Killing by Degradation: Regulation of Apoptosis by the Ubiquitin-Proteasome-System

Apoptosis is a cell suicide process that is essential for development, tissue homeostasis and human health. Impaired apoptosis is associated with a variety of human diseases, including neurodegenerative disorders, autoimmunity and cancer. As the levels of pro- and anti-apoptotic proteins can determine the life or death of cells, tight regulation of these proteins is critical. The ubiquitin proteasome system (UPS) is essential for maintaining protein turnover, which can either trigger or inhibit apoptosis. In this review, we will describe the E3 ligases that regulate the levels of pro- and anti-apoptotic proteins and assisting proteins that regulate the levels of these E3 ligases. We will provide examples of apoptotic cell death modulations using the UPS, determined by positive and negative feedback loop reactions. Specifically, we will review how the stability of p53, Bcl-2 family members and IAPs (Inhibitor of Apoptosis proteins) are regulated upon initiation of apoptosis. As increased levels of oncogenes and decreased levels of tumor suppressor proteins can promote tumorigenesis, targeting these pathways offers opportunities to develop novel anti-cancer therapies, which act by recruiting the UPS for the effective and selective killing of cancer cells.


Apoptosis (Programmed Cell Death)
Apoptosis is a morphologically and mechanistically distinct cell death program that is essential for the elimination of unwanted and damaged cells during development and tissue homeostasis [1][2][3]. Abnormal regulation of this process is associated with a wide variety of human diseases, including immunological and developmental disorders, neurodegeneration and cancer [2][3][4]. Apoptosis is executed by caspases, enzymes that are activated following cleavage from their inactive pro-caspase form [5]. Caspase-dependent cell death is the hallmark of apoptosis. However, several alternative modes of non-apoptotic cell death have been described that do not involve caspases [6][7][8][9][10][11]. Apoptosis is regulated by two major pathways: the extrinsic pathway and the intrinsic (mitochondrial) pathway [1,12]. The extrinsic pathway is initiated when apoptotic-inducing ligands bind to death receptors [13]. The intrinsic pathway is mainly induced by internal apoptotic signals, such as DNA damage, as well as by certain external stimuli, such as nerve growth factor (NGF) withdrawal-induced cell death [14]. There is an essential crosstalk between the extrinsic and intrinsic pathways, for example, via caspase-induced-cleavage of BID. Truncated Bid (tBID) is initially cleaved in the extrinsic pathway by caspase-8 and can amplify the apoptotic signal by translocating to the mitochondria and inducing mitochondrial outer membrane permeabilization (MOMP) [15][16][17]. This in turn leads to the release of Cytochrome c and formation of the apoptosome complex, which activates caspase-9 and, subsequently, downstream effector caspases [5,[16][17][18][19]. This initiates a proteolytic cascade, which culminates in the cleavage of substrate proteins, leading to the disassembly of the cell [17]. In living cells, caspases are kept controlled by inhibitors of apoptosis (IAP) proteins. The IAP proteins prevent cell death by binding to and inhibiting active caspases [20]. XIAP (X-linked IAP), is the best studied IAP, which has three Baculovirus IAP Repeat the MDM2 RING domain binds the internal ribosome entry site (IRES) region in XIAP mRNA, thereby preventing homodimerization of MDM2 and its autoubiquitylation. This results in increased MDM2 stability and XIAP translation, leading to cell survival [157,165]. Moreover, p53 promotes the expression of the ubiquitin domain-containing 1 (UBTD1) protein, which in turn induces MDM2 degradation through a positive feedback loop mechanism [166]. Although MDM2 is considered the primary E3 ligase of p53, other E3 ligases were shown to control the stability of p53. Some E3 ligases can promote the degradation of p53 independently of MDM2, such as Pirh2, COP1, ARF-BP1, CHIP, TOPORS, Synoviolin and Carps (see Table 1 for a complete list). Three additional E3 ligases, TRIM28, RNF2 and Cul4a, can promote p53 degradation by interacting with MDM2.  Regulation of apoptosis by the UPS also involves other proteins that modulate E3 ligase activity and thereby control the levels of pro-apoptotic proteins. MDM2 is an unstable protein that undergoes ubiquitylation and degradation in an autocatalytic manner [152,153]. MDM2 exhibits increased E3 ubiquitin activity and self-ubiquitylation when it forms homo-or hetero-dimers [154,155]. Moreover, MDM2 stability increases when cer-  Regulation of apoptosis by the UPS also involves other proteins that modulate E3 ligase activity and thereby control the levels of pro-apoptotic proteins. MDM2 is an unstable protein that undergoes ubiquitylation and degradation in an autocatalytic manner [152,153]. MDM2 exhibits increased E3 ubiquitin activity and self-ubiquitylation when it forms homo-or hetero-dimers [154,155]. Moreover, MDM2 stability increases when cer- Figure 2. Each of these feedback loops consists of proteins whose levels and function are influenced by the activation or inhibition of their E3 ligase. Arrows indicate stimulatory interactions, whereas horizontal bars denote inhibitory influences. p53-induced RING-H2 protein (Pirh2) is an important transcriptional target of p53. Pirh2 can physically interact with p53, inducing its ubiquitylation and degradation independently of MDM2, which initiates a negative feedback loop ( Figure 2) [167][168][169]. Various kinds of cancers express high levels of Pirh2, which is associated with poor prognosis and survival rate [169][170][171]. Using hepatocellular carcinoma samples, the SCYL1-binding protein 1 (SCYL1BP1) was found to promote Pirh2 degradation and thereby restore p53 levels [172]. The COP1 protein contains an N-terminal RING finger and can ubiquitylate p53 independently of MDM2 and Pirh2 under stressed and unstressed conditions [173]. However, COP1 hypo/− mice do not exhibit any dysregulation in p53 levels, suggesting that COP1 does not play a role in regulating the levels of p53 [174]. ARF-BP1 (HUWE1, MULE) is a HECT E3 ligase that can also regulate p53 levels independently of MDM2 [175]. Silencing of the ARF-BP1 expression by RNAi in U2OS cells resulted in p53-dependent apoptosis [175]. The TRIM protein family has an N-terminal RING finger domain and at least one B-box zinc finger domain [176]. TRIM28 can induce p53 ubiquitylation and degradation via its interaction with MDM2 [177]. Yet, silencing TRIM28 caused an increase in the expression of p53 target genes [177,178]. U-box E3 ligases can also induce p53 ubiquitylation and degradation. On example is CHIP (carboxyl terminus of Hsc70-interacting protein), a chaperone-interacting protein that has E3 ligase activity and is responsible for the ubiquitylation and degradation of various proteins via its C-terminal U-box [179,180]. CHIP can promote p53 degradation by interacting with Hsc70 [179,180]. CHIP ubiquitylates and degrades both wild-type and mutant p53 via both the proteasome and the lysosome pathways [181,182]. Since p53 regulates various cellular processes, it has been suggested that each E3 ligase is assigned for regulating p53 under certain conditions or in specific tissues or cell types [183]. p53 promotes apoptosis mainly by inducing the transcriptional upregulation of proapoptotic proteins, such as the death receptor 5 (DR5), TNFR1 and Fas, which results in the activation of caspase-8 [184,185]. In addition, p53 can induce the activation of pro-apoptotic Bcl-2 family proteins, such as BAX, PUMA, BAD, BID, BAK and NOXA [186][187][188][189][190]. The induction of BID, BAK and BAX promotes the permeabilization of the outer mitochondrial membrane and amplifies the caspase activation process [191,192]. p53 also increases the transcription of the pro-apoptotic XIAP-antagonist, ARTS, which relieves caspases from inhibition by XIAP, leading to the cleavage of BID and MOMP [190,193].
In living cells, Bax is mainly present in the cytosol. However, in response to apoptotic stimuli, Bax is activated by undergoing a conformational change that causes translocation to the mitochondrial outer membrane (MOM) [234][235][236]. There, BAX binds to BAK to initiate MOMP, allowing the release of pro-apoptotic proteins. These include Smac/Diablo (Smac) and Cytochrome c (Cyto c), which normally reside in the mitochondrial inner membrane space (IMS) [237]. Since activation of Bax and its translocation to the MOM act as a major regulatory checkpoint in apoptosis, Bax protein levels are tightly controlled by the UPS [238][239][240]. The in-between RING (IBR) domain containing 2 (IBRDC2) E3 ligase induces the degradation of BAX in response to p53-mediated apoptosis [241]. IBRDC2 is highly specific for BAX, as it does not bind to BAK, PUMA or NOXA [241]. Johnson et al. identified Parkin as another E3 ligase that can ubiquitylate BAX, thereby limiting the mitochondrial pool of BAX under non-apoptotic and stress conditions ( Figure 1) [242].

Targeting the UPS for Apoptosis-Induced Cancer Therapy
Cancer cells engage the UPS for degrading pro-apoptotic proteins [307,308]. Hence, evading apoptosis by tipping the balance between pro-and anti-apoptotic proteins may allow initiation of tumorigenesis [4,[309][310][311][312]. Intensive efforts are being made to use the UPS for killing cancer cells. The original strategy was to develop general proteasome inhibitors for the treatment of multiple myeloma, which lead to several FDA-approved drugs for the treatment of multiple myeloma (see below) [313][314][315]. This success has encouraged additional efforts to develop new cancer therapeutics by targeting UPS-mediated protein degradation [316][317][318].

Proteasome Inhibitors
At this time, three proteasome inhibitors have been approved by the FDA for the treatment of multiple myeloma: Bortezomib (Velcade, PS341), its second-generation derivative Carfilzomib (Kyprolis) and Ixazomib (MLN9708, Ninlaro) [313,314]. Bortezomib was the first proteasome inhibitor to be approved by the FDA in 2003 [319][320][321][322]. Bortezomib inhibits the 20S proteasome subunit, affecting several vital cellular pathways, including the NF-κB signaling, thereby promoting apoptosis [320,323]. Moreover, Bortezomib increases the levels of the pro-apoptotic protein NOXA [324]. Carfilzomib was the second proteasome inhibitor to be approved by the FDA in 2012 [325,326]. Carfilzomib is a more potent inhibitor of the proteasome when compared to Bortezomib, and it is effective against Bortezomib-resistant multiple myeloma [325,326]. Carfilzomib is thought to induce apoptosis by increasing NOXA levels, which results in the activation of capase-3 and -7 [327]. Unfortunately, despite its effectiveness, Carfilzomib shows dose-limiting toxicities [112].
Ixazomib is the first orally administered proteasome inhibitor; it is as effective as Bortezomib, with respect to inhibiting the proteasome, but has better pharmacokinetic properties. Ixazomib was approved by the FDA and is administered in combination with lenalidomide and dexamethasone in patients with relapsed and refractory myeloma [328,329]. Significantly, Ixazomib can induce apoptosis in Bortezomib-resistant multiple myeloma patients [330]. Delanzomib is another orally administered proteasome inhibitor that inhibits NF-κB signaling and can promote apoptosis in multiple myeloma [331]. Delanzomib is also more effective than Bortezomib in treating normal human epithelial bone marrow progenitor and bone marrow-derived stromal cancer cells [331]. Despite its relative effectiveness, phase II clinical trials with Delanzomib were terminated due to considerable toxicity [332]. Another proteasome inhibitor is Oprozomib, an oral tripeptide epoxyketone. Oprozomib has a longer half-life than Bortezomib and causes activation of caspases-9, -3 and -7 and apoptosis [333][334][335]. Finally, Marizomib is the first natural proteasome inhibitor derived from Salinosporamide tropica, a marine actinomycete bacterium [336]. Marizomib causes irreversible inhibition of the 20S proteasome when tested in in vitro and in vivo models [336]. Both Oprozomib and Marizomib underwent clinical trials as a single agent or in combination with other drugs [329]. Despite the relatively high efficacy of general proteasome inhibitors to treat multiple myeloma, their long-term use is limited due to aqcuired resistance towards these compounds [337].

Cancer Therapies Targeting p53 and MDM2 for Degradation
Because of their vital role in maintaining protein turnover, E3 ligases are also complicit in assisting tumorigenesis. Therefore, E3 ligases present promising drug targets for cancer treatments. The inhibition of E3 ligases is supposed to be more target-specific and show lower toxicity compared to general proteasome inhibitors [316].
MDM2 is essential for restricting the levels of p53. Various cancers overexpress MDM2 to hinder the p53-mediated pathway, thus resulting in tumor progression [338,339]. Therefore, MDM2 became an emerging target for developing cancer treatments [316]. The Nutlin small molecules are a family of cis-imidazoline analogs first described as selective and potent inhibitors of MDM2 [340,341]. Nutlins can occupy the binding site of p53 in MDM2 and allow p53 to escape MDM2-mediated ubiquitylation and degradation [340].
Amongst the Nutlin family members, only the enantiomer Nutlin-3a exhibited a potent binding ability to MDM2. Yet, these molecules were not effective enough to be further examined in clinical trials [342]. Resolving the crystal structure of Nutlin-3a facilitated the discovery of better MDM2-p53 inhibitors, such as RG7112, which is currently in phase I clinical trials [343][344][345]. Other small molecules that target the interaction between MDM2 and p53 have been developed and are under various stages of clinical trials. These include AMG-232, APG-115, BI-907828, CGM097, RG7388, DS-3032b and HDM201 [346][347][348].
More recently, intense efforts have been made to develop drugs that mediate protein degradation, rather than just binding and inhibiting target protein activity. A major advantage of this approach is the ability to target what were classically considered "nondruggable proteins" [354][355][356][357]. Proteolysis Targeting Chimeras (PROTACs) are small molecules that can promote the ubiquitylation of target proteins by directing specific E3 ligases to specific substrates [318,358,359]. Mechanistically, PROTACs form a ternary complex by binding the protein of interest and an E3 ligase, resulting in the ubiquitylation and degradation of the target protein [359]. For example, Bromodomain Extra-Terminal chimeric molecules (BET-PROTACs), such as ARV-825 and ARV-771, are capable of binding specific target proteins and inducing their ubiquitylation and degradation [360]. Recently, Bcl-xL, Mcl-1 and Bcl-2 PROTACs have been developed [361][362][363][364][365]. DT2216 is the most promising Bcl-xL-specific PROTAC and brings Bcl-xL to the Von Hippel-Lindau (VHL) E3 ligase for degradation by the proteasome [362]. DT2216 is derived from the ABT-263 and showed higher selectivity to kill Bcl-xL-dependent cancer cells than ABT-263 [361,362]. The dMcl1-2 and C3 PROTACs induce Mcl-1 degradation by bringing it into close proximity to cereblon (CRBN) cullin-4A RING E3 ligases [364,365]. In addition, C5 PROTAC was shown to potently and selectively induce the ubiquitylation and proteasomal degradation of Bcl-2 [364].

Cancer Therapies Targeting IAPs for Degradation
IAPs are overexpressed in various tumors, making them attractive drug targets for cancer therapy [366][367][368]. In the past few years, efforts have been made to target IAPs, and specifically XIAP, by small molecules "Smac-mimetics" (SMs) [280,281,[369][370][371][372][373]. SMs are small molecules that were based on the evolutionary conserved tetra-peptide IAP-Binding Motif (IBM, AVPI/F). This motif was originally observed in the Drosophila IAPantagonists Reaper, Hid, Grim and is also found in the mammalian IAP-antagonists Smac and Omi [270,271,273,274,280,370]. There are two classes of SMs: monovalent, which contain one AVPI binding motif and bivalent, which contain two AVPI binding motifs and are more potent than the monovalent [371,374]. SMs were originally designed to target and inhibit XIAP [269,372,[375][376][377]. Although SMs can bind XIAP, they are not very effective in degrading it [247,373,374,378]. On the other hand, SMs efficiently promote the degradation of cIAPs via the UPS [283,379,380]. The SM-mediated inhibition of cIAPs causes apoptosis through the inhibition of NF-κB signaling [250,251,[380][381][382][383][384][385][386][387]. SM130 and SM114 primarily target cIAPs for degradation but have reduced affinity towards XIAP [373]. TL32711 (Birinapant) is a bivalent molecule that works particularly well against cIAP1 and is well tolerated at doses that sustain target inhibition [388][389][390][391]. Unfortunately, most cancer cell lines are resistant to SMs [383,392,393]. Therefore, combinations with other anti-cancer drugs are being explored in an effort to overcome resistance [394][395][396]. Historically, SMs were developed to target XIAP, but until recently, no potent XIAP-only inhibitors were available [34,356]. Mamriev et al. reported of small molecule ARTS-mimetics that can bind XIAP and promote its degradation via the UPS [397]. This compound was identified in a virtual screen for small molecules with the highest docking affinity to the specific and unique binding site of ARTS within the BIR3 domain of XIAP [397]. These smallmolecule ARTS-mimetics can induce apoptosis in a wide range of cancer cells but not in healthy PBMC (Peripheral Blood Mononuclear Cells). ARTS-mimetic small molecules bind specifically to XIAP, but not cIAP1, and promote the degradation of both XIAP and Bcl-2 through the UPS [397]. ARTS-mimetics act as PROTACs by bringing the E3 ligase XIAP to its target Bcl-2, thereby inducing the degradation of both these proteins [397]. ARTS-mimetics provide a promising novel platform for developing highly specific and potent anti-cancer drugs by targeting XIAP-and Bcl-2 for degradation [397].
Another strategy to target IAPs for degradation is by a series of chimeric molecules termed specific and non-genetic inhibitor of apoptosis protein (IAP)-dependent protein erasers (SNIPERs). SNIPERs consist of three distinct parts: a target protein ligand, an E3 ligase ligand and a linker between them [398][399][400][401]. They were shown to recruit IAPs and promote their targeted protein degradation [398,399,402,403]. Unlike the traditional PROTACs, SNIPERs induce simultaneous degradation of IAPs, such as cIAP1 and XIAP, along with their target proteins [399,404]. Although PROTACs and SNIPERs exhibited promising results in degrading target proteins, there are still some challenges to overcome before these compounds can be used in the clinic [401]. For example, most PROTACs do not obey Lipinski's rule of five (RO5) because of their relatively high molecular weight [405]. In addition, PROTAC's toxicity, bioavailability, distribution and metabolism still need to be determined [401]. For now, two PROTACs have entered phase I/II clinical trials-the PROTAC ARV-110 for the treatment of prostate cancer and ARV-741 for the treatment of breast cancer [401].

Future Directions and Challenges
The UPS is considered to be a major target for developing novel types of anti-cancer drugs. This was initiated with the use of Proteasome inhibitors, which were proven to be effective for patients with hematological malignancies (Mantle cell lymphoma and multiple myeloma) [406]. However, the efficiency of proteasome inhibitors is seriously compromised due to innate and acquired drug resistance [407,408]. Numerous studies have helped uncover the pathways responsible for drug resistance, making it easier to predict which patients can benefit from specific proteasome inhibitor therapy [409][410][411][412]. One of the approaches proposed to overcome drug resistance is combination therapy [406,412]. This might also help with treating malignancies, which present a limited response to proteasome inhibitors. In many cases, upregulation of E3 ligases is responsible for drug resistance [58,192,413]. Fortunately, many E3 ligases, such as cIAP, XIAP, MDM2 and others, have become popular targets for drug development, including specific small molecules, targeting E3-ligases. Nevertheless, there are still challenges to overcome, including the vast diversity of E3 ligases and the fact that E3 ligases can have various different substrates, including tumor suppressors and oncogenes [414]. The combination of small molecules with other cancer therapies showed better efficacy than monotherapies [414]. Besides small molecules, the protein-targeting chimeras (PROTACs) were developed to overcome drug resistance via hijacking the UPS mechanism. While PROTACs are showing promise in providing a novel approach to overcome drug resistance, many challenges await to be resolved regarding their drug design and possible clinical applications.
In summary, the UPS plays a major role in regulating key apoptotic proteins. As many as 20 E3 ligases alone are known to control the levels of the p53. This illustrates the importance of regulated protein degradation and governing the activity of this major tumor-suppressor protein. High levels of Bcl-2, XIAP and cIAPs are characteristic of many types of cancers and hence make these proteins attractive drug targets [366][367][368]415,416]. Interestingly, far fewer E3 ligases control the levels of these proteins compared to p53. A possible reason for this difference is that these apoptosis-suppressing proteins are regulated by direct binding to neutralizing proteins (such as Bax in the case of Bcl-2) or IAP-antagonists [35, 263,417]. Perhaps this regulation through protein-protein interactions can complement any possible limitations resulting from the relatively small number of specific E3 ligases dedicated to these proteins. Degrading the target protein rather than binding and blocking their function has significant advantages; mainly, it reduces the load of the elevated expression of the target protein, which are often inhibitory proteins. Moreover, it can reduce systemic drug concentrations and, hence, possible cytotoxic side effects. Recently, major efforts have been devoted both in academia and by pharma companies to develop therapies that recruit the UPS for promoting apoptosis in cancer cells. Compounds that specifically target proteins for degradation resulting in effective tumor killing may dramatically improve the success of cancer therapy.