Residual complexes containing SMARCA2 (BRM) underlie the oncogenic drive of SMARCA4 (BRG1) mutation

Abstract

Collectively, genes encoding subunits of the SWI/SNF (BAF) chromatin remodeling complex are mutated in 20% of all human cancers, with the SMARCA4 (BRG1) subunit being one of the most frequently mutated. The SWI/SNF complex modulates chromatin remodeling through the activity of two mutually exclusive catalytic subunits, SMARCA4 and SMARCA2 (BRM). Here, we show that a SMARCA2-containing residual SWI/SNF complex underlies the oncogenic activity of SMARCA4 mutant cancers. We demonstrate that a residual SWI/SNF complex exists in SMARCA4 mutant cell lines and plays essential roles in cellular proliferation. Further, using data from loss-of-function screening of 165 cancer cell lines, we identify SMARCA2 as an essential gene in SMARCA4 mutant cancer cell lines. Mechanistically, we reveal that Smarca4 inactivation leads to greater incorporation of the nonessential SMARCA2 subunit into the SWI/SNF complex. Collectively, these results reveal a role for SMARCA2 in oncogenesis caused by SMARCA4 loss and identify the ATPase and bromodomain-containing SMARCA2 as a potential therapeutic target in these cancers.

FIG 1

Residual SWI/SNF complexes exist in SMARCA4 mutant cells and are essential for proliferation. (A) SMARCC1 associates with other SWI/SNF subunits in SMARCA4 mutant cancer cell lines. Immunoblots of SWI/SNF subunits before (input, 10%) and after precipitation (IP) with SMARCC1, SMARCA4, and IgG antibodies. Nuclear extracts used in these experiments are derived from the SMARCA4 mutant cell lines NCI-H1299 and A549. (B) SMARCB1 knockdown impairs proliferation of SMARCA4 mutant cancer cells. shRNA-mediated knockdown of SMARCB1 in SMARCA4 mutant NCI-H1299 and A549 cell lines is shown. Immunoblotting was used to determine the efficiency of knockdown. Cell proliferation was determined by cell count. Data are represented as means ± SEM from three biological replicates. (C) Reexpression of SMARCB1 rescues the growth defect. Shown is a colony formation assay following reexpression of a mouse derivative of SMARCB1 (mSMARCB1) that is not targeted by human SMARCB1 shRNAs. Images are from duplicate biological replicates. Immunoblotting was used to determine the efficiency of SMARCB1 knockdown.

FIG 2

SMARCA2 is differentially essential in SMARCA4 mutant cancer cells. (A) Achilles shRNA screen identifies SMARCA2 as the number one vulnerability in SMARCA4-inactivated cancer cell lines. Gene scores from comparison of inactivating SMARCA4 mutant cancer cell lines to SMARCA4 wild-type cell lines are shown. (B) Cell lines with inactivating mutations in SMARCA4 are most sensitive to reduced levels of SMARCA2. Waterfall plot of Achilles SMARCA2 shRNA scores. Cell lines containing all SMARCA4 mutations are displayed in the left panel, and cell lines with inactivating SMARCA4 mutations (nonsense, frameshift, and large focal deletions) are displayed in the right panel. Red columns indicate SMARCA4 mutant cancer cell lines. (C) Reduced levels of SMARCA2 do not affect the growth of SMARCA4 wild-type cell lines. SMARCA2 knockdown in SMARCA4 wild-type cell lines HCC827, NCI-H2122, and H460. Immunoblotting is used to determine the efficiency of knockdown. Cell proliferation is evaluated using colony formation assays. Representative images are shown from one of two biological replicates for each cell line. (D) Reduced levels of SMARCA2 impair the growth of SMARCA4 mutant cell lines. SMARCA2 knockdown in SMARCA4 mutants NCI-H1299 and A549 leads to decreased proliferation. Immunoblotting is used to determine efficiency of knockdown. Cell proliferation is evaluated using colony formation assays. Representative images are shown from one of two biological replicates for each cell line.

FIG 3

SWI/SNF complex assembles in the absence of SMARCA2 and SMARCA4. (A) SMARCC1 associates with other SWI/SNF subunits following SMARCA2 knockdown in SMARCA4 mutant cells. Immunoblots of SWI/SNF subunits before (input, 10%) and after precipitation (IP) with SMARCC1 and IgG antibodies are shown. Nuclear extracts used in these experiments are derived from the SMARCA4 mutant cell line A549 treated with SMARCA2 shRNAs or nonsilencing control shRNAs. A representative image from duplicate biological replicates is shown. (B) SMARCC1 does not associate with the Polycomb-group protein BMI1. Immunoblots of inputs (input, 5%) and SMARCC1 and IgG precipitates (IP) following SMARCA2 knockdown in the SMARCA4 mutant cell line A549 are shown.

FIG 4

Reciprocal compensatory functions for SMARCA2 and SMARCA4. (A) Smarca4 inactivation in primary cells leads to specific upregulation of Smarca2 and increased association of SMARCA2 with other SWI/SNF subunits. An immunoblot of SWI/SNF subunits before (input, 10%) and after precipitation (IP) with SMARCC1 is shown. Nuclear extracts used in these experiments are derived from wild-type and Smarca4-deficient mouse embryonic fibroblasts. A representative image from duplicate biological replicates is shown. (B) Smarca2 transcript levels are elevated following Smarca4 inactivation. A plot of Smarca2 transcript expression levels in wild-type and Smarca4-deficient mouse embryonic fibroblasts is shown. **, P = 0.029. (C) SMARCA2 knockdown leads to SMARCA4 upregulation and increased association of SMARCA4 with other SWI/SNF subunits. Immunoblots of SWI/SNF subunits before (input, 10%) and after precipitation (IP) with SMARCC1 and IgG antibodies. Nuclear extracts used in these experiments are derived from the SMARCA4 wild-type cell lines NCI-H2122 and H460 treated with SMARCA2 shRNAs or nonsilencing control shRNAs. Representative images from duplicate biological replicates are shown.