Therapeutic Strategies to Inhibit MYC

Each year, an estimated 450,000 Americans are diagnosed with MYC-dependent cancer. The development of direct-acting MYC inhibitors remains a challenge; indirect modulators of MYC may have more potential.

Abstract

MYC is a master regulator of stem cell state, embryogenesis, tissue homeostasis, and aging. As in health, in disease MYC figures prominently. Decades of biological research have identified a central role for MYC in the pathophysiology of cancer, inflammation, and heart disease. The centrality of MYC to such a vast breadth of disease biology has attracted significant attention to the historic challenge of developing inhibitors of MYC. This review will discuss therapeutic strategies toward the development of inhibitors of MYC-dependent transcriptional signaling, efforts to modulate MYC stability, and the elusive goal of developing potent, direct-acting inhibitors of MYC.

MYC (v-myc myelocytomatosis viral oncogene homolog) is a nuclear DNA-binding transcription factor that orchestrates transcriptional regulatory pathways underlying cell growth, cell-cycle progression, metabolism, and survival (Meyer and Penn 2008; Conacci-Sorrell et al. 2014). MYC was first described 30 years ago in the context of viral oncogenesis (Duesberg and Vogt 1979; Hu and Vogt 1979; Sheiness and Bishop 1979), and subsequently a human homolog was identified and characterized (Dalla-Favera et al. 1982; Taub et al. 1982) Since its discovery, MYC has emerged as a master regulator of numerous biological and disease processes.

MYC is the most frequently amplified oncogene in human cancers, and MYC alteration is observed in a wide range of tissue types including breast, lung, and prostate cancer (Beroukhim et al. 2010). MYC overexpression occurs in 30% of all human cancers and frequently predicts for a poor clinical outcome, aggressive biological behavior, increased likelihood of relapse, and advanced stage of disease (Gamberi et al. 1998; Nesbit et al. 1999). An estimated 450,000 Americans each year are diagnosed with a MYC-dependent cancer. Deregulated expression of MYC is a hallmark feature of cancer (Conacci-Sorrell et al. 2014; Gabay et al. 2014), uncoupling physiologic, growth factor–dependent proliferation. Deregulated expression of MYC in cancer occurs through gene amplification (Sauter et al. 1995), chromosomal translocation (Klein 1983; Leder et al. 1983), focal enhancer amplification (Herrmann et al. 2012), germline enhancer polymorphism (Easton et al. 2007; Greenman et al. 2007; Kiemeney et al. 2008; Pomerantz et al. 2009; Wright et al. 2010), or commonly through constitutive activation of upstream signaling pathways. Each of these mechanisms serves to uncouple physiologic, growth factor–dependent proliferation.

MYC has been multiply validated as essential for tumor initiation and maintenance in numerous tumor histologies (Huang and Weiss 2013; Roussel and Robinson 2013; Gabay et al. 2014; Schmitz et al. 2014). Studies in transgenic mouse models identify that MYC inactivation leads to prompt tumor regressions, often associated with phenotypes of differentiation and cellular senescence with or without apoptosis (Felsher and Bishop 1999; Pelengaris et al. 1999, 2002; Jain et al. 2002; Flores et al. 2004; Soucek et al. 2008). In models of osteosarcoma, even brief MYC inactivation significantly improved survival rates owing to terminal differentiation of tumor cells (Jain et al. 2002).

The pleiotropic role of MYC in developmental biology and tissue homeostasis has prompted reasonable concern that targeting MYC may provoke severe, untoward toxicity. Indeed, genetic ablation of MYC in mice is embryonic lethal at an early stage of gestation (between days 9.5 and 10.5), with failed hematopoietic development (Davis et al. 1993; Hurlin 2013). Although subsequent studies have shown tolerance of tissue-specific MYC loss or down-regulation, such as in the liver and small intestine (Bettess et al. 2005; Hurlin 2013), concerns remain that pharmacologic inhibition of MYC may feature a narrow therapeutic index. For example, tolerance of MYC loss in intestinal epithelium could be conferred by compensatory up-regulation of MYCN (Bettess et al. 2005; Hurlin 2013), suggesting that selective inhibition of MYC-family isoforms may be required to firmly establish a therapeutic index.

Recently, a powerful genetic instrument developed by Evan and colleagues has provided compelling evidence that a therapeutic index does exist for targeting MYC. Structure-guided design allowed the optimization and characterization of a 93 residue, dominant-negative peptide comprising the basic helix-loop-helix leucine zipper (bHLHZ) domain (see Conacci-Sorrell et al. 2014) called Omomyc (Soucek et al. 2008; Savino et al. 2011). Four amino acid substitutions in the extended leucine zipper domain relieve repulsive charge interactions allowing for a coiled-coil-binding event to occur, presumably with endogenous MYC. Thus, Omomyc functions to competitively bind MYC in a manner preventing MYC:MAX heterodimerization, and expression of Omomyc prompts rapid growth arrest and down-regulation of MYC target genes (Soucek et al. 1998, 2002). Tetracycline-inducible expression of Omomyc has allowed a simulation of what pharmacologic modulation of MYC function may hold, in vitro and in vivo. As described in Gabay et al. (2014), constitutive induction of Omomyc prompts tonic inhibition of MYC:MAX heterodimerization associated with significant toxicity in proliferative tissue compartments in rodents (bone marrow and bowel). However, periodic induction of Omomyc has proven well tolerated. Dosing regimens have already been identified that provoke unprecedented responses in aggressive murine models of cancer, notably including the KRAS:TP53 model of non-small-cell lung carcinoma (Soucek et al. 2013). This research further validates MYC as a therapeutic target in cancer, importantly downstream from other intractable protein targets and tumor suppressors, and establishes a therapeutic index for MYC inhibition, in vivo.

Beyond cancer, MYC is also a positive effector of tissue inflammation. Activation of MYC is observed in immune cells during inflammatory cell expansion (Wang et al. 2011). MYC amplifies the transcriptional response to inflammatory transcription factor signaling in conditions such as rheumatoid arthritis (Pap et al. 2004). MYC function has also been implicated in the pathophysiology of heart failure, during tissue remodeling associated with hypertrophy and dilatation (Ahuja et al. 2010). Hence, the ability to inhibit MYC could extend beyond cancer therapy.

TOWARD DIRECT INHIBITION OF MYC

In the postgenomic era, the emergent paradigm of drug discovery is increasingly target directed. Disease biology identifies a critical effector mediating the pathophysiology of illness or the maladaptive response, and target-directed ligand discovery is undertaken. Innovation in the science of therapeutics over the last three decades has generated diverse strategies, which may be deployed to realize direct-acting inhibitors of protein targets. Still, the prioritization of targets is heavily influenced by perceptions of success and the horizons of supported research (Rask-Andersen et al. 2011; London et al. 2013; Swinney 2013).

As a class, transcription factors have proven particularly evasive in the discipline of ligand discovery (Darnell 2002; Berg 2008). Transcription factors are master regulators of cell state, controlling gene expression programs driving cell type specification and contributing to the pathophysiology of a broad range of diseases (Ptashne and Gann 1997, 2002). As a class, transcription factors are highly desirable proteins for therapeutic targeting. Yet, also as a class, transcription factors biophysically associate through extensive interfacial associations featuring large surface areas, a lack of hydrophobic invaginations, and noncontiguous contacts. These features are at odds with traditional binding models for small organic drug molecules (Fletcher and Hamilton 2007). The number of transcription factors successfully approached to date with discovery chemistry is therefore rather small. It has been our experience, developing the first direct-acting inhibitors of the NOTCH1 transactivation complex for probe and therapeutic development in T-cell acute lymphoblastic leukemia, that significant challenges exist in each phase of assay development, small-molecule optimization, biophysical characterization, mechanistic validation, and cellular permeability (Moellering et al. 2009). For these and other reasons, despite a confluence of detailed mechanistic insights and unmet medical needs, direct inhibition of MYC remains a historic challenge (Meyer and Penn 2008). Indeed, MYC has emerged as arguably the prototypical example of an “undruggable target,” a term reserved for protein targets that have proven intractable to coordinated efforts in ligand discovery.

MYC presents specific, significant obstacles to discovery chemistry. MYC lacks enzymatic activity, limiting many effective approaches to direct inhibition. Rather, MYC functions via protein–protein interactions, which remain a technical and psychological barrier to organized efforts in drug discovery (Wells and McClendon 2007). Structurally, MYC lacks globular functional domains, which might be approached with structure-based or empirical biochemical screening. MYC is comprised of a largely unstructured amino-terminal region involved in transactivation and protein stability (Conacci-Sorrell et al. 2014; Farrell and Sears 2014). Four highly conserved functional modules exist at the amino terminus, termed MYC boxes (MBI–IV). MBI possesses a degron targeted by FBW-7 (Welcker et al. 2004; Yada et al. 2004). MBII recruits a chromatin-modifying complex containing TRRAP, GCN-5, and TIP60 (McMahon et al. 1998; Hann 2014). MBIII and MBIV contribute to MYC-specific phenotypes, such as MYC-induced apoptosis, but are not clearly relevant for MYC-induced cell proliferation and survival (Herbst et al. 2005; Cowling et al. 2006). Among the MYC boxes, structural data is only available for MBI. Arrowsmith and colleagues have reported a structural model for MBI binding to the SH3 domain of the BIN-1 tumor suppressor (Andresen et al. 2012). However, this interaction does not present a compelling therapeutic rationale, and the structure is limited to 13 residues of MYC (Fig. 1A).

Figure 1.

Structural studies of MYC reveal challenges of developing direct-acting inhibitors of (A) MYC Box I (MBI as red ribbon and BIN-1 as white space-fill; PDB: 1MV0) and (B) the MYC:MAX heterodimer (MYC as red ribbon and MAX as white ribbon on white DNA; PDB: 1NKP).

The carboxyl terminus of MYC encodes a 100-residue basic helix-loop-helix-leucine-zipper (bHLH-LZ) DNA-binding domain. The leucine zipper forms a coiled-coil heterodimer with a homologous region on the transcriptional repressor MAX, which together engage E-box DNA-binding sites (see Conacci-Sorrell et al. 2014). Localization of the heterodimer to promoter and enhancer regions positively regulates transcription of proliferation-associated genes through control of transcription elongation (Eilers and Eisenman 2008; Meyer and Penn 2008; Rahl et al. 2010). Heterodimerization of MYC and MAX is required for transcriptional activation and oncogenic transformation, validating this interaction for further study. A seminal structural study by Nair and Burley (Nair and Burley 2003), however, elegantly displays the extended interaction between the two proteins that offers no apparent site for positioning a small-molecule inhibitor (Fig. 1B). The structurally and functionally similar MYCN and MYCL are comparably organized, posing then identical challenges for direct-acting therapeutic development in diseases attributable to these two oncogenes.

Intrinsically disordered proteins, such as MYC, comprise >30% of eukaryotic proteins (Wright and Dyson 1999; Metallo 2010; Andresen et al. 2012). Many such proteins are found at signaling nodes, where rapid physiologic response and dynamic cell state changes demand finely tuned posttranslational regulatory mechanisms that may be facilitated by intrinsic disorder (Gsponer and Babu 2009). Disordered proteins are often encoded by transcripts featuring short poly(A) tails, which reduce mRNA stability, and feature amino-terminal degrons and internal PEST sequences allowing for precise control of protein half-life (Gsponer et al. 2008). As discussed in detail in Farrell and Sears (2014), MYC has an ephemeral half-life (20–30 min), and is tightly regulated by E3 ubiquitin ligase recruitment and proteasomal degradation through at least two complexes (SKP-2 and FBW-7) and an evolutionarily conserved PEST domain (Herbst et al. 2004; Cowling and Cole 2006). Perhaps the intrinsic disorder of MYC and additional counter-regulatory measures (short protein and mRNA half-lives, transcriptional attenuation) are then ancient evolutionary safeguards to having such a dangerous oncogene in the genome.

Despite the apparent obstacles involved in studying the function of intrinsically disordered proteins, several research groups have organized around the challenge of developing first direct-acting inhibitors of MYC. The concept underlying these studies is theoretical in nature, arguing that the inherent, intrinsic disorder of native MYC (Nair and Burley 2003; Follis et al. 2008; Hammoudeh et al. 2009) establishes an opportunity for small-molecule-mediated induction of a neomorphic-binding site, a perhaps extreme example of induced fit. Induced fit implies a conformational change in the protein target provoked by a ligand, which may be accompanied by target inhibition or functional modulation as has been characterized extensively in enzymology and protein–nucleic acid interactions (Williamson 2000; Johnson 2008). The concept of induced fit applied to intrinsically disordered proteins possesses additional considerations. The lack of secondary structure in proteins such as MYC limits competing native intramolecular interactions, which are often low affinity, yet is complicated by the need for significant binding energy to overcome entropic penalties (Metallo 2010), and the likelihood of multiple binding states. This latter consideration presents challenges in establishing modes of molecular recognition, capturing low-energy assembly states for structural characterization, or even rigorously validating inhibitory biochemistry. As such, current described inhibitors of the MYC:MAX interaction serve as early, provocative examples of prototypical compounds, but, as a group, lack properties of target specificity biophysical characterization, and show utility in cell and model organism assays of MYC biology, which would define these compounds as chemical probes of MYC function and bona fide leads for therapeutic development (Frye 2010).

To our knowledge, the first efforts to inhibit MYC:MAX heterodimerization were reported by Vogt and colleagues in 2002 (Berg et al. 2002). Using a CFP/YFP fluorescence resonance energy transfer (FRET) assay, the bHLH-LZ domains of MYC and MAX were studied for inhibition of dimerization by a library of 7000 small molecules. These studies identified a series of related, substituted isoindolines (e.g., IIA4B20) (Fig. 2), which showed weak but reproducible inhibitory activity in FRET and ELISA assays (75–210 μm). Soon thereafter, Prochownik and colleagues (Yin et al. 2003), studied a commercial library of 10,000 compounds using a Gal-4-based yeast two-hybrid assay and the minimal bHLH-LZ interacting domains. Emerging from these studies were a series of structurally diverse compounds that impaired MYC:MAX binding by electrophoretic mobility shift assay (EMSA), including compounds that have been subsequently studied in cellular models of MYC biology (10074-A4, 10074-G5, and 10058-F4) (Fig. 2). A second study by Yang and coworkers reported curcuminoid inhibitors of MYC:MAX binding, in the context of a study aiming to establish binding constants for the heterodimeric interaction on DNA (Park et al. 2004). By measuring binding of MYC:MAX heterodimers to a fluorescein-labeled oligonucleotide, Berg and colleagues screened 17,298 compounds to discover pyrazolo-pyrimidine inhibitors with measurable biochemical inhibitory activity for MYC (Fig. 2) (e.g., Mycro1; IC50 = 30 μm) and a narrow window of selectivity for MAX:MAX homodimers (IC50 = 72 μm) (Kiessling et al. 2006, 2007). Additional efforts to target the MYC: MAX heterodimerization event have been undertaken subsequently, including efforts using further FRET screens (Xu et al. 2006), cellular transformation assays (Shi et al. 2009), and further EMSA screens (Jeong et al. 2010). These and other studies have led to a list of structurally diverse compounds with reported activity against the MYC:MAX:DNA-binding event (Fig. 2) (Yap et al. 2013).

Figure 2.

Small-molecule disruptors of MYC:MAX heterodimerization.

As described above, this research comprises more than a decade of study to identify the first qualified chemical probes of MYC function, for mechanistic use in cellular biology. Medicinal chemistry efforts directed at improving the limited potency of these compounds are described in the published and patent literature (WO2010083404 A2). However, many of the active species belong to structural classes annotated as nonselective ligands in pharmaceutical screening libraries, so-called pan-assay interference compounds (PAINS) (Baell and Holloway 2010; Whitty 2011), establishing important first directions for follow-up chemistry. Perhaps consistent with this behavior, incisive and important studies have observed multiple binding events of bioactive compounds along the MYC protein by circular dichroism (Hammoudeh et al. 2009), nuclear magnetic resonance (NMR) (Hammoudeh et al. 2009), and most recently by mass spectrometry (Harvey et al. 2012). The relatively low potency, demonstrable selectivity, and durable biostability of lead compounds have prevented use in MYC model systems in vivo (Guo et al. 2009; Clausen et al. 2010), but insights are emerging from this work and structural modeling. Still needed are direct-acting MYC inhibitors with properties commensurate with a bona fide chemical probe, such as defined by Frye (2010). There exists, therefore, an opportunity to revisit this biochemistry with new screening methodologies, new types of expansive chemical libraries, and new measurements to characterize the mode of molecular recognition of candidate inhibitors.

Beyond the MYC:MAX heterodimerization interface, functional studies highlight MBII as a putative target for chemical probe development. The requirement for MBII in MYC-specific transactivation and cellular transformation highlights the emerging validation of this protein–protein interaction. Unfortunately, limited biochemical and no structural information is available for MBII. Focused biophysical and structural studies of MBII may pave the way to miniaturized assays capable of interrogating MBII with discovery chemistry.

INHIBITION OF MYC-DEPENDENT TRANSCRIPTIONAL SIGNALING

MYC-dependent transcription requires the assembly and function of critical transcriptional and chromatin-modifying enzyme complexes (Fig. 3) (Dang et al. 2006; Eilers and Eisenman 2008; also see Hann 2014). Communication of transcriptional impulses is initiated by promoter- and enhancer-bound MYC, which serves to organize histone acetyltransferases (such as TIP60, GCN5, and CBP/P300) and the mediator complex. These events promote the preparation and stabilization of accessible euchromatin, as well as bridging interactions with initiated RNA Pol II. Recently, the Young laboratory described a pivotal role of MYC also in genome-wide transcriptional elongation mediated by proximal promoter pause release (Rahl et al. 2010). MYC importantly functions to promote CDK-9-dependent phosphorylation of RNA Pol II, triggering the release of paused polymerase to execute definitive target gene transcription (see Levens 2013; Rahl and Young 2014; Sabò and Amati 2014).

Figure 3.

Chromatin-dependent MYC transcriptional signaling. MYC and MAX bind at E-box motifs (boxes) in enhancers and promoters. Protein complex formation causes looping to engage promoters adjacent to MYC target genes (arrow to gene body), inducing RNA Pol II (RNA POL-2) loading to initiation site. Recruitment of TRRAP and histone acetyltransferases (HAT) leads to covalent modification of nucleosomes (clustered gray spheres) with lysine side-chain acetylation (Kac; small black circles). Acetylated, open chromatin is bound by bromodomain-containing proteins, including HATs (CBP, EP300, and GCN5) and BET-family coactivators (BET; yellow). BET bromodomains recruit the P-TEFb elongation complex (cyclin T and CDK-9; cyan), phosphorylates (small green circles) the carboxy-terminal domain of RNA POL-2, causing pause release and leading to elongation.

These insights support a model of targeting MYC via coactivator proteins critical to MYC-specific initiation and elongation. The pronounced effect of MYC on local chromatin acetylation (via HAT recruitment) (McMahon et al. 2000; Knoepfler et al. 2006), and the emerging role for MYC in elongation (via recruitment of CDK-9), prompted our consideration that the bromodomain and extraterminal domain (BET) family of coactivator proteins (BRD-2, BRD-3, BRD-4, and BRDT) figure prominently in MYC-mediated transcriptional pause release. BET bromodomains bind to polyacetylated histone tails via molecular recognition of ε-acetyl lysine (Kac) by an antiparallel bundle of four α-helices that forms a hydrophobic pocket with a conserved asparagine and an organized network of deep water molecules (Mujtaba et al. 2004, 2007; Filippakopoulos et al. 2012; Prinjha et al. 2012). Prior research established a role for BET bromodomains in cell-cycle progression and mitotic memory (Dey et al. 2003). Biochemical studies in virology by Verdin and coworkers importantly established that BET bromodomains possess a high-affinity binding motif for the CDK-9-containing positive transcription elongation factor complex (P-TEFb) at the distal carboxyl terminus (Bisgrove et al. 2007). We thus hypothesized that BET bromodomains may function to recruit CDK-9 to sites of hyperacetylated chromatin in cancer cells, thereby facilitating MYC-specific transcription elongation.

Having developed and characterized the first bromodomain inhibitor (Filippakopoulos et al. 2010), targeting the BET family of bromodomains, we performed a chemical genetic study of BET inhibition on MYC transactivation in vitro and in vivo (Delmore et al. 2011). Mechanistic studies performed in translational models of multiple myeloma (a MYC-dependent plasma cell malignancy) revealed that chemical displacement or genetic silencing of BET bromodomains results in the selective, coordinate down-regulation of the MYC transcriptional program. Concurrent, collaborative studies of BET inhibition in acute myeloid leukemia with Vakoc and Lowe comparably showed pronounced MYC pathway inhibition associated with marked down-regulation of transcription of the MYC gene itself (Zuber et al. 2011). In both studies, abrogation of chromatin-dependent MYC gene transcription via BET bromodomain inhibition elicited meaningful therapeutic responses in murine models of these aggressive hematological malignancies. Targeting MYC function via BET bromodomain inhibition has now been validated by studies in Burkitt lymphoma (Mertz et al. 2011), B- and T-cell acute lymphoblastic leukemia (Ott et al. 2012), non-small-cell lung carcinoma (Shimamura et al. 2013), and diffuse large B-cell lymphoma (Chapuy et al. 2013). In model systems of neuroblastoma, MYCN transcription and function was found to be comparably sensitive to BET inhibition. This hypothesis is presently being assessed in clinical trials of BET bromodomain inhibitors in MYC-dependent solid and hematopoietic tumors.

Another noteworthy approach to targeting MYC transcription derives from a unique and well-characterized feature of the MYC locus, namely, a predisposition for formation of G-quadruplex DNA (Siddiqui-Jain et al. 2002; Mathad et al. 2011). Small-molecule stabilizers of G-quadruplex DNA have been developed as tool compounds, which establish an impediment to RNA Pol II transcription of MYC. Further optimization is required before these compounds can be clinical deployed, but MYC-specific biological consequences have been observed with these compounds in experimental systems (Han et al. 1999a,b, 2000, 2001; Han and Hurley 2000; Hurley et al. 2006).

Beyond BET bromodomains, additional MYC-associated coactivator proteins may also then be considered for therapeutic development. As shown in Figure 4 (and as described in Conacci-Sorrell et al. 2014; Hann 2014) the MYC interactome features known critical mediators of transcriptional output, among them protein targets amenable to small-molecule discovery. Through MBII, MYC recruits the TRRAP complex that includes the histone acetyltransferases TIP60 and GCN5. The successful development of P300 histone acetyltransferase inhibitors by Cole and colleagues (Zheng et al. 2005; Bowers et al. 2010), suggests that discovery chemistry directed at these validated effectors of MYC-dependent chromatin remodeling and transcription is within reach. As P300 is a known binding partner of MYC via interactions in the transactivation domain (Faiola et al. 2005), the C646 inhibitor of P300 could contribute to the further understanding of this enzyme in MYC stability and function. Downstream from enhancer-bound factors such as these, MYC-mediated recruitment of mediator and TFIIH suggests consideration for the development of selective inhibitors of transcriptional kinases, such as CDK-7, CDK-8, and CDK-9, as MYC-directed tools and prototype therapeutics. Small-molecule inhibitors of transcriptional kinases are being intensively pursued in academia and industry, and a partial list of emerging compounds is presented in Table 1.

Figure 4.

The MYC protein interactome (STRING).

Table 1.

Small molecules linked to MYC-pathway inhibition

Compound name	Class	Target	References
Flavopiridol	CDK inhibitor	Cdk-9	Chen et al. 2005; Rahl et al. 2010
Purvalanol A	CDK inhibitor	Cdk-1	Goga et al. 2007
SU9516	CDK inhibitor	Cdk-2, Cdk-9	Gao et al. 2006
PHA 767491 HCI	CDK inhibitor	Cdc-7 and Cdk-9	Montagnoli et al. 2008; Natoni et al. 2011
SNS-032	CDK inhibitor	Cdk-2, Cdk-7, and Cdk-9	Walsby et al. 2011
JQ1	BET bromodomain inhibitor	Brd-4, Brd-3, Brd-2	Filippakopoulos et al. 2010; Delmore et al. 2011
SGI-1776	PIM kinase inhibitor	Pim-1	Zippo et al. 2007, 2009
EPZ004777	Dot1 L inhibitor	Dot1 L	Daigle et al. 2011
C464	p300/CBP ACTfrase inhibitor	p300	McMahon et al. 1998
SAHA	HDAC inhibitor
Triptolide	TFIIH/XPB	XPB	Titov et al. 2011
Nutlin-3a	p53-MDM2 inhibitor	p53-MDM-2	Felsher et al. 2000
SB220025	MAPK inhibitor	p38	Zhu et al. 2008
LY294002	PI3K inhibitor	PI3K	Zhu et al. 2008; Liu et al. 2011; Muellner et al. 2011
VAV-939	Wnt inhibitor	Tankyrase 1,2	He et al. 1998
LY-411575	γ-Secratase inhibitor	Notch 1	Moellering et al. 2009
VX-680	Aurora kinase inhibitor	Aurora kinases	Yang et al. 2011
NSC71948	BAG1 inhibitor	BAG-1	Zhang et al. 2011
ABT263	BCL2 inhibitor	BCL-2	Zhang et al. 2011
VER-155008	HSP70 inhibitor	Hsp70/Hsc70	Zhang et al. 2011
KW-2478	HSP90 inhibitor	Hsp90	Nakashima et al. 2010
SB 218078	Chk1 inhibitor	Hck-1	Jiang et al. 2011
Leflunomide	DHODH inhibitor	DHODH	White et al. 2011

Insights from developmental and disease biology suggest additional opportunities to target MYC function at the level of nuclear chromatin. Functioning alongside MYC-dependent transcription are parallel transcriptional networks that may be approached for synthetic lethality. Eisenman and colleagues discovered and characterized a functional interaction between Drosophila Myc and the Trithorax group protein Little imaginal discs (Lid) (Secombe et al. 2007), a homolog of the human proto-oncogene JARID1. They hypothesize a role for JARID1 in MYC-dependent neoplasia (Secombe and Eisenman 2007; de Rooij et al. 2013), supported by recent evidence that the NUP98-JARID1A oncogene in acute myeloid leukemia (AML) functions, in part, through an essential interaction between a plant homeodomain (PHD) finger encoded by JARID1A and histone 3 lysine 4, providing functional validation of JARID1A in a MYC-dependent tumor (Wang et al. 2009; Li et al. 2010).

Beyond chromatin-dependent MYC transcriptional signaling, other chromatin-associated factors such as the PIM-1 kinase may contribute to high global transcriptional output at MYC target genes (Breuer et al. 1989; van Lohuizen et al. 1989). PIM-1 phosphorylates histone H3 and, interestingly and similar to MYC, PIM-1 regulates transcriptional elongation (Zippo et al. 2007, 2009). PIM-1 inhibitors are in development as antitumor agents, including the ADP-competitive inhibitor SGI-1776. Additional small molecules targeting MYC-associated transcriptional regulators such as HDAC1, HDAC2 (Peart et al. 2005; Frumm et al. 2013), TFIIH (Titov et al. 2011), P53-MDM2 (Felsher et al. 2000), and DOT1L (Daigle et al. 2011) are referenced in Table 1.

TARGETING PATHWAYS UPSTREAM OF MYC

As an end-effector of numerous upstream signal transduction pathways, it is highly likely that effective targeting of dominant growth factor–independent signaling networks will mute upstream activation of MYC. Often, mitogenic signaling pathways activate intermediary effector transcription factors that integrate upstream surface-to-nuclear signaling with the common transcriptional language of MYC-dependent cell proliferation. A classic example is KRAS activation, which is among the most prevalent alterations in human cancer. KRAS activation in cancer drives MYC activation. Both genetic (Omomyc) (Soucek et al. 2013) and chemical (BET inhibition with JQ1) (Shimamura et al. 2013) perturbation of MYC function arrest proliferation downstream from KRAS and validate MYC as a target in KRAS-dependent tumors.

Indeed, many other upstream signaling pathways lead to deregulated MYC function, including the MAPK, PI3K, Wnt/β-catenin, and Notch pathways (see Eilers and Eisenman 2008; Meyer and Penn 2008; Conacci-Sorrell et al. 2014). The Wnt/β-catenin pathway regulates MYC expression through signaling to the Tcf-4 transcription factor that binds tissue-specific enhancers upstream of the MYC gene (Wright et al. 2010). Stabilization of NOTCH1 by juxtamembrane or PEST domain alteration in T-ALL leads to deregulation of MYC expression and function. Targeting the Notch pathway with upstream inhibition of γ-secretase (Weng et al. 2006), or direct inhibition using stabilized α-helical peptides (Moellering et al. 2009), leads to prompt down-regulation of MYC associated with antiproliferative consequences. Exemplary, additional inhibitors of upstream signal transduction pathways, such as MAPK and PI3K are presented in Table 1. As targeted agents are brought forward for clinical trials, preclinical studies assessing combinations of upstream and downstream modulators of MYC function should be pursued.

MODULATING MYC STABILITY

MYC has a naturally short half-life (Andresen et al. 2012). Regulation and preservation of the short transcript and protein half-life is essential to long-term survival without cancer. MYC protein homeostasis is considered in detail in Farrell and Sears (2014) and elsewhere by Sears and colleagues (2000). Here, we describe only briefly the opportunity to target MYC protein stability using therapeutic approaches. The transactivating domain is subject to phosphorylation, acetylation, and ubiquitylation, which serve to regulate stability and binding interactions (Cowling and Cole 2006; Andresen et al. 2012). Alterations in MYC and associated protein homeostatic pathway effectors are observed in human tumors, leading to the inappropriate posttranslational processing and stability of the MYC oncoprotein.

The facile development of kinase inhibitors warrants consideration of approaches to target MYC degradation, as MYC degradation is heavily influenced by serine and threonine phosphorylation. Biochemical and functional studies have established that phosphorylation of serine 62 (pSer62) by ERK pathway kinase activity stabilizes MYC and promotes proliferation. In contrast, phosphorylation of threonine 58 (pThr58) by GSK-3β destabilizes MYC. Notably, mutations of Thr58 are observed in Burkitt lymphoma resulting in constitutive MYC stabilization. Thus, therapeutic approaches to MEK-ERK pathway inhibition, or efforts to inhibit GSK-3β such as inhibition of PI3K-AKT signaling, may favorably modulate MYC stability for therapeutic benefit. This may be particularly true in tumors with up-regulated MEK-ERK signaling (KRAS or EGFR activated) or phosphatase and tensin (PTEN) loss. Table 1 lists exemplary compounds, which may be used as chemical probes of pathway function.

Opposing MYC phosphorylation at serine 62 is protein phosphatase 2 (PP2A), which functions as a heterotrimeric enzyme complex to indirectly and directly modulate MYC protein stability. The Sears group has shown that pSer62 is a direct target of PP2A, supporting a model in which PP2A loss (a common tumor suppressor loss in cancer) promotes MYC stability. Presently, pharmaceutical compounds targeting PP2A do not exist. However, in collaborative research with the laboratories of Thomas Look, Jon Aster, and Hanno Steen, we recently discovered PP2A as the antiproliferative target of perphenazine (and other phenothiazines), by ligand-affinity chromatography (Gutierrez et al. 2014). As small molecules targeting and activating tumor suppressors are rarely described, we are interested now to optimize these compounds for enhanced PP2A binding and activation, to destabilize MYC and modulate other critical PP2A target proteins. Recently, Eilers and colleagues reported a novel approach to destabilize MYCN via inhibition of Aurora A kinases (Brockmann et al. 2013). MYCN is targeted for degradation by the FBXW7 ubiquitin ligase. MYCN protein stability is promoted by an interaction with the Aurora A mitotic kinase. Although the kinase activity of AURKA is not required for MYCN protein stability, two ATP-competitive azepine inhibitors of AURKA (MLN-80545 and MLN-8237) disrupted binding to MYCN. In mechanistic and translational experiments performed in MYCN-dependent neuroblastoma models, AURKA inhibition prompted MYCN degradation and an antiproliferative effect.

A final opportunity for therapeutic targeting of MYC protein stability is presented by MYC-associated histone acetyltransferases. Beyond their established role in promoting chromatin remodeling, bromodomain recruitment, and gene expression, acetylation of MYC by CBP/P300, GCN5, and TIP60 promotes MYC stability and transcriptional activity (Cowling and Cole 2006). As above, inhibition of these acetyltransferase enzymes may display favorable activity in cancer model systems, in support of further clinical development.

MYC AND SYNTHETIC LETHALITY

MYC-addicted cancer cells may harbor unique dependencies that a normal cell may not require. Such synthetic lethal interactions with MYC may prompt the development of context-specific inhibitors. Two kinase targets have been identified that are required for cell survival in the context of enforced MYC expression: Aurora kinase B (Yang et al. 2010) and Cdk-1 (Goga et al. 2007). Inhibition of these kinases with cell-permeable small molecules has shown compelling evidence of a leveraged antiprolifeartive effect in MYC-dependent cancer models, albeit with a narrow therapeutic index. Emerging research approaches using highly parallel reverse genetic screens (RNA interference; RNAi), has reinvigorated efforts to identify MYC-specific tumor dependencies. Recently, two laboratories performed RNAi screening in a MYC-dependent context leading to the suggestion of CSNK1e kinase (Toyoshima et al. 2012) and the SAE1/2 SUMO-activating enzyme (Kessler et al. 2012) as putative targets for the development of synthetic lethal small molecules. Rational approaches to synthetic lethality have included consideration of metabolic adaptation to MYC-mediated tumorigenesis. Observing therapy-associated cellular senescence in treated Eµ-MYC cells, Schmitt and colleagues targeted glucose uptake and autophagy with impressive antiproliferative results (Dorr et al. 2013). Together, these studies establish a new path to MYC-specific therapeutic development, revealed by powerful new technologies and a detailed mechanistic understanding of MYC biology at the interface of transcription, translation, and metabolism.

CONCLUDING REMARKS

The centrality of MYC to the pathogenesis of cancer, illustrated in granular detail in the accompanying cited literature, demands the creativity and commitment of the allied fields of ligand discovery. In this review, we provide a rationale for targeting MYC that balances high-risk, high-reward approaches of direct MYC:MAX modulation with near-term opportunities to approach MYC transcriptional signaling via chromatin-dependent processes and pathways of protein homeostasis. Still, direct inhibition of MYC remains an elusive goal in cancer medicine—an inadequacy of current therapeutic science perhaps at the heart of our inability to eradicate this disease.