Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Dec 15.
Published in final edited form as: Cell Rep. 2014 Oct 2;9(1):248–260. doi: 10.1016/j.celrep.2014.08.069

Transformation resistance in a premature aging disorder identifies a tumor-protective function of BRD4

Patricia Fernandez 1,7, Paola Scaffidi 1,2,3,7, Elke Markert 4,5, Ji-Hyeon Lee 6, Sushil Rane 6, Tom Misteli 1,*
PMCID: PMC4194066  NIHMSID: NIHMS627470  PMID: 25284786

Abstract

Advanced age and DNA damage accumulation are strong risk factors for cancer. The premature-aging disorder Hutchinson Gilford Progeria Syndrome (HGPS) provides a unique opportunity to study the interplay between DNA damage and aging- associated tumor mechanisms, since HGPS patients do not develop tumors despite elevated levels of DNA damage. Here, we have used HGPS patient cells to identify a protective mechanism to oncogenesis. We find that HGPS cells are resistant to neo- plastic transformation. This resistance is mediated by the bromodomain protein BRD4, which exhibits altered genome-wide binding patterns in transformation- resistant cells leading to inhibition of oncogenic de-differentiation. BRD4 also inhibits, albeit to a lower extent, the tumorigenic potential of transformed cells from healthy individuals and BRD4-mediated tumor protection is clinically relevant, since a BRD4 gene signature predicts positive clinical outcome in breast and lung cancer. Our results demonstrate a protective function for BRD4 and suggest tissue-specific functions for BRD4 in tumorigenesis.

Introduction

Neoplastic transformation is a multistep process whereby normal cells acquire a distinct set of cellular properties and develop into malignant derivatives (Hanahan and Weinberg, 2000). Transformation often involves an oncogenic reprogramming process during which cells de-differentiate and take on functional properties of tumor-initiating cells (Scaffidi and Misteli, 2011; Schwitalla et al., 2013; Vicente-Duenas et al., 2013). Common triggers for oncogenic transformation include intrinsic and extrinsic insults such as replication stress, irradiation, and exposure to environmental chemicals, all of which lead to genetic and epigenetic alterations (Hanahan and Weinberg, 2000). A prominent cause of oncogenic transformation is the accumulation of DNA damage as indicated by the markedly increased susceptibility to cancer in diseases caused by mutations in the DNA repair machinery (Hoeijmakers, 2009; Venkitaraman, 2002). Interestingly, several of these diseases, such as Werner Syndrome and Bloom Syndrome, manifest themselves as premature aging disorders (Brosh and Bohr, 2007), highlighting the complex relationship between DNA damage, cancer and aging.

A striking example of a premature aging disease characterized by dramatically elevated levels of DNA damage is Hutchinson-Gilford Progeria Syndrome (HGPS). HGPS is caused by a mutation in the LMNA gene, which encodes lamin A and lamin C, two major architectural components of the cell nucleus (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003; Scaffidi et al., 2005). The HGPS mutation leads to the accumulation of an alternatively spliced variant of lamin A, named progerin, which acts in a dominant gain-of-function fashion and induces nuclear defects, including chromatin changes and high levels of constitutive DNA damage (Scaffidi et al., 2005; Scaffidi and Misteli, 2006). HGPS patients display numerous symptoms of accelerated and normal aging, including cardiovascular defects, bone abnormalities and lipodystrophy and the disease is invariably fatal (Gordon et al., 2014). Low levels of progerin are also expressed in healthy individuals and induce age-related nuclear abnormalities similar to those observed in HGPS patients, suggesting relevance of HGPS to the normal aging process (Cao et al., 2007; McClintock et al., 2007; Scaffidi and Misteli, 2006). Remarkably, despite exceedingly high levels of DNA damage, HGPS patients do not develop cancers (Gordon et al., 2014).

In this study, we sought to identify the molecular basis for the observed resistance to cancer in HGPS. We find that HGPS patients are protected from cancer by a cell intrinsic mechanism, which inhibits neoplastic transformation and de-differentiation of HGPS cells towards a malignant stem-cell like state. Using a functional genomics approach, we identify the general transcriptional regulator BRD4 as a key mediator of resistance to transformation. We show that BRD4 is redistributed on chromatin in transformation resistant cells and activates tumor-protective cellular pathways. Importantly, BRD4 also protects cells from healthy individuals and is clinically relevant in breast and lung cancer.

Results

HGPS fibroblasts are resistant to experimental transformation

We hypothesized that the observed absence of tumors in HGPS patients, despite high loads of cellular DNA damage (Fig. S1A), is caused by the presence of intrinsic tumor resistance mechanisms in HGPS cells rather than the limited lifespan of HGPS patients. To directly test this hypothesis, we experimentally assessed the transformation potential of HGPS patient cells. Primary skin fibroblasts from multiple HGPS patients and age- matched control wild-type individuals were challenged in a standard transformation assay by retroviral introduction of TERT (T), V12-HRAS (R) and SV40 large and small T antigens (S) (Hahn et al., 1999). Both wild-type and HGPS cells expressing the transforming factors (referred to as TRS-WT and TRS–HGPS, respectively) underwent morphological changes typically observed upon transformation (Fig. S1B), proliferated at comparable rates, and, as expected, faster than control cells expressing telomerase only (Fig. S1C). When tested in soft agar assays, TRS-HGPS cells showed reduced clonogenic capacity compared to TRS-WT cells. While TRS-WT cells efficiently formed colonies at a frequency of 1:2–1:5, the colony formation frequency of TRS-HGPS was only 1:33–1:108 (p<0.01; Fig. 1A,B). The reduced clonogenic capacity of TRS-HGPS was not due to differences in proliferation rates or levels of exogenous TERT, SV40 T-antigens or HRAS mRNA or protein (Fig. S1C–E). Resistance of HGPS cells to transformation was confirmed in vivo by transplantation of TRS cells into nude mice (Fig. 1C,D). While 11 out of 16 injections of TRS-WT cells induced tumors, only 1 out of 16 injections of TRS-HGPS cells from two different patients generated a small tumor which appeared 3–4 weeks later than the wild type tumors and grew more slowly (p<0.01, Fig. 1C,D). Differences in tumorigenicity were corroborated in immunocompromised NSG mice, excluding the possibility of host effects (TRS-WT: 16/16 outgrowths, TRS-HGPS 0/8; Fig. S1F).

Figure 1. HGPS fibroblasts are resistant to experimental transformation.

Figure 1

Open in a new tab

A, B. Soft agar assay using transformed fibroblasts from two HGPS patients (TRS- HGPS1 and TRS-HGPS2) and aged-matched wild-type individuals (TRS-WT1 and TRS- WT2). Representative images of MTT-stained colonies (A) and quantification of colony formation (B). Values represent mean ± SEM (n=3). Statistical significance is indicated by one asterisk (p<0.01, TRS-HGPS cell lines versus TRS-WT2 cell line, one-way ANO- VA followed by Dunnett’s test). C,D, In vivo transplantation assays into nude mice (2.5 million injected cells). Image of a representative mouse in which TRS-WT (left flank) and TRS-HGPS (right flank) cells were injected (C) and quantification of tumor growth (D). Values of tumor volume represent mean ± SEM from 8 injections for each cell line. Statistical significance is indicated by one asterisk (p=0.0044, Student’s t test). E, Heat map representing relative expression of differentially expressed genes (DEG, fold ≥2, p<0.05) in TRS-WT cells compared to TRS-HGPS cells; expression of these genes is shown for all cell lines before and after experimental transformation. Red and blue represent the highest and lowest values of each gene among all samples, respectively. F, Venn diagram showing number of DEG during transformation in WT cells and HGPS cells (TRS-cells compared to TERT-cells). G, Relative expression of fibroblast-related genes in TRS-WT and TRS-HGPS cells compared to expression in their corresponding TERT-cell lines (dashed line). Average values from multiple probesets from RMA normalized microarray data were determined for each gene. Values represent mean ± SEM from 2 cell lines in each group. H, GSEA of indicated signatures on TRS-HGPS versus TRS-WT cells. NES=Normalized enrichment score. FDR q-val=false discovery rate q- value.

Transformation resistance of HGPS cells was supported by global gene expression analysis. While expression of 786 genes changed more than 2-fold during transformation of control wild type cells, only 15% of those responded to the transforming factors in HGPS cells (Fig. 1E,F). In contrast to TRS-WT cells, TRS-HGPS cells failed to activate several oncogenic pathways including RAS, VEGF, ERB2, EGFR and gene signatures of cytokine and chemokine activities (Fig. S2), as determined by gene set enrichment analysis (GSEA). On the other hand, TRS-HGPS cells failed to downregulate fibroblast- related pathways during the transformation process, such as those involved in extracellular matrix (ECM) organization (p<10−13), collagen fibril organization (p<10−9), and skin development (p<10−8) (Fig. S2D). While several collagens and other ECM components, which are highly expressed by fibroblasts, showed up to 80% downregulation in TSR- WT cells compared to untransformed cells, no significant reduction was seen in TRS- HGPS cells (Fig. 1G), indicating that, unlike wild type cells, HGPS cells retain their fibroblast identity upon experimental transformation. Moreover, gene signatures associated with human stem cells, such as mammary (Pece et al., 2010) and mesenchymal stem cells (Secco et al., 2009), were enriched in tumorigenic TRS-WT cells, but not in transformation-resistant TRS-HGPS cells (Fig. 1H, Fig S2B; see below). We have previously shown that, in line with reprogramming into a stem-cell-like state, fibroblasts acquire differentiation ability upon transformation (Scaffidi and Misteli, NCB 2011). While TRS-WT cells were able to efficiently differentiate into adipocytes in vitro, TRS-HGPS failed to do so, further confirming their inability to reprogram (Fig. S2E). We conclude that HGPS cells are refractory to oncogenic challenges and fail to undergo efficient de- differentiation and neoplastic reprogramming required for acquisition of tumor- initiating ability (Scaffidi and Misteli, 2011; Schwitalla et al., 2013).

Progerin is necessary and sufficient for tumor protection

The observed transformation resistance of HGPS cells may be due to the action of progerin or may be a secondary, adaptive response of HGPS cells. To distinguish between these possibilities, we performed knockdown and overexpression experiments. Silencing of progerin in TRS-HGPS cells (Fig. S3A) resulted in a 6-fold increase in the number of colonies formed in soft agar compared to control cells expressing shGFP (p<0.01) (Fig. 2A) and restored the ability to form tumors in vivo (3/3) (Fig. 2B), demonstrating that progerin is required for tumor protection. Conversely, induction of ectopic expression of progerin in TRS-WT cells induced a 2-fold reduction in the number of colonies in vitro (Fig. 2C, p<0.01) and completely suppressed their ability to form tumors in nude mice (0/5) (Fig. 2D). As a control, expression of wild-type lamin A did not impair tumor formation (4/5) (Fig. 2D). These results indicate that progerin is both necessary and sufficient to protect from oncogenic transformation. Furthermore, they show that the progerin-dependent protective mechanism can be reactivated in already transformed cells.

Figure 2. Progerin suppresses tumorigenicity of transformed fibroblasts.

Figure 2

Open in a new tab

A, Soft agar assay using TRS-HGPS cells in which progerin has been stably knocked- down. TRS-HGPS cells expressing shRNAs against LMNA (shLMNA) or GFP (shCtrl) were used as controls. Values represent mean ± SEM (n=4–6). Statistical significance is indicated by one or two asterisks (p<0.05, p<0.01, respectively, one-way ANOVA followed by Dunnett’s test). B, In vivo transplantation assays into NSG mice using TRS- HGPS cells in which progerin has been stably knocked-down (2.5 million injected cells). Values of tumor volume represent mean ± SEM from 3–4 injections for each cell line. Statistical significance is indicated by one asterisk (p<0.01, two-way ANOVA followed by Bonferroni post-test). C, Soft agar assay using TRS-WT cells expressing GFP-wild type lamin A (TRS-WT-GFP-wtLaminA) or GFP-progerin (TRS-WT-GFP-progerin). The number of colonies formed by each induced cell line (On) is normalized to their respective non-induced cell lines (Off). Values represent mean ± SEM (n=4). Statistical significance is indicated by two asterisks (p<0.001, Student’s t test). Representative images of colonies growing with or without induction of GFP-wtLamin A or GFP-progerin are shown. D, In vivo transplantation assays into nude mice using induced cell lines expressing GFP-lamin A or GFP-progerin (2.5 million injected cells). Values of tumor volume represent mean ± SEM from 5 injections for each cell line. Statistical significance is indicated by one asterisk (p<0.01, two-way ANOVA followed by Bonferroni post-test).

Identification of BRD4 as a mediator of transformation resistance

To identify factors that mediate transformation resistance, we performed an RNAi- based loss-of-function screen. TRS-HGPS cells were transduced with a pooled genome- wide shRNA library and cells were grown in soft agar for 4 weeks (see Methods). shRNAs restoring clonogenic capacity in TRS-HGPS cells were recovered from colonies and identified by hybridization on microarrays (Fig. 3A). Three independent screens were performed and 167 hits showing enrichment over the control population (trans- duced cells grown for 3 days as monolayer) in all three experiments were identified (Table S1). We focused on DNA-binding proteins and transcription regulators, since they represented the most prominent class of proteins based on GO analysis (p<3*10−3) (Fig. 3B) and because chromatin defects have been strongly implicated in HGPS (Scaffidi et al., 2005). 16 of 58 selected candidates were confirmed in vitro by soft agar assays (Fig. 3C and Table S2) and 3/7 of the top hits were positive in transplantation as- says in NSG mice (Fig. 3D).

Figure 3. A genome wide shRNA screen identifies BRD4 as a mediator of HGPS cells oncogenic resistance.

Figure 3

Open in a new tab

A, Outline of the performed functional genome-wide shRNA screen. B, Gene ontology analysis of positive hits by DAVID. The top enriched cluster and individual p values corresponding to the specific GO terms are shown. C, Soft agar assay using TRS-HGPS cells in which selected hits have been stably knocked-down. TRS-HGPS cells expressing the indicated shRNAs against gene candidates, GFP (shCtrl) and progerin (shProgerin) were assessed for colony formation. Values represent mean ± SEM (n=4–6). D, In vivo transplantation assays into NSG mice using TRS-HGPS cells in which the indicated proteins (in vitro validated positive hits) have been stably knocked- down (2.5 million injected cells). Untransduced TRS-HGPS cells were used as control. Values inside parentheses indicate number of outgrowths per injections. Values of tumor volume represent mean ± SEM from 4 injections for each cell line. Statistical significance is indicated by one asterisk (p<0.05, two-way ANOVA followed by Bonferroni post-test). E, Soft agar assay using TRS-HGPS cells in which BRD4 has been stably knocked-down. TRS-HGPS cells expressing the indicated shRNAs against BRD4 and GFP (shCtrl) were assessed for colony formation. Values represent mean ± SEM (n=4–6). Statistical significance is indicated by two asterisks (p<0.01, one way ANOVA followed by Dunnett’s test). F, In vivo transplantation assays into NSG mice using TRS-HGPS cells in which BRD4 has been stably knocked-down using the indicated shRNA (2 million injected cells). Untransduced TRS-HGPS cells were used as control cells. Values of tumor volume represent mean ± SEM from 3–4 injections for each cell line. Statistical significance is indicated by two asterisks (p<0.01, two-way ANOVA followed by Bonferroni post-test).

The hit showing the strongest effect in vivo was the bromodomain-containing 4 protein (BRD4). BRD4 is a member of the BET family of bromodomain-containing proteins that binds acetylated histones and modulates gene expression by recruiting transcriptional regulators to specific genomic locations (Wu and Chiang, 2007). BRD4 has been implicated in cellular growth, gene bookmarking and postmitotic transcription (Dey et al., 2003; Dey et al., 2009; Mochizuki et al., 2008; Zhao et al., 2011). BRD4 has a strong cancer promoting role in hematopoietic malignancies (Dawson et al., 2011; Delmore et al., 2011; Mertz et al., 2011; Ott et al., 2012; Zuber et al., 2011), and has been implicated in breast cancer as an anti-metastatic protein and in colon cancer as an anti-proliferative factor (Alsarraj et al., 2011; Crawford et al., 2008; Rodriguez et al., 2012). Reduction of BRD4 by ~60% using multiple shRNAs in TRS-HGPS cells increased colony formation 4–5 fold (Fig. 3E; Fig. S3C) and promoted efficient tumor formation (4/4 tumors for shBRD4-1; 3/3 for shBRD4-3) (Fig. 3D, F).

Given the reported role of BRD4 in promoting hematological cancers, we also assessed the tumorigenic potential of transformed B-lymphocytes from HGPS patients and age- matched healthy individuals. As expected, due to the very low expression levels of lam- in A, and thus progerin, in lymphocytes (Gerner and Sauermann, 1999; Rober et al., 1990) (Fig. S3G,H,J), transformed HGPS B-lymphocytes did not display typical progerin-induced nuclear defects (Fig. S3G,I) and formed colonies at similar frequency as wild type controls (p>0.05) (Fig. S3K).

Genome-wide redistribution of BRD4 in transformation resistant cells

An obvious potential mechanism for the protective effect of BRD4 is its upregulation in HGPS cells. This was not the case, since BRD4 levels were similar in TRS-HGPS cells compared to TRS-WT cells (Fig. S3D–F). Strikingly, however, genome-wide mapping by ChIP-Seq in multiple cell lines showed differential genome binding patterns of BRD4 in TRS-WT compared to TRS-HGPS cells (Fig. 4; Fig. S4A,B). In agreement with the notion that BRD4 regulates transcription by binding to both gene promoters and intergenic enhancers (Loven et al., 2013), ~9% (350 sites) and ~43% (1621 sites) of shared BRD4 binding sites between TRS-WT and TRS-HGPS cells localized at promoters or intergenic regions, respectively (Fig. 4B). TRS-WT-specific BRD4 binding sites showed a similar genomic distribution (~11% at promoters, ~34% at intergenic regions). In contrast, only 1% of TRS-HGPS-specific binding sites were in promoter regions whereas ~43% of the binding sites were at intergenic, indicating genomic redistribution of BRD4 (Fig. 4A–C). Amongst BRD4 binding sites shared between TRS-WT and TRS-HGPS cells, over 300 regions showed enhanced binding (≥2-fold) in TRS-HGPS cells (Fig. 4D), indicating that redistribution of BRD4 entailed both binding to distinct genomic sites and accumulation at locations occupied at low levels in TRS-WT cells. Interestingly, numerous BRD4 binding sites are in proximity to genes which differentially respond to transformation in TRS- HGPS and TRS-WT cells, including several collagen genes, FN1 and metallopeptidases (Fig. 1G; Table S3). Knock-down of progerin in TRS-HGPS cells restored TRS-WT-like BRD4 binding patterns (Fig. 4E). We conclude that BRD4 exhibits altered binding pat- terns in TRS-HGPS cells, as a result of both redistribution onto distinct genomic regions and accumulation at sites bound at lower levels in TRS-WT cells.

Figure 4. Distinct genomic distribution of BRD4 binding events in TRS-HGPS cells.

Figure 4

Open in a new tab

A, Distribution of TRS-WT- and TRS-HGPS-specific binding sites around gene transcription start sites (TSS). B, Genomic distribution of the indicated sets of BRD4 binding sites. The percentage of sites in each genomic class is indicated. C, D, Tracks of BRD4 binding sites in the indicated genomic regions. TRS-WT-specific site in the promoter region of HDAC1 (C, left) and TRS-HGPS-specific sites in an intergenic region of chromosome 2 (C, right). Examples of BRD4 binding sites common to both TRS-WT and TRS- HGPS cells, but showing enhanced binding (≥2-fold) in TRS-HGPS cells (D). E, Analysis of cell type-specific BRD4 binding to the indicated regions by ChIP-qPCR. Percentage of input was normalized to H3 ChIP-qPCR. Negative BRD4 binding was assessed by NANOG TSS qPCR and mean value is represented by the dashed line. Bars represent mean ± SEM (n=3–4). Statistical significance is indicated by one asterisk (p<0.05, Student’s t test, TRS-HGshProgerin versus TRS-HGPS).

In line with its genome-wide redistribution, BRD4 exhibited distinct cellular behavior in TRS-WT compared to TRS-HGPS cells. While BRD4 was distributed throughout the nucleus of TRS-WT cells, it accumulated in distinct foci in TRS-HGPS cells together with acetylated histone H4 (H4Ac), a major binding substrate (Fig. 5A,B, Fig. S5C). TRS- HGPS cells showed altered histone modification patterns, displaying large accumulations of H4Ac in regions of the nucleus but overall reduced levels of the modification compared to TRS-WT cells (Fig. 5A–D, Fig. S4D). Similar differences were observed for acetylated H3K9, another major BRD4 binding substrate (Fig. S4D). Alterations in his- tone modification patterns in TRS-HGPS are in agreement with observations in primary HGPS cells (McCord et al., 2013; Pegoraro et al., 2009; Scaffidi and Misteli, 2005; Shumaker et al., 2006) and confirm that the presence of progerin leads to global chromatin changes. Using quantitative Fluorescence Recovery After Photobleaching to measure chromatin binding of BRD4 in living cells, Cherry-BRD4 had a significantly slower recovery kinetics in TRS-HGPS cells, compared to TRS-WT cells, indicating stronger binding to chromatin (p<0.01) (Fig. 5E,F, Fig. S4E). Slowed recovery indicative of more persistent binding was particularly pronounced in H4Ac foci (T50foci = 2.5 sec T50nucleoplasm = 9 sec, Fig. 5F). Binding to chromatin of BRD4 was affected by progerin since induction of the protein in TRS-WT cells significantly reduced the mobility of Cherry-BRD4 to levels measured in TSR-HGPS cells, demonstrating that progerin affects BRD4 chromatin binding properties (Fig. 5E). We conclude that while there are not significant differences in BRD4 expression level in HGPS cells, BRD4 genome distribution and binding properties are altered.

Figure 5. Distinct histone modifications patterns in TRS-HGPS cells alter BRD4 sub- nuclear distribution and chromatin binding dynamics.

Figure 5

Open in a new tab

A, Immunofluorescence microscopy on TRS-WT1 and TRS-HGPS1 cells. Cells were either immunostained with an anti-BRD4 antibody (upper panels) or an anti-H4Ac antibody (bottom panels) or imaged as live cells upon transduction with a construct expressing a Cherry-BRD4 (Ch- BRD4) fusion protein (middle panels). BRD4 accumulates in large nuclear foci in TRS- HGPS cells. The graphs in green are line scans of fluorescence intensity in the regions marked by the red bars. Scale bar: 5 μm. B, Immunofluorescence microscopy on a TRS- HGPS1 cell expressing Ch-BRD4. Cells were fixed and stained with an anti-H4Ac. Co- localization of Ch-BRD4 foci and H4Ac foci is indicated by arrows and analyzed by line scan of fluorescence intensity (red bar). C, Distributions of the size of H4Ac foci in TRS- WT1 and TRS-HGPS1 cells (n>1000, p < 0.001, Kolmogorov-Smirnov test). Foci with area greater than 20 pixels were defined as large foci. D, Global H4-acetylation (H4Ac) was assessed by Western blot on purified histones from TRS-WT and TRS-HGPS cell lines. Total H3 expression level was used for normalization. A representative immunob- lot is shown. Relative densitometry values are indicated. E, FRAP analysis (Half-FRAP, see methods) of Ch-BRD4 in the indicated cell lines. Quantitative analysis of recovery of the fluorescence signal in the entire bleached area (half nucleus). Values are averages from 10–14 cells. Similar results were obtained in three independent experiments. For clarity, error bars are omitted; typical standard errors were below 10%. Statistical significance is indicated by two asterisks (p < 0.01) and was estimated by trying to fit the experimental points measured in the TRS-HGPS cell lines using the best fit model of the experimental points measured in TRS-WT1 cells (TRS-HGPS1 reduced χ2 = 11.6 ; TRS- WT1-progerin reduced χ2 = 145.5). F, FRAP analysis (Spot-FRAP, see methods) of Ch- BRD4 in living TRS-HGPS cells. Quantitative analysis of recovery of the fluorescence signal in the bleached area (either a Ch-BRD4 focus or a nucleoplasmic region of identical size). Values are averages from at least 5 cells ± SE. Statistical significance is indicated by two asterisks (p < 0.01) and was estimated by Student t-test comparing the relative fluorescence values 4 sec after bleaching.

Identification of BRD4-sensitive genes involved in transformation

To characterize cellular pathways involved in BRD4-mediated tumor resistance, we performed gene expression analysis. Consistent with the observed failure to activate various transformation pathways, comparison of expression profiles of TRS-HGPS-shBRD4 cells with that of TRS-WT and TRS-HGPS cells confirmed reversal of the resistant phenotype in HGPS cells upon BRD4 silencing. Knock-down of BRD4 in TRS-HGPS cells resulted in differential expression (fold change >2, FDR adjusted p<0.05) of 256 genes, of which 43% (111; 66 upregulated, 45 downregulated) overlapped with genes differentially expressed between TRS-WT cells TRS-HGPS cells (Fig. 6A). GSEA indicated that transformation-sensitive gene sets, defined as differentially expressed in TRS-WT cells versus T-WT cells (Table S3), were re-activated in the absence of BRD4 (Fig. S5B). Moreover, TRS-HGPS-shBRD4 cells re-expressed oncogenic gene networks, including the Ras signature, similarly to wild-type cells (Fig. S5B,C).

Figure 6. BRD4 protects from de-differentiation towards a stem-cell like state during transformation.

Figure 6

Open in a new tab

A, Heat map representing relative expression of DEG in TRS-HGPS cells compared to TRS-WT cells, which respond to BRD4 knock-down. B, GSEA plots showing enrichment of indicated gene sets in TRS-HGPS-shBRD4 versus TRS-HGPS cells. NES=Normalized enrichment score. FDR q-val=false discovery rate q-value. C, Relative expression of fibroblast-related markers in TRS-WT and TRS-HGPS-shBRD4 cells compared to expression in TRS-HGPS-cells. Average values from individual probesets from RMA normalized microarray data were determined for each gene. Values represent mean ± SEM (n=2–3). D, Sphere formation assay of indicated cell lines. Values represent mean ± SEM (n=4–6). Statistical significance is indicated by two asterisks (p<0.01, one-way ANOVA followed by Dunnett’s t test).

ChIP-seq analysis of BRD4 binding sites in TRS-WT and TRS-HGPS cells had revealed numerous binding sites near genes which respond differentially to transformation in TRS-HGPS cells compared to TRS-WT cells (Fig. 1G; Table S3), suggesting regulation of these genes by BRD4. In addition, comparative analysis of ChIP-seq BRD4 binding sites in TRS-HGPS cells with mis-regulated genes in TRS-HGPS-shBRD4 cells showed a significant overlap between BRD4 binding sites and components of the reactivated oncogenic pathways (Table S4). BRD4 binding sites were found near multiple genes associated with RAS signatures (IL8, CXCL1, PLAU, TFPI2), EGFR and VEGF oncogenic signatures (CXCL2, IL6, CXCL6, SOD2) and transformation-sensitive genes (COL1A2, COL5A2, COL8A1, CDH11) (Table S4).

BRD4 protects from de-differentiation during transformation

A prominent class of gene expression profiles identified in the GSEA analysis were stem cell gene signatures, which were enriched in the tumorigenic TRS-HGPS-shBRD4 and TRS-WT cells, but not in transformation-resistant TRS-HGPS cells (Fig. 6B, Fig. S5B), suggesting that loss of BRD4 enabled reactivation of stem cell gene signatures required to overcome the block in de-differentiation observed in TRS-HGPS cells. In line with this interpretation, TRS-HGPS-shBRD4 cells showed down-regulation of fibroblast- related genes that were unaffected in TRS-HGPS cells, suggesting successful de- differentiation of TRS-HGPS cells upon BRD4 silencing (Fig. 6C). In agreement, shRNA knock-down of BRD4 in TRS-HGPS significantly increased the percentage of cells able to withstand anoikis and to grow as spheres in suspension, a hallmark of normal and malignant stem cells (Pece et al., 2010) (Fig. 6D). Furthermore, treatment of TRS-HGPS cells with the BRD4 inhibitor JQ1 (100 nM) reduced cell viability of the total cell population but increased the sphere formation ability of the resistant population (Fig. S5D–F). These observations, combined with our finding that TRS-HGPS cells retain fibroblast- like features, suggest a role of BRD4 in protecting cells from de-differentiation during transformation.

BRD4 protection in human cancer

To explore a possible more general protective role of BRD4 in human cancer, we knocked down BRD4 in wild-type skin fibroblasts and assayed for transformation efficiency. The clonogenic ability of wild-type cells was increased ~ 2-fold compared to controls, suggesting a role for BRD4 in tumor protection in non-HGPS cells (Fig. S6B). Consistent with this notion, analysis using COSMIC identified 111 reported mutations in BRD4 in a collection of 9624 cancer samples, with 72% of mutations representing missense mutations likely leading to heterozygous loss of function. This mutation rate is comparable to that of well-established tumor suppressors including BLM, BRCA1, CDH11 or ETV6. To more directly probe the role of BRD4 in human cancer, we analyzed the expression of BRD4-sensitive genes identified in HGPS cells in samples from several types of human cancer including acute myeloid leukemia (AML) (Metzeler dataset (Metzeler et al., 2008), 163 samples), diffuse large B-cell lymphoma (DCBCL) (Hummel dataset (Hummel et al., 2006), 166 samples), in which BRD4 has previously been reported to act as a tumor promoting factor, breast cancer, where BRD4 has antimetastatic roles (van’t Veer dataset (van’t Veer et al., 2002), 295 samples; Hatzis dataset (Hatzis et al., 2011), 508 samples, HER2-negative only), and lung adenocarcinoma (Kohno dataset (Okayama et al., 2012), 246 samples), in which variable effects of BRD4 inhibition have been reported (Lockwood et al., 2012). We defined a BRD4-loss-of- function signature (BRD4-KD), consisting of genes that are significantly upregulated in TRS-HGPS-shBRD4 cells compared to TRS-HGPS cells (Table S3; see Methods), and measured its expression in patients using GSEA. In all cancers, patients clustered based on the similarity of their expression profiles with the BRD4-KD signature (Fig. 7, Table S5). Importantly, the BRD4-KD signature was predictive of patient survival in a cancer- dependent manner. In line with a tumor-protective role of BRD4, breast cancer patients expressing BRD4-sensitive genes similarly to tumor-resistant TRS-HGPS cells, had a significantly better outcome, both in a general clinical set of breast cancers (p=0.006) and in a set of HER2-negative cancers (p=0.003), whereas patients with patterns similar to tumorigenic TRS-HGPS-shBRD4 cells had shorter survival and were enriched in ER/PR negative cases (p<1.0 e-06) and radiation-resistant cases (p=1.0 e-04) (Fig. 7, Fig. S6A). Similarly, lung cancer patients with TRS-HGPS-like profiles had a longer survival (p=0.004), showed a lower rate of relapse (p=0.048), had a longer relapse-free time (p=7.7e-03) and lower stage tumors (p=2.2e-02) than patients with a TRS-HGPS- shBRD4-like signature (Fig. 7, Fig. S6A). Furthermore, in both breast and lung cancer datasets the BRD4-KD signature positively correlated with proliferation gene signatures (Table S7). The tumor-protective effect of BRD4 in these cancer types was experimentally confirmed in colony formation assays, which demonstrated a ~7-fold and ~2-fold increase, respectively, in the clonogenic capacity of T47D breast cancer and H358 lung adenocarcinoma cells upon knockdown of BRD4 using two different shRNAs (Fig. S6B), further supporting a tumor protective function of BRD4. In striking contrast, in lymphoma, where BRD4 has tumor promoting activity, patients characterized by a TRS- HGPS-shBRD4-like profile had a better outcome compared with the HGPS-like profile (p=0.03) (Fig. 7) and the BRD4-KD signature negatively correlated with proliferation gene signatures (Table S7). In AML, BRD4-sensitive genes identified in HGPS cells had no prognostic value (p=0.44) (Fig. 7), suggesting differential gene target specificity for BRD4 in this disease. In agreement, genes controlled by BRD4 in AML, including c- MYC, a major mediator of the BRD4 pro-tumorigenic effect, are insensitive to BRD4 knock-down in TRS-HGPS fibroblasts (Fig. S7). Together with our finding that loss of BRD4 in wild-type cells enhances experimental transformation (Fig. S6B), these results demonstrate a protective effect of BRD4 in some human cancer types. The differential behavior of BRD4 target genes in breast and lung cancer compared to the hematopoietic cancers and their different prognostic value suggest that BRD4 functions in a cancer- and tissue-dependent manner.

Figure 7. BRD4-KD gene signature predicts clinical outcomes.

Figure 7

Open in a new tab

Kaplan-Meier analysis of patients affected by the indicated cancer types, with clustering based on the BRD4- KD signature defined using HGPS cells. Each plot set corresponds to the data set indicated in the top line. Signature expression and clustering of the samples are illustrated in the heatmap bars, with a red (blue) color indicating significant positive (negative) association with the signature (black color: no significant association, see methods). Patients showing TRS-HGPS-like expression profiles (absence of the BRD4-KD signature, green cluster) have a better outcome in breast and lung cancer, in line with a tumorprotective role of BRD4, whereas patients showing TRS-HGPS-shBRD4-like expression profiles (presence of the BRD4-KD signature, red cluster) have a better outcome in lymphoma, in line with a tumor-promoting role for BRD4. Clustering of patients in AML based on the BRD4-KD signature has no prognostic value. Statistical significance (logrank test) is indicated in the graph plots.

Discussion

We demonstrated here the existence of a tumor protective mechanism in a rare premature aging disease and provide evidence for its relevance in human cancer. We show that the presence of progerin renders HGPS cells refractory to oncogenic challenges by preventing their reprogramming to acquire tumor-initiating ability. We find that ectopic expression of progerin in tumorigenic wild type cells is sufficient to override the effect of both HRAS pathway activation and p53 and pRB inhibition, indicating that the progerin-dependent protective mechanism can be reactivated in already transformed cells. In line with our findings, the normal lamin A precursor pre-lamin A has been re- ported to have a tumor-protective function (de la Rosa et al., 2013). We identify BRD4 as a downstream effector of progerin in mediating resistance to oncogenic transformation. BRD4 acts independently of p53 or pRB, which are both inhibited in our experimental system, as well as telomeres, since telomerase is ectopically expressed, and thus represents a tumor protective pathway distinct from cellular senescence, which has been suggested to limit the proliferative potential of HGPS cells (Kudlow et al., 2008). Importantly, while progerin strongly enhances the protective function of BRD4 in HGPS patients, BRD4-mediated tumor protection is also active in normal individuals, albeit much less potent, as indicated by the observed increased transformation efficiency of wild-type cells upon BRD4 knock-down and the prognostic value of BRD4-sensitive genes in breast and lung cancer.

Our ChIP-seq findings suggest that BRD4 exerts its protective function in HGPS cells by progerin-induced redistribution of its binding sites throughout the genome. HGPS cells have been demonstrated to have altered histone modification landscapes including loss of H3K9me3 and altered acetylation patterns (McCord et al., 2013; Pegoraro et al., 2009; Scaffidi and Misteli, 2005; Shumaker et al., 2006). These changes are possibly, at least in part, brought about by progerin-dependent changes to chromatin remodeling and modification factors. For example the histone deacetylase HDAC1, the NURD chromatin remodeling complex and the structural heterochromatin protein HP1 are downregulated in HGPS cells altering the global chromatin landscape (McCord et al., 2013; Pegoraro et al., 2009; Scaffidi and Misteli, 2005; Shumaker et al., 2006). In line with these observations, we find altered acetylation patterns in transformed HGPS cells. The altered his- tone modification landscape in HGPS cells creates a distinct set of binding sites for the transcriptional regulator BRD4, leading to altered gene expression programs. Consistent with the notion that the protective effect of BRD4 seen in HGPS cells is an exaggerate form of pathway which is also active in healthy individuals, low levels of progerin are present in healthy individuals (Cao et al., 2007; McClintock et al., 2007; Scaffidi and Misteli, 2006) and we find that numerous binding sites strongly bound by BRD4 in HGPS cells are also occupied at lower levels in wild type cells.

Based on analysis of cancer patient cohorts, our observations on tumor-protection in HGPS are of general clinical relevance for various cancers. We show that BRD4 target genes, which we find to mediate tumor resistance, have prognostic value in breast and lung cancer. We demonstrate that BRD4 loss-of-function gene signatures correlate with poor outcome, in line with a tumor protective role of the protein. Our observations are in agreement with previous reports showing that the BRD4 gain-of function signatures are associated with reduced risk of metastasis in breast cancer (Crawford et al., 2008) and that restoration of normal levels of BRD4 in colon cancer cells reduces in vivo tumor growth (Rodriguez et al., 2012). On the other hand, in agreement with the finding that BRD4 is critically required for disease maintenance in some hematological cancers and BRD4 inhibition has a strong anti-leukemic effect, we show that the BRD4 loss-of- function gene signature identified in HGPS cells correlates with good prognosis in lymphoma.

These considerations suggest that BRD4 has a dual role in cancer and can either exert a tumor-promoting or tumor-protective role depending on the cellular context. Precedents for other factors showing antithetic functions in different cancer types exist (Dawson and Kouzarides, 2012; Ikushima and Miyazono, 2010; Ntziachristos et al., 2014). The Notch signaling pathway exerts an oncogenic function in prostate, colon and mammary tissue, whereas it has a tumor-suppressive role in several leukemias (AML, B-ALL, chronic myelomonocytic leukemia) and lung carcinomas (small cell and squamous cell carcinoma) (Ntziachristos et al., 2014). Similarly, TGF-β acts as a classical tumor suppressing factor for most epithelial cells, but promotes proliferation of certain mesenchymal and cancer cells, and in late stage cancers it favors metastasis (Ikushima and Miyazono, 2010). Given the dependence of BRD4 function on chromatin status, maybe the most relevant bi-functional cancer modulator is the polycomb protein EZH2, a histone methyltransferase. EZH2 is overexpressed in various cancers and high levels of the protein correlate with poor prognosis in prostate and breast cancer, whereas recurrent loss-of-function mutations in EZH2 gene have been described in the myeloid malignancies and T-ALL (Dawson and Kouzarides, 2012). The mechanistic basis of the dual roles of these factors is largely uncharacterized, but a common theme, which similarly applies to BRD4, is the presence of numerous context-dependent factors that determine the final cellular response to the pleiotropic functions of these proteins. A possible explanation for the dual role of BRD4 is that cell type-specific chromatin landscapes may dictate their target genes. Histone acetylation profiles have been shown to stratify cancer cell lines according to cancer types, with breast and lung cancer cell lines clustering at the opposite side of the spectrum compared to leukemic cell lines (Leroy et al., 2013). In agreement, inhibition of BRD4 alters selective sets of genes in a tissue- specific manner (Dawson et al., 2011; Loven et al., 2013; Nicodeme et al., 2010). Gene- specific targeting may also be determined by different binding partners or by BRD4 phosphorylation status (Shi et al., 2014; Wu et al., 2013). Moreover, the existence of multiple BRD4 isoforms generated by alternative splicing may contribute to the tissue- specific mode of action of the protein (Alsarraj et al., 2011; Floyd et al., 2013).

Our findings indicate that human cells have the potential to counteract oncogenic stimuli through activation of a BRD4-dependent protective transcriptional program. This pathway is enhanced in HGPS cells due to the presence of progerin, but is also active in healthy individuals. The observation of a tumor protective role of BRD4 in lung and breast cancer extends earlier observations, which demonstrate a tumor promoting function of BRD4 in leukemias and lymphoma. Taken together, these results point to a highly context-dependent and tissue-specific function of BRD4. Given the intense interest in using interference with BRD4 function as a therapeutic approach, BRD4-targeted therapeutic strategies may therefore need to be cautiously evaluated depending on the cancer type so as to minimize potentially deleterious effects of these strategies in normal tissues.

Experimental procedures

The following experimental procedures are described in supplemental information: Cell lines and plasmids, microarrays and gene expression analysis, ChIP-sequencing and ChIP- qPCR, quantitative immunofluorescence, Western blot and FRAP analysis, and meta- analysis of published cancer data sets.

Soft-agar and non-adherent sphere-formation assays

Assays were carried out as previously described (Hahn et al., 1999; Scaffidi and Misteli, 2011). Briefly, for soft-agar assays cells were plated in six-well plates (5,000 per well) in 0.35% SeaPlaque Agar (Lonza) in MEM. For sphere formation, cells were plated in Knockout DMEM (Invitrogen) supplemented with Knockout Serum Replacement (Invitrogen) in uncoated Petri dishes at clonogenic density (1,000 cells ml−1). Plates were scanned after incubation with MTT (Sigma) at concentrations of 10mg/well for soft agar assays and 100mg/Petri dish for sphere assays. Colonies and spheres were quantified by using ImageJ software. JQ1 (kindly provided by J. Bradner) was dissolved in DMSO and tested at final concentration of 100 nM in the sphere formation assay. Media and fresh drug were replaced every 48 hours for a period of 8 days.

Transplantation assays

Six-week-old male immunodeficient BALB/cAnNCr-nu/nu mice (from the Animal Production Program, NCI-Frederick) and NOD/SCID/interleukin 2 receptor γnull mice (NSG, from The Jackson Laboratory) were maintained in pathogen-free conditions. For generation of tumors, cells (0.5–2.5×106 per injection in 50–100μl of PBS) were injected intradermally into the flanks of mice. NSG mice were locally shaved with a depilatory cream one day before injection. Tumor growth was assessed twice a week, using a digital caliper, for up to 10 weeks after injection. Tumor volume was calculated according to the formula d2D/2, where d and D are the shortest and the longest diameter, respectively.

Genome-wide shRNA screen

Three independent shRNA screens were performed. For each screen, TRS-HGPS cells (TRS-HGPS2) (6 × 106) were transduced with lentiviral shRNA particles contained in the GeneNet Human 50k shRNA Library (System Biosciences, SBI). The shRNA library comprises 200,000 probes that target ~50,000 human transcripts. Three days after infection, 10 × 106 infected cells were frozen and used as control samples (see below), while 10 × 106 cells were plated in 0.8% SeaPlaque Agar in MEM medium in 170 15-cm2 plates and incubated for colony formation assay. Colonies grown after 4 weeks (~600) were picked and pooled into groups of ~150 colonies each. Total RNA from each pool of colonies or control cells was extracted with Trizol reagent (Invitrogen) and processed according to the GeneNet manual to amplify and biotin-label the shRNAs. Labeled RNA was hybridized onto Human Genome U133 plus 2.0 Affymetrix microarrays. Signal intensities were retrieved from microarray CEL files according to SBI instructions. Independent control samples (whole population of transduced cells, n=2) were processed in each screen. 95% of shRNAs comprising the library were detected in control cell samples indicating good representation of the library complexity. For each shRNA sequence, fold change was calculated by dividing intensity value of colony samples over control samples (mean value of 2 samples) for each of the screens. We selected shRNAs showing ≥2-fold enrichment in each independent screen and considered as positive hits the gene targets of the selected shRNAs. For validation, individual shRNA were expressed in TRS-HGPS cells.

Statistical Analysis

Results are presented as mean ± SEM, unless otherwise stated, and derived from a minimum of three independent experiments. Statistical tests were performed using GraphPad Prism software package (version 5.0). In general, we used Student’s t test for comparisons between two experimental groups, one-way ANOVA followed by Dunnett’s test or Bonferroni multiple t test for comparisons of more than two groups, and two-way ANOVA to compare difference between groups at multiple time points. Additional statistical tests are described in figure legends. P value ≤0.05 was considered statistically significant.

Supplementary Material

1

Figure S1. Characterization of wild type and HGPS cells after experimental transformation. Related to Figure 1.

A, Immunofluorescence microscopy of primary and transformed (TRS) cell lines. Merge between γH2AX signal (red) and DAPI staining (blue) is shown. A representative image is shown for each cell population. Number of foci was quantified in 200 cells for each cell line and results are shown in the images (mean±SEM).

B, Morphology of TRS-WT and TRS-HGPS cells is shown in representative images.

C, Cell growth kinetics of the indicated cell lines assessed by proliferation assay (CellTiter 96 AQueous, Promega). Absorbance is proportional to the number of cells. Fold increased in absorbance was calculated for each cell line relative to the value measured at t=0. Log 2 values of fold are represented as mean ± SEM (n=3).

D, Western blot analysis of the indicated cell lines showing protein expression levels of SV40 Large T antigen, RAS (the antibody detects both endogenous and exogenous protein), and lamin A/C and β-actin as loading controls. TERT expression level was assessed by semi-quantitative RT-PCR. A TERT-immortalized wild type cell line (TERT-WT1) is used as control.

E, Analysis of TERT, SV40 Large T antigen and exogenous HRAS-V12 mRNA expression levels by qRT-PCR in TRS-WT and TRS-HGPS cell lines. Cyclophilin A and TBP mRNA were used for normalization.

F, In vivo transplantation assays into NSG mice. Values indicate number of outgrowths per injections detected 10 weeks after injection.

Figure S2. Oncogenic, inflammatory and stem cell signatures are downregulated in TRS-HGPS cells. Related to Figure 1.

A–C, GSEA of TRS-HGPS and TRS-WT cells using the indicated gene ontology sets (A), custom-generated or literature-based gene sets (B) and oncogenic signatures (C). Name and number of genes (size) of gene sets are shown in the left column. Normalized enrichment score (NES) of statistically significant upregulated or downregulated gene sets in TRS-HGPS cells are shown in red or blue color, respectively. FDR q-val=false discovery rate q-value. Statistical significance is given for FDR q-val<0.05. Asterisk indicates signatures shown in Figure 1.

D, Enriched GO terms identified by DAVID analysis of DEG during transformation in WT cells and HGPS cells (TRS-cells compared to TERT-cells). GO terms with p value <0.001, false discovery rate (FDR) <0.05 are represented.

E, TRS-HGPS cells fail to differentiate into adipocytes. Adipocyte differentiation was induced in TRS-WT and TRS-HGPS cell lines as previously described (Scaffidi and Misteli, Nature cell biology, 2011). Representative images of Oil Red O-stained cells are shown. Values represent mean±SEM of Oil-red O incorporation measurements, quantified by optical density (O.D.) at 490nm and normalized per number of cells.

Figure S3. Related to Figures 2 and 3.

Analysis of progerin, lamin A and BRD4 expression levels (A–L).

A–C, Quantitative RT-PCR analysis of TRS-HGPS cells in which progerin (A), lamin A (B) or BRD4 (C) were stably knocked-down. Relative mRNA levels are determined by comparison to control cells (TRS-HGPS-shCtrl).

D, Quantitative RT-PCR analysis of BRD4 expression level in TRS-HGPS compared to TRS-WT cells.

A–D, Values represent mean ± SEM (n=4–6). Cyclophilin A and TBP are used as housekeeping genes for normalization. Statistical significance is indicated by one (p<0.01, Student’s t test) or two (p<0.01, one-way ANOVA followed by Dunnett’s t test) asterisks.

E–F, Western blot analysis of BRD4 protein levels in TRS-WT (WT1, WT2) and TRS-HGPS (HG1, HG2) cell lines using two antibodies from Abcam and Bethyl laboratories. Hsc70 was used for normalization. Representative images are shown in E and quantification of relative protein levels is shown in F. Values represent mean±SEM (n=5–8 samples in each group).

Absence of progerin-mediated phenotypes in HGPS B-Lymphocytes (G–L).

G, Immunofluorescence microscopy on TRS-HGPS fibroblasts and transformed B-lymphocytes from two HGPS patients (LHG506 and LHG344) and one wild-type control (LWT508). Cells were stained with the indicated antibodies. Merge between green, red and DAPI channels is shown. Asterisks mark lymphocytes with abnormal A-type lamins staining in the endoplasmic reticulum (ER). Scale bar: 10 μm.

H, I, Quantification of cells showing high levels of A-type lamins and containing γH2AX foci. Most lymphocytes, from both HGPS patients and wild type control, express low levels of A-type lamins. The quantification includes cells with abnormal lamin A staining in the ER and thus overestimates the percentage of lymphocytes with high levels of properly localized A-type lamins. No lymphocytes containing γH2AX foci were observed. Statistical significance of the differences between the group of fibroblasts and the group of lymphocytes is indicated.

J, Low levels of lamin A in WT and HGPS lymphocytes were corroborated by qRT-PCR. Values represent mean ± SEM (n=3). Cyclophilin A and TBP are used as housekeeping genes for normalization.

K, Clonogenic ability of WT and HGPS lymphocytes was analyzed by colony formation in soft agar. Colonies were quantified after 14 days in culture. Values represent mean ± SEM (n=6–9). No statistically significant difference was observed between the HGPS and WT groups (unpaired t test, p>0.05).

L, Cell proliferation kinetics of the indicated cell lines assessed by MTS assay (CellTiter 96 AQueous, Promega). Values represent number of cells and are shown as mean ± SEM (n=3). Individual cell line differences in colony formation are likely due to differential proliferative kinetics.

Figure S4. Distinct genomic distribution, nuclear localization and binding dynamics of BRD4 in TRS-HGPS compared to TRS-WT cells. Related to Figures 4 and 5.

A, Heat map plot of pair-wise correlation generated using DiffBind package for BRD4 ChIP-seq analysis. TRS-WT and TRS-HGPS cell lines clustered into two distinct groups. Peaks were detected using SICER with cutoff of FDR 1E-5.

B, Genomic distribution of BRD4 binding events commonly occurred in TRS-WT and TRS-HGPS groups (two cell lines in each group).

C, Live-cell imaging of TRS-WT and TRS-HGPS cells expressing a Cherry-BRD4 (Ch-BRD4) fusion protein 3 days after infection. Representative images for each cell line are shown. Scale bar: 5 μm.

D, Quantitative analysis of H4panAc and H3K9Ac immunodetection in TRS-WT and TRS-HGPS cells by microscopy. For each box, the boundary of the box closest to zero indicates the 25th percentile of the distribution, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers indicate the 90th and 10th percentiles and the open circles the 95th and 5th percentiles. Statistical significance of the differences between each is indicated by one or two asterisks (p < 0.05, p < 0.01, respectively, TRS-HGPS cell lines versus TRS-WT2 cell line) and was estimated by Mann-Whitney test.

E, FRAP analysis of Ch-BRD4 in living TRS-HGPS cells. Quantitative analysis of recovery of the fluorescence signal in the bleached area (either a Ch-BRD4 focus or a nucleoplasmic region of identical size). Values are averages from at least 5 cells ± SE. Statistical significance is indicated by two asterisks (p < 0.01) and was estimated by Student t-test comparing the relative fluorescence values 4 sec after bleaching.

Figure S5. Oncogenic signatures, inflammatory gene sets and stem cell signatures are upregulated in TRS-HGPS cells after knockdown of BRD4. Related to Figure 6.

A–C, GSEA of the indicated pairs of cell lines using the indicated gene ontology sets (A), custom-generated or literature-based gene sets (B) and oncogenic signatures (C). Name and number of genes (size) of gene sets are shown in the left column. Normalized enrichment score (NES) of statistically significant upregulated or downregulated gene sets are shown in red or blue color, respectively. FDR q-val=false discovery rate q-value. Statistical significance is attributed for FDR q-val<0.05. Asterisk indicates signatures shown in Figure 6.

D–F, Sphere formation assay on TRS-HGPS cells treated with JQ1 (100 nM) or vehicle (DMSO).

D, Number of spheres was measured at day 8. Fold increase in sphere number over control cells is represented as mean ± SEM (n=6–8). Statistical significance is indicated by one asterisk (p<0.05, one-way ANOVA followed by Dunnett’s t test).

E, Cell viability was measured after 96 hours by MTT test.

F, Fold change in sphere number over control cells normalized to cell viability. Bars represent mean ± SEM. Statistical significance is indicated with two asterisks (p<0.001), as determined by Student’s t test.

Figure S6. BRD4 protective function in breast and lung cancer. Relates to Figure 7.

A, Clinical characterization of patients in data sets where BRD4 function is associated with a benign outcome: Heat maps show clustering of patients according to the association with BRD4-KD signature (bottom bar), and the resulting distribution of clinical variables (middle bars). Clusters are annotated in colors (top bar) and percentages of patients presenting a variable or average of variable for each cluster are listed on the right in the corresponding colors. Information on clinico-pathological variables across data sets is limited (in breast cancer data sets, neither stage, grade nor nodal status were annotated). The existing significant variables are shown here. P value showing statistical differences between the two groups was determined by Fischer’s exact test.

B, Knockdown of BRD4 increases clonogenic ability in human cancer cell lines: Soft agar assay using the indicated cell lines in which BRD4 had been stably knocked-down. Cells expressing shRNAs against BRD4 (shBRD4-1/-2) and GFP (shCtrl) were assessed for colony formation. Values represent mean ± SEM (n=3). Statistical significance is indicated by one asterisks (p<0.05, one way ANOVA followed by Dunnett’s test).

Fig. S7. Genes controlled by BRD4 in hematological cancers are insensitive to BRD4 knock-down in transformed fibroblasts. Relates to Figure 7.

Normalized microarray signals for the indicated genes measured in TRS-WT, TRS-WT-shBRD4, TRS-HGPS (TRS-HG) and TRS-HGPS-shBRD4 (TRS-HGshBRD4) cells. For each gene, probe signals were averaged to obtain a single gene expression value. While inhibition of BRD4 represses all these genes in hematological cancers (Dawson et al. Nature 2011), the majority of genes are unaffected in transformed fibroblasts by BRD4 knock-down, indicating differential target specificity of BRD4 in different cell types. Values represent mean ± SEM (n=2–3). Statistically significant differences between knocked-down cells and their respective controls are indicated by an asterisk (p<0.05, Student t-test).

2. Supplementary Table S1.

Hits identified through three independent genome-wide shRNA screens. Related to Figure 3

3. Supplementary Table S2.

shRNA sequences used in this study. Related to Figures 2 and 3

4. Supplementary Table S3.

Transformation gene sets and BRD4-KD signature. Related to Figures 1, 4, 6 and 7.

Transformation signatures include upregulated (Transformation UP) or downregulated (Transformation DOWN) genes when expression profiles of TRS-WT and TERT-WT cells are compared (fold change > 2, p value <0.05). BRD4-KD signature includes upregulated genes when expression profiles of TRS-HGPS-shBRD4 and TRS-HGPS cells are compared (fold change > 2, FDR corrected p value <0.05)

5. Supplementary Table S4.

Overlap between BRD4-sensitive genes in TRS-HGPS cells and TRS-HGPS BRD4 binding sites (ChIP-seq analysis). Related to Figures 4 and 6

Bold font indicates transformation-sensitive genes. Red font indicates genes upregulated in RAS oncogenic signatures (BILD HRAS SIGNATURE, KRAS LUNG BREAST UP, KRAS DF UP; see Fig S5). Underline indicates genes upregulated in the following oncogenic signatures: VEGF A UP, EGFR UP, HINATA NFKB MATRIX/IMMU INF, RAF UP (see Fig S5).

6

Supplementary Table S5. Distribution and significance of BRD4-KD signature association across data sets. Related to Figure 7

First column shows the mean of BRD4-KD signature scores for each set (scores normalized for numbers of genes on chip and in signature). Score distributions were normal with Lilliefors-test values <1.0e-03. Second column shows variance of normalized scores for each set. Third column lists the percentage of samples in each set whose signature score was not significant (p<0.05) as measured by random permutation test (random gene sets of same size). Fourth column lists the median of the p-values of score association across the set (ie half of all samples had significant scores with p-values below this median). These numbers depend on the layout of the array: the fewer signature genes are represented on the microarray chip, the lower statistical significance of score assignments.

Supplementary Table S6. Distribution of BRD4 expression across data sets. Related to Figure 7

The table compares the mean values and variance of expression (logscale) of the probes for BRD4, with the average mean value and average variance taken over all probes on the chip (mean of means). For each data set, the top line referenced as “Average on chip” shows the mean of means of all probes (ie the “expected mean”), and the mean of variances of all probes on the chip (“expected variance”). The third column annotates the 90th percentile, i.e. the value for which 90 percent of probes on the chip have an equal or lower mean expression. The following lines show the means and variances of the probes specific for BRD4 on the chip. Notably, with the exception of two probes in the Kohno set, which had exceptionally low signal or expression, BRD4 expression was within comparable ranges in both breast and lung as well as lymphoma and AML sets.

Supplementary Table S7. Correlation between BRD4KD signature and Proliferation signature. Related to Figure 7

Values in red indicate positive and significant correlation, values in blue indicate negative and significant correlation

Acknowledgments

We thank Keiko Ozato and Kent Hunter for helpful discussions, James Bradner, Peter Howley, Robert Weinberg, Thijn Brummelkamp, Anup Dey and William Lockwood for providing reagents, Vassilis Roukos and Nard Kubben for support with imaging, Anand Merchant for help with microarray analysis, Han Si and Margaret Cam for ChIP- seq analysis. Fluorescence imaging was performed at the NCI Fluorescence Imaging Facility and the NCI High-throughput Imaging Facility. P.F. was supported by the Instituto de Salud Carlos III (Spain). This project was funded, in part, by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research, and a grant from the Ellison Medical Research Foundation.

Footnotes

Author contributions

PS generated and characterized TRS-WT and TRS-HGPS cells lines, performed initial in vitro and in vivo tumorigenicity assays, in vitro differentiation assays, FRAP experiments and BRD4 nuclear localization analysis. PF performed the RNAi screen, in vitro and in vivo validation, generation of shRNA-stable cell lines and all other BRD4-related experiments. PF performed microarray data analysis, GSEA and ChIP experiments. PF and PS analyzed and interpreted data. EM performed analysis of cancer data sets. JL and SR helped with initial in vivo experiments. PF, PS and TM designed experiments and wrote the manuscript with input from EM. PS and TM conceived and supervised the study.

Accession numbers

The NCBI Gene Expression Omnibus accession numbers for the microarray data and ChIP-seq data reported in this paper are GEO67697, GEO67698, GEO72541.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Alsarraj J, Walker RC, Webster JD, Geiger TR, Crawford NP, Simpson RM, Ozato K, Hunter KW. Deletion of the proline-rich region of the murine metastasis susceptibility gene Brd4 promotes epithelial-to-mesenchymal transition- and stem cell-like conversion. Cancer research. 2011;71:3121–3131. doi: 10.1158/0008-5472.CAN-10-4417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brosh RM, Jr, Bohr VA. Human premature aging, DNA repair and RecQ helicases. Nucleic acids research. 2007;35:7527–7544. doi: 10.1093/nar/gkm1008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Cao K, Capell BC, Erdos MR, Djabali K, Collins FS. A lamin A protein isoform overexpressed in Hutchinson-Gilford progeria syndrome interferes with mitosis in progeria and normal cells. Proceedings of the National Academy of Sciences of the United States of America. 2007;104:4949–4954. doi: 10.1073/pnas.0611640104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Crawford NP, Alsarraj J, Lukes L, Walker RC, Officewala JS, Yang HH, Lee MP, Ozato K, Hunter KW. Bromodomain 4 activation predicts breast cancer survival. Proceedings of the National Academy of Sciences of the United States of America. 2008;105:6380– 6385. doi: 10.1073/pnas.0710331105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Dawson MA, Kouzarides T. Cancer epigenetics: from mechanism to therapy. Cell. 2012;150:12–27. doi: 10.1016/j.cell.2012.06.013. [DOI] [PubMed] [Google Scholar]
  6. Dawson MA, Prinjha RK, Dittmann A, Giotopoulos G, Bantscheff M, Chan WI, Robson SC, Chung CW, Hopf C, Savitski MM, et al. Inhibition of BET recruitment to chromatin as an effective treatment for MLL-fusion leukaemia. Nature. 2011;478:529– 533. doi: 10.1038/nature10509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. de la Rosa J, Freije JM, Cabanillas R, Osorio FG, Fraga MF, Fernandez-Garcia MS, Rad R, Fanjul V, Ugalde AP, Liang Q, et al. Prelamin A causes progeria through cell-extrinsic mechanisms and prevents cancer invasion. Nature communications. 2013;4:2268. doi: 10.1038/ncomms3268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. De Sandre-Giovannoli A, Bernard R, Cau P, Navarro C, Amiel J, Boccaccio I, Lyonnet S, Stewart CL, Munnich A, Le Merrer M, et al. Lamin a truncation in Hutchinson- Gilford progeria. Science. 2003;300:2055. doi: 10.1126/science.1084125. [DOI] [PubMed] [Google Scholar]
  9. Delmore JE, Issa GC, Lemieux ME, Rahl PB, Shi J, Jacobs HM, Kastritis E, Gilpatrick T, Paranal RM, Qi J, et al. BET bromodomain inhibition as a therapeutic strategy to target c-Myc. Cell. 2011;146:904–917. doi: 10.1016/j.cell.2011.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dey A, Chitsaz F, Abbasi A, Misteli T, Ozato K. The double bromodomain protein Brd4 binds to acetylated chromatin during interphase and mitosis. Proceedings of the National Academy of Sciences of the United States of America. 2003;100:8758–8763. doi: 10.1073/pnas.1433065100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dey A, Nishiyama A, Karpova T, McNally J, Ozato K. Brd4 marks select genes on mitotic chromatin and directs postmitotic transcription. Molecular biology of the cell. 2009;20:4899–4909. doi: 10.1091/mbc.E09-05-0380. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Eriksson M, Brown WT, Gordon LB, Glynn MW, Singer J, Scott L, Erdos MR, Robbins CM, Moses TY, Berglund P, et al. Recurrent de novo point mutations in lamin A cause Hutchinson-Gilford progeria syndrome. Nature. 2003;423:293–298. doi: 10.1038/nature01629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Floyd SR, Pacold ME, Huang Q, Clarke SM, Lam FC, Cannell IG, Bryson BD, Rameseder J, Lee MJ, Blake EJ, et al. The bromodomain protein Brd4 insulates chromatin from DNA damage signalling. Nature. 2013;498:246–250. doi: 10.1038/nature12147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gerner C, Sauermann G. Nuclear matrix proteins specific for subtypes of human hematopoietic cells. Journal of cellular biochemistry. 1999;72:470–482. [PubMed] [Google Scholar]
  15. Gordon LB, Rothman FG, Lopez-Otin C, Misteli T. Progeria: a paradigm for translational medicine. Cell. 2014;156:400–407. doi: 10.1016/j.cell.2013.12.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hahn WC, Counter CM, Lundberg AS, Beijersbergen RL, Brooks MW, Weinberg RA. Creation of human tumour cells with defined genetic elements. Nature. 1999;400:464– 468. doi: 10.1038/22780. [DOI] [PubMed] [Google Scholar]
  17. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100:57–70. doi: 10.1016/s0092-8674(00)81683-9. [DOI] [PubMed] [Google Scholar]
  18. Hatzis C, Pusztai L, Valero V, Booser DJ, Esserman L, Lluch A, Vidaurre T, Holmes F, Souchon E, Wang H, et al. A genomic predictor of response and survival following taxane-anthracycline chemotherapy for invasive breast cancer. JAMA : the journal of the American Medical Association. 2011;305:1873–1881. doi: 10.1001/jama.2011.593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hoeijmakers JH. DNA damage, aging, and cancer. The New England journal of medicine. 2009;361:1475–1485. doi: 10.1056/NEJMra0804615. [DOI] [PubMed] [Google Scholar]
  20. Hummel M, Bentink S, Berger H, Klapper W, Wessendorf S, Barth TF, Bernd HW, Cogliatti SB, Dierlamm J, Feller AC, et al. A biologic definition of Burkitt’s lymphoma from transcriptional and genomic profiling. The New England journal of medicine. 2006;354:2419–2430. doi: 10.1056/NEJMoa055351. [DOI] [PubMed] [Google Scholar]
  21. Ikushima H, Miyazono K. TGFbeta signalling: a complex web in cancer progression. Nature reviews Cancer. 2010;10:415–424. doi: 10.1038/nrc2853. [DOI] [PubMed] [Google Scholar]
  22. Kudlow BA, Stanfel MN, Burtner CR, Johnston ED, Kennedy BK. Suppression of proliferative defects associated with processing-defective lamin A mutants by hTERT or inactivation of p53. Molecular biology of the cell. 2008;19:5238–5248. doi: 10.1091/mbc.E08-05-0492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Leroy G, Dimaggio PA, Chan EY, Zee BM, Blanco MA, Bryant B, Flaniken IZ, Liu S, Kang Y, Trojer P, et al. A quantitative atlas of histone modification signatures from human cancer cells. Epigenetics & chromatin. 2013;6:20. doi: 10.1186/1756-8935-6-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lockwood WW, Zejnullahu K, Bradner JE, Varmus H. Sensitivity of human lung adenocarcinoma cell lines to targeted inhibition of BET epigenetic signaling proteins. Proceedings of the National Academy of Sciences of the United States of America. 2012;109:19408– 19413. doi: 10.1073/pnas.1216363109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Loven J, Hoke HA, Lin CY, Lau A, Orlando DA, Vakoc CR, Bradner JE, Lee TI, Young RA. Selective inhibition of tumor oncogenes by disruption of super- enhancers. Cell. 2013;153:320–334. doi: 10.1016/j.cell.2013.03.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. McClintock D, Ratner D, Lokuge M, Owens DM, Gordon LB, Collins FS, Djabali K. The mutant form of lamin A that causes Hutchinson-Gilford progeria is a biomarker of cellular aging in human skin. PloS one. 2007;2:e1269. doi: 10.1371/journal.pone.0001269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McCord RP, Nazario-Toole A, Zhang H, Chines PS, Zhan Y, Erdos MR, Collins FS, Dekker J, Cao K. Correlated alterations in genome organization, histone methylation, and DNA-lamin A/C interactions in Hutchinson-Gilford progeria syndrome. Genome research. 2013;23:260–269. doi: 10.1101/gr.138032.112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Mertz JA, Conery AR, Bryant BM, Sandy P, Balasubramanian S, Mele DA, Bergeron L, Sims RJ., 3rd Targeting MYC dependence in cancer by inhibiting BET bromodomains. Proceedings of the National Academy of Sciences of the United States of America. 2011;108:16669–16674. doi: 10.1073/pnas.1108190108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Metzeler KH, Hummel M, Bloomfield CD, Spiekermann K, Braess J, Sauerland MC, Heinecke A, Radmacher M, Marcucci G, Whitman SP, et al. An 86-probe-set gene- expression signature predicts survival in cytogenetically normal acute myeloid leukemia. Blood. 2008;112:4193–4201. doi: 10.1182/blood-2008-02-134411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mochizuki K, Nishiyama A, Jang MK, Dey A, Ghosh A, Tamura T, Natsume H, Yao H, Ozato K. The bromodomain protein Brd4 stimulates G1 gene transcription and promotes progression to S phase. The Journal of biological chemistry. 2008;283:9040–9048. doi: 10.1074/jbc.M707603200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nicodeme E, Jeffrey KL, Schaefer U, Beinke S, Dewell S, Chung CW, Chandwani R, Marazzi I, Wilson P, Coste H, et al. Suppression of inflammation by a synthetic histone mimic. Nature. 2010;468:1119–1123. doi: 10.1038/nature09589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Ntziachristos P, Lim JS, Sage J, Aifantis I. From Fly Wings to Targeted Cancer Therapies: A Centennial for Notch Signaling. Cancer cell. 2014;25:318–334. doi: 10.1016/j.ccr.2014.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Okayama H, Kohno T, Ishii Y, Shimada Y, Shiraishi K, Iwakawa R, Furuta K, Tsuta K, Shibata T, Yamamoto S, et al. Identification of genes upregulated in ALK-positive and EGFR/KRAS/ALK-negative lung adenocarcinomas. Cancer research. 2012;72:100–111. doi: 10.1158/0008-5472.CAN-11-1403. [DOI] [PubMed] [Google Scholar]
  34. Ott CJ, Kopp N, Bird L, Paranal RM, Qi J, Bowman T, Rodig SJ, Kung AL, Bradner JE, Weinstock DM. BET bromodomain inhibition targets both c-Myc and IL7R in high-risk acute lymphoblastic leukemia. Blood. 2012;120:2843–2852. doi: 10.1182/blood-2012-02-413021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pece S, Tosoni D, Confalonieri S, Mazzarol G, Vecchi M, Ronzoni S, Bernard L, Viale G, Pelicci PG, Di Fiore PP. Biological and molecular heterogeneity of breast cancers correlates with their cancer stem cell content. Cell. 2010;140:62–73. doi: 10.1016/j.cell.2009.12.007. [DOI] [PubMed] [Google Scholar]
  36. Pegoraro G, Kubben N, Wickert U, Gohler H, Hoffmann K, Misteli T. Ageing-related chromatin defects through loss of the NURD complex. Nature cell biology. 2009;11:1261–1267. doi: 10.1038/ncb1971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rober RA, Sauter H, Weber K, Osborn M. Cells of the cellular immune and hemopoietic system of the mouse lack lamins A/C: distinction versus other somatic cells. Journal of cell science. 1990;95(Pt 4):587–598. doi: 10.1242/jcs.95.4.587. [DOI] [PubMed] [Google Scholar]
  38. Rodriguez RM, Huidobro C, Urdinguio RG, Mangas C, Soldevilla B, Dominguez G, Bonilla F, Fernandez AF, Fraga MF. Aberrant epigenetic regulation of bromodomain BRD4 in human colon cancer. J Mol Med (Berl) 2012;90:587–595. doi: 10.1007/s00109-011-0837-0. [DOI] [PubMed] [Google Scholar]
  39. Scaffidi P, Gordon L, Misteli T. The cell nucleus and aging: tantalizing clues and hopeful promises. PLoS biology. 2005;3:e395. doi: 10.1371/journal.pbio.0030395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Scaffidi P, Misteli T. Reversal of the cellular phenotype in the premature aging disease Hutchinson-Gilford progeria syndrome. Nature medicine. 2005;11:440–445. doi: 10.1038/nm1204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Scaffidi P, Misteli T. Lamin A-dependent nuclear defects in human aging. Science. 2006;312:1059–1063. doi: 10.1126/science.1127168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Scaffidi P, Misteli T. In vitro generation of human cells with cancer stem cell properties. Nature cell biology. 2011;13:1051–1061. doi: 10.1038/ncb2308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Schwitalla S, Fingerle AA, Cammareri P, Nebelsiek T, Goktuna SI, Ziegler PK, Canli O, Heijmans J, Huels DJ, Moreaux G, et al. Intestinal tumorigenesis initiated by dedifferentiation and acquisition of stem-cell-like properties. Cell. 2013;152:25–38. doi: 10.1016/j.cell.2012.12.012. [DOI] [PubMed] [Google Scholar]
  44. Secco M, Moreira YB, Zucconi E, Vieira NM, Jazedje T, Muotri AR, Okamoto OK, Verjovski-Almeida S, Zatz M. Gene expression profile of mesenchymal stem cells from paired umbilical cord units: cord is different from blood. Stem cell reviews. 2009;5:387–401. doi: 10.1007/s12015-009-9098-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Shi J, Wang Y, Zeng L, Wu Y, Deng J, Zhang Q, Lin Y, Li J, Kang T, Tao M, et al. Disrupting the Interaction of BRD4 with Diacetylated Twist Suppresses Tumorigenesis in Basal-like Breast Cancer. Cancer cell. 2014;25:210–225. doi: 10.1016/j.ccr.2014.01.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Shumaker DK, Dechat T, Kohlmaier A, Adam SA, Bozovsky MR, Erdos MR, Eriksson M, Goldman AE, Khuon S, Collins FS, et al. Mutant nuclear lamin A leads to progressive alterations of epigenetic control in premature aging. Proceedings of the National Academy of Sciences of the United States of America. 2006;103:8703–8708. doi: 10.1073/pnas.0602569103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. van’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature. 2002;415:530–536. doi: 10.1038/415530a. [DOI] [PubMed] [Google Scholar]
  48. Venkitaraman AR. Cancer susceptibility and the functions of BRCA1 and BRCA2. Cell. 2002;108:171–182. doi: 10.1016/s0092-8674(02)00615-3. [DOI] [PubMed] [Google Scholar]
  49. Vicente-Duenas C, Romero-Camarero I, Cobaleda C, Sanchez-Garcia I. Function of oncogenes in cancer development: a changing paradigm. The EMBO journal. 2013;32:1502–1513. doi: 10.1038/emboj.2013.97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wu SY, Chiang CM. The double bromodomain-containing chromatin adaptor Brd4 and transcriptional regulation. The Journal of biological chemistry. 2007;282:13141–13145. doi: 10.1074/jbc.R700001200. [DOI] [PubMed] [Google Scholar]
  51. Wu SY, Lee AY, Lai HT, Zhang H, Chiang CM. Phospho switch triggers Brd4 chromatin binding and activator recruitment for gene-specific targeting. Molecular cell. 2013;49:843–857. doi: 10.1016/j.molcel.2012.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhao R, Nakamura T, Fu Y, Lazar Z, Spector DL. Gene bookmarking accelerates the kinetics of post-mitotic transcriptional re-activation. Nature cell biology. 2011;13:1295–1304. doi: 10.1038/ncb2341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zuber J, Shi J, Wang E, Rappaport AR, Herrmann H, Sison EA, Magoon D, Qi J, Blatt K, Wunderlich M, et al. RNAi screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. Nature. 2011;478:524–528. doi: 10.1038/nature10334. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1

Figure S1. Characterization of wild type and HGPS cells after experimental transformation. Related to Figure 1.

A, Immunofluorescence microscopy of primary and transformed (TRS) cell lines. Merge between γH2AX signal (red) and DAPI staining (blue) is shown. A representative image is shown for each cell population. Number of foci was quantified in 200 cells for each cell line and results are shown in the images (mean±SEM).

B, Morphology of TRS-WT and TRS-HGPS cells is shown in representative images.

C, Cell growth kinetics of the indicated cell lines assessed by proliferation assay (CellTiter 96 AQueous, Promega). Absorbance is proportional to the number of cells. Fold increased in absorbance was calculated for each cell line relative to the value measured at t=0. Log 2 values of fold are represented as mean ± SEM (n=3).

D, Western blot analysis of the indicated cell lines showing protein expression levels of SV40 Large T antigen, RAS (the antibody detects both endogenous and exogenous protein), and lamin A/C and β-actin as loading controls. TERT expression level was assessed by semi-quantitative RT-PCR. A TERT-immortalized wild type cell line (TERT-WT1) is used as control.

E, Analysis of TERT, SV40 Large T antigen and exogenous HRAS-V12 mRNA expression levels by qRT-PCR in TRS-WT and TRS-HGPS cell lines. Cyclophilin A and TBP mRNA were used for normalization.

F, In vivo transplantation assays into NSG mice. Values indicate number of outgrowths per injections detected 10 weeks after injection.

Figure S2. Oncogenic, inflammatory and stem cell signatures are downregulated in TRS-HGPS cells. Related to Figure 1.

A–C, GSEA of TRS-HGPS and TRS-WT cells using the indicated gene ontology sets (A), custom-generated or literature-based gene sets (B) and oncogenic signatures (C). Name and number of genes (size) of gene sets are shown in the left column. Normalized enrichment score (NES) of statistically significant upregulated or downregulated gene sets in TRS-HGPS cells are shown in red or blue color, respectively. FDR q-val=false discovery rate q-value. Statistical significance is given for FDR q-val<0.05. Asterisk indicates signatures shown in Figure 1.

D, Enriched GO terms identified by DAVID analysis of DEG during transformation in WT cells and HGPS cells (TRS-cells compared to TERT-cells). GO terms with p value <0.001, false discovery rate (FDR) <0.05 are represented.

E, TRS-HGPS cells fail to differentiate into adipocytes. Adipocyte differentiation was induced in TRS-WT and TRS-HGPS cell lines as previously described (Scaffidi and Misteli, Nature cell biology, 2011). Representative images of Oil Red O-stained cells are shown. Values represent mean±SEM of Oil-red O incorporation measurements, quantified by optical density (O.D.) at 490nm and normalized per number of cells.

Figure S3. Related to Figures 2 and 3.

Analysis of progerin, lamin A and BRD4 expression levels (A–L).

A–C, Quantitative RT-PCR analysis of TRS-HGPS cells in which progerin (A), lamin A (B) or BRD4 (C) were stably knocked-down. Relative mRNA levels are determined by comparison to control cells (TRS-HGPS-shCtrl).

D, Quantitative RT-PCR analysis of BRD4 expression level in TRS-HGPS compared to TRS-WT cells.

A–D, Values represent mean ± SEM (n=4–6). Cyclophilin A and TBP are used as housekeeping genes for normalization. Statistical significance is indicated by one (p<0.01, Student’s t test) or two (p<0.01, one-way ANOVA followed by Dunnett’s t test) asterisks.

E–F, Western blot analysis of BRD4 protein levels in TRS-WT (WT1, WT2) and TRS-HGPS (HG1, HG2) cell lines using two antibodies from Abcam and Bethyl laboratories. Hsc70 was used for normalization. Representative images are shown in E and quantification of relative protein levels is shown in F. Values represent mean±SEM (n=5–8 samples in each group).

Absence of progerin-mediated phenotypes in HGPS B-Lymphocytes (G–L).

G, Immunofluorescence microscopy on TRS-HGPS fibroblasts and transformed B-lymphocytes from two HGPS patients (LHG506 and LHG344) and one wild-type control (LWT508). Cells were stained with the indicated antibodies. Merge between green, red and DAPI channels is shown. Asterisks mark lymphocytes with abnormal A-type lamins staining in the endoplasmic reticulum (ER). Scale bar: 10 μm.

H, I, Quantification of cells showing high levels of A-type lamins and containing γH2AX foci. Most lymphocytes, from both HGPS patients and wild type control, express low levels of A-type lamins. The quantification includes cells with abnormal lamin A staining in the ER and thus overestimates the percentage of lymphocytes with high levels of properly localized A-type lamins. No lymphocytes containing γH2AX foci were observed. Statistical significance of the differences between the group of fibroblasts and the group of lymphocytes is indicated.

J, Low levels of lamin A in WT and HGPS lymphocytes were corroborated by qRT-PCR. Values represent mean ± SEM (n=3). Cyclophilin A and TBP are used as housekeeping genes for normalization.

K, Clonogenic ability of WT and HGPS lymphocytes was analyzed by colony formation in soft agar. Colonies were quantified after 14 days in culture. Values represent mean ± SEM (n=6–9). No statistically significant difference was observed between the HGPS and WT groups (unpaired t test, p>0.05).

L, Cell proliferation kinetics of the indicated cell lines assessed by MTS assay (CellTiter 96 AQueous, Promega). Values represent number of cells and are shown as mean ± SEM (n=3). Individual cell line differences in colony formation are likely due to differential proliferative kinetics.

Figure S4. Distinct genomic distribution, nuclear localization and binding dynamics of BRD4 in TRS-HGPS compared to TRS-WT cells. Related to Figures 4 and 5.

A, Heat map plot of pair-wise correlation generated using DiffBind package for BRD4 ChIP-seq analysis. TRS-WT and TRS-HGPS cell lines clustered into two distinct groups. Peaks were detected using SICER with cutoff of FDR 1E-5.

B, Genomic distribution of BRD4 binding events commonly occurred in TRS-WT and TRS-HGPS groups (two cell lines in each group).

C, Live-cell imaging of TRS-WT and TRS-HGPS cells expressing a Cherry-BRD4 (Ch-BRD4) fusion protein 3 days after infection. Representative images for each cell line are shown. Scale bar: 5 μm.

D, Quantitative analysis of H4panAc and H3K9Ac immunodetection in TRS-WT and TRS-HGPS cells by microscopy. For each box, the boundary of the box closest to zero indicates the 25th percentile of the distribution, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers indicate the 90th and 10th percentiles and the open circles the 95th and 5th percentiles. Statistical significance of the differences between each is indicated by one or two asterisks (p < 0.05, p < 0.01, respectively, TRS-HGPS cell lines versus TRS-WT2 cell line) and was estimated by Mann-Whitney test.

E, FRAP analysis of Ch-BRD4 in living TRS-HGPS cells. Quantitative analysis of recovery of the fluorescence signal in the bleached area (either a Ch-BRD4 focus or a nucleoplasmic region of identical size). Values are averages from at least 5 cells ± SE. Statistical significance is indicated by two asterisks (p < 0.01) and was estimated by Student t-test comparing the relative fluorescence values 4 sec after bleaching.

Figure S5. Oncogenic signatures, inflammatory gene sets and stem cell signatures are upregulated in TRS-HGPS cells after knockdown of BRD4. Related to Figure 6.

A–C, GSEA of the indicated pairs of cell lines using the indicated gene ontology sets (A), custom-generated or literature-based gene sets (B) and oncogenic signatures (C). Name and number of genes (size) of gene sets are shown in the left column. Normalized enrichment score (NES) of statistically significant upregulated or downregulated gene sets are shown in red or blue color, respectively. FDR q-val=false discovery rate q-value. Statistical significance is attributed for FDR q-val<0.05. Asterisk indicates signatures shown in Figure 6.

D–F, Sphere formation assay on TRS-HGPS cells treated with JQ1 (100 nM) or vehicle (DMSO).

D, Number of spheres was measured at day 8. Fold increase in sphere number over control cells is represented as mean ± SEM (n=6–8). Statistical significance is indicated by one asterisk (p<0.05, one-way ANOVA followed by Dunnett’s t test).

E, Cell viability was measured after 96 hours by MTT test.

F, Fold change in sphere number over control cells normalized to cell viability. Bars represent mean ± SEM. Statistical significance is indicated with two asterisks (p<0.001), as determined by Student’s t test.

Figure S6. BRD4 protective function in breast and lung cancer. Relates to Figure 7.

A, Clinical characterization of patients in data sets where BRD4 function is associated with a benign outcome: Heat maps show clustering of patients according to the association with BRD4-KD signature (bottom bar), and the resulting distribution of clinical variables (middle bars). Clusters are annotated in colors (top bar) and percentages of patients presenting a variable or average of variable for each cluster are listed on the right in the corresponding colors. Information on clinico-pathological variables across data sets is limited (in breast cancer data sets, neither stage, grade nor nodal status were annotated). The existing significant variables are shown here. P value showing statistical differences between the two groups was determined by Fischer’s exact test.

B, Knockdown of BRD4 increases clonogenic ability in human cancer cell lines: Soft agar assay using the indicated cell lines in which BRD4 had been stably knocked-down. Cells expressing shRNAs against BRD4 (shBRD4-1/-2) and GFP (shCtrl) were assessed for colony formation. Values represent mean ± SEM (n=3). Statistical significance is indicated by one asterisks (p<0.05, one way ANOVA followed by Dunnett’s test).

Fig. S7. Genes controlled by BRD4 in hematological cancers are insensitive to BRD4 knock-down in transformed fibroblasts. Relates to Figure 7.

Normalized microarray signals for the indicated genes measured in TRS-WT, TRS-WT-shBRD4, TRS-HGPS (TRS-HG) and TRS-HGPS-shBRD4 (TRS-HGshBRD4) cells. For each gene, probe signals were averaged to obtain a single gene expression value. While inhibition of BRD4 represses all these genes in hematological cancers (Dawson et al. Nature 2011), the majority of genes are unaffected in transformed fibroblasts by BRD4 knock-down, indicating differential target specificity of BRD4 in different cell types. Values represent mean ± SEM (n=2–3). Statistically significant differences between knocked-down cells and their respective controls are indicated by an asterisk (p<0.05, Student t-test).

NIHMS627470-supplement-1.pdf (2.2MB, pdf)
2. Supplementary Table S1.

Hits identified through three independent genome-wide shRNA screens. Related to Figure 3

NIHMS627470-supplement-2.docx (16.6KB, docx)
3. Supplementary Table S2.

shRNA sequences used in this study. Related to Figures 2 and 3

NIHMS627470-supplement-3.xlsx (585.3KB, xlsx)
4. Supplementary Table S3.

Transformation gene sets and BRD4-KD signature. Related to Figures 1, 4, 6 and 7.

Transformation signatures include upregulated (Transformation UP) or downregulated (Transformation DOWN) genes when expression profiles of TRS-WT and TERT-WT cells are compared (fold change > 2, p value <0.05). BRD4-KD signature includes upregulated genes when expression profiles of TRS-HGPS-shBRD4 and TRS-HGPS cells are compared (fold change > 2, FDR corrected p value <0.05)

NIHMS627470-supplement-4.xlsx (542.9KB, xlsx)
5. Supplementary Table S4.

Overlap between BRD4-sensitive genes in TRS-HGPS cells and TRS-HGPS BRD4 binding sites (ChIP-seq analysis). Related to Figures 4 and 6

Bold font indicates transformation-sensitive genes. Red font indicates genes upregulated in RAS oncogenic signatures (BILD HRAS SIGNATURE, KRAS LUNG BREAST UP, KRAS DF UP; see Fig S5). Underline indicates genes upregulated in the following oncogenic signatures: VEGF A UP, EGFR UP, HINATA NFKB MATRIX/IMMU INF, RAF UP (see Fig S5).

NIHMS627470-supplement-5.xlsx (23.4KB, xlsx)
6

Supplementary Table S5. Distribution and significance of BRD4-KD signature association across data sets. Related to Figure 7

First column shows the mean of BRD4-KD signature scores for each set (scores normalized for numbers of genes on chip and in signature). Score distributions were normal with Lilliefors-test values <1.0e-03. Second column shows variance of normalized scores for each set. Third column lists the percentage of samples in each set whose signature score was not significant (p<0.05) as measured by random permutation test (random gene sets of same size). Fourth column lists the median of the p-values of score association across the set (ie half of all samples had significant scores with p-values below this median). These numbers depend on the layout of the array: the fewer signature genes are represented on the microarray chip, the lower statistical significance of score assignments.

Supplementary Table S6. Distribution of BRD4 expression across data sets. Related to Figure 7

The table compares the mean values and variance of expression (logscale) of the probes for BRD4, with the average mean value and average variance taken over all probes on the chip (mean of means). For each data set, the top line referenced as “Average on chip” shows the mean of means of all probes (ie the “expected mean”), and the mean of variances of all probes on the chip (“expected variance”). The third column annotates the 90th percentile, i.e. the value for which 90 percent of probes on the chip have an equal or lower mean expression. The following lines show the means and variances of the probes specific for BRD4 on the chip. Notably, with the exception of two probes in the Kohno set, which had exceptionally low signal or expression, BRD4 expression was within comparable ranges in both breast and lung as well as lymphoma and AML sets.

Supplementary Table S7. Correlation between BRD4KD signature and Proliferation signature. Related to Figure 7

Values in red indicate positive and significant correlation, values in blue indicate negative and significant correlation

NIHMS627470-supplement-6.xlsx (98.1KB, xlsx)

RESOURCES