Tilting MYC toward cancer cell death

Abstract

MYC oncoprotein promotes cell proliferation and serves as the key driver in many human cancers; therefore, considerable effort has been expended to develop reliable pharmacological methods to suppress its expression or function. Despite impressive advances, MYC-targeting drugs have not reached the clinic. Recent advances suggest that within a limited expression range unique to each tumor, MYC oncoprotein can have a paradoxical, pro-apoptotic function. Here we introduce a counterintuitive idea that modestly and transiently elevating MYC levels could aid chemotherapy-induced apoptosis and thus benefit the patients as much, if not more than MYC inhibition.

Introduction

MYC has long been considered a major cancer driver, on par with activated K-Ras. For that reason, therapeutic targeting of this oncogene has become an article of faith in precision medicine. The overall rationale for targeting initiating oncogenes is based on the phenomenon known as “oncogene addiction” where tumor cells become dependent on one protein or signaling pathway in a way their normal counterparts never are [1]. Thus, the former can be specifically killed while leaving the latter mostly unharmed, which is not the case for cytotoxic therapies like chemotherapy. The concept of addiction to MYC has been amply validated in genetically engineered mouse models [2], and several MYC-targeting compounds are now in clinical development or in clinical trials. They range from broad blockers of MYC transcription such as Brd4 inhibitors (BRD4i) [3, 4] and G-quadruplex-promoting compounds [5] to very specific blockers of MYC protein function such as 10074-G5 [6], MYCi361 [7], and Omomyc [8]. However, all these compounds are yet to produce tangible successes in the clinic, and in fact some of them such as MYC-targeting siRNA [9] resulted in terminated clinical trials.

One could argue that this is because the MYC locus frequently finds ways to bypass transcriptional inhibition, while specific and potent inhibitors of MYC itself are proving to be elusive. Interestingly, even though BRD4i are ineffective against most MYC-overexpressing cancers, they show acceptable potency against bona fide Brd4-driven cancers such as midline carcinoma [10], arguing that Brd4i are in fact good drugs, just not necessarily for MYC-driven cancers. The alternative explanation – unproven, but hard to rule out - is that genetically complex human cancers are far less dependent on MYC than commonly thought.

The second concern is that in both preclinical and clinical settings, Brd4i, Omomyc, and their brethren are sometimes tested as monotherapies, setting aside their interactions with existing standards of care in such scenarios. Thus, in parallel to inhibiting MYC expression or function as an anti-growth strategy, a complementary approach would be to exploit unique vulnerabilities associated with high MYC expression as an anti-survival strategy. One such vulnerability is the propensity of MYC to engage cell death pathways. In this review, we highlight this paradoxical pro-apoptotic function of MYC and the often-overlooked fact that MYC-driven tumors live (and sometimes die) by a Goldilocks Principle, according to which the levels of this lethal oncoprotein have to be just right: not too low, but not too high either. The corollary of this balancing act is that transiently elevating MYC levels aids chemotherapy-induced apoptosis and thus could directly benefit patients and potentially overcome chemotherapy resistance.

The many facets of MYC

The MYC gene is deregulated in over half of all cancers, making it the most frequently altered oncogene [11]. Gene amplification of MYC and its paralogs MYCN and MYCL alone was observed in almost 30% of all The Cancer Genome Atlas (TCGA) samples. MYC deregulation also has been found to occur by various other means including point mutations, activation of upstream signaling pathways resulting in elevated transcription and protein stabilization, and more recently – by enhancer hijacking within core regulatory circuitries, which are especially important for MYCN activation [12, 13]. Breakthroughs in deciphering the mechanisms of MYC-driven tumorigenesis came from the characterization of MYC protein domains and identification of MYC target genes and gene networks. MYC (formerly referred to as c-Myc) is an oncoprotein and a nuclear transcription factor that belongs to a family that also includes N-MYC and L-MYC. It has distinct structural modules and several key phosphorylation sites, which control its function and well as turnover via proteasomal degradation (see Figure 1 for details) (highlighted in [14, 15]).

Figure 1. MYC protein structure and regulation.

MYC protein is comprised of 5 major domains: MBI and MBII (MYC Box I and II) in the N-terminus, and the NLS (nuclear localization signal), bHLH and LZ domains (basic helix-loop-helix and leucine zipper) in the C-terminus. This C-terminal domain also allows MYC to hetero-dimerize with its binding partner Max and associate with E-box DNA sequences (CACGTG), which is essential for its transcriptional and transforming activity. Another box (MB0) has been recently identified and shown to positively control transcription elongation. There are also multiple protein-protein interactions, which either enhance (blue stars) or counteract (yellow stars) MYC function. Thus, inhibitors of CDK9 and Erk, activators of PP2A, and the synthetic peptide Omomyc (which disrupts MYC:MAX binding) bring about either reduced MYC levels or impaired MYC function. Conversely, inhibitors of GSK3 and E3 ligases (Fbxw7, Skp2, etc.) stabilize MYC protein levels and boost its function.

In addition to its well-established role as a transcriptional activator, MYC has been shown to function as a transcriptional repressor, which in most cases involves complex formation with another nuclear protein Miz-1 [16]. Its opposite effects on gene expression notwithstanding, MYC has been shown to bind the majority of actively transcribed human genes and regulate both protein-coding and non-coding RNA genes, suggesting that it acts as a global transcriptional regulator (reviewed in [17, 18]). A large proportion of MYC-regulated genes aid tumor growth and upkeep. This set includes genes that control cell cycle and growth, metabolism, protein synthesis, cell migration, angiogenesis, and chromosomal instability. All these pathways contribute to MYC-mediated transformation of normal cells into cancerous ones (reviewed in [19]). More recently, the role of MYC in creating the immunosuppressive microenvironment (for example, through its effects on CD47 and PDL1 expression) has come to the fore [20]. Thus, it comes as no surprise that MYC is one of the few oncogenes that could single-handedly drive rapid neoplastic growth. Consequently, in many systems MYC inactivation led to regression of established tumors [21–24]. Numerous subsequent studies have confirmed that MYC plays an important role not only in tumor initiation, but also in tumor maintenance in multiple organ systems (reviewed in [25]). For this reason, considerable efforts have been expended to learn how to inhibit MYC activity.

Therapeutic targeting of MYC: opportunities and challenges

Transcription factors are notoriously challenging to target pharmacologically due to their lack of hydrophobic pockets and large interaction surface areas, which are at odds with the standard binding models of small drug molecules [26]. Only very recently some advances have been achieved with direct MYC inhibitors [7]. As an alternative to disrupting MYC with small molecules, the MYC dominant-negative peptide Omomyc has been developed to inhibit its function [27–29]. This 92-amino acid polypeptide homodimerizes and binds to both MYC and Max, thus preventing MYC:Max heterodimerization. While Omomyc has displayed anti-tumor activity in experimental models of non-small cell lung cancer, its efficacy in the patient setting is currently being determined (ClinicalTrials.gov Identifier: NCT04808362) [8].

Thus, complementary efforts were expanded to inhibit MYC expression and function at the mRNA and protein levels and to target signaling pathways that activate MYC. There are approaches to inhibit MYC function at just about every level of its regulation in the cell. Numerous upstream signaling pathways deregulate MYC activity, such as Notch, WNT, PI3K, and MAPK pathways [30]. Many small molecule inhibitors exist for each of these pathways and are currently undergoing pre-clinical and clinical trials.

One recently emerged strategy to target MYC function is the inhibition of BET bromodomain family members. BET inhibition was first shown to down-regulate the MYC transcriptional program [31]. Following the discovery that the BET transcriptional regulator BRD4 can bind to the MYC promoter and regulate MYC expression, investigators began to explore the therapeutic use of BRD4 inhibitors such as JQ1 to treat MYC-driven cancers [32–34]. There are currently several BRD4 inhibitors in clinical trials as monotherapies for a variety of cancer types. Despite the promising pre-clinical studies, the clinical trial results have thus far been mixed. Positive anti-tumor effects have been observed, but so have detrimental side effects at below-efficacy doses (reviewed in [35]). Additionally, pre-clinical studies have identified resistance mechanisms to BET inhibition, such as increased expression of anti-apoptotic or autophagy proteins, that could likewise end up leading to BET inhibitor resistance in patient tumors [36]. It is also notable that BET inhibition affects key MYC-independent oncogenic pathways (see for instance [37]), making interpretation of results less straightforward.

In addition to BRD4, another targetable protein that affects the function of MYC is Aurora kinase A. This enzyme was first identified to enhance N-MYC stability in neuroblastoma, and since then has been tested pre-clinically as a druggable target for MYC-driven cancers [38–40]. These findings led to clinical trials of Aurora kinase inhibitors as monotherapy in solid and hematologic tumors; these inhibitors were unable to produce durable responses in solid tumors but were somewhat more effective at treating the hematologic malignancies (reviewed in [41]). As the mixed responses in clinical trials would suggest, BET and Aurora kinase inhibition may only be suitable for certain subsets of MYC-driven cancers.

A complementary approach to MYC destabilization has been to boost the activity of the PP2A phosphatase using compounds collectively known as SMAPs (small-molecule activators of PP2A) [42].

Together with mTOR inhibition, this approach has shown efficacy in pre-clinical models of highly aggressive pancreatic ductal adenocarcinoma [43], but SMAPs are yet to be tested in human patients.

Finally, there is considerable evidence that some therapeutic benefits could be reaped from inactivating MYC targets. One salient example is ornithine decarboxylase (ODC). In the 2005 proof-of-principle paper, deleting the ODC gene or inhibiting the enzyme with difluoromethyl-ornithine (DFMO) delayed lymphoma development in the Eμ-MYC mouse model [44]. However, in the realm of experimental oncology, this approach appears to be re-directed towards N-MYC-driven tumors, such as neuroblastoma [45]. In summary, while multiple independent strategies are concurrently being pursued to inhibit MYC either directly or indirectly, each has its own limitations, which make it difficult to ascertain whether any of them would ever emerge as blockbuster drugs to successfully treat MYC-driven malignancies.

MYC-driven apoptosis: a potential vulnerability

Given that the efforts to inhibit pro-oncogenic activities of MYC are yet to come to fruition clinically, unorthodox approaches might be in order. One such approach would be to exploit the long-known function of MYC to promote programmed cell death, or apoptosis, during normal development. Indeed, in a recent study MYC was found to be highly expressed in young tissues that were exquisitely primed to undergo apoptosis, and loss of a MYC allele resulted in a reduced response to apoptotic stimuli [46]. Rather paradoxically, oncogenic MYC also retains the conflicting functions of driving proliferation and cell death. The anti-survival effects of MYC first reported 30 years ago are achieved by a variety of mechanisms (comprehensively reviewed in [47]) in both immortalized and cancerous cells lines as well as in vivo tumor models.

In fact MYC-dependent apoptosis is a hallmark of many cancers such as Burkitt lymphoma, in which tumor cells are highly proliferative but at the same time also display high levels of apoptosis [48]. In the realm of solid tumors such as lung, liver, and ovarian cancers, where MYC is usually activated via copy number gains [49], activation of MYC was reported to correlate with higher apoptotic indices and/or focal amplification of anti-apoptotic genes [50–52]. Admittedly, correlation does not prove causation, but there is also a large body of evidence demonstrating direct pro-apoptotic effects of MYC, as discussed below.

Early studies revealed that stimuli such as low serum, cytokine withdrawal, or T-cell activation induce apoptosis, and that this cell death response was dependent on MYC [53–55]. Not long after it was discovered that MYC triggers apoptosis via ARF upregulation and ensuing activation of the tumor suppressor p53; this p53-driven apoptotic response was found to require the activity of the DNA damage response protein ATM [56, 57]. Subsequently, it was noted that MYC-driven murine transgenic lymphomas and other MYC-driven tumors inactivate this pro-apoptotic pathway by disabling Arf or p53 [58]. The inactivation of MYC-driven apoptosis is thought to be crucial for tumor initiation and progression, because upon loss of the ARF-MDM2-p53 pathway in mouse models of MYC-driven cancers, there was a robust acceleration of tumorigenesis [59, 60]. Two mutant forms of MYC commonly found in Burkitt lymphoma (P57S and T58A) are unable to induce apoptosis owing to their failure to upregulate BIM and inhibit Bcl-2 function, further demonstrating how critical it is for cancer to evade MYC-driven apoptosis during tumorigenesis [61]. Of note, adjacent serine and proline residues are recurrently mutated in other histotypes as well, including endometrial salivary and gland cancers ([62]; in the COSMIC database T58 is annotated as T73, owing to the existence of the longer CTG codon-initiated MYC Isoform 2 with the Uniprot Identifier P01106–2).

While it is well established that p53 is the main effector of MYC-mediated apoptosis, p53-independent mechanisms have also been identified (reviewed in [47]). It has been often proposed that MYC exerts its pro-apoptotic function through direct modulation of gene transcription. Later it was found that numerous intrinsic and extrinsic apoptotic genes (Figure 2) are in fact direct transcriptional targets of MYC and are bound by MYC at their promoters. In Eμ-MYC murine lymphomas MYC activity led to the suppression of the anti-apoptotic proteins BCL-2 and BCL-XL [63, 64], and in melanoma MYC-driven transcription of NOXA, a pro-apoptotic protein in the Bcl-2 family of proteins, was observed [65]. One of the major pro-apoptotic proteins, Bax, is also a direct transcriptional target of MYC [66], although non-transcriptional regulation of Bax activity by MYC has been described as well [67]. Additionally, MYC has been shown to bind to and activate the promoters of the pro-apoptotic proteins BIM and BID and in doing so contribute to the priming of mitochondria to respond to apoptotic stimuli. This was observed in multiple models including transformed rodent fibroblasts [67, 68] and B-cell neoplasms [69, 70], but also pancreatic [71], and breast, ovarian, and colon [66, 72] carcinomas. MYC was also found to bind the promoter of mtCLIC, a pro-apoptotic mitochondrial chloride ion channel; when the expression of this gene was suppressed, so was MYC-driven apoptosis, suggesting that it is yet another player in MYC-mediated cell death [73].

Figure 2. Regulation of apoptosis by MYC.

MYC can promote apoptosis through both intrinsic (mitochondrial) and extrinsic (death receptor-mediated) mechanisms by regulating many key components of these pathways. Common anti-cancer drugs (e.g., chemotherapeutics) often trigger both apoptotic pathways, thereby increasing their reliance on MYC. Additionally, the extrinsic pathway can be specifically engaged by death receptor agonists or DR4-activating antibodies (e.g., mapatumumab).

In addition to amplifying intrinsic, or mitochondrial cell death pathway, MYC has also been found to modulate extrinsic, or death-receptor mediated cell death (Figure 2). Early studies demonstrated that MYC could sensitize cells to signaling though the death receptor CD95/Fas [74]. MYC expression was also found to greatly sensitize cells to apoptosis triggered by the extrinsic ligand TRAIL [75]. Just as with intrinsic apoptosis, many extrinsic apoptosis genes are direct transcriptional targets of MYC. This includes genes like death receptor-4 (DR4) and the extrinsic ligand FasL which are both positively regulated by MYC, as well as CFLAR/FLIP, the negative regulator of Caspase 8, whose transcriptional expression is repressed by MYC [76–78]. The TRAIL receptor DR5 was found to be upregulated at the cell surface upon MYC activation [79, 80]. But do these well-established axes represent valid therapeutic targets in MYC-driven tumors? Some preliminary answers are beginning to emerge from studies exploiting the concept of synthetic lethality: perturbations that uniquely affect cells with a certain genetic background, in this case MYC dysregulation.

Lessons from synthetic lethality and correlative studies

Genome-wide screens have proved an invaluable tool for global assessment of MYC-synthetic lethality genes. Using shRNA, siRNA, or CRISPR pooled libraries screens, numerous genes and gene networks have emerged as synthetically lethal with high MYC expression. Some of these gene networks include components of RNA polymerase complexes, transcription initiation complexes, DNA repair and cell cycle checkpoint proteins, metabolic enzymes, and notably – components of the apoptotic pathways (reviewed in [81, 82]). For example, CDK1 knockdown was identified from RNAi screens to be synthetically lethal with MYC overexpression, and inhibiting CDK1 in MYC-driven lymphoma and neuroblastoma models lead to apoptosis and decreased tumor growth owing to dysregulation of a direct CDK1 target BIRC5 (a.k.a. survivin), one of the Inhibitor of Apoptosis (IAP) gene family members and a caspase 3/7 inhibitor (Figure 2) [83]. Of note, survivin is now considered a good drug target, with multiple compounds in clinical development [84].

Admittedly, many other MYC synthetic lethal genes emerging from genome-wide screens have no known direct connection to cell death pathways and are involved instead in a range of other functions ranging from cell cycle regulation (checkpoint kinase 1) to metabolism (glutaminase) [85]. However, there is little overlap between targets identified in independent screens, questioning the broad application of specific synthetic lethal targets across different tumor types. In contrast, targeted synthetic lethality approaches aimed at apoptotic pathways might be more enlightening. For example, it has long been appreciated that MYC sensitizes cells to apoptosis through the extrinsic death receptor Fas, and subsequently it was found that the expression of high MYC lead to cells being sensitized to the extrinsic ligand TRAIL via MYC-dependent upregulation of death receptor-5 (DR5) [79]. Another recent study similarly identified that while high MYC expression drives brain metastases of breast cancer, it also generates a synthetically lethal interaction with TRAIL [86]. Through these studies, the field has been able to identify unique vulnerabilities of MYC-driven cancers, many of which are now being pursued in the pre-clinical or clinical settings ([87] and references therein).

However, for the majority of tumor types, new investigational drugs targeting specific genetic lesions are tested in combination with standards of care [88]. Thus, the complex effects of MYC on pro-survival and pro-apoptotic pathways might be particularly relevant in the context of chemotherapy, where MYC-directed and conventional therapies would need to function together. There is considerable evidence that they might. Almost 20 years ago, it was demonstrated that in human colorectal cancers, MYC amplification (combined with wild type p53 expression) increases susceptibility to 5-fluorouracil in vivo [89]. Subsequently, sensitizing effects of MYC were observed in fibroblasts [90] and in a non-lymphoid hematologic malignancy with dismal outcomes: multiple myeloma [91]. More recently, the presence of high MYC was shown to sensitize multiple cancer types to anti-mitotic chemotherapy agents through upregulation of pro-apoptotic BH3 proteins and suppression of BCL-XL [92]. Similarly, in breast cancer the anti-tumor activity of Bcl-2 inhibitors combined with AMPK activators was found to be dependent on high MYC expression [93]. In yet another study using an in vivo model of MYC-driven small cell lung cancer, tumors were sensitized to Aurora kinase inhibitors which synergized with chemotherapy to induce apoptosis [94]. Conversely, numerous studies on transformed fibroblasts, B-cell lymphoma, adrenocortical cancer and other cancer types have demonstrated that the loss of MYC expression confers resistance to chemotherapeutic drugs like doxorubicin, etoposide, and paclitaxel [68, 95–97] as well as the proteasome inhibitor bortezomib [98]. In fairness, there are several cell lines where inactivation of MYC was reported to render cells more susceptible to chemotherapy, e.g., M14 melanoma [99, 100] and MCF-7 breast carcinoma [101]. This complexity indicates that the role of MYC in chemotherapy is either narrowly histotype-specific or more likely of bi-phasic nature, where for the tumor to withstand the onslaught of anticancer drugs MYC levels have to be just right: not too low, but not too high either. This latter Goldilocks scenario has broad translational implications.

Boosting MYC-dependent therapeutic apoptosis

The intrinsic ability of MYC to drive apoptosis is clearly insufficient to offset high cell proliferation rates found in most MYC-driven tumors. However, by employing strategies to enhance MYC-mediated apoptosis it should be possible to tip the scale in favor of cancer cell death and ensuing tumor regression. Sophisticated in vivo studies with finely controlled MYC alleles have suggested that there are thresholds for MYC activity: a modest increase in the level of MYC led to oncogenesis, but yet higher levels were required to trigger apoptosis [92, 102]. This would indicate that in the context of cancer, there could be a certain threshold for MYC that when surpassed, could activate cell death.

Conceptually, there are precedents for the counterintuitive overexpression of oncogenes as a therapeutic strategy. As early as in 1998, it was demonstrated that overexpression of adenovirus 5 E1A oncogene suppresses tumor growth in vivo via increased apoptosis [103]. This concept has been advanced to Phase I clinical studies using intratumoral E1A cancer gene therapy [104, 105]. Although these formulations are currently not in clinical development, recent studies demonstrated the validity of this approach in genetically engineered mouse models, where transgenically expressed E1A blocked chemical skin carcinogenesis [106]. However, up until recently, it remained to be determined how exactly the pro-apoptotic function of MYC could be re-engaged in various cancers.

Several laboratories have reported that strengthening the CD19-PI3K-AKT axis is a reliable method to boost MYC protein stability in B-lymphoid cells [107–111]. This finding is consistent with the propensity of glycogen synthase kinase 3 beta (GSK-3β), which is inhibited by Akt, to phosphorylate MYC at Thr-58, marking MYC for recognition by the E3 ubiquitin ligase Fbxw7 and subsequent degradation [112–115] [reviewed in [116] (Figure 1). Given that activation of the PI3K pathway is one of the most frequent alteration in human cancers [11, 117], and that it is a known suppressor of MYC-induced apoptosis [118, 119]it would need to be manipulated only transiently, for example with short-lived GSK-3 inhibitors.

Further work demonstrated that transient (90–120 min) stabilization of MYC protein using the GSK-3β inhibitor CHIR99021 [120] (see Table 1) is sufficient to sensitize therapy-resistant B-cell lymphomas to chemotherapy and direct engagers of the extrinsic apoptotic pathway such as TRAIL or DR4 agonistic antibodies [121] (Figure 2). Several lines of evidence suggest that the bulk of these sensitizing effects was MYC-dependent. First, inhibition of GSK-3 had no discernable effect on sensitivity to chemotherapy in Epstein-Barr virus-transformed B-lymphoid P493–6 cells with tet-repressible MYC alleles [122] when they were maintained in the MYC^OFF state (with the caveat that these cells were also non-proliferating and potentially refractory to genotoxic stresses.) Second, in Burkitt lymphoma cell lines the effects of GSK3i were cancelled following JQ1–mediated MYC downregulation. Lastly, no chemosensitizing effects were observed in lymphoma cells bearing MYC Thr-58 mutations, where MYC protein levels are no longer regulated by GSK-3β nor transiently increased by GSK3i. Of note, while this mutation is relatively frequent in Burkitt’s lymphoma, it is not common in human cancers in general. For example, COSMIC Cancer Mutation Census estimates its frequency to be 2.3×10⁻³ and recurrence - 144 (https://cancer.sanger.ac.uk/cmc/gene/myc).

Table 1.

MYC-stabilizing compounds in clinical development

Inhibitor	Mechanism of action	Indication	Stage of development	References
CHIR99021	GSK-3 inhibitor	various xenograft models	Preclinical	Harrington et al., 2019; O’Flaherty et al., 2019
tideglusib	GSK-3 inhibitor	Alzheimer’s disease, myotonic dystrophy	Phase II completed	Lovestone et al., 2015
LY2090314	GSK-3 inhibitor	acute leukemia, metastatic solid cancers	Phase II trials	Rizzieri et al., 2016; Gray et al., 2015
9-ING-41	GSK3-inhibitor	pediatric and adult cancers	Ongoing Phase I/II	Ugolkov et al., 2018
LiCl	GSK-3 inhibitor	bipolar disorders	FDA approved	O’Brien and Klein, 2009
Compound A	Skp2 inhibitor	hematologic malignancies	Preclinical	Chen et al., 2008
C1/C2	Skp2 inhibitor	soft tissue sarcoma	Preclinical	Wu et al., 2012; Li et al., 2020
C25	Skp2 inhibitor	T-ALL, other cancers	Preclinical	Chan et al., 2013; Rodriguez et al., 2020
Dioscin	Degradation of SKP2	colorectal carcinoma	Preclinical	Zhou et al., 2020

Beyond B-cell malignancies, GSK-3 inhibition has been shown to induce apoptosis via MYC-dependent mechanisms in certain cancer types, such as KRAS-mutant pancreatic adenocarcinoma, neuroblastoma, and glioma even as a monotherapy [123–125]. Other groups demonstrated that GSK-3 inhibition can increase the sensitivity of melanoma cells to the extrinsic ligand TRAIL and to inhibitors of mutant BRAF [126, 127]. Although in these studies the authors attributed the sensitization to the activation of the Wnt pathway, the caveat is that MYC itself is a well-known transcriptional target of that pathway [128], making it difficult to invoke MYC-independent effects. Additionally, the role of GSK-3 in cancer is complicated by contradictory findings that this protein promotes apoptosis in some cell lines while inhibiting it in others. Yet recent studies by several groups provide rationale and supporting evidence for the therapeutic targeting of GSK-3 in cancer ([129], also [130] and references therein, ).

Whatever pharmacological approach one might take, timing will be of essence. Certainly, long-term stabilization of MYC must be avoided, not only because of its pro-growth properties, but also because of well documented immunosuppressive effects of MYC, for example via controlling PD-L1 expressing (first reported in [131]; reviewed in [132]). In several murine models, inhibition of MYC allowed recruitment of immune effector cells, which contributed to anti-tumor effects (see for instance [7]). Curiously, in yet other mouse models, including the above-referenced AMPK study, pharmacological reactivation of MYC increased, not suppressed susceptibility to anti-PD-1 immunotherapy [93]. One possible explanation for this is that high levels of MYC have been documented to increase genomic instability [133], which at least in principle could lead to the increased expression of neoantigens. But regardless of the exact effects of MYC on the immune system, adaptive immune responses unravel slowly, within days, and are unlikely to be affected by two hour-long elevation of MYC levels, as achieved with GSK-3i [121]. In fact, one can envision a bi-phasic form of therapy where long-term MYC inhibition (to sustain anti-tumor responses) is combined with metronomic burst of short-term MYC induction scheduled to coincide with cycles of chemotherapy.

Repeated, “metronomic” administration of MYC agonists might also avoid another commonly encountered problem: that of clonal selection. First, short-term nature of MYC induction removes continuous pressure needed to select for MYC-resistant clones. Second, as drug selection often favors quiescent or dormant cells, which are inevitably characterized by low MYC expression, MYC agonists could reverse or at least ameliorate this unfavorable trend.

Concluding remarks

Work by many laboratories have shown that the ability of MYC to engage cell death pathways is the core feature of this oncoprotein. The pro-apoptotic activity of MYC could be leveraged for improved treatment outcomes even in chemoresistant tumors, if one considers that tumor cells require enough MYC to initiate tumorigenesis and sustain growth, but cannot have too much MYC, which would induce cell death. What constitutes “too much MYC” is likely going to be different among individual tumors, depending at least in part on the genetic make-up of each particular neoplasm (TP53 status, BCL2 expression, etc.) One can envision a simple scenario where tumors A, B, and C have already achieved their own maximum tolerated levels of MYC, and further increases, however small, would push them over the edge when combined with chemotherapy (Figure 3).

Figure 3. The role of MYC in development and cancer.

During normal development, MYC levels start out high and decline in aging tissues, with a concomitant decline in both proliferation and apoptosis (left). During the development of cancer, MYC levels are increased and so is neoplastic growth, which compensates for the increase in apoptosis and results in tumorigenesis (right). Both proliferative and apoptotic thresholds are not absolute and vary in individual neoplasms, resulting in inter-tumor heterogeneity. Here tumors A, B, and C have already achieved their own maximum tolerated levels of MYC, and further increases, however small, would push them over the edge when combined with chemotherapy.

Indeed, a recent study has demonstrated that transiently increasing MYC levels immediately prior to chemotherapy through GSK-3β inhibition with CHIR99021 [120] improves apoptotic response in p53-mutated, highly chemoresistant lymphomas [121]; similar data exist for solid cancers such as lung adenocarcinoma as well [129]. Besides CHIR99021, there is a handful of GSK-3 inhibitors currently in clinical trials [134], such as tideglusib which has undergone trials for Alzheimer’s disease [135], and LY2090314 which is in phase II trials for the treatment of acute leukemia [136] and has been tested in phase I clinical trials in many other cancer types [137]. The potential to re-purpose the FDA approved GSK-3β inhibitor lithium chloride [138] makes this adjuvant therapy strategy particularly viable as transitioning this psychiatric drug to cancer therapy would be relatively unchallenging. Of note, long-term usage of lithium chloride is not correlated with an increase in cancer incidence and appears to be a safe adjuvant [139]. Interestingly, inhibition of GSK-3α was recently shown to sensitize drug-resistant leukemias to asparaginase by a mechanism at least partly dependent of WNT signaling, which is the upstream regulator of MYC [140] and relies on it for some of its key functions in several organ systems [141].

In addition to GSK-3β, there are many other regulators of MYC protein stability that could be considered as therapeutic targets. Besides FBXw7, MYC is regulated by a host of E3 ubiquitin ligases including but not limited to Skp2 [reviewed by [116]. Thus, therapeutic inhibition of Skp2 [142–147] and other degradation complexes could potentially be another way to boost MYC expression and re-engage apoptosis. Furthermore, it was recently discovered that CDK9, which was already known to be required for transcription of the MYC gene, positively regulates MYC protein stability and prevents its degradation [148]. Although CDK9 agonists are yet to be developed, it is clear that there are numerous avenues to manipulating MYC levels for therapeutic benefit.

Finally, given that much of MYC pro-apoptotic effects are realized through the extrinsic apoptotic pathway, one could also revisit the use of death receptor agonists as anti-cancer agents. For example, TRAIL and its analogs had undergone extensive Phase I/II clinical trials where they well tolerated but displayed minimal efficacy against tumors [149]. Similar fate befell death receptor agonist antibodies such as mapatumumab [150]; however, the addition of a GSK-3β inhibitor could well bring these compounds back into clinical relevance. More broadly speaking, re-designing MYC synthetic lethality screens to incorporate standards of care appears to be a highly promising approach.

In fairness, many questions pertaining to MYC stabilization therapies remain unanswered (see Outstanding Questions Box). Thus, the goal of this review is not to advocate for the immediate use of GSK3i or similar molecules in the clinic, but rather to promote further research on the subject in preclinical models utilizing both chemotherapeutic drugs and death receptor agonists.

Outstanding Questions

• Are pro-apoptotic effects of MYC limited to certain tumor types, such as hematologic malignancies, or are they a hallmark of most, if not all, human cancers?

• What other small molecules, beyond GSK3i, could be used to hyperactivate MYC?

• How would tumors with stabilized MYC respond to radiation therapy?

• What effects would MYC stabilization therapy have on anti-tumor immunity and more specifically - on the efficacy of immune checkpoint inhibitors?

• Could tumors become resistant to MYC-and-chemotherapy-driven apoptosis?

• Could inhibitors and activators of MYC be combined into single therapeutic regimens to alternatily target cell proliferation and cell survival?

Highlights

• As genomic and transciptomic profiling of human cancers is becoming a routine diagonstic and prognostic tool, the MYC oncogene has emerged as a pervasive force in human cancers, whose gain-of-function alterations are apparent in the plurality of tumor types and specimens

• Thus, inhibitors of the MYC oncoprotein are under active development and show promise in model systems, but are yet to prove their utility in clinical settings

• On the other hand, a considerable body of evidence indicates that expressing too much MYC can be counterproductive for neoplastic growth, as it drives tumor cell apoptosis and correlates with favorable responses to chemotherapy

• Indeed, some very recent studies demonstrate that as antagonists of MYC degradation pathways (such as small molecule inhibitors of the GSK3 kinase) transiently elevate MYC levels, they confer chemosensity within that narrow window.