Getting MAD at MYC

The MYC protooncogene family (MYC, MYCN, and MYCL), known variously as super, master, or global transcription factors, affects the expression of an estimated 15% of the entire genome (1), which may in fact be an underestimate, given that Myc accumulates not just at specific DNA target sites but also in the promoter regions of most actively transcribed genes and leads to their amplification (2). Myc deregulation is associated with an astonishing 50% or more of human cancers and is often a predictor of aggressive disease and poor clinical outcome, making it an extraordinarily desirable therapeutic target (3, 4). An obstacle to this goal is that transcription factors generally associate with other proteins over large surface areas lacking contiguous interacting sites and hydrophobic pockets that enhance binding, making it difficult to directly target them with traditional small-molecule inhibitors (5). Although there have been numerous clinical trials evaluating means of blocking Myc (such as inhibition of MYC transcription or translation, the assembly of a transcriptionally active complex, and Myc protein stability), at this time, there is no therapy approved by the Food and Drug Administration that directly targets Myc (4). In PNAS, Li et al. (6) identify a molecular mechanism for the regulation of c-MYC expression and function and describe a small molecule that inhibits MYC by enhancing the expression of a functional MYC antagonist in the transcriptional regulatory network.

Myc has low affinity for its DNA response elements, referred to as “E-box” regulatory sequences. To be active, it must dimerize with another protein, Max, via a homotypic interaction between their common basic helix-loop-helix leucine zipper (bHLH/LZip) regions (7). The transactivating complex is antagonized by another heterodimer between bHLH/LZip-containing proteins, in this case, between Max and Mad1. Whereas the Myc/Max heterodimer recruits chromatin-modifying transcriptional activators including acetyl transferases, Mad/Max recruits a corepressor complex containing mSin3, N-Cor, and histone deacetylases. The functional outcome of Myc and Mad1 expression (e.g., proliferation vs. differentiation) is the result of a balance between activating (Myc/Max) and repressing (Mad/Max) heterodimers (8).

Given its central importance to so many cellular processes, it is no surprise that the expression of the proteins involved in Myc activity are tightly controlled; they must be turned “on” and “off ” in a highly temporally and quantitatively precise manner. The off in this case is achieved by limiting the stability of Myc and Mad1 proteins, each having a half-life on the order of minutes (9, 10). Their brief existence is predominantly controlled by the ubiquitin–proteasome system (UPS): Lysine (K)48-linked polyubiquitin chains are added by protein ubiquitin ligases (E3s) in concert with ubiquitin-conjugating enzymes (E2s), the modified molecules being destroyed in proteasomes. Countering this are ubiquitin-specific proteases that remove ubiquitin chains and prolong protein survival. The best-studied E3 involved in ubiquitination of Myc itself is Skp1-Cullin1-Fbox (SCF)^Fbxw7, although a variety of other ubiquitin protein ligases have also been implicated (10). As of this report, the situation with Mad1 is simpler, with a single E3 having been identified: cIAP1 (also known as BIRC2). cIAP1 belongs to the Inhibitor of Apoptosis (IAP) family, the members of which contain the defining Baculovirus IAP Repeat (BIR) protein-interaction motif (11). cIAP1 and the closely related cIAP2 also contain a C-terminal RING domain, which confers ubiquitin protein ligase activity. It is by virtue of their enzymatic function, rather than their extremely modest antiapoptotic activity, that cIAPs participate in a variety of signaling pathways, such as NF-κB activation, by promoting polyubiquitination of their targets, as well as by regulating their own levels by auto- or transubiquitination and degradation (11–14). In 2007, Yuan and coworkers (15) identified Mad1 as such a target in vitro and in cells and reported that cIAP1-dependent loss of Mad1 expression cooperated with Myc to promote tumor cell line colony formation. They proposed that cIAP1 could act as a tumor promoter by shifting the prevalence of Max dimers from those containing Mad1 to ones containing Myc, favoring cell survival and growth.

The study by Li et al. (6) uses the functional interplay between cIAP1 and MAD1 as a starting point to develop reagents that reverse the cell growth-promoting effects of MYC. They began by using so-called Smac mimetics, small molecules that bind cIAP1/2 to cause their dimerization, activation, autoubiquitination, and degradation (16–18). This resulted in dose-dependent up-regulation of c-MYC protein, but not mRNA, that correlated nicely with loss of cIAP1 expression. Recalling that Smac mimetics are also known as IAP antagonists, among the possible mechanisms compatible with this outcome are two simple but diametrically opposed Smac mimetic effects: (i) the loss of cIAP protein due to autoubiquitination/degradation, which would be consistent with the possibility that cIAPs are as-yet-unidentified E3s for MYC, or (ii) the initial increase of cIAP E3 activity resulting indirectly in the destabilization of MYC. To distinguish between these, the compounds were tested in cells in which cIAP expression was reduced or absent. Arguing against the first possibility, loss of cIAP1 did not increase c-MYC levels, but actually decreased them, something that was not seen when cIAP2 levels were manipulated. Moreover, in the absence of cIAP1, the Smac mimetics no longer increased c-MYC expression, nor was it increased when an E3-inactive cIAP1 mutant was overexpressed. These results indicate that it was the rapid Smac mimetic-induced increase in cIAP1 enzymatic activity that led to c-MYC up-regulation. Attention next focused on the known cIAP1 target, MAD1. It was unsurprising that MAD1 protein levels fell after treatment with Smac mimetics, but the finding that knockdown of MAD1 rendered the cells resistant to MYC up-regulation was intriguing. In fact, direct manipulation of MAD1 levels revealed an inverse relationship with c-MYC levels that correlated with changes in c-MYC ubiquitination. These and other observations led to the conclusion that c-MYC is shielded from cIAP1 when complexed with MAX, and its displacement by MAD1 results in rapid ubiquitination/degradation.

It follows that if one could decrease MAD1 degradation, there should be an increase in the MAD1-to-MYC ratio, favoring MAD1 in its competition for limiting MAX, which would then be exacerbated by the loss of “free” MYC. To this end, Li et al. (6) devised a high-throughput screen for inhibitors of cIAP1 autoubiquitination and selected one hit, named D19, that inhibited cIAP1 autoubiquitination and, importantly, transubiquitination of MAD1 for further study. Treatment with D19 increased MAD1 and decreased c-MYC levels, with a corresponding change in the ratio of c-MYC/MAX to MAD1/MAX. An insight into its mechanism of action was provided by what might at first be perceived as a paradoxical result: Rather than inhibiting cIAP1 binding to its cognate E2, which would be a simple mechanism of E3 inhibition, D19 bound to the RING domain and increased the stability of the interaction. To understand the inferred mechanism of inhibition, one must consider the dynamics of ubiquitination. Polyubiquitination is an iterative process in which a ubiquitin-charged E2 cooperates with the E3 to transfer a single ubiquitin moiety, disengages to be charged once again by an E1, and then rebinds the E3 to continue the process (19, 20). Preventing or delaying E2 dissociation from the E3 would impede multiple rounds of activity and thus decrease polyubiquitination. The D19-mediated change in cIAP1–E2 binding was selective, given that the drug had no obvious effect on global protein ubiquitination or p53 ubiquitination by MDM2, and its effect on MAD1 required the presence of cIAP1. Finally, D19 and its more potent derivative, D19-14, inhibited c-MYC–dependent gene transcription and cell transformation in cell lines and reduced c-MYC levels and tumor cell growth in a number of model systems and mice. Therefore, blocking cIAP1-mediated MAD1 ubiquitination and degradation decreases c-MYC activity, initially by altering the competition between the two for MAX, and then by destabilizing c-MYC itself, with therapeutic benefit in experimental systems (Fig. 1).

Fig. 1.

A model of how manipulating cIAP1 E3 activity alters the balance between MYC/MAX and MAD1/MAX based on the study of Li et al. (6). (A) At steady state, the normal balance between c-MYC/MAX-activating and MAD1/MAX-repressing heterodimers is maintained by cIAP1-mediated ubiquitination/degradation of MAD1, the latter indicated by a dotted outline and loss of opacity. (B) Smac mimetics (SM) enhance cIAP1 E3 activity, resulting in massive ubiquitination and degradation of MAD1, leaving MAX free to dimerize with c-MYC and tipping the balance to MYC/MAX transactivation. Note that this depiction reflects early consequences of cIAP1 activation. What effect the longer-term depletion of cIAP1 might have on MAD1 and MYC levels was not determined. (C) D19 stabilizes the interaction of cIAP1 with its cognate E2, preventing the addition of long ubiquitin chains, resulting in accumulation of inhibitory MAD1/MAX heterodimers.

A novel aspect of the D19 family of cIAP1 antagonists is that it is not enzymatic activity that is inhibited, but rather the dissociation of the cognate E2, rendering the E2–E3 complex unable to act processively in the creation of polyubiquitin chains. Although not directly demonstrated, one might predict that, in the presence of D19, there would be a decrease not only in the number of ubiquitinated MAD1 molecules but also in the average length of the ubiquitin chains that did form. It is worth contrasting this to Smac mimetics, also known as IAP antagonists. In actuality, Smac mimetics are rapid and potent agonists of cIAP1 E3 activity that induce NF-κB activation and TNF-α production, which in susceptible tumor cells, results in caspase activation and apoptosis. They are antagonists in a functional sense, however, in that the result of their agonist activity is the UPS-mediated degradation of the cIAPs themselves, which would of course prevent cIAP-dependent events in the longer term. Interestingly, Smac mimetics were found to increase c-MYC levels in apoptosis-resistant tumor cell lines, suggesting a possible mechanism for their refractoriness. D19, on the other hand, was effective at reducing c-MYC levels in cells resistant to Smac mimetics.

It has been said that MYC is undruggable because of its structure and cellular localization; thus, efforts to counter its role in neoplastic progression have focused on strategies that reduce its expression or interaction with other proteins (4). The study by Li et al. (6) has identified a molecular mechanism of MYC regulation, one that is indeed druggable, and whose manipulation provides an interesting strategy to attack MYC-dependent cancers.