Targeting the mTOR-DEPTOR Pathway by CRL E3 Ubiquitin Ligases: Therapeutic Application

Abstract

The mammalian target of rapamycin (mTOR), an evolutionarily conserved serine/threonine protein kinase, integrates both intracellular and extracellular signals and serves as a central regulator of cell metabolism, growth, proliferation, survival, and autophagy. The mTOR pathway is frequently activated in many human cancers, mainly resulting from alterations in the upstream regulators, such as phosphoinositide 3-kinase (PI3K)/AKT activation, PTEN loss or dysregulation of mTOR-negative regulators (e.g., TSC1/2), leading to uncontrolled proliferation. Thus, inhibiting the PI3K/AKT/mTOR pathways is widely considered as an effective approach for targeted cancer therapy. Recently, we and others found that DEPTOR, a naturally occurring inhibitor of both mTORC1 and mTORC2, was degraded by SCF (Skp1-Cullin-F box proteins) E3 ubiquitin ligase, the founding member of cullin-RING-ligases (CRLs), resulting in mTOR activation and cell proliferation. In addition to DEPTOR, previous studies have demonstrated that several other negative regulators of mTOR pathway are also substrates of CRL/SCF E3s. Thus, targeting CRL/SCF E3s is expected to cause the accumulation of these mTOR signal inhibitors to effectively block the mTOR pathway. In this review, we will discuss mTOR signaling pathway, how DEPTOR regulates mTOR/AKT axis, thus acting as a tumor suppressor or oncogene in some cases, how DEPTOR is ubiquitinated and degraded by SCF^β-TrCP E3, and how MLN4924, a small-molecule indirect inhibitor of CRL/SCF E3 ligases through blocking cullin neddylation, might be useful as a novel approach of mTOR pathway targeting for cancer therapy.

The mTOR and mTOR Signaling Pathway: An Introduction

The mammalian target of rapamycin (mTOR), an evolutionarily conserved serine/threonine protein kinase, integrates both intracellular and extracellular signals and serves as a central regulator of cell metabolism, growth, proliferation, survival, and autophagy [1,2]. In mammalian cells, mTOR forms two structurally and functionally distinct complexes, namely mTORC1 and mTORC2 [3,4]. Whereas mTORC1 consists of mTOR, Raptor, mLST8/GβL, PRAS40, and DEPTOR, mTORC2 is composed of mTOR, Rictor, GβL, Protor, Sin1, and DEPTOR [1,2]. It is well known that mTORC1 mainly promotes protein translation and cell growth by phosphorylating S6K1 and 4E-BP1, whereas mTORC2 regulates cytoskeletal organization [5] as well as cell survival through directly phosphorylating and activating AKT [3,4]. In addition, as a negative feedback loop, mTORC1 through S6K inhibits AKT by downregulating the expression of IRS-1/2 to block phosphoinositide 3-kinase (PI3K) activation [3,4]. Furthermore, mTORs are well-established negative regulators of autophagy [6,7]; whereas mTORC1 inhibits autophagosome formation by phosphorylating Ulk1/2 and Atg13 [8,9], mTORC2 represses the expression of some autophagy-related genes and other autophagy regulators [10–13]. Likewise, rapamycin, a potent inhibitor of mTOR, is commonly used as an autophagy inducer [6,14].

In response to multiple signals, including growth factors, nutrients, energy, and stress, mTOR is activated or inactivated, resulting in altered cellular processes [1,5,15]. On exposure to growth factors, mTOR is activated through 1) the PI3K-AKT pathway, leading to inactivation of TSC2, a negative regulator of mTOR [16], which complexes with TSC1 to inactivate Rheb GTPase, an mTORC1 activator [16–18]; and 2) the RAS-ERK-RSK pathway, which also phosphorylates and inhibits TSC2 [19,20]. We recently showed that RSK1 could directly target the mTOR complex by phosphorylating DEPTOR, an mTOR inhibitor [21], thereby promoting its degradation [22] (see below). In response to hypoxia, HIF-1 is induced to block mTORC1 through the HIF-1-REDD1 axis, which activates the TSC1–TSC2 complex through less well-defined mechanisms or a mechanism involving 14-3-3 [23,24]. Furthermore, TSC2 is activated by the cellular energy sensor AMP-activated protein kinase (AMPK). When the ratio of AMP/ATP is increased, AMPK is activated to phosphorylate and activate TSC2, leading to mTORC1 inactivation [25]. AMPK also inhibits mTOR by directly phosphorylating and inactivating Raptor in a TSC2-independent manner [26]. By phosphorylating and activating AMPK, the tumor suppressor LKB1 also significantly abrogates mTOR signaling [27]. Finally, mTORC1 can be activated by amino acids through Rag GTPases [28,29]. However, the upstream activators of mTORC2 are ill defined, but it is generally thought that mTORC2 is activated directly or indirectly only by growth factors [2]. Figure 1 briefly summarizes the signaling pathways that respond to various stimuli, leading to mTOR activation and subsequent cell growth and proliferation, survival, and autophagy inhibition.

Figure 1.

The mTOR signaling pathway. Arrows stand for activation; bars represent inhibition. represents phosphorylation. For details, see text.

DEPTOR, an mTOR Inhibitor: A Tumor Suppressor or an Oncogene?

Recently, Sabatini's group discovered a novel mTOR binding partner, designated as DEPTOR in reference to the presence of two tandem N-terminal DEP (disheveled, egl-10, pleckstrin) domains with unknown function [30] and its specific interaction with mTOR [21]. DEPTOR also contains a C-terminal PDZ (postsynaptic density 95, discslarge, zonula occludens-1) domain [31], which is often involved in protein-protein interaction. Indeed, DEPTOR binds to FAT domain on mTOR through its PDZ domain [21]. Significantly, DEPTOR binds to both mTORC1 and mTORC2 and inhibits their activities, as measured by in vitro kinase assay as well as in cell-based assays. Nevertheless, the messenger RNA (mRNA) and protein levels of DEPTOR are negatively regulated by the activities of both mTORC1 and mTORC2, establishing a regulatory feedback loop [21].

By blocking the activity of mTOR, DEPTOR acts, in general, as a tumor suppressor through inhibiting protein synthesis, cell proliferation, and survival. Indeed, small interfering RNA knockdown of DEPTOR increased the activities of both mTORC1 and mTORC2, as evidenced by increased S6K1/AKT phosphorylation, and thereby promoting cell proliferation and survival [1,21] (Figure 2A). However, under certain circumstances, DEPTOR could act as an oncogene [21,32]. High levels of DEPTOR inhibited mTORC1 and activated AKT by relieving the feedback inhibition from S6K1 to PI3K, as demonstrated by reduced S6K1 phosphorylation and increased AKT phosphorylation, thus promoting the survival of cancer cells [21,22] (Figure 2B). It is not surprising that the levels of DEPTOR are downregulated inmost cancers because the mTOR pathway negatively regulates DEPTOR mRNA expression and protein stability [21]. In a subset (∼28%) of human multiple myelomas, however, DEPTOR was overexpressed. In these cells, high levels of DEPTOR were required for the maintenance of PI3K/AKT activation and small interfering RNA knockdown of DEPTOR induced apoptotic cell death [21]. Consistently, increased DEPTOR rendered cancer cells more resistant to chemotherapeutic drugs [22,33,34]. Finally, DEPTOR overexpression was found in hepatocellular carcinoma and differentiated thyroid carcinoma, which correlated with the poor prognosis of patients [35,36]. These conflicting data suggest a complex role of DEPTOR in the regulation of mTOR signaling pathways as well as other yet-to-be-identified pathways in a tissue-specific and cell context-dependent manner. The in vivo physiological role of DEPTOR as a tumor suppressor or oncogene will only be clearly demonstrated in mouse knockout studies by determining the effect on tumorigenesis on Deptor deletion (total or tissue specific) alone or in combination with, for example, Ras activation or Pten loss.

Figure 2.

DEPTOR has tumor suppressive and oncogenic properties. (A) DEPTOR acts as a tumor suppressor. (B) DEPTOR acts as an oncogene. For details, see text.

DEPTOR Ubiquitination and Degradation by SCF^β-TrCP E3 Ubiquitin Ligase

Although it was established that mTOR promotes DEPTOR degradation on mitogen stimulation [21], the E3 ubiquitin ligase, responsible for DEPTOR degradation, was unknown until most recently. Three independent groups, including ours, identified SCF^β-TrCP E3 ubiquitin ligase is the E3 that promotes the ubiquitination and degradation of DEPTOR in response to serum stimulation [22,33,37].

Cullin-RING ligases (CRLs) with the founding member, SCF E3, are the largest E3 ubiquitin ligase family. CRLs seem to account for ∼20% of total protein degradation by the ubiquitin-proteasome system [38]. By promoting ubiquitination and degradation of a variety of key substrates, CLRs control a number of important biologic processes including cell cycle progression, signal transduction, gene transcription, embryonic development, genomic integrity, and tumor suppression [39–41]. In mammalian cells, there are a total of eight members of the cullin family of proteins (CUL1, 2, 3, 4A, 4B, 5, 7, and PARC) [42], which individually binds to an adaptor protein (such as SKP1 and DDB1) and a substrate receptor protein (such as F-box protein) at the N-terminus, and a RING protein, RBX1 or RBX2 (also known as ROC2 or SAG) at the C-terminus [43]. It is well established that the substrate specificity of CRLs/SCF E3s is determined by substrate receptor proteins, such as F-box proteins [44], whereas the core ligase activity is possessed by the cullins-RBX1/2 complex [45].

Now how does SCF^β-TrCP promote DEPTOR ubiquitination and degradation? It is well established that in order for F-box proteins to bind to its substrates for subsequent ubiquitination and degradation, the substrates have to be phosphorylated first at the serine or threonine residue(s) on a consensus-binding motif of a given F-box protein [39]. In the case of DEPTOR, we did identify an evolutionarily conserved consensus β-TrCP binding motif (DSGXXS) on residues 286 to 291 (SSGYFS). The only mismatch here is the replacement of the conserved aspartic acid (D) by a serine (S) (see Figure S1A in Zhao et al. [22]), which, nevertheless, behaves as aspartic acid on phosphorylation. The exact same sequence of the binding motif was identified on BimEL, a previously characterized substrate of SCF^β-TrCP [46]. We further identified that kinases S6K1, a downstream target of mTORC1, and RSK1 are responsible for DEPTOR phosphorylation at the S²⁸⁶-S²⁸⁷-S²⁹¹ residues, with S6K1 acting as a prime kinase, although which kinase actually phosphorylates which serine residue remains to be defined [22,47]. Conversely, two other groups reported that mTOR and CK1α are the responsible kinases with mTOR acting as a prime kinase to facilitate DEPTOR phosphorylation on three serine residues at the binding motif (²⁸⁶SSGYFS²⁹¹) by CK1α [33,37]. To further clarify this discrepancy, we performed an in vitro kinase assay for DEPTOR phosphorylation on these three serine residues (Ser^286,287,291), using a specific phosphor antibody raised against a peptide with all three serine residues phosphorylated [22]. As shown in Figure 3A, immune-purified FLAG-tagged DEPTOR was phosphorylated by mTOR (slightly) and CK1α alone or in combination or by the combination of S6K1 and RSK1. Thus, the combination of either mTOR/CK1α or S6K1/RSK1 was able to phosphorylate this binding motif in vitro. The in vivo DEPTOR phosphorylation for its targeted degradation, however, will likely be triggered by a different set of kinases activated in response to a given stimulus. Taken together, these combined results indicate that mTOR positively regulates its own activity by inducing the ubiquitination and degradation of DEPTOR, one of its own endogenous inhibitors, either directly or through its downstream target S6K1 [22,33,37,47]. On phosphorylation on β-TrCP binding motif, DEPTOR is bound to β-TrCP, which recruits E2-loaded RBX1-Cullin 1 complex, catalyzing ubiquitin transfer from E2 to DEPTOR. Multiple runs of this E1/E2/E3-mediated chain reaction cause DEPTOR polyubiquitination, which is then recognized by the 26S proteasome for subsequent degradation (Figure 3B) [22,33,37].

Figure 3.

DEPTOR is ubiquitinated by SCF^β-TrCP E3 ligase on phosphorylation. (A) In vitro phosphorylation of DEPTOR on the consensus-binding motif. FLAG-tagged DEPTOR was transfected into 293 cells, purified by IP using bead-conjugated anti-FLAG Ab, and incubated with active CK1 or mTOR alone or in combination, as well as the combination of active S6K1 and RSK1 in a kinase reaction mixture. DEPTOR phosphorylation on Ser^286,287,291 was detected by immunoblotting using a DEPTOR phosphor-specific Ab [22]. (B) Schematic summary of DEPTOR ubiquitination by SCF^β-TrCP E3 ubiquitin ligase and degradation by 26S proteasome. In response to mitogen, RSK1 and S6K1 are activated to phosphorylate DEPTOR at β-TrCP degron site [22]. Conversely, DEPTOR is phosphorylated by mTOR at non-degron site, which acts as primer for subsequent phosphorylation at degron site by CK1α [33,37]. Phosphorylated DEPTOR is then recognized by β-TrCP for subsequent ubiquitination by SCF E3 and degradation by proteasome.

CRLs/SCF E3 Ligases Degrade Multiple Components of mTOR Pathway

In addition to DEPTOR, a variety of components, both upstream and downstream, of mTOR signaling pathway were found to be the substrates of CRLs/SCF E3 ligases (Figure 4A). First of all, mTOR itself was reported to undergo ubiquitin-mediated degradation by Cul1-Skp1-Fbw7 E3 ligase, although the kinase(s) responsible for prerequisite phosphorylation was not defined [48]. Second, TSC2 was identified as the substrate of Cul4-DDB1-Fbw5 E3 ligase that mediated TSC2 protein stability and TSC complex turnover [49]. Third, REDD1 was rapidly degraded by Cul4A-DDB1-β-TrCP E3 ligase with GSK-3β as a corresponding kinase for prerequisite phosphorylation [50]. Fourth, HIF-1α was a well-characterized substrate of Cul2-Elongin BC-VHL E3 ligase [51,52]. Fifth, IRS1 can be ubiquitinated and degraded either by Cul7-Skp1-Fbw8 in a manner dependent of mTOR and S6K [53] or by Cul1-Skp1-Fbxo40 E3 ligase with ligase activity enhanced by increased tyrosine phosphorylation of IRS1 [54]. Sixth, PHLPP1, a protein phosphatase that dephosphorylates activated AKT, was found to be substrate of SCF^β-TrCP on phosphorylation by CK1 and GSK-3β [55]. Finally, PDCD4, a downstream target of mTOR/S6K1, was identified as a substrate of Cul1-Skp1-β-TrCP E3 ligase on prerequisite phosphorylation at Ser67 by S6K1 [56]. Thus, given the fact that all these CRLs E3 substrates, except mTOR itself and IRS1, are the negative regulators of mTOR pathway, the small-molecule inhibitor of CRLs/SCF E3 ligases would likely cause the accumulation of these negative regulators to shut down mTOR pathways, leading to suppression of cell proliferation and survival.

Figure 4.

CRLs/SCF E3 ubiquitin ligases regulate mTOR pathway. (A) Schematic summary of CRLs/SCF E3 substrates in mTOR signaling pathway. (B) Schematic summary of cullin neddylation. (C) Schematic summary of ubiquitination reaction catalyzed by SCF E3 ubiquitin ligases and potential targeting points for inhibition. For details, see text.

Targeting mTOR Pathway by Inactivating CRL/SCF E3s: Therapeutic Application

The mTOR pathway is frequently and constitutively activated in many types of human cancer mainly through activation of PI3K/AKT and inactivation of PTEN in the upstream signaling pathways [3,5,57,58]. The PI3K/AKT signaling is activated through a variety of mechanisms, including mutations or amplifications of cell surface receptor tyrosine kinases (RTKs), such as EGFR [59–61] and IGFR [62,63], the gain-of-function mutations in the PIK3CA gene itself, which encodes the p110a catalytic subunit of PI3K [64–66], and loss of expression of the tumor suppressor PTEN, which encodes a lipid and protein phosphatase [67]. PTEN inactivation is found in many cancers, such as carcinomas of breast, endometrium, thyroid, and prostate as well as melanoma, leukemia, lymphoma, and glioblastoma, through several mechanisms, including mutations, deletions, loss of heterozygosity, methylation, and microRNA silencing [67]. Furthermore, mTOR activation was also found in several benign tumor syndromes, including tuberous sclerosis (TSC1/2 mutations) [25], Peutz-Jeghers syndrome (LKB1 mutations) [27], Cowden syndrome (PTEN mutations) [68], and neurofibromatosis (NF1 mutations) [69]. Thus, PI3K/AKT/mTOR pathways are considered as attractive targets for effective cancer treatment [64,70–74].

Given the association of mTOR with a wide variety of diseases including cancer, metabolic disorders, neurologic diseases, and autoimmune diseases [2], mTOR inhibitors may therefore have a broad application in the treatment of these diseases. Rapamycin (sirolimus) and its analogs (rapalogs), which include temsirolimus (CCI-779) [75], everoliums (RAD001) [76], and deforolimus (AP23573) [77], are the first generation of mTOR inhibitors [73], which form a complex with the immunophilin FKBP12 and inhibit mTORC1 activity through binding to FRB (FKBP12-rapamycin-binding) domain on mTOR [78]. Numerous clinical trials are ongoing to test their efficacy as anticancer agents, but with limited success [73,79,80]. The cancer resistance to rapamycin type of drugs could be attributable to the existence of the negative feedback mechanism by which inactivation of mTORC1/S6K1 releases S6K1 inhibition on IRS-PI3K, leading to activation of AKT for cancer cell survival [73]. Thus, inhibition of mTORC1 reduces cell growth but could improve the survival [81]. Recently, ATP-competitive inhibitors, such as Torin1 [82], PP242 and PP30 [83], which directly target the catalytic site of mTOR kinase and inhibit both mTORC1 and mTORC2 kinase activity, were developed as the second generation of mTOR inhibitors, currently in early stage of clinical trials with a better anticancer efficacy than rapalogs in preclinical settings [80,84].

An alternative approach of blocking mTOR or mTOR pathway, which has never been tested before, would be through modulating the stability of mTOR-negative regulators. Given the fact that most mTOR signal blockers, such as DEPTOR, HIF-1α, REDD1, TSC2, and PDCD4, are substrates of CRLs/SCF E3 ligases, small-molecule inhibitors of the CRL/SCF E3s are expected to cause their accumulation, thus abrogating mTOR pathway, particularly in human cancers with constitutive mTOR activation. Unfortunately, specific small-molecule inhibitor against the multicomponent CRL/SCF E3 ligases has not been discovered to test this hypothesis. However, MLN4924, a newly discovered small-molecule inhibitor of NEDD8-activating Enzyme (NAE) [38], could act as an indirect inhibitor of CRL/SCF E3s, thus to be used as an alternative option.

Like ubiquitination, neddylation requires E1 NEDD8-activating enzyme, E2 NEDD8-conjugating enzyme (UBC12), and E3 NEDD8 ligase to catalyze the transfer of NEDD8 to a target molecule [85,86] (Figure 4B). Inhibition of either enzyme in the reaction would abrogate substrate neddylation. MLN4924 is a newly discovered small-molecule inhibitor of NAE [38]. MLN4924 binds to NAE to create a covalent NEDD8-MLN4924 adduct, which cannot be further used in subsequent intraenzyme reactions, thus blocking NAE enzymatic activity [87] (Figure 4B). Because the activity of CRL/SCF E3s requires cullins neddylation, which disrupts inhibitory binding by CAND1 and retains the CRLs/SCFs in an active conformation [88–91], MLN4924 became the first-in-class indirect inhibitor of CRL/SCF E3s [38]. By inactivating CRL/SCF E3s, MLN4924 causes the accumulation of a number of SCF E3 substrates to inducing apoptosis [38,92–95] and senescence [96–98], thus inhibiting tumor growth both in vitro and in vivo. We used MLN4924 as an alternative mTOR pathway inhibitor and found that it effectively blocked the activity of mTORC1 in both time- and dose-dependent manners in almost all tested human cancer cell lines derived from carcinomas of breast, prostate, colon, brain, and cervix (unpublished observations). Thus, antitumor activity of MLN4924 could be attributable to its inhibition of the mTOR pathway. It is worth noting that among these reported substrates of CRL/SCF E3s in the mTOR pathway, HIF-1α, DEPTOR, and PDCD4 were most frequently accumulated, whereas mTOR itself, TSC2, and REDD1 were never accumulated on MLN4924 treatment in all tested cell lines under our assay conditions (unpublished observations). The differential MLN4924 responses, in substrate accumulation among tested cancer lines, are likely attributable to cell line- and cell context-dependent availability of F-box/adaptor proteins as well as of kinases prerequisite for substrate phosphorylation.

It should be clearly noted that, although MLN4924 demonstrated a universal inhibitory activity against mTORC1, it suffers from a major specificity issue as an mTOR pathway blocker. First of all, there are thousands of CRL/SCF substrates, given the fact that CRLs can be assembled into more than 200 different E3 enzymes by different combinations of its multiple components [39,99]. Apparently, some of the substrates are oncogenes, whereas others are tumor suppressors [40,100]. The net biologic outcome of MLN4924 action (such as induction of apoptosis, senescence, or even autophagy; unpublished observation) would likely depend on the interaction of these substrates in a cell context-, temporal-, and spatial-dependent manner. Second, MLN4924 is an NAE inhibitor and would likely inhibit other cellular neddylation reactions (much to be learned) [85,86], although cullins are the only physiological substrates [85,101]. Third, it has been recently reported that cancer cells could develop the resistance to MLN4924 by selection of rare clones with heterozygous mutations in targeting enzyme NAEβ [102,103]. Nevertheless, MLN4924 is the first-in-class and only indirect inhibitor of SCF E3 ligases, which demonstrated potent anticancer activity with limited toxicity in mice in preclinical settings [38,104] and has been advanced to several phase 1 clinical trials [105,106].

Future Perspectives on DEPTOR Biology and CLR/SCF E3 Inhibitors

Although activation of PI3K-AKT-mTOR pathway, resulting from PI3K/AKT mutation and/or PTEN loss, plays a key role in promoting cell proliferation/survival and tumorigenesis in many human cancers [64,67,70,107], it is totally unknown how this oncogenic pathway is regulated by DEPTOR under in vivo physiological settings, particularly given the fact that DEPTOR could act as a tumor suppressor or oncogene in cell culture settings [1,21]. Several DEPTOR biologic questions remain to be addressed: 1) Is Deptor a tumor suppressor or an oncogene in vivo when acting alone? 2) Can Deptor affect tumorigenesis triggered by activation of PI3K-AKT-mTOR axis as a result of Ras activation or Pten loss? 3) Can targeting DEPTOR degradation suppress tumorigenesis in vivo? The answers to these fundamental questions by the use of genetically modified mouse models and human tumor xenograft models would certainly further our current understanding of DEPTOR function and its regulation of PI3K-AKT-mTOR axis, and provide molecular basis for rational drug design in targeting this oncogenic axis.

With regard to specific inhibitors of CRL/SCF E3s, which have been validated as the attractive anticancer targets [108,109], encouraging progresses are being made. Examples include the discovery of Cpd A that disrupted Skp2-p27 binding to abrogate p27 degradation [110] and discovery of few SCF-Cdc4 inhibitors [111,112] and a Cdc34 inhibitor [113]. Other options include 1) disruption of E2 binding with RING proteins, RBX1/RBX2; 2) disruption of receptor binding with tumor suppressive substrates; and 3) inhibition of kinases that phosphorylate tumor suppressor substrates for targeted degradation (Figure 4C). It is hoped that specific inhibitors targeting CRL/SCF E3s will be eventually discovered and further developed as novel anticancer agents for the treatment of human cancers with activated CRL/SCF E3s.