G
,
Barisione
G
,
Ferrini
S
, et al
The biology of uveal melanoma
.
Cancer Metastasis Rev
2017
;
36
:
109
40
.
3.
Onken
MD
,
Worley
LA
,
Ehlers
JP
,
Harbour
JW
. 
Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death
.
Cancer Res
2004
;
64
:
7205
9
.
4.
Park
JJ
,
Diefenbach
RJ
,
Joshua
AM
,
Kefford
RF
,
Carlino
MS
,
Rizos
H
. 
Oncogenic signaling in uveal melanoma
.
Pigment Cell Melanoma Res
2018
;
31
:
661
72
.
5.
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
.
6.
Shain
AH
,
Pollack
JR
. 
The spectrum of SWI/SNF mutations, ubiquitous in human cancers
.
PLoS One
2013
;
8
:
e55119
.
7.
Shi
J
,
Whyte
WA
,
Zepeda-Mendoza
CJ
,
Milazzo
JP
,
Shen
C
,
Roe
J-S
, et al
Role of SWI/SNF in acute leukemia maintenance and enhancer-mediated Myc regulation
.
Genes Dev
2013
;
27
:
2648
62
.
8.
Hohmann
AF
,
Martin
LJ
,
Minder
JL
,
Roe
J-S
,
Shi
J
,
Steurer
S
, et al
Sensitivity and engineered resistance of myeloid leukemia cells to BRD9 inhibition
.
Nat Chem Biol
2016
;
12
:
672
9
.
9.
Mathur
R
,
Alver
BH
,
Roman
AKS
,
Wilson
BG
,
Wang
X
,
Agoston
AT
, et al
ARID1A loss impairs enhancer-mediated gene regulation and drives colon cancer in mice
.
Nat Genet
2016
;
49
:
296
302
.
10.
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
2016
;
49
:
289
95
.
11.
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
.
12.
Jagani
Z
,
Chenail
G
,
Xiang
K
,
Bushold
G
,
Bhang
H-EC
,
Li
A
, et al
In-depth characterization and validation in BRG1-mutant lung cancers define novel catalytic inhibitors of SWI/SNF chromatin remodeling
.
bioRxiv
2019
;
812628
.
13.
Papillon
JPN
,
Nakajima
K
,
Adair
CD
,
Hempel
J
,
Jouk
AO
,
Karki
RG
, et al
Discovery of orally active inhibitors of brahma homolog (BRM)/SMARCA2 ATPase activity for the treatment of brahma related gene 1 (BRG1)/SMARCA4-mutant cancers
.
J Med Chem
2018
;
61
:
10155
72
.
14.
Demichelis
F
,
Greulich
H
,
Macoska
JA
,
Beroukhim
R
,
Sellers
WR
,
Garraway
L
, et al
SNP panel identification assay (SPIA): a genetic-based assay for the identification of cell lines
.
Nucleic Acids Res
2008
;
36
:
2446
56
.
15.
Hoffman
GR
,
Rahal
R
,
Buxton
F
,
Xiang
K
,
McAllister
G
,
Frias
E
, et al
Functional epigenetics approach identifies BRM/SMARCA2 as a critical synthetic lethal target in BRG1-deficient cancers
.
Proc Natl Acad Sci U S A
2014
;
111
:
3128
33
.
16.
Banaszynski
LA
,
Chen
L-C
,
Maynard-Smith
LA
,
Ooi
AGL
,
Wandless
TJ
. 
A rapid, reversible, and tunable method to regulate protein function in living cells using synthetic small molecules
.
Cell
2006
;
126
:
995
1004
.
17.
Shoemaker
RH
. 
The NCI60 human tumour cell line anticancer drug screen
.
Nat Rev Cancer
2006
;
6
:
813
23
.
18.
Corces
MR
,
Trevino
AE
,
Hamilton
EG
,
Greenside
PG
,
Sinnott-Armstrong
NA
,
Vesuna
S
, et al
An improved ATAC-seq protocol reduces background and enables interrogation of frozen tissues
.
Nat Methods
2017
;
14
:
959
62
.
19.
McDonald
ER
,
de Weck
A
,
Schlabach
MR
,
Billy
E
,
Mavrakis
KJ
,
Hoffman
GR
, et al
Project DRIVE: a compendium of cancer dependencies and synthetic lethal relationships uncovered by large-scale, deep RNAi screening
.
Cell
2017
;
170
:
577
92
.
20.
Tsherniak
A
,
Vazquez
F
,
Montgomery
PG
,
Weir
BA
,
Kryukov
G
,
Cowley
GS
, et al
Defining a cancer dependency map
.
Cell
2017
;
170
:
564
76
.
21.
Szerlong
H
,
Hinata
K
,
Viswanathan
R
,
Erdjument-Bromage
H
,
Tempst
P
,
Cairns
BR
. 
The HSA domain binds nuclear actin-related proteins to regulate chromatin-remodeling ATPases
.
Nat Struct Mol Biol
2008
;
15
:
469
76
.
22.
Zhao
K
,
Wang
W
,
Rando
OJ
,
Xue
Y
,
Swiderek
K
,
Kuo
A
, et al
Rapid and phosphoinositol-dependent binding of the SWI/SNF-like BAF complex to chromatin after T lymphocyte receptor signaling
.
Cell
1998
;
95
:
625
36
.
23.
Pan
J
,
McKenzie
ZM
,
Mashtalir
N
,
Lareau
CA
,
Pierre
RS
,
Wang
L
, et al
The ATPase module of mammalian SWI/SNF family complexes mediates subcomplex identity and catalytic activity-independent genomic targeting
.
Nat Genet
2019
;
51
:
618
26
.
24.
Khavari
PA
,
Peterson
CL
,
Tamkun
JW
,
Mendel
DB
,
Crabtree
GR
. 
BRG1 contains a conserved domain of the SWI2/SNF2 family necessary for normal mitotic growth and transcription
.
Nature
1993
;
366
:
170
4
.
25.
Dunaief
JL
,
Strober
BE
,
Guha
S
,
Khavari
PA
,
Alin
K
,
Luban
J
, et al
The retinoblastoma protein and BRG1 form a complex and cooperate to induce cell cycle arrest
.
Cell
1994
;
79
:
119
30
.
26.
Mizutani
T
,
Ito
T
,
Nishina
M
,
Yamamichi
N
,
Watanabe
A
,
Iba
H
. 
Maintenance of integrated proviral gene expression requires Brm, a catalytic subunit of SWI/SNF complex
.
J Biol Chem
2002
;
277
:
15859
64
.
27.
Marathe
HG
,
Watkins-Chow
DE
,
Weider
M
,
Hoffmann
A
,
Mehta
G
,
Trivedi
A
, et al
BRG1 interacts with SOX10 to establish the melanocyte lineage and to promote differentiation
.
Nucleic Acids Res
2017
;
45
:
6442
58
.
28.
Vachtenheim
J
,
Ondrusová
L
,
Borovanský
J
. 
SWI/SNF chromatin remodeling complex is critical for the expression of microphthalmia-associated transcription factor in melanoma cells
.
Biochem Biophys Res Commun
2010
;
392
:
454
9
.
29.
Mehta
G
,
Kumarasamy
S
,
Wu
J
,
Walsh
A
,
Liu
L
,
Williams
K
, et al
MITF interacts with the SWI/SNF subunit, BRG1, to promote GATA4 expression in cardiac hypertrophy
.
J Mol Cell Cardiol
2015
;
88
:
101
10
.
30.
la Serna de
IL
,
Ohkawa
Y
,
Higashi
C
,
Dutta
C
,
Osias
J
,
Kommajosyula
N
, et al
The microphthalmia-associated transcription factor requires SWI/SNF enzymes to activate melanocyte-specific genes
.
J Biol Chem
2006
;
281
:
20233
41
.
31.
Laurette
P
,
Strub
T
,
Koludrovic
D
,
Keime
C
,
Le Gras
S
,
Seberg
H
, et al
Transcription factor MITF and remodeller BRG1 define chromatin organisation at regulatory elements in melanoma cells
.
Elife
2015
;
4
:
e06857
.
32.
Tolstorukov
MY
,
Sansam
CG
,
Lu
P
,
Koellhoffer
EC
,
Helming
KC
,
Alver
BH
, et al
Swi/Snf chromatin remodeling/tumor suppressor complex establishes nucleosome occupancy at target promoters
.
Proc Natl Acad Sci U S A
2013
;
110
:
10165
70
.
33.
Ho
L
,
Ronan
JL
,
Wu
J
,
Staahl
BT
,
Chen
L
,
Kuo
A
, et al
An embryonic stem cell chromatin remodeling complex, esBAF, is essential for embryonic stem cell self-renewal and pluripotency
.
Proc Natl Acad Sci U S A
2009
;
106
:
5181
6
.
34.
Onken
MD
,
Makepeace
CM
,
Kaltenbronn
KM
,
Kanai
SM
,
Todd
TD
,
Wang
S
, et al
Targeting nucleotide exchange to inhibit constitutively active G protein α subunits in cancer cells
.
Sci Signal
2018
;
11
:
eaao6852
.
35.
Vangamudi
B
,
Paul
TA
,
Shah
PK
,
Kost-Alimova
M
,
Nottebaum
L
,
Shi
X
, et al
The SMARCA2/4 ATPase domain surpasses the bromodomain as a drug target in SWI/SNF-mutant cancers: insights from cDNA rescue and PFI-3 inhibitor studies
.
Cancer Res
2015
;
75
:
3865
78
.
36.
Inoue
D
,
Chew
G-L
,
Liu
B
,
Michel
BC
,
Pangallo
J
,
D'Avino
AR
, et al
Spliceosomal disruption of the non-canonical BAF complex in cancer
.
Nature
2019
;
574
:
432
6
.
37.
Herpel
E
,
Rieker
RJ
,
Dienemann
H
,
Muley
T
,
Meister
M
,
Hartmann
A
, et al
SMARCA4 and SMARCA2 deficiency in non-small cell lung cancer: immunohistochemical survey of 316 consecutive specimens
.
Ann Diagn Pathol
2016
;
26
:
47
51
.
38.
Karnezis
AN
,
Wang
Y
,
Ramos
P
,
Hendricks
WP
,
Oliva
E
,
D'Angelo
E
, et al
Dual loss of the SWI/SNF complex ATPases SMARCA4/BRG1 and SMARCA2/BRM is highly sensitive and specific for small cell carcinoma of the ovary, hypercalcaemic type
.
J Pathol
2015
;
238
:
389
400
.
39.
Farnaby
W
,
Koegl
M
,
Roy
MJ
,
Whitworth
C
,
Diers
E
,
Trainor
N
, et al
BAF complex vulnerabilities in cancer demonstrated via structure-based PROTAC design
.
Nat Chem Biol
2019
;
15
:
672
80
.