Role of specialized composition of SWI/SNF complexes in prostate cancer lineage plasticity

Advanced prostate cancer initially responds to hormonal treatment, but ultimately becomes resistant and requires more potent therapies. One mechanism of resistance observed in around 10–20% of these patients is lineage plasticity, which manifests in a partial or complete small cell or neuroendocrine prostate cancer (NEPC) phenotype. Here, we investigate the role of the mammalian SWI/SNF (mSWI/SNF) chromatin remodeling complex in NEPC. Using large patient datasets, patient-derived organoids and cancer cell lines, we identify mSWI/SNF subunits that are deregulated in NEPC and demonstrate that SMARCA4 (BRG1) overexpression is associated with aggressive disease. We also show that SWI/SNF complexes interact with different lineage-specific factors in NEPC compared to prostate adenocarcinoma. These data point to a role for mSWI/SNF complexes in therapy-related lineage plasticity, which may also be relevant for other solid tumors.


Gene expression levels of the REST gene upon SMARCA4 knock-down, assessed by RNA-seq in LNCaP and 22Rv1 cells.
There is no statistically significant difference in REST expression between the SMARCA4 knock-down condition and the control (Scrambled siRNA). Wilcoxon test. Representation of a two-tailed test.  2) The effects of SMARCA4 knock-down on cell growth in LNCaP cells are not entirely abrogated by p53 or Rb loss. (a) Cell growth curves in control cells (Scrambled gRNA), RB1negative cells, p53-negative cells or double Rb1/p53 negative cells upon SMARCA4 knockdown, SMARCA2 knock-down or treatment with Scrambled siRNA; n=3 biologically independent experiments. Data are presented as mean values +/-SEM and analyzed using two-way ANOVA test. Adjustments were made for multiple comparisons. *p<0.05, **p<0.001, ****p<0.0001 (for Scrambled gRNA: ***p=0.0003, for RB1: ****p<0.0001, for TP53: ***p=0.0007, for RB1/TP53: *p=0.021). (b) Immunoblot validation of CRISPR-Cas9 mediated loss of p53 and/or Rb and of siRNA-mediated knock-down of SMARCA4 (BRG1) or SMARCA2 (BRM). Source data are provided as a Source Data file.

Immunoblot showing BRG1 (SMARCA4) and BRM (SMARCA2) expression levels in 22Rv1 cells transduced with lentiviral vectors and treated for 24h with MG-132 or DMSO (control).
The expression levels of the p21 protein are shown as control for the MG-132 treatment. N=1 (one pilot experiment assessing feasibility). Samples shown in the Brg1 and Brm rows derive from the same experiment, but the images come from two separate blots, which were ran and processed in parallel. Equilibrated loading in both blots was confirmed using the loading control marker (vinculin) on each blot. Source data are provided as a Source Data file.

The effects of BAF53B or BAF45B shRNA-mediated knock-down on cell growth of a CRPC-NE patient tumor organoid-derived 2D cell line (WCM155).
The immunoblots show knock-down efficiency control (one representative experiment). Each growth curve shows pooled results from three independent experiments (bars: standard error). Source data are provided as a Source Data file. Figure 21 (related to Fig 3) Box plots comparing SMARCA4 mRNA levels and SMARCA4 knock-down signature score values across three PCa patient cohorts. Each dot represents a sample. SMARCA4 mRNA expression levels are consistent with the predicted signature score (samples with lower SMARCA4 expression show higher SMARCA4 knock-down signature scores, and vice versa). Mann-Whitney Wilcoxon test. TCGA N= 495 samples, of which 124 were classified SMARCA4 high. SU2C N= 332 samples, of which 83 were classified SMARCA4 low and 83 cases as SMARCA4 high. WCMN=49 samples, of which 12 were classified as SMARCA4 low and 12 as SMARCA4 high. The box plots represent the median values with upper and lower quartiles; the whiskers represent the range outside the quartiles and the outliers are plotted as the individual points.

Supplementary
SMARCA4 knock-down signature vs Metastasis in the JHMI cohort. Patients were stratified based on quartiles (a) or on deciles (b) of SMARCA4 knock-down signature scores, to test for associations between SMARCA4 knock-down signature scores and metastasis-free survival. There is a trend of lower SMARCA4 knock-down signature scores being associated with worse metastasis-free survival, but it did not reach statistical significance. Kaplan-Meier analysis and Cox proportional hazard model used; p values shown in (a) and (b) pertain to statistical analysis comparing groups with lowest (green) and highest (red) SMARCA4 knock-down signature scores.

Genes encoding factors that show differential binding to SWI/SNF between CRPC-NE and prostate adenocarcinoma cell lines are also differentially expressed at the transcript level across prostate cancer cell lines. Graphs show gene expression levels assessed by RT-PCR.
Relative mRNA levels of each gene shown were normalized to the expression of the average of housekeeping genes GAPDH and ACTB. Each graph represents three replicates and bars show standard deviation (SD). Figure 25 (related to Fig 4)

Confirmation of factors associating with BAF155 (SMARCC1) in NCI-H660 and in LNCaP-AR cells by co-immunoprecipitation (co-IP) and immunoblotting.
(a) Co-IP of BAF155 (SMARCC1) followed by immunoblotting for BRG1 (SMARCA4), CHD4, VGF, NKX2.1 (TTF1), MTA1 and H3 in nuclear extracts from the CRPC-Adeno cell line LNCaP-AR and from the CRPC-NE cell line NCI-H660. In NCI-H660, MTA1, NKX2.1, VGF and CHD4 were pulled down in the SMARCC1 Co-IP condition, but were either absent or only weakly expressed in the IgG Co-IP (in accordance with mass spectrometry findings). In LNCaP-AR, most of these factors were detectable in the nuclear fraction, but were not detected in the SMARCC1 Co-IP. SMARCA4 (positive control) was pulled down in the Co-IP experiment for both cell lines. C1:IP indicates SMARCC1 Co-IP; IgG indicates control Co-IP with isotype antibody; SN indicates supernatant and serves as control to indicate how much protein was pulled down; H3 serves as nuclear fraction control. (b) Nuclear extract Co-IP of SMARCC1 and of SMARCA4 followed by immunoblotting for SMARCA4, SMARCC1, HOXB13, NKX3.1, REST and H3 in the CRPC-Adeno cell line LNCaP-AR. The immunoblot confirms that HOXB13 co-immunoprecipitated with SMARCC1 and SMARCA4 (in accordance with mass spectrometry findings). Coimmunoprecipitation of NKX3.1 and REST with SMARCA4 or SMARCC1 could not be confirmed in this immunoblot. Source data are provided as a Source Data file.

Supplementary tables
Supplementary  2). Analysis of association between SMARCA4 (BRG1) and SMARCA2 (BRM) protein expression (strong vs. moderate/weak/negative) and patient's overall survival adjusted for single covariates (factors with known impact on PCa prognosis). Cox proportional hazards models were executed (univariable and multivariable), all p-values were two-sided with statistical significance evaluated at the 0.05 alpha level. Adjustments were not made for multiple comparisons due to the exploratory nature of the analyses.