BRD4 and MYC: Power couple in Transcription and Disease

Abstract

The MYC proto-oncogene and BRD4, a BET family protein, are two cardinal proteins that have a broad influence in cell biology and disease. Both proteins are expressed ubiquitously in mammalian cells and play central roles in controlling growth, development, stress responses and metabolic function. As chromatin and transcriptional regulators, they play a critical role in regulating the expression of a burgeoning array of genes, maintaining chromatin architecture and genome stability. Consequently, impairment of their function or regulation leads to many diseases, with cancer being the most predominant. Interestingly, accumulating evidence indicates that regulation of the expression and functions of MYC are tightly intertwined with BRD4 at both transcriptional and post transcriptional levels. Here, we review the mechanisms by which MYC and BRD4 are regulated, their functions in governing various molecular mechanisms and the consequences of their dysregulation that lead to disease. We present a perspective of how the regulatory mechanisms for the two proteins could be entwined at multiple points in a BRD4-MYC nexus that leads to the modulation of their functions and disease upon dysregulation.

Graphical Abstract

MYC and BRD4 play central roles as chromatin and transcriptional regulators in controlling growth, development, stress responses, metabolism, and genome stability. This review discusses MYC and BRD4 regulation, their overlapping roles in various cellular processes, and how their functional or regulatory impairment leads to disease. Understanding this complex interplay between the two proteins can inform treatment modalities that will effectively counter their pathological effects while preserving normal cellular function.

Introduction

BRD4 and MYC have well-documented roles in controlling numerous cellular functions and are involved in a broad range of diseases, especially in various cancers. While both proteins are primarily labelled as transcription regulatory factors, they are known to play critical roles in chromatin regulation and post-translational regulation of many proteins involved in a multitude of biological processes. The functions and regulation of MYC have been studied for several decades. On the other hand, the functions of BRD4 are still being defined and the regulatory mechanisms controlling it are yet to be fully deciphered. Our understanding of the mechanisms regulating both proteins have continued to evolve and indicate a complex, multi-level web of regulatory steps consistent with the broad reach of these proteins in biology and disease. Amidst this complexity, an intriguing pattern of crosstalk between the regulatory mechanisms of these two important proteins is slowly emerging. In the following sections we present an overview of BRD4 and MYC individually and their regulation followed by a discussion of biological processes where there is an overlap of their regulation and function. We place a special emphasis on how dysregulation of BRD4 or MYC and the common pathways they regulate manifests in disease.

BRD4- a complex protein with intrinsic kinase and histone acetyltransferase activities

BRD4 is a bromodomain and extra terminal (BET) protein that has been described as a protein with many roles. It has been variously defined as a transcription cofactor, chromatin reader and regulator, recruiting and scaffolding factor, a DNA damage response factor, and a mitotic bookmark. Befitting its multifunctional role, BRD4 is involved in a significantly large number of biological processes that include transcription, DNA replication, cell cycle and signaling during stress responses. Accordingly, BRD4 dysfunction has been linked to a multitude of diseases including various types of cancers, heart, kidney, lung and inflammatory diseases, auto immune diseases, and neuro-degenerative disorders [1, 2]. While BRD4 is expressed ubiquitously in most tissues, BRD4 RNA and protein levels increase in response to cell proliferative signals and decline with growth arrest [3].

BRD4 exists as two isoforms- a long isoform comprising the complete protein and a short isoform with only the N-terminal half, which was originally named as HUNK1 [4]. Unless otherwise stated, the focus of the study will be on the long form of the BRD4. The N-terminus of BRD4 has clearly defined structural domains while the C-terminal domain consists of a long intrinsically disordered domain or IDR [5, 6] [Fig. 1A]. BRD4 contains two bromodomains (BD1 and BD2) in tandem and an extra terminal (ET) domain in its N-terminus. The bromodomains are conserved sequences that interact with acetylated lysine residues, typically in histones, but often in other proteins as well. The ET domain mediates BRD4 interactions with a wide range of cellular co-activators and viral proteins, engaging in binding-induced folding [7, 8]. Besides these domains, BRD4 also contains highly conserved regions such as the Motif A, Motif B, basic residue interaction domain (BID) and the Ser/Glu/Asp rich region (SEED). Importantly, the N-terminus of BRD4 contains a kinase domain that spans BD2-B-BID region [5]. Furthermore, unlike other BET proteins, BRD4 has a divergent C-terminal tail that encompasses a histone acetyl transferase (HAT) catalytic domain that confers HAT enzymatic activity to BRD4 [9]. We have recently reported that the full length BRD4 dimerizes through its B motif [5], a finding later confirmed by another group using partial BRD4 fragments encompassing the B motif [10]. The dimerization of BRD4, in combination with its extended conformation, structured N-terminus and its disordered C-terminus confers a high degree of flexibility. Such structural flexibility in BRD4 helps create a scaffold for the recruitment and interaction of multiple binding factors. The unique structural features of BRD4 influences its functions. Motif A is important for its binding with the C-terminal domain of the largest subunit of RNA Pol II CTD (RNAPII CTD). Interestingly, the C-terminal tail of BRD4 binds MYC, which inhibits BRD4 HAT activity [9, 11].

Figure 1: Schematic models of BRD4 and MYC.

Organization of functional domains in BRD4 and MYC. (A) Schematic showing the known functional domains in the BRD4 long isoform (BRD4L) and BRD4 short isoform (BRD4S). All domains notated on BRD4S exist, but some have not been functionally characterized (B) Schematic showing the functional domains of MYC. Phosphorylation sites regulating MYC stability are indicated in red letters. Numbers on the schematics represents amino acids.

BRD4 was originally reported as a passive scaffolding factor that recruited several key chromatin and transcriptional regulatory factors. The discovery of two critical enzymatic functions intrinsic to BRD4, namely kinase and HAT activity, redefined it as a protein that dynamically regulates its interacting partners. BRD4’s enzymatic activities set it apart from other BET proteins. Through its kinase activity, BRD4 phosphorylates many of its interacting partners like transcription factors MYC and TAF7, transcription elongation factor PTEFb and most importantly, the RNAPII CTD [12] thereby directly regulating transcription. Through its HAT activity, BRD4 acetylates nucleosomes and regulates chromatin architecture [9]. Thus, through its unique binding domains and distinct enzymatic activities BRD4 plays critical roles in directly regulating local chromatin architecture and the transcription machinery.

Besides regulating transcription at promoters and the gene body, BRD4 also functions as a critical regulatory factor at active enhancers and super-enhancers (SEs) genome-wide and has been shown to directly control the Myc SE [13–15]. The finding that BRD4 could regulate Myc transcript levels led to tremendous interest in BRD4 as a conduit to regulate MYC, a critical oncogene that is notoriously challenging to control. Regulation of Myc levels by BRD4 was facilitated by the discovery of small-molecule BET inhibitors such as JQ1 [16] and I-BET [17] which mimic the acetyl moiety, specifically bind BET proteins to occlude their acetyl lysine-binding pocket and potently inhibit growth of malignant cells by reducing expression of oncogenes such as Myc [18].

MYC- a global regulator under multi-layered control

The MYC oncoprotein functions as a master transcription factor that is estimated to control the expression of a significant proportion of expressed genes in the human genome [19]. While there are three cellular Myc gene paralogs – c-Myc, N-Myc, and L-Myc, the focus of this review will be on the most predominantly expressed and well-studied paralog c-Myc, herein referred to as Myc. MYC forms the crux of numerous growth-promoting signal transduction pathways. Consistent with its transcriptional regulation of a wide range of genes, MYC plays a direct or indirect role in a broad variety of cellular functions including cell cycle, survival, differentiation, metabolism, cell fate decisions, ribosome biogenesis and translation. It also plays a pivotal role in numerous physiological responses such as programmed cell death, stress response, and immune regulation [20]. The role of MYC and its dysregulation in cancer has been extensively reviewed elsewhere [21, 22]. However, it is underappreciated that MYC is a multifunctional protein which also affects nuclear organization as well as the stability of the whole genome [23] and could therefore play a role in many diseases.

MYC is a 439aa protein with well-defined domains that are critical for its function and stability [Fig. 1B]. It has a large unstructured N-terminal region containing conserved regions known as MYC boxes (MBI, II, III, and IV). MBI and MBII overlap with a transactivation domain (TAD), a 143-amino-acid acidic domain, that is responsible for transcriptional activation by MYC along with MBIV [24, 25]; while MBIII is required for the transcriptional repression by MYC [26]. A basic helix–loop–helix (B-HLH) domain at the C-terminal end allows MYC to dimerize with its cofactor MAX and other proteins [27] which allows it to bind to E-box DNA sequences (5’-CACGTG-3’) through its leucine zipper (LZ) domain [28]. E-boxes are frequently found in active promoter-proximal regions and required for transcription activation at these sites [29, 30]. The chromatin-modifying complex consisting of TIP60, TRRAP, TIP48, and GCN5 recruited by the DNA-bound MYC/MAX heterodimer propels transcription further by modifying chromatin into a more accessible form [20]. MYC can also regulate transcription initiation and elongation through its interaction with cofactors like BRD4, PTEFb, PAF1, Mediator and the RNA polymerase II complex [11, 31–33].

Befitting its role in numerous biological functions, Myc expression is tightly regulated at multiple steps, from transcription initiation to post-translational modifications. The Myc locus can be regulated at the DNA level through alternate non-B DNA structures [34]. Its transcription is modulated by at least four promoters- P0, P1, P2 and P3, multiple initiation sites, two different polyadenylation sites and several antisense transcription products [35, 36]. Besides, it is also regulated through transcription elongation control [37]. Numerous transcription factors and chromatin regulators responding to many different signals have been demonstrated to regulate Myc expression [38]. A series of enhancers both upstream and downstream of Myc play an important role in precisely regulating Myc expression. These include several tissue-specific enhancers that respond to various stimuli and super-enhancers [39].

MYC levels are also regulated through its mRNA stability [40, 41]. Myc mRNA stabilization has been linked to oncogenic activity in lymphoma cells [42]. Increased Myc mRNA stability in tumorigenic cells is caused by the binding of the Coding Region Determinant-Binding Protein (CRD-BP) to Myc mRNA destabilizing elements [43]. Its stability can also be modulated by various microRNAs such as let-7, miR-34, and miR-145 [44]. Notably, let-7 acts as a strong tumor suppressor which reduces cancer aggressiveness, chemoresistance, and radio-resistance through its repression of Myc levels [45]. miR-34 represses Myc levels during oncogene-induced senescence, while miR-145 functions as a tumor repressor by suppressing Myc levels in colon and breast cancers [46, 47]. Several lncRNAs are also known to regulate Myc expression. Some lnc RNAs simply act as sponges or sinks for MYC-regulating miRNAs while others can directly or indirectly regulate Myc transcription, mRNA stability, translation, and activity [44].

MYC is considered a relatively weak transcriptional modulator whose effect is amplified through its interaction with a broad range of transcription factors, co-activators and recruitment of chromatin modifiers [48]. One study suggested that increasing MYC levels have very little effect on increasing the number of genes bound by MYC. Instead, there is an increase in levels of MYC at MYC target promoters and occupancy of enhancers of actively transcribed genes, thus causing transcriptional amplification [49]. Other studies have shown that rather than regulate a specific set of target genes, increasing MYC levels can annex most active elements in the genome and mediate their transcription, leading to MYC being defined as a global “transcriptional amplifier” [50, 51].

MYC binding is predominantly associated with transcription activation, but genome-wide studies have demonstrated that MYC represses almost as many target genes as it activates. While the mechanisms involved in transcriptional repression by MYC are yet to be fully elucidated, MYC interaction with the transcription factor MIZ-1 is thought to create a DNA-bound complex that displaces coactivators and recruits co-repressors [52, 53]. More recently, it was also shown that MYC interaction with PAF1C forms a repressive complex preventing PAF1C from functioning as an elongation factor. MYC degradation and turnover are required to release PAF1C to generate productive elongation [31]. MYC-mediated gene repression has also been linked to its ability to activate microRNAs [54].

The broad range of MYC functions makes its dysregulation responsible for a wide range of pathologies, the most well studied being cancer. In contrast to many other oncogenes such as Src and Ras where mutations within the gene result in tumorigenesis, the oncogenic effects of MYC are primarily due to its dysregulated expression levels. MYC is one of the most highly amplified oncogenes across numerous human cancers, and deregulation of MYC levels has been shown to contribute to more than half of human cancers [55, 56]. Gene amplifications, chromosomal translocations and retroviral insertions are the three most widely reported causes for dysregulated Myc gene expression [57]. Importantly, even small increases in MYC abundance (<2 fold) are sufficient to drive cell proliferation. Since aberrant proliferation is the hallmark of all cancers, MYC has been predominantly associated with the initiation and maintenance of tumorigenesis.

Regulation of MYC and BRD4 structure and function through post-translational modifications

Consistent with its ubiquitous presence throughout the cell cycle, the expression level of BRD4 is stable (add citation here). Several post translational modifications (PTMs) have been shown to play critical roles in BRD4 function and stability. BRD4 PTMs include phosphorylation, ubiquitination, acetylation, polyribosylation, and proline isomerization and hydroxylation. Phosphorylation of BRD4 occurs in two potential clusters of sites broadly designated as NPS (N-terminal cluster of phosphorylation sites) and CPS (C-terminal cluster of phosphorylation sites) (Fig. 1). Casein kinase 2 (CK2) phosphorylates all seven potential sites in the NPS and mediates a phospho-regulatory mechanism leading to BRD4 association with p53 and acetylated chromatin [58]. CK2-mediated phosphorylation of BRD4 has been shown to facilitate BRD4 dimerization [10]. As CK2 activity is elevated in most cancer cells, the CK2-mediated phospho-switch mechanism regulating BRD4 function could be responsible for dysregulating BRD4 in these cancers. Interestingly, aberrant hyperphosphorylation of the BRD4-NUT fusion protein in NUT midline carcinoma (NMC) enhances oncogenic potential of the BRD4-NUT fusion protein. CDK9 has been reported to act as a potent kinase of the BRD4-NUT fusion protein leading to its hyperphosphorylation [59]. Since BRD4 can activate CDK9 kinase activity [60], this raises the question as to whether it potentiates cancer causation in NMCs. Further extending the notion that phosphorylation contributes to the oncogenic potential of BRD4, a recent study showed that in colorectal cancer cells, interleukin 6/8-mediated JAK2 signaling induces BRD4 BD1 phosphorylation at tyrosine 97/98. This phosphorylation results in BRD4 stabilization through its interaction with the deubiquitinase UCHL3, resulting in increased binding to chromatin, resistance to BET inhibitors, and supporting a tumor-promoting transcriptional program through interaction with Signal transducer and transcription activator (STAT3) [61]. Apart from JAK2 signaling, the c-jun-N-terminal kinases (JNK) pathway has also been suggested to affect BRD4 association with mitotic chromatin [62]. However, whether JNK kinases directly contribute to BRD4 phospho-regulation has not been reported yet.

BRD4 stability is positively regulated by peptidyl-prolyl isomerase Pin1, a phosphorylation-directed proline isomerase which binds to BRD4 phosphorylated at Thr204. BRD4 is isomerized at the adjacent Pro205 residue by Pin1 which facilitates BRD4-CDK9 interaction and transcriptional activity of BRD4 in gastric cancer cells [63]. The kinase that phosphorylates Thr204 remains to be identified. It is possible that a positive feedback loop exists, where CDK9 phosphorylates Thr204 on BRD4 leading to BRD4 stabilization and isomerization by Pin1 to promote CDK9 recruitment by BRD4. BRD4 protein levels are also controlled by a polyubiquitination/deubiquitination switch called the SPOP-DUB3 switch. BRD4 is ubiquitinated by E3 ubiquitin ligase SPOP leading to its proteasomal degradation in a process believed to be critical to maintain the homeostatic levels of BRD4. SPOP mutations, which are frequent in prostate malignancies, fail to interact with BRD4 and prevent BRD4 degradation thereby leading to resistance to BRD4 inhibition.[64]. On the other hand, deubiquitinase DUB3 can directly bind BRD4 and deubiquitinate it, leading to BRD4 stabilization and elevated levels of BRD4. Since elevated levels of BRD4 confer resistance to BET inhibitors (references), DUB3 mediated BRD4 stabilization results in resistance to BRD4 inhibition in cancer [65].

In acute myeloid leukemia cells, BRD4 transcriptional activity and cell proliferation have been linked to proline hydroxylation at position 536 by the prolyl hydroxylase domain protein PHD2. [66]. Although BRD4 Hyp-P536 is not important for its stability, it affects BRD4 interaction with pTEFb components CDK9 and Cyclin T1, CDK1, and the DNA replication licensing factor MCM5 [67]. BRD4 is also poly ADP-ribosylated at its C-terminus by Poly(ADP-ribose)polymerase-1 (PARP1) [68]. Poly-ADP ribosylation (PARylation) is a modification that can occur on the Asp(D), Glu(E) and Lys(K) residues. In cardiac hypertrophy, BRD4 PARylation allows it to bind to the TSS of hypertrophic genes, resulting in enhanced RNAPII phosphorylation and hypertrophic gene activation. [68]. It remains to be seen if this modification has a broader regulatory role in other diseases where BRD4 plays a pathogenic role. Evidence of BRD4 acetylation was also recently discovered. p300/CBP-associated factor (PCAF) mediates acetylation of BRD4 at Lys332 promoting its association with the transcription factor intestine-specific homeobox (ISX). This leads to the expression of genes critical for epithelial–mesenchymal transition and cancer metastasis [69].

Like BRD4, MYC protein is subject to numerous PTMs such as phosphorylation, acetylation, methylation, ubiquitination, sumoylation and glycosylation. These PTMs control both the stability and activity of MYC and mutations in the sites of these modifications can lead to dysregulated MYC levels and function. Consistent with the tight regulation of MYC levels, the MYC protein is rapidly degraded following its synthesis with a half-life of about 15–20 min [70]. In many tumors, stabilization of MYC protein contributes to its deregulation [71, 72].

MYC levels are controlled through a complex phosphorylation and cis-trans isomerization process that leads to its degradation by the ubiquitin–proteasome system [73] [Fig. 2]. MYC contains two phosphorylation sites within its TAD domain, Threonine 58 (Thr-58), and Serine 62 (Ser-62), that are highly conserved across all mammalian MYC isoforms [74]. Phosphorylation of the Ser-62 site by the ERK1 and CDK2 kinases leads to stabilization of MYC while Thr-58 phosphorylation by the GSK-3β or BRD4 kinases results in degradation of MYC through the ubiquitin–proteasome pathway [72]. A complex signaling pathway regulates these phosphorylations [75]. Upon cell growth stimulation, MYC is stabilized by Ser-62 phosphorylation and becomes functionally activated. Ser-62 phosphorylation is also required to prime subsequent phosphorylation at Thr-58 by the GSK-3β kinase [76]. Although ERK1 and GSK-3β are the prominent MYC regulatory kinases, they are dependent on signaling pathways. Cell/mitogenic stimulatory signals activate the Ras/Raf/ERK1 kinase cascade to stabilize MYC. Ras also induces activation of the PI3K and Akt kinases. ERK1, PI3K and Akt regulate the activity of GSK-3β kinase which when phosphorylated, is inactivated and exported into the cytoplasm, preventing degradation of MYC by GSK-3β. A decrease in the levels of Ras allows GSK-3β to reenter the nucleus, phosphorylate MYC and trigger its degradation [77].

Figure 2: Overview of MYC-BRD4 cross-regulation.

BRD4 regulates transcription of Myc gene though its binding at both Myc promoter and enhancers resulting in increased expression of Myc. BRD4’s intrinsic HAT activity drives Myc transcription by facilitating chromatin decompaction primarily through H3K122 acetylation at the dyad axis of nucleosomes, while BRD4 kinase activity regulates RNA Pol II and PTEFb. Mature Myc transcript is regulated by the CRD-BP protein and several microRNAs including miRNA let-7, miR-34 and miR-145. On the other hand, MYC protein inhibits BRD4 HAT activity, limiting chromatin remodeling and its own transcription. BRD4 also maintains homeostatic levels of MYC protein. MYC protein stability is regulated by two different mechanisms depending on the physiological state of cell. Under homeostatic conditions, BRD4 dimers phosphorylate MYC at Thr58, leading to its degradation. During cell stimulatory signals ERK1 and GSK3β regulate the levels of MYC. While ERK1 phosphorylation of MYC Ser62 leads to its stabilization, MYC Thr58 phosphorylation by GSK3β directs it for degradation. Pin1 also regulates MYC levels by either stabilizing it through trans to cis isomerization or leading to its degradation by cis-trans isomerization that promotes the dephosphorylation of Ser62 by PP2A phosphatase leading to MYC degradation. In addition to Thr58, phosphorylation of Thr244 and Thr248 by a currently unknown kinase also leads to MYC degradation. ‘Ac’ represents acetylation of histones and ‘p’ represents phosphorylation of Pol II and PTEFb.

Like BRD4, Pin1 also influences MYC degradation. Pin1 interacts with pSer-62 MYC to isomerize proline 63 from trans to cis conformation to activate MYC DNA binding and transcriptional activity. Subsequently, once MYC has performed its function and is marked for degradation through phosphorylation at Thr-58, Pin1 reverses proline 63 isomerization from cis to trans to facilitate Protein Phosphatase 2A (PP2A)-mediated dephosphorylation of Ser-62, leaving only the pT58 mark that is then recognized by the SCF^Fbw7 E3 ubiquitin ligase and degraded by the 26S proteasome [75, 78, 79]. Thus, the balance between Ser-62 and Thr-58 phosphorylations plays a critical role in regulating MYC expression during cell proliferation. Mutations of MYC Thr-58 and Ser-62, prevalently found in Burkitt lymphoma, are associated with stabilized and sustained levels of MYC leading to tumorigenesis [80]. Additionally, human B-cell lymphomas harbor point mutations in the MYC binding site of the SCF^FBW7 ubiquitin ligase, resulting in MYC stabilization in these cancers [81]. Supporting the importance of increased MYC stability in oncogenesis, point mutations in SCF^Fbw7 are also linked to lymphoid and myeloid leukemias [82, 83]. Interestingly, a recent study reported the presence of a second phospho-degron composed of Thr-244 and Thr-248 which synergizes with the Thr-58/Ser-62 phospho-degron for binding to SCF^FBW7 ubiquitin ligase dimers [Fig. 2]. In contrast to previous studies, this study also presented evidence that the pSer-62 mark does not prevent SCF^FBW7 from binding MYC rather may improve their interaction [84]. While these latest findings remain to be further validated, particularly with respect to the roles of Pin1 and PP2A, they indicate that the MYC degradation pathway is more complex than the currently accepted models suggest. In summary, both BRD4 and MYC are regulated by many similar post-translational modifications and BRD4 is an important driver of pThr-58, a PTM critical for MYC stability.

Cross–regulation between BRD4 and MYC

In addition to the multiple PTMs discussed above that modify BRD4 and MYC functions, their functions are also influenced by each other through intricate mechanisms. For two decades, the Ras/ERK1/GSK-3β MYC regulatory pathway was the only known mechanism to trigger ubiquitin-mediated MYC degradation. However, our recent discovery that BRD4 also phosphorylates MYC at Thr-58 and leads to its ubiquitination and degradation helps broaden our understanding of the mechanisms involved in MYC degradation [11]. BRD4 phosphorylation of MYC Thr-58 addresses critical gaps in the Ras/ERK1/GSK-3β MYC degradation pathway model while also raising new questions. GSK-3β is localized almost entirely in an inactive form in the cytoplasm, which implies that an alternative pathway controls the degradation of MYC when cell stimulatory signals are abated. In this context, BRD4, being a nuclear kinase, can phosphorylate MYC Thr-58 and maintain homeostatic MYC levels independent of any signaling pathway. However, MYC can respond to dynamic stimulatory signals even in the presence of BRD4 as ERK1 strongly inhibits BRD4 kinase activity [11]. Moreover, GSK-3β can only phosphorylate MYC that is already phosphorylated at Ser-62 by ERK1. BRD4 phosphorylation of MYC overcomes this requirement, as unlike GSK-3β, BRD4 does not require a priming pSer-62 mark on MYC to phosphorylate Thr-58. Notably, phosphorylation of MYC Thr-58 triggers its isomerization by Pin1 and subsequent degradation. While MYC stability is thus affected by Pin1, BRD4 is also known to be isomerized and stabilized by Pin1 [63]. Thus, Pin1 could also modulate MYC levels through its stabilization of BRD4. Besides, BRD4 phosphorylation by CK2 facilitates BRD4 dimerization [85] which in turn is essential for MYC Thr 58 phosphorylation in vivo but not in vitro. This suggests a critical structural requirement for BRD4-MYC complex formation in vivo. The fact that MYC binds to the unstructured C-terminal region of BRD4, unlike other BRD4 kinase substrates such as RNAPII CTD and TAF7, further strengthens the notion of a complex cross regulation of structure and function of both BRD4 and MYC. Thus, multiple factors influence BRD4’s ability to modulate MYC stability and could impact MYC levels leading to cancer and other diseases.

Besides regulating MYC protein levels, BRD4 also plays a major role in regulating Myc transcription. BRD4 acts as a mitotic bookmark and plays a critical role in the reactivation of Myc transcription at the end of mitosis [15, 18, 86, 87]. BRD4 binds the Myc promoter and is involved in its transcriptional upregulation [18]. Interestingly, BRD4 kinase activity has little or no role in regulating Myc transcript levels independent of HAT activity. Overexpression of BRD4 with intact kinase activity, but lacking HAT activity, does not alter Myc transcript levels. BRD4 affects Myc transcript levels through its HAT activity by increasing accessibility of the Myc promoter [9] suggesting that BRD4 regulation of Myc transcription is almost entirely through its HAT activity. In this context, it is notable that dysregulated Myc expression is largely a result of large clusters of enhancers called super-enhancers (SE) that are occupied by BRD4. BET inhibition specifically disrupts expression of Myc and many other genes associated with these SEs [88]. Thus, by activating Myc transcription on one hand through its HAT activity while degrading MYC through its kinase activity on the other, BRD4 maintains homeostatic MYC levels [Fig. 2].

While BRD4 regulates MYC stability, MYC can also modulate BRD4 function by inhibiting BRD4 HAT activity and thus its ability to remodel chromatin [11]. However, MYC phosphorylated by either ERK1 or BRD4 is unable to inhibit BRD4 HAT activity, suggesting a possible inbuilt mechanism to modulate MYC regulation of BRD4 function. Interestingly, chromatin-bound BRD4 is unable to phosphorylate MYC [11]. This is explained by our recent findings that the BRD4 kinase activity spans its BD2-B-BID domain [5]. The overlap of the BRD4 bromodomain 2 (BD2) with its kinase domain probably prevents it from phosphorylating MYC when bound to chromatin. In fact, MYC phosphorylation by ERK1 could be important in regulating BRD4 HAT activity on the chromatin. In support of this, we have demonstrated that BRD4, MYC and ERK1 exist in a trimeric complex in both chromatin-bound and non-chromatin bound fractions of the nucleus. In addition, ERK1 inhibits BRD4 kinase activity by directly binding to its kinase domain. Thus, ERK1 prevents MYC degradation by BRD4, imparts a stabilizing phosphorylation on MYC and controls MYC’s ability to regulate BRD4 HAT activity [11]. These interactions and phosphorylations between BRD4, MYC and ERK1 create an additional level of complexity to BRD4-MYC dynamics and the crosstalk between them.

Role of BRD4 and MYC in transcriptional and co-transcriptional regulation

After transcription initiation, Pol II traverses about 20–50 bp from the TSS and then pauses as a regulatory step due to the binding of 2 factors: the negative elongation factor (NELF) and DRB sensitivity–inducing factor (DSIF). PTEFb-mediated phosphorylation of these factors as well as Pol II CTD Ser2 is required for pause release (reviewed in [89]). BRD4 recruits PTEFb from the HEXIM1/7SKsnRNA repressive complex to active transcription sites [90] and regulates PTEF-b complex and thus influences pause release [Figs. 3, 4]. While the broadly cited role of BRD4 in transcriptional regulation is its recruitment of PTEFb, BRD4 can also regulate transcription initiation and elongation independent of PTEFb recruitment arguing for a PTEFb-independent role for BRD4 in transcription [91, 92]. Consistently, BRD4 phosphorylates the RNAPII CTD Ser2 through its BD2-B-BID kinase domain while binding to RNAPII CTD through motif A [5]. BRD4 also regulates the kinase activity of PTEFb (CDK9) by phosphorylating it at two sites - phosphorylation at Thr186 activates CDK9 kinase activity while phosphorylation at Thr29 inhibits it [60, 93]. Furthermore, BRD4 has been reported to play a central role in the crosstalk between the primary RNAPII CTD kinases CDK9, CDK7 (TFIIH) and BRD4, through its phosphorylation of TAF7, a general transcription factor that has been demonstrated to regulate the activities of all three kinases. The phosphorylation of TAF7 by BRD4 affects its interaction with each CTD kinase differentially in a manner that could modulate their activity in a coordinated manner [60].

Figure 3: Crosstalk between MYC and BRD4 during transcription.

BRD4 binds and acetylates nucleosomes both at promoters and SEs while recruiting the mediator complex that promotes the link between SE and Promoter. BRD4 and MYC both interact with PTEFb promoting RNAPII CTD phosphorylation. While BRD4 activates TOP I activity through its RNAPII CTD phosphorylation, MYC directly promotes TOP I/II activation. Together, this leads to pause release and productive elongation. During elongation, BRD4 and MYC both regulate alternative splicing by directly interacting to splicing factors. In addition, MYC also regulates the expression of several splicing factors. BRD4 plays additional roles in elongation, 3’end processing and termination by directly interacting with elongation, 3’end processing and termination factors. Green and yellow circles represent RNAPII CTD Ser5 and Ser2 phosphorylations respectively. Grey, orange, and blue circles represent distinct activators. CTD; carboxy terminal domain of RNA polymerase II.

Figure 4: BRD4 and MYC act as master regulators of nuclear processes.

BRD4 and MYC directly or indirectly regulate several nuclear events. These include chromatin remodeling, super-enhancer function and transcription. They are also involved in formation of condensate and ecDNA hubs and regulate genomic stability.

Another way in which BRD4 promotes Pol II pause release is by recruiting JMJD6 to distal enhancers designated anti-pause enhancers [Fig. 5] [94]. JMJD6 is an arginine demethylase that removes the methyl cap of the 7SK snRNA resulting in the disassembly of the HEXIM1/7SK snRNA PTEFb inhibitory complex while also demethylating H4R3Me2- a repressive methyl mark that binds the 7SK snRNA. BRD4-mediated JMJD6 recruitment at enhancers results in removal of the H4R3me2 mark that is responsible for recruiting 7SK snRNA while also directly decapping and demethylating the 7SK snRNA at target genes. This removes the PTEFb inhibitory complex while both BRD4 and JMJD6 interact with the active form of PTEFb complex and facilitate productive transcription elongation. Furthermore, BRD4 regulates elongation by activating Topoisomerase I (Top I) through its phosphorylation of RNA Pol II CTD [Figs. 2–5]. Activated Top I facilitates the relaxation of DNA supercoils created by the movement of the RNA Pol II, allowing passage of paused RNA pol II [95]. Consistent with BRD4’s role in pause release and elongation, JQ1 treatment leads to elongation defects in transcription of Myc and other genes preferentially associated with super-enhancers [88]. Thus, BRD4 loss affects pause-release and transcription elongation at the Myc oncogene providing yet another way for BRD4 to control Myc transcription and oncogenesis.

Figure 5: Intricate network of BRD4 and MYC regulated processes leads to Myc dysregulation and oncogenesis.

Chart showing multiple interactions and effects of BRD4 and MYC leading to oncogenesis. Thick arrows represent BRD4 enzymatic activities while double-sided arrows represent interactions. Green boxes represent major cellular effects of BRD4 and MYC activity resulting in increases in Myc transcription, BRD4/MYC translation, or decreased MYC degradation (red boxes) ultimately promoting oncogenesis.

MYC also contributes to these processes by enhancing the translation of TFIIH and PTEFb by increasing their mRNA CAP methylation [Fig. 5] [96]. This results in increasing global RNA pol II Ser5 and Ser 2 phosphorylation. Besides, MYC also plays a role in recruiting TFIIH [96] and PTEFb [97] to the TSS of target genes [Fig. 5]. In fact, MYC-mediated PTEFb recruitment at MYC target genes is required for transactivation of these genes [Fig. 3] [98]. In embryonic stem cells, MYC regulates transcription of its target genes primarily at the level of pause release [99]. MYC also affects pause release in cancer cells [49]. Similar to BRD4, MYC affects DNA supercoiling by directly binding to topoisomerases TOP1 and TOP2 and enhancing their activity for the elevated transcription of MYC-driven genes [Fig. 3, 5] [100]. Additionally, MYC turnover is necessary for transcriptional pause release [31]. Therefore, it is plausible that BRD4 -mediated degradation of MYC may play a role in pause release at MYC-regulated genes. As BRD4 directly interacts with MYC and is involved in upregulation of MYC target gene, BRD4 may function with MYC to properly modulate RNA pol II initiation, topoisomerase function, pause-release, and elongation at its target genes.

BRD4 also plays a role in co-transcriptional processes. We have recently shown that BRD4 regulates alternative splicing during different stages in thymocyte development and in T-cell acute lymphoblastic leukemia (T-ALL) cells. BRD4 travels with the elongation complex and modulates exon usage by directly interacting with splicing factors especially the core U1 spliceosome component U1–70 and regulatory splicing factors such as FUS and HNRNPM [Fig. 3] [101]. BRD4 can also negatively regulate intron retention upon heat stress. BRD4 is recruited to nuclear speckle bodies by the heat shock factor 1 (HSF1) and enhances SatIII RNA transcription, thereby influencing heat-induced splicing regulation [102]. Additionally, BRD4 interacts with several 3’ processing and termination factors such as the subunits of cleavage and polyadenylation specificity factor (CPSF) and cleavage stimulation factor (CstF) [Fig. 3]. The impaired recruitment of these factors upon BRD4 depletion has been suggested to be responsible for 3’ processing and termination defects [103].

MYC also regulates RNA splicing of several transcripts primarily related to spliceosome and alternative pre-mRNA splicing machinery. MYC modulates exon choice in splicing regulatory proteins where increased MYC levels result in exclusion of exons with premature stop codons, avoiding nonsense-mediated decay of the RNA [104]. Such exon-skipping results in elevated protein levels leading to aberrant splicing in MYC-overexpressing cells. Dysregulated MYC expression has been shown to affect splicing in several cancers [105, 106]. Overall, MYC and BRD4 can induce cell proliferation and oncogenesis by regulating multiple steps in transcription as well as in co-transcriptional processes.

Role of BRD4 and MYC in chromatin regulation

a). BRD4 is a histone acetyltransferase

BRD4 is recruited to chromatin by binding to acetylated histones through its bromodomains [Fig. 5] [86]. It preferentially associates with euchromatic and transcriptionally active regions and is excluded from transcriptionally inactive heterochromatic regions like centromeres [3, 86]. Work from our lab showed that BRD4 has intrinsic histone acetyl transferase (HAT) activity [9]. BRD4 is a true histone acetyltransferase in that it acetylates only histones, unlike lysine acetyltransferases like p300 that acetylate a variety of other proteins and are broadly referred to as KATs. Importantly, BRD4 acetylates H3 at K122, a site which maps to the dyad axis of the nucleosome where histone binding to DNA is the strongest [107–109], leading to the dissociation of the targeted nucleosome [Fig. 2]. Acetylation of H3 at K122 is enriched in transcriptionally active chromatin, is sufficient for nucleosome clearance and transcription activation, and co-occurs with marks associated with transcription start sites and distal enhancers. BRD4 also acetylates the N-terminal tails of H3 at H3K4/9/18/27 and H4 at H4K5/8/12/16 [9] but the role of these tail residue acetylations by BRD4 remains to be fully understood. The BRD4 C-terminal domain that encompasses its HAT domain is required for maintenance of chromatin structure and the loss of this region results in chromatin compaction and fragmentation [110]. In addition, H3K122Ac is excluded from condensed chromatin [109]. Interestingly, the short isoform of BRD4, which does not have the HAT domain, insulates chromatin from DNA damage by facilitating chromatin compaction through recruitment of the condensin II chromatin remodeling complex [111]. Thus, the presence of BRD4 HAT activity positively correlates with active chromatin architecture and transcription [Fig. 5] while its absence correlates with chromatin compaction and reduced access to the transcription machinery. Significantly, BET inhibitors led to decreased levels of H3 and H4 acetylation [17].

b). BRD4 associates with chromatin modifiers and remodelers

BRD4 can also modify the chromatin environment by its interaction with other chromatin modifiers like arginine demethylase JMJD6, lysine methyltransferase Nsd3, acetyltransferases p300/CBP [112], and ATP-dependent chromatin remodellers like SWI/SNF, CHD8 and CHD4 [94, 113, 114]. NSD3 and JMJD6 are recruited to BRD4 target genes in a BRD4-dependent fashion and serve to upregulate transcription of these genes [113]. Knockdown of either NSD3 or BRD4 reduces H3K36Me3 levels at these genes. This histone mark is associated with active transcription and productive transcription elongation. Interestingly, a short isoform of NSD3 (NSD3-short) that lacks the methyltransferase domain (but retains binding to BRD4, CHD8 and H3K36-methylated peptides) mediates AML proliferation. NSD3-short is recruited to chromatin by BRD4 and functions in the pathogenesis of AML by recruiting chromatin remodeler CHD8 [114]. Loss of either NSD3 or CHD8 decreases expression of MYC mRNA and protein and leads to terminal myeloid differentiation of leukemic cells. Consistent with this, BRD4, NSD3 and CHD8 co-occupy Myc SE as well as SEs of other oncogenic drivers. Treatment with JQ1 results in loss of all 3 proteins from the Myc SE in a variety of leukemic cell lines. BRD4 interactor JMJD6, an arginine demethylase, removes the repressive H4R3Me2 mark in distal enhancers, promotes pause release and upregulates transcription in a BRD4-dependent manner [94].

BRD4 binds P300/CBP through its bromodomains resulting in enhanced H3K27Ac at ESC enhancers [112]. BET inhibition decreases H3K27Ac while reducing recruitment of chromatin remodeler Brg1 as well. BRD4 can also bind p300/CBP through oncogenic fusion proteins like BRD4-NUT that occur in NUT midline carcinomas [115]. In these fusions, the N-terminal of the BRD4 protein (with the bromo- and ET domains but lacking the C-terminal domain along with the HAT activity) is fused with the NUT protein. This fusion oncoprotein binds chromatin through BRD4 bromodomains and recruit p300 through NUT. Binding of the BRD4-NUT fusion protein leads to large regions of hyperacetylation and activation of stemness-promoting transcription factors like MYC, SOX2 and TP63. MYC is overexpressed in NUT midline carcinomas where its expression is dependent on the fusion protein and is required to keep the cells in a dedifferentiated state [116]. The effects of the BRD4-NUT fusion may be dependent on its recruitment of the full-length BRD4 [115] and hence may require its HAT activity. This is supported by the observations that full-length BRD4 is found associated with the fusion protein.

BRD4 and SWI/SNF bind the Myc promoter and its downstream SE to upregulate Myc transcription. Myc expression is dependent on Brg1, an ATPase subunit of chromatin remodeller SWI/SNF in different types of leukemic cells but not in mESCs, melanoma or breast cancer cells suggesting a context-specific role for Brg1 [117]. While BRD4 can bind SWI/SNF, in this study, they did not depend on each other for recruitment to chromatin. Instead, Brg1 was shown to be required for looping interactions between Myc promoter and SE, and for transcription factor association with the SE. Therefore, it is plausible that BRD4 and SWI/SNF collaborate to create a conducive chromatin environment to facilitate Myc overexpression in leukemias. Interestingly, BRD4 interacts with CHD4, a component of the NuRD chromatin remodeling complex, which has also been shown to interact with SWI/SNF. CHD4 localizes to SEs and regulates SE accessibility to promote tumorigenesis [118]. It is required for leukemic cell growth and tumorigenesis but not for growth of normal hematopoetic cells. Critically, CHD4 loss leads to repression of Myc expression [119], suggesting a role for CHD4 in regulating the Myc SE in leukemic cells in conjunction with BRD4.

c). MYC influences chromatin by associating with chromatin modifiers and remodelers

Like BRD4, MYC preferentially associates with genes having positive histone marks [Fig. 5] [120]. MYC further influences the chromatin environment through its binding partners [Fig. 5]. The MYC transactivation domain (TAD), binds several histone acetyltransferase (HAT) complexes like TRRAP-GCN5, GCN5/PCAF, Tip60, and P300/CBP [121–125]. MYC can also transcriptionally upregulate GCN5 gene expression [126] further reinforcing positive histone marks. MYC recruitment at target genes correlates with increased histone acetylation and gene activation [126–128]. Consistent with these observations, MYC inactivation led to global reduction in H4 acetylation and H3K4Me3, and an increase of H3K9Me3 in an osteosarcoma cell line, and a global reduction in total histone H4 acetylation in transplanted hepatocellular carcinomas [129]. Additionally, MYC loss was accompanied by a decrease in nuclear volume and histone acetylation, increased H3K9Me2, heterochromatinization and decreased nuclease sensitivity [126]. MYC-driven prostate cancers show reduced global levels of H3K27Me3 [130] where H3K27Me3 levels inversely correlated with disease aggressiveness. Reducing MYC levels increased H3K27Me3 levels in multiple cancer cell lines including prostate cancer and B-cell Burkitt’s lymphoma cell lines. Thus, MYC indirectly plays an important role in regulating chromatin structure.

Both, BRD4 and MYC, are recruited to regions rich in acetylated residues- BRD4 is recruited by acetylated lysines through its bromodomains while MYC is recruited through direct DNA binding. Both proteins can also be recruited by their interaction with other proteins including with each other. Thus, BRD4 and MYC could reinforce each other’s recruitment to open chromatin regions. BRD4 increases acetylation levels through its own HAT activity while MYC does the same through its association with HATs including BRD4 [Fig. 4]. Thus, together they provide a chromatin environment conducive to increased transcription of MYC target genes resulting in oncogenesis [Fig. 5].

BRD4 and Myc in super-enhancers and bio-condensates

Super-enhancers (SEs) are large clusters of transcriptional enhancers that regulate gene expression of cell-type specific master transcription factors and thus define cell identity. The enhancer elements and the genes they regulate lie in close proximity in the SE clusters. SEs are characterized by the binding of high levels of transcriptional factors, coactivators, chromatin regulators, Mediator and RNA polymerase II [131]. As a result, they drive high levels of transcription from genes associated with them. SE-associated biomolecules can congregate to form liquid-liquid phase-separated condensates [132] that are defined by rapid recovery after photobleaching and sensitivity to 1,6-hexanediol. These condensates can be nucleated by a single event and can dissolve when DNA elements or factors involved in their formation are removed or depleted. Furthermore, intrinsically disordered regions (IDRs) have been implicated in condensate formation. Mediator subunit MED1 and BRD4 have large IDRs and participate in the formation of SE condensates. Disruption of SE condensates by hexanediol results in loss of both Mediator and BRD4 at SEs and a loss of RNA pol II at SEs and SE-driven genes. Such condensate formation at SEs serves to compartmentalize and concentrate biomolecules to facilitate transcription. MYC which itself has a large IDR can also phase separate (see review [133]) but its role in SE function remains to be determined.

Furthermore, condensate formation relies on multivalent interactions between proteins, DNA, RNA and through PTMs on proteins and modifications on RNA [134–136]. BRD4 modifies chromatin through its HAT activity [9] and phosphorylates itself as well as RNA pol II CTD, both important components of SE condensates, through its kinase activity [137]. It remains to be seen if BRD4 enzymatic activities facilitate multivalent interactions involved in the formation of SEs. BRD4 kinase activity is also regulated by two other RNA pol II CTD kinases involved in transcription, CDK7 and CDK9 [60]. Inhibition of CDK7 specifically affects transcription of SE-associated genes in MYC-driven tumor cells as compared to transcription from genes driven by typical enhancers [138] while Cyclin T1’s ability to form condensates is important for CDK9 hyperphosphorylation of RNA Pol II CTD as well as its transcriptional activity [139]. Prephosphorylation by CDK7 enhanced incorporation of RNA pol II into PTEFb condensates. This raises the possibility that the cross-regulation of CDK7, CDK9 and BRD4, and their phosphorylation of RNA pol II CTD may influence SE condensate formation and function [Fig. 5]. BRD4’s association with acetylated histones through its BDs is also important for phase separation [140]. Finally, BRD4’s ability to dimerize and its interaction with a multitude of proteins involved in transcription could influence SE condensate formation [Fig. 4].

SEs drive transcription dysregulation and are associated with SNPs related to a variety of diseases [131] and with oncogenes in various cancers [88]. In fact, cancer cells are thought to acquire SEs at oncogenes during tumor pathogenesis. SEs are found around the Myc gene in cancer cells and not in their normal counterparts. In fact, the Myc SE is required for cancer causation but is dispensable for normal development and homeostasis [141]. Significantly, BRD4 is associated with Myc SE. Disruption of BRD4 binding by JQ1 inhibition results in loss of BRD4 as well as coactivators Mediator and PTEFb, underlining the role of BRD4 in recruiting these proteins to SEs [Fig. 5]. The major function of Mediator is to act as a bridge between transcription factors and RNA polymerase. This helps relay signals from enhancer-bound activators to the pre-initiation complex at transcriptional start sites (TSS). Mediator also serves to recruit PTEFb as part of the super-elongation complex (SEC) [142]. The loss of coactivators triggered by disruption of BRD4 binding to chromatin happens preferentially at SEs, including the Myc SE, in cancer cells [88]. This results in selective repression of transcription of oncogenic drivers, most notably, Myc. Cancers that are addicted to MYC, including multiple myeloma (MM), Burkitt’s lymphoma (BL), acute myeloid leukemia (AML), and acute lymphoblastic leukemia (ALL), can be preferentially targeted by BET inhibitors that disrupt SEs. Thus, BRD4-mediated recruitment of transcriptional cofactors to SEs results in increased expression of Myc and drives oncogenesis [Figs. 4, 5]. It is thought that the preferential action of BET inhibitors at SEs is due to the highly cooperative nature of binding of transcription factors where SEs that have higher levels of transcriptional activators are most sensitive to disruption of factor binding [88]. A recent study showed that JQ1, through its physiochemical properties selectively partitions into SE condensates [143]. This could explain why BRD4 inhibition selectively affects oncogenes with large SEs like Myc.

Translocated SE regions, commonly referred to as enhancer hijacking, also result in dysregulated Myc expression [144]. An example of this is chromosomal translocation t(8;14), frequently seen in multiple myeloma, which moves the strong immunoglobulin H (IgH) super-enhancer at chromosome 14 to the breakpoint at 8q24 near the Myc loci. This foreign SE drastically increases Myc expression, resulting in aggressive multiple myeloma [145]. BRD4’s involvement in this process is highlighted by the fact that its expression positively correlates with disease progression in multiple myeloma (MM) patients, often correlating with BRD4 amplification, with BRD4 upregulation aided by bone marrow stromal cells [15]. BRD4 binds IgH enhancers as well as the endogenous Myc promoter and transcription start site (TSS) to drive Myc overexpression [Fig. 5]. This explains why MM with a wide range of MYC lesions are sensitive to BET inhibition. JQ1 induces downregulation of Myc transcription, decreases chromatin binding of MYC resulting in repression of its target genes leading to cell cycle arrest and cellular senescence. Cell proliferation can be restored by MYC supplementation demonstrating a direct role for MYC in the pathogenesis of MM. JQ1 treatment decreases tumor burden and increases overall survival. In another example of enhancer hijacking, in diffuse large B cell lymphoma (DLBCL), a chromosomal translocation t(3;8)(q27;q24) creates a fusion of the Myc and BCL6 gene with its super-enhancer leading to enhanced activation of Myc and oncogenesis [146].

The Myc gene has a CTCF site about 2 kb upstream of its TSS in the gene desert region that assembles SEs in cancer cells [144]. CTCF-binding sites have been shown to correlate with strong enhancers and active promoters [147] and occupy unmethylated regions. The Myc −2kb CTCF site acts as an enhancer-docking site that interacts with multiple tumor SEs. These cancer SEs depend on the Myc CTCF site for enhancer-promoter looping and for increased Myc mRNA expression [144]. Loss of the CTCF binding site or targeted methylation of the region results in decreased expression of Myc. Some of the SE domains overlap with DNA methylation valleys (DMV) [131]. It was recently reported that a protein called QSER1 protects these valleys from DNA hypermethylation [148]. Significantly, QSER1 has been shown to interact with BET domain proteins including BRD4. This leads to the intriguing possibility that besides helping maintain a hyperacetylated chromatin environment by its own HAT activity (as well as by contributing to the recruitment of p300), BRD4 might also help keep the Myc SE, including the −2kb CTCF site, open by counteracting repressive DNA methylation marks through recruitment of proteins like QSER1 [Fig. 5]. This could be another possible mechanism by which BRD4 contributes to maintaining Myc expression.

Modulation of BRD4 and MYC levels/function by the RNA methyltransferase machinery

METTL3 is the catalytic subunit of a methyltransferase complex (MTC) that modifies RNA at N6-methyladenosine (m6A) (reviewed in [149]). M6A is an important mRNA modification that influences multiple cellular processes like cell self-renewal, differentiation, invasion, and apoptosis through its effects on mRNA translation, degradation, splicing, export and folding. METTL3 exerts its oncogenic effects in several cancers including gastric cancer, lung cancer, bladder cancer, ovarian cancer, colorectal cancer, pancreatic cancer, and AML, through its effects on genes involved in tumor progression and apoptosis. METTL3 has been shown to modify multiple genes involved in cancer including Myc [150] and BRD4 [151]. This promotes transcription [150] and translation of Myc [152] while enhancing the translation of BRD4 through BRD4 mRNA looping and ribosome recycling through the interaction of METTL3 with eIF3h and promoting oncogenesis [Fig. 5] [151].

In AML, METTL3 overexpression results in increased translation and decreased degradation of Myc. Besides, m6A reader YTHDC1 is highly expressed in myeloid leukemia [153]. Myc and MYC targets are strongly downregulated on YTHDC1 depletion. This protein forms liquid-liquid phase condensates (LLPC) through binding m6A-RNA, including Myc. Sequestration of the m6A-modified Myc RNA in the condensates protects it from degradation through the PAXT-exosome complex. Significantly, a third of the m6A reader colocalizes with BRD4 SE condensates that enhance MYC transcription. This colocalization reveals coordination of the transcriptional upregulation of MYC with preventing degradation of MYC RNA resulting in leukemogenesis [Fig. 4, 5] [153]. Besides mRNA, YTHDC1 also bind m6A-modified eRNAs that similarly phase separate into transcriptional condensates that overlap with BRD4 condensates. M6A eRNAs preferentially associate with active enhancers i.e. regions with high levels of H3K27Ac, Mediator and BRD4. Depletion of the m6A reader YTHDC1 resulted in compromised BRD4 recruitment to enhancers, loss of BRD4 condensates and decreased transcription [154] emphasizing the importance of the m6A modification in enhancer and SE function. Thus, the m6A writer-reader machinery orchestrates amplified Myc transcription through enhancing BRD4 and MYC translation, preventing degradation of Myc RNA and influencing formation and functioning of BRD4 SEs that are required for MYC-mediated oncogenic transformation [Fig. 5].

Role of BRD4 and MYC in ecDNA hubs

Extrachromosomal, circular DNA (ecDNAs) are prevalent in cancers and known to drive cancer progression [155, 156]. They carry amplified oncogenes and can contribute to tumor heterogeneity by unequal segregation into daughter cells. ecDNAs connect extensively with the genome especially at promoters and enhancers. Because ecDNAs are enriched in RNA Pol II and SEs, they function as mobile enhancers that work in trans to amplify transcription in other ecDNAs as well as at chromosomal loci [157, 158].

ecDNAs can also cluster to form ecDNA hubs. These hubs have been recently shown to be held together by BRD4 in a MYC-amplified colorectal cancer cell line [158]. They are distinct from BRD4 SE condensates as they are not sensitive to hexanediol. Most of the Myc transcription in these cells originates in the ecDNA hubs and the clustering of ecDNAs was shown to be important for the high levels of Myc transcription. Interestingly, BRD4 associates with the ecDNA enhancers, and inhibiting binding of BRD4 resulting in disintegrating of these hubs preferentially inhibits ecDNA-derived-oncogene transcription [Figs. 4,5]. Whether these ecDNA hubs are held together by MYC-BRD4 interactions or are influenced by BRD4 homodimerization remains to be seen. BRD4 could facilitate formation of intermolecular ecDNA-ecDNA or ecDNA-chromosomal contacts through its interaction with chromatin mediated by its bromodomains and through its various protein-protein contacts including with MYC and with itself.

MYC and BRD4 in genomic instability

Deregulated MYC expression leads to chromosomal and extrachromosomal gene copy number increases in several genes, recombination events, long-range chromosomal rearrangements and alters the stability of multiple genes including those coding for regulatory RNA. These findings support a direct role for MYC in genomic instability. Importantly, these gene amplifications and alterations result in increased cell proliferation, metabolic changes, increased metastatic potential and resistance to drugs and radiation (reviewed in [23]). MYC is known to regulate replication initiation [159] and dysregulated MYC can lead to multiple replication initiations per replication fork per cell cycle [160]. MYC recruitment to sites of replication is affected by its association with several proteins involved in replication [161]. Many of the effects of MYC on genomic instability are transcription independent. However, there are also transcription-dependent roles for MYC in DNA damage and repair. MYC can cause DNA breaks resulting from reactive oxygen species formed due to MYC-driven deregulated gene expression. MYC also directly suppresses homologous and non-homologous and mismatch repair pathways [162]. So, several MYC-driven cancers have deficient DNA damage responses (DDRs) especially in conjunction with other deficiencies like loss of ATM or p53.

BRD4 also ensures genomic integrity by playing an important role in DNA damage repair (reviewed in [163]). BRD4 is involved in spatiotemporal control of transcription and replication by preventing the formation of RNA:DNA hybrids called as R loops thereby preventing DNA damage due to transcription-replication conflicts [164]. BRD4 also plays a role in DNA replication checkpoint signaling by regulating the pre-replication factor CDC6 [165]. Additionally, BRD4 is critical for the recruitment of repair factor 53BP1 and for the repair of Activation Induced cytidine Deaminase (AID)-mediated breaks in Class Switch Recombination during immunoglobulin isotype switching in B lymphocytes [166]. Double strand breaks (DSBs) in DNA leads to increased histone H4 acetylation as well as phosphorylation of H2AX (γH2AX) at the break sites. BRD4 is recruited to the break sites through the acetylation marks but it remains to be seen if the BRD4 HAT contributes to histone acetylation during DNA damage or if its kinase activity modulates function of proteins involved DNA damage response. Interestingly, a BRD4 isoform that does not have the HAT domain associates with the condensin II complex involved in chromatin compaction and reduces the formation of γH2AX foci indicating impaired DNA damage response [111]. JQ1 treatment increases ionizing radiation-induced H2AX phosphorylation in multiple cancer cell lines. In prostate cancer cells, BRD4 is required for repair of ionizing radiation-induced DNA DSBs and promotes gene rearrangements [167].

Thus, both MYC and BRD4 play important roles in maintaining genomic stability with their dysregulation resulting in catastrophic outcomes for the cell resulting in either cell death or oncogenesis [Fig. 4].

BRD4 inhibitors: a silver bullet or unicorn in the quest to regulate MYC in cancers?

The inhibition of the BET family proteins has been one of the most prominent emerging strategies in treating cancers and several other diseases in recent years. Its intimate involvement with MYC regulation resulted in the successful use of BET inhibitors in targeting MYC-addicted cancers [168]. BET bromodomain inhibitors induce cell cycle arrest and apoptosis in a variety of cancer cells especially leukemias, lymphomas and myelomas. They potently inhibit growth of malignant cells by reducing expression of oncogenes such as Myc [91]. In myeloma and lymphoma cell lines, BET inhibition directly downregulates Myc transcription and expression of its target genes resulting in growth arrest [15, 18]. Consistently, BET inhibition also results in significant anti-tumor activity and increased survival in animal xenograft models of Burkitt’s lymphoma and AML. Similarly, leukemia cells are also dependent on BRD4 to maintain MYC expression and cell proliferation [169]. In AML cells, a physical association of BRD4 with Mediator was shown to be important in upregulation of leukemia-specific genes [170]. Treatment with JQ1 led to downregulation of these genes- with the most striking effect on Myc mRNA levels- leading to the differentiation of leukemic blasts into macrophage-like cells.

The discovery that the inhibition of BRD4 by JQ1, one of the first BET inhibitors developed, could lead to a drastic reduction in MYC levels in cancer cells sparked a race to develop better and more specific BET inhibitors. Due to its role in regulating MYC, BRD4 has emerged as the primary BET target for new inhibitor design, with specificity and affinity for BRD4 being the benchmark. Current BRD4 inhibitors can be divided into four major classes based on their targeting mechanism. These first two categories are acetylated lysine mimetics and non-acetylated lysine mimetics, both of which bind and block the bromodomains. The third category of BRD4 inhibitors are non-bromodomain targeting inhibitors that target other regions of BRD4. The fourth class of inhibitors are BRD4 degraders known as Proteolysis Targeted Chimeras (PROTACs) which are designed to recruit BRD4 to a E3 ubiquitin ligase, Cereblon, which triggers BRD4 degradation by the proteasome. Significantly, while all these inhibitors lead to the reduction of MYC levels eventually, they could have different effects on MYC leading to subtle differences in therapeutic targeting and thus, efficacy. For example, while bromodomain inhibitors such as JQ1 can release BRD4 from the chromatin and decrease Myc transcript levels, they can result in increased degradation of MYC protein through increased MYC T58 phosphorylation by the free BRD4 released from chromatin. On the other hand, PROTAC inhibitors such as MZ1 which degrade BRD4 would also reduce Myc transcripts but increase MYC protein half-life through loss of BRD4-mediated MYC T58 phosphorylation [11]. Thus, these findings suggest that the role of BRD4 inhibitors in regulating MYC levels is highly context dependent and may indeed have negative consequences in the quest to regulate MYC levels in certain cancers. Indeed, it is no coincidence that although over a dozen different BET inhibitors have entered clinical trials, none have been particularly successful, and several trials have been terminated due to adverse effects [171]. While BET inhibitors were initially designed to directly suppress the previously ‘undruggable’ MYC, the complex crosstalk between BRD4, MYC and the pathways that regulate them warrants further study. Additionally, it is important that the nexus between the two proteins in influencing overlapping pathologies needs to be carefully considered before pharmacologically targeting either protein.

Summary

In this review, we have discussed how BRD4 and MYC function far beyond their recognized role as transcription factors. The two proteins have multiple overlapping effects [Fig. 4, 5] and directly regulate each other [Figs. 2,3]. BRD4 helps maintain homeostatic levels of MYC both by activating Myc transcription and influencing MYC stability. MYC also regulates BRD4 function by inhibiting its HAT activity. The combined effects of BRD4 structure (its bromodomains and its intrinsically disordered regions), its ability to interact with key transcription factors, coactivators, chromatin remodelers/modifiers as well as its enzymatic activities contribute to its pleiotropic effects on transcription regulation and stability of the MYC oncoprotein.

BRD4 regulates Myc transcription by nucleosome clearance at the Myc locus through its HAT activity and by phosphorylation of RNAPII CTD at Ser2 which leads to topoisomerase activation. It also regulates pause release by recruitment of PTEFb or JMJD6. Additionally, BRD4 regulates enhancer and super-enhancer activity, including Myc super-enhancers (SE), and is involved in recruitment of both Mediator and PTEFb to SEs. Phosphorylation of RNAPII CTD can increase its incorporation into SE condensates, possibly enhancing Myc transcription. Furthermore, BRD4’s recruitment of QSER1 could counteract methylation of important regulatory regions in SEs resulting in elevated Myc expression. Additionally, BRD4 is instrumental in the formation of ecDNA hubs that drive elevated Myc expression. Like BRD4, MYC also affects transcription of its target genes. MYC can regulate transcription either by affecting chromatin remodeling through recruiting HATs and Chromatin Remodeling Machines (CRMs) or by directly interacting with RNAPII and critical transcription regulators including BRD4, PTEFb, and Mediator. MYC enhances translation of TFIIF and PTEFb by increasing their mRNA CAP methylation. It also plays an important role in pause release through PTEFb and enhances transcription through its activation of topoisomerases (Top I & II). Notably, the m6A RNA writer- reader machinery can enhance MYC levels by increasing Myc transcription, sequestering m6A-Myc RNA in condensates and preventing their degradation, while also affecting translation of both MYC and BRD4. Thus, through their interactions with multiple influencers and modifiers of transcription and translation, BRD4 and MYC orchestrate cellular programs leading to elevated Myc expression and oncogenesis. An important caveat in correlating and understanding the extensive range of reports on BRD4 and MYC is the lack of attention to the different isoforms, domain structure and modifications of the two proteins in these studies. It is becoming increasingly apparent that the different domains of the two proteins and post translational modifications in these domains has an enormous impact on the ultimate functions of BRD4 and MYC.

BRD4 and MYC dysregulation have been linked to numerous diseases. Despite a growing understanding of the broad range of their functions and promising preclinical results, attempts to therapeutically target these proteins haven’t achieved clinical success. The multifaceted, interwoven regulatory pathways that both proteins are involved in [Fig. 5] is likely to be the primary reason for these setbacks. Understanding the complex interplay between MYC and BRD4 is essential to design treatment modalities that will effectively counter the pathological effects of both proteins while balancing the need to preserve their function for normal cellular processes. While we have tried to summarize the currently known overlap between BRD4 and MYC in their functions and regulation, much remains to be unearthed in this unique and powerful partnership.

Abbreviations

BET

Bromodomain and extra terminal

BD

Bromodomain

ET

Extra terminal

BID

Basic residue interaction domain

SEED

Ser/Glu/Asp rich region

HAT

Histone acetyl transferase

RNAPII CTD

RNA polymerase II carboxy-terminal domain

SE

Super-enhancer

MB

MYC box

TAD

Transactivation domain

B-HLH

Basic helix-loop-helix

LZ

Leucine zipper

CRD-BP

Coding region determinant-binding protein

NPS and CPS

N- and C- terminal cluster of phosphorylation sites

NMC

NUT midline carcinoma

DMV

DNA methylation valleys

m6A

N6-methyladenosine

MTC

methyl transferase complex

LLPC

Liquid-liquid phase condensates

ecDNA

Extrachromosomal circular DNA

DDR

DNA damage response

PROTACs

Proteolysis targeted chimeras

CRMs

Chromatin remodeling machines