The SWI/SNF ATPases Are Required for Triple Negative Breast Cancer Cell Proliferation

Abstract

The Brahma (BRM) and Brahma-related Gene 1 (BRG1) ATPases are highly conserved homologs that catalyze the chromatin remodeling functions of the multi-subunit human SWI/SNF chromatin remodeling enzymes in a mutually exclusive manner. SWI/SNF enzyme subunits are mutated or missing in many cancer types, but are overexpressed without apparent mutation in other cancers. Here, we report that both BRG1 and BRM are overexpressed in most primary breast cancers independent of the tumor's receptor status. Knockdown of either ATPase in a triple negative breast cancer cell line reduced tumor formation in vivo and cell proliferation in vitro. Fewer cells in S phase and an extended cell cycle progression time were observed without any indication of apoptosis, senescence, or alterations in migration or attachment properties. Combined knockdown of BRM and BRG1 showed additive effects in the reduction of cell proliferation and time required for completion of cell cycle, suggesting that these enzymes promote cell cycle progression through independent mechanisms. Knockout of BRG1 or BRM using CRISPR/Cas9 technology resulted in the loss of viability, consistent with a requirement for both enzymes in triple negative breast cancer cells.

Figure 1. BRG1 and BRM are overexpressed in primary breast tumors

(A) Western blots showing BRG1 and BRM protein levels in primary breast tumors compared to adjacent normal breast tissues from grade II, ER+, PR+, HER2- patients. N; normal adjacent tissue. T; tumor tissue. Coomassie staining of a duplicate gel was used to demonstrate equal loading between normal and tumor samples from each patient. (B) Western blots showing BRG1 and BRM protein levels in primary breast tumors compared to adjacent normal breast tissues from patients with grade III, mixed ER/PR/HER2 status tumors. N; normal adjacent tissue. T; tumor tissue. Lamin B levels were used as a loading control. (C) Immunohistochemistry of BRG1 and BRM expression in adjacent normal breast tissue (left) and primary tumor (right) sections. This patient was Grade II, ER+, PR+, HER2−. (D) Images of grade I, II, and III tumors stained for BRG1 or BRM extracted from the Biomax 1503a tissue arrays.

Figure 2. BRG1 or BRM reduction in MDA-MB-231 metastatic breast cancer cells reduced colony formation in culture and tumor growth in vivo

(A) Western blot indicating BRG1 and BRM levels in MDA-MB-231 cells inducibly expressing a control (Scram) shRNA or shRNA targeting BRG1 or BRM. GAPDH levels were monitored as a loading control. DOX; doxycycline. (B) Colony forming assay demonstrates that BRG1 or BRM knockdown reduces the ability of MDA-MB-231 cells to form colonies when plated at low density. (C) Quantification of colony forming assay. Quantification is the average from 3 independent experiments; error bars, SD. ***P<0.001. (D) BRG1 or BRM knockdown cells formed fewer and smaller tumors after implantation in the mammary fat pad of SCID/NCr mice. Each bar represents averaged results, n=12; error bars, SD. ***P<0.001. (E) Images of representative tumors.

Figure 3. Knockdown of BRG1 and/or BRM reduces triple negative breast cancer cell proliferation

(A) Cell number was monitored over time to examine cell proliferation of each of the cell lines in the induced and uninduced state. DOX; doxycycline. (B) Corresponding western blot measuring the level of BRG1 and/or BRM knockdown. GAPDH is shown as a loading control. Data represent the average of three independent experiments; error bars, SD. (C) MDA-MB-231 cells were treated with one of three distinct siRNAs targeting BRG1 (Imbalzano et al., 2013) or with a pool of siRNAs targeting BRM or with both siRNA pools. Cell proliferation was monitored, and western blots confirmed knockdown of the target protein(s). (D) MDA-MB-468 cells were treated and experimentally evaluated exactly as in (C). (E) Increasing levels of BRG1 or BRM cDNAs were introduced to double knockdown MDA-MB-231 cells and proliferation rate was monitored. Results are presented as fold-change relative to the proliferation rate of the scram siRNA treated cells. (F) A representative western blot measuring BRG1 and BRM levels from the experiments described in (E).

Figure 4. BRG1 and/or BRM knockdown results in a longer cell cycle

(A) Knockdown of BRG1 resulted in fewer cells in S-phase as measured by BrdU incorporation. Representative images are presented. (B) Quantified data are the average of three independent experiments; three fields of at least 100 nuclei were counted per experiment. error bars, SD. **P<0.01. (C) The mitotic marker histone H3 phospho-serine 10 was reduced in BRG1 and/or BRM knockdown cells as measured by Western blotting. (D) Proliferation curves of control and BRG1, BRM, and double knockdown cells as determined by time-lapse videography.

Figure 5. BRG1 and/or BRM knockdown in MDA-MB-231 cells did not alter apoptosis, senescence, migration, or attachment to different collagen substrates

(A) BRG1, BRM, or double knockdown did not cause apoptosis as judged by Annexin V staining. (B) BRG1, BRM, or double knockdown did not induce senescence as measured by senescence associated-β-galactosidase staining. (C) BRG1, BRM, or double knockdown did not change the ability of cells to fill the gap scratched into a monolayer of cells in a standard wound healing assay. (D) BRG1, BRM, or double knockdown did not affect attachment to collagen I- or collagen IV- coated plates.

Figure 6. Loss of BRG1 or BRM impaired cell viability

(A) Western blots demonstrate the efficacy of CRISPR/Cas9 mediated knockout of BRG1 or BRM in MDA-MB-231 cells. Cell lines were derived using each of three different guide RNAs for BRG1 (187, 189, 191) or three different guide RNAs for BRM (181, 183, 185). GAPDH was monitored as a loading control. (B) Dynamic cell proliferation measured by xCELLigence RTCA system. 10,000 cells of each line were plated on E-Plate 8-well plates. Cell proliferation was monitored in 30-minute intervals over 220 hours. Cell Index values for all cell lines were calculated and plotted with the RTCA Software.