SNF5/INI1 deficiency redefines chromatin remodeling complex composition during tumor development

Abstract

Malignant rhabdoid tumors (MRT), a pediatric cancer that most frequently appears in the kidney and brain, generally lack SNF5 (SMARCB1/INI1), a subunit of the SWI/SNF chromatin-remodeling complex. Recent studies have established that multiple SWI/SNF complexes exist due to the presence or absence of different complex members. Therefore, the effect of SNF5 loss upon SWI/SNF complex formation was investigated in human MRT cells. MRT cells and primary human tumors exhibited reduced levels of many complex proteins. Furthermore, reexpression of SNF5 increased SWI/SNF complex protein levels without concomitant increases in mRNA. Proteomic analysis, using mass spectrometry, of MRT cells before and after SNF5 reexpression indicated the recruitment of different components into the complex along with the expulsion of others. IP-Western blotting confirmed these results and demonstrated similar changes in other MRT cell lines. Finally, reduced expression of SNF5 in normal human fibroblasts led to altered levels of these same complex members. These data establish that SNF5 loss during MRT development alters the repertoire of available SWI/SNF complexes, generally disrupting those associated with cellular differentiation. These findings support a model where SNF5 inactivation blocks the conversion of growth-promoting SWI/SNF complexes to differentiation-inducing ones. Therefore, restoration of these complexes in tumors cells provides an attractive approach for the treatment of MRTs.

Implications: SNF5 loss dramatically alters SWI/SNF complex composition and prevents formation of complexes required for cellular differentiation.

Figure 1. Decreased SWI/SNF complex protein but not mRNA in MRT cell lines

(A) Total cellular proteins (30 μg) isolated from MRT and non-MRT cell lines were separated on a 4% to 12% SDS-polyacrylamide gel, transferred to PVDF membranes and probed with antiserum against SWI/SNF complex members. We used Ku80 as the loading control. (B) RNA was extracted from the indicated cell lines. The mRNA levels were measured for each gene by qPCR and normalized to D98OR β-actin expression. Blue- MRT cell lines; Red- non-MRT cell lines. Columns – mean of four independent experiments; Error bars – standard error.

Figure 2. Decreased SWI/SNF complex proteins in primary tumor samples

(A) Nuclear protein (3 μg) isolated from primary tumors were separated on a 4 to 12% Bis-Tris polyacrylamide gel, transferred to PVDF membranes and probed with anti-BRG1, anti-BAF57, anti-SNF5, anti-BAF180, or anti-H3. WT = Wilms’ tumor; RMS = Rhabdomyosarcoma; MRT = Malignant Rhabdoid Tumor. (B) Densitometry was carried out using Bio-Rad Imagelab 4.1. All values expressed as a ratio to H3. The values were then aggregated by tumor types and averaged. Columns, mean of ratio of each tumor type; Error bars – standard error *, P < 0.05 relative to MRT; #, P > 0.05 relative to MRT using Student’s T-test.

Figure 3. hSNF5 Re-expression in MRT Cell Lines Increases SWI/SNF Complex Component Levels

(A) Cells were harvested at time 0 and 24 hours after infection with Ad-CMV, Ad-GFP or Ad-SNF5-GFP. Total cellular proteins (30 μg) were separated on a 4% to 12% Bis-Tris polyacrylamide gel, transferred to PVDF membranes and probed with indicated antibody. Ku served as the loading control. UN, uninfected control. (B) Cells were harvested at time 0 and 24 hours after infection with Ad-CMV, Ad-GFP or Ad-SNF5-GFP and RNA extracted from the indicated cell lines. The mRNA levels were measured for each gene by qPCR and normalized to β-ACTIN expression at time 0. Because of the undetectable levels of SNF5 in the G401 and TTC549 cell lines, a log scale was used to graph SNF5 expression. Blue – G401; Red – TTC549; Green – TTC642; Columns, mean of four independent experiments; Error bars- standard error.

Figure 4. Characterization of SNF5-inducible MRT cell lines

Cells were induced with doxycycline (DOX) at Time 0 and samples were harvested at 24 hours after induction. Total cellular proteins (30 μg) were separated on a 4% to 12% Bis-Tris polyacrylamide gel, transferred to PVDF membranes and probed with indicated antibody. KU served as the loading control. UN, untreated control

Figure 5. SWI/SNF complex composition changes after SNF5 re-expression

(A) The relative abundance of SWI/SNF complex components in the absence (untreated) and presence (Dox) of SNF5 in the TTC642 cell line was determined by label-free quantitative mass spectrometry. The identity of BCL7B could not unambiguously be determined versus BCL7C due to shared peptides. (B) The indicated cell lines (A–C) were induced with doxycycline, and the samples were immunoprecipitated with either rabbit IgG or anti-BRG1. IP samples were then separated on a 4% to 12% Bis-Tris polyacrylamide gel, transferred to PVDF membranes and probed with the indicated antibodies.

Figure 6. Reduction of SWI/SNF complex protein levels after inhibition of SNF5 in normal human fibroblasts and the effects of MG132

(A) SNF5 knockdown cells were generated by infecting normal human fibroblasts (NHF-1) with lentivirus expressing 2 different shRNAs against SNF5 (TRCN 39585 and TRCN 39587) or a non-targeting shRNA control (pLKO.1) and immediate selection with puromycin. After 72 hrs. on selective medium, total cellular protein (30 μg), was separated on a 4% to 12% Bis-Tris polyacrylamide gel, transferred to PVDF membranes and probed with the indicated antibodies. Ku served as the loading control. (B) NHF-1 hTERT cells were infected with indicated lentivirus and placed in selective media for 3 days. Six hours before harvesting for protein isolation, half of the samples were treated with the proteasome inhibitor MG132. After protein isolation, levels for the indicated proteins were detected by SDS-PAGE and Western blotting. c-FOS served as a positive control for proteasome inhibition while Ran, unaffected by MG132 treatment, served as a negative control. Ku served as the loading control.

Figure 7. Proteasome inhibition causes increase protein levels for a subset of SWI/SNF complex components in MRT cell lines

The TTC642 pINDUCER20-fSNF5-HA, G401 pINDUCER20-fSNF5-HA and TTC549 pINDUCER20-fSNF5-HA cell lines, were treated with DMSO, 1μg/ml doxycycline or 10uM MG132 for 6 hrs. After protein isolation, levels for the indicated proteins were detected by SDS-PAGE and Western blotting as previously described (Chai et al., 2005a). c-FOS served as a positive control for proteasome inhibition while Ran, unaffected by MG132 treatment, served as a negative control. Ku served as the loading control.