The roles of E3 ubiquitin ligases in cancer progression and targeted therapy

2.1 RING finger family E3 ligases

The RING finger E3 ligases make up the largest E3 family that contains the RING or U-box catalytic domain. The canonical RING finger is a cysteine-rich domain with the sequence of Cys-X₂-Cys-X_(9-39)-Cys-X_(1-3)-His-X_(2-3)-Cys-X₂-Cys-X_(4-48)-Cys-X₂-Cys (where X represents any amino acid). It uses the cysteine and histidine residues to coordinate two zinc ions in an eight ligands cross based structure.^{30, 31} RING finger E3 ligases do not bind ubiquitin directly but mediate the transfer of ubiquitin from bound E2 (E2-Ub) to the target substrate.³² Ring finger E3 ligases may function independent of any auxiliary protein, as a monomer or homodimer, such as TRIM (Tripartite Motif containing E3 ligases), TRAF6 (TNF Receptor Associated Factors 6), cIAP (cellular Inhibitor of Apoptosis), XIAP (X-linked inhibitor of apoptosis), RING finger (RNF) containing proteins and MDM2 (Murine Double Minute 2)^33-37 or associate with another protein and function as a heterodimer or multisubunit complex. Heterodimeric ring finger E3 ligases include TRAF heterodimers,³⁸ MDM2/MDMX,³⁹ BRCA1/BARD1⁴⁰ and RING1B/BMI-1.⁴¹ MDMX, BARD1 and BMI-1 in the above heterodimers have ring domain, but do not possess E3 ligase activity therefore they associate with the E3 ligase ring domain counterpart to promote substrate ubiquitination.

Multisubunit ring finger E3 ligases include the Cullin Ring Finger ligases (CRL) and Anaphase Promoting Complex/cyclosome (APC/C). The CRL is a large family, containing Cullin scaffold protein, E2 binding RING-box protein (Rbx1 and Rbx2), adaptor protein and substrate recognition protein.⁴² The substrate recognition proteins bind adaptors while the Cullin forms a central scaffold that bridges Rbx1/2 and the adaptor-substrate recognition protein. Differences in Cullin type (Cul1, Cul2, Cul3, Cul4A, Cul4B, Cul5, Cul7 and Cul9) form the basis for different groups of the CRL subfamily.⁴³ The specificity of substrate recruitment is determined by substrate recognition proteins with over 400 kinds identified.⁴⁴ SCF (SKP1-Cullin1-F-box) E3 ligase is the largest CRL family with SKP1 as the adaptor and F-box proteins as the substrate recognition unit.⁴⁵ There are 69 F-box proteins encoded in human genome which can be grouped into three subfamilies based on their substrate recruiting domain: F-box with WD-40 domain (FBXW), F-box with leucine rich repeat (FBXL) and F-box with other domains (FBXO).^{46, 47} F-box proteins are important regulators of diverse cell functions; an example is SKP2 (also called FBXL1), a popular oncoprotein that mediates the degradation of CDK inhibitors to enable G1 to S phase cell cycle progression.⁴⁸ The other ring finger complex, APC/C, functions in cell cycle mitotic progression. They contain 19 subunits that are grouped into three sub-complexes: scaffolding platform, ring finger containing catalytic core and the tetratricopeptide repeat (TPR).⁴⁹ APC/C requires co-activator proteins, CDC20 and CDH1 for substrate recruitment to the APC/C complex. The co-activator enters the APC/C complex by binding APC3 and APC8 in the TPR sub-complex,^{50, 51} recognises and recruits substrate containing KEN-box or D-box to the APC/C complex for ubiquitination.²⁹

2.2 HECT family E3 ligases

The HECT families of E3 ligase bear HECT catalytic domain at their Carboxyl terminal lobe. A flexible link with a cysteine active site connects the N terminal and C terminal lobes.⁵² This cysteine active site, confers on them a distinct mechanism of ubiquitination from the ring finger families. In HECT-mediated ubiquitination, ubiquitin from E2 first forms an intermediate thioester bond with the cysteine before being transferred to the target substrate.⁵³ Twenty-eight HECT E3 ligases have been identified in human and are grouped into three different subfamilies based on N terminal domain including the NEDD4 family containing N-terminal WW and C2 domain, the HERC family containing N-terminal RCC-like domain (RLD) and other HECT with variable N-terminal domain. HECT members are actively involved in several cellular processes that drive cancer progression. For example, E6AP, a member of other HECT family drives cervical cancer by associating with E6 protein of human papillomavirus to promote proteasome degradation of p53^{54, 55}; the NEDD4 family members WWP1 and NEDD4 promote PI3K/AKT signalling by catalysing the ubiquitin-mediated degradation of PTEN⁵⁶; ITCH, another NEDD4 member, is an important regulator of immune response.^{57, 58}

2.3 RBR family E3 ligases

The RBR ligases are considered the smallest E3 family with only 14 members among which some well-known E3 ligases such as Parkin, HHARI, TRIAD1, HOIP, HOIL-IL and RNF144 are included.^{59, 60} Canonical RBRs consist of three components: RING1 with an E2 binding domain, RING2 with a catalytic cysteine residue and an In-between Ring domain (IBR).⁶¹ RBRs are described as RING–HECT hybrid because the RING1 is structurally similar to Ring type E3 ligase, while the RING2 bears an exposed catalytic cysteine residue that forms intermediate thioester linkage with ubiquitin during substrate ubiquitination just like the HECT family.⁶² Traditional RBRs exist in the form of N-RING1–IBR–RING2-C; however, they can contain additional domains at the N terminal, middle or C terminal. These outside domains confer a characteristic auto-inhibitory action on RBR, for example, the C-terminal Ariadne domain of HHARI blocks RING2 active site leading to HHARI autoinhibition.⁶⁰ Parkin, a tumour suppressor RBR implicated in neurodegeneration disease and innate immune response, is kept in an autoinhibition state by three outside domains, ubiquitin like domain (Ubl), Ring0 domain and a REP domain.^{63, 64} REP and Ubl mask E2 binding site on RING1, while Ring0 binds to RING2 and blocks its catalytic cysteine residue thereby keeping Parkin in an autoinhibition state.^{65, 66} Phosphorylation of the Ubl domain by PTEN-induced kinase 1 (PINK1) and the binding of phosphorylated ubiquitin to Parkin disrupt the autoinhibition state, releasing Parkin for activity.^{67, 68} LUBAC (Linear ubiquitin chain assembly complex), another RBR ubiquitin ligase active in innate immune response, is composed of HOIP, HOIL-IL and SHARPIN that generates linear polyubiquitin chains.^{19, 60} HOIP bears ubiquitination catalytic activity but is kept in the inactive auto-inhibited state by the ubiquitin-associated (UBA) domain at its N terminus. To distort the autoinhibition, the ubiquitin like domain of either HOIL-IL or SHARPIN interacts and forms a complex with the UBA domain of HOIP thereby activating HOIP for M1 linear ubiquitination.^{69, 70} In addition to the established E3 ligase catalytic activity of HOIP, HOIL-1L has been identified as E3 ligase that catalyses oxyester linked self-monoubiquitination and monoubiquitination of protein components of Toll-like receptor (TLR).¹²

3.1 Role of E3 ubiquitin ligases in cell cycle progression

Cancer cells abrogate cell cycle regulation to sustained proliferation and progression. The cell cycle is a series of incidents that progress through four phases: gap1 phase (G1 phase), DNA synthesis phase (S phase), gap2 phase (G2 phase) and mitosis phase (M phase). The cyclin-dependent kinases (CDKs) are key regulators of the cell cycle that drive cell division by forming complex with cyclins with distinct CDK/cyclin complex operating at different phases.^{72, 73} Cyclins are short-lived proteins degraded during cell cycle by E3 ligases,⁷⁴ while CDKs are relatively stable; however, their activities can be inhibited by other cell cycle regulatory proteins among which E3 ligases represent specific functions.⁷⁵ The classical E3 ligases regulating cell cycle include APC/C and SCF containing the substrate recognition proteins such as SKP2, β-TrCP and FBXW7 (Figure 1 and Table 1).^{76, 77}

APC/C mainly mediates K11 polyubiquitination and degradation of its substrate. It requires its co-activators CDC20 and CDH1 to regulate the cell cycle between the M phase and early G1 phase.⁵¹ APC/C^CDC20 is activated in the M phase by CDK1-mediated phosphorylation while APC/C^CDH1 is simultaneously phosphorylated and inhibited until the late M phase when there is a low mitotic kinase level.^78-80 The activated APC/C^CDC20 is inhibited by the mitotic checkpoint complex at metaphase. This induces metaphase arrest to ensure the correct attachment of sister chromatids to bipolar spindle before transitioning into anaphase.^{50, 81} Upon mitotic checkpoint complex satisfaction, APC/C^CDC20 catalyses the ubiquitination and proteasomal degradation of securin (a separase inhibitor) to promote chromatid segregation, as well as the degradation of cyclin A and cyclin B to promote anaphase onset.⁴⁹ The degradation of cyclin B weakens CDK1 and abates CDK1-mediated inhibition of CDH1 during anaphase. This activity coupled with CDC14 phosphatase-induced dephosphorylation of CDH1 promotes the activation of APC/C^CDH1.^{82, 83} Active APC/C^CDH1 then mediates the proteasome degradation of CDC20, Polo-like kinase 1 (Plk1) and Aurora kinases (Aurora A and B) at late mitosis to ensure mitotic exit and the degradation of CDC25A (a phosphatase activator of CDK2) and SKP2 at early G1 to lower CDK accumulation.^{84, 85} In late G1 phase, early mitotic inhibitor 1 (EMI1) inhibits APC/C and disrupts APC/C^CDH1-mediated degradation of its substrates. Although EMI1 is an F-box protein, its cell cycle regulating activity is F-box domain independent, to this end, EMI1 regulates APC/C^CDH1 by antagonising the APC/C^CDH1 generation the of K11 polyubiquitin chain.⁸⁶ This inhibitory action of EMI1 continues through the S phase and G2 phase until it is degraded by SCF^β-TrCP to allow cell cycle progression through mitosis.⁸⁷ Additionally, SCF^β-TrCP promotes the degradation of CDC25A to suppress cell cycle progression and degrades WEE1 (CDK1 inhibitor) and transcription factor FOXO3 for cell cycle progression.^88-91

SKP2 is activated in the late G1 phase by CDK2, which phosphorylates and protects it from APC/C^CDH1-mediated degradation.⁹² Once activated, SCF^SKP2 facilitates G1/S transition by promoting the ubiquitination and degradation of Kip/Cip members of CDK inhibitors (p21^CIP1, p27^KIP1 and p57^KIP2).^{48, 93} During the G1 phase, both SCF^SKP2 and SCF^FBXW7 target cyclin E for proteasomal degradation^94-96 and act as E3 ligases of the oncogenic transcription factor c-MYC but with distinct functions. SCF^SKP2 stabilises and promotes c-MYC transcriptional activity, while SCF^FBXW7 mediates proteasome degradation of c-MYC.^{97, 98}

In addition to the classical regulators mentioned above, many other E3 ligases including Cul3 E3 ligase complex, Cul4 E3 ligase complex and the SCF containing FBXW8, FBXO4, FBXO7, FBXO31 and FBXL2 have been identified as cell cycle regulators that directly target cyclins and CDK inhibitors as shown in Figure 1.^99-101 More important also is MDM2, the key E3 ligase targeting p53 for proteasome degradation.¹⁰²

Deregulation of cell cycle-related E3 ligases has been reported in several cancers. SCF^SKP2 complex is a positive regulator of cell cycle considered as proto-oncogene because it targets CDK inhibitors and other tumour suppressors. The oncogenic role of SKP2 reconciles with its overexpression in various human cancers including ovarian adenocarcinoma, breast cancer, lung cancer, colorectal cancer, prostate cancer, leukaemia and squamous cell carcinoma.^103-106 Interestingly, deletion of SKP2 has been shown to compensate for anti-tumour and cell safeguarding deficiency in p53 deleted cancer cells due to elevated p27 levels. Zhao and co-workers demonstrated this using SKP2 knockout mice and found that loss of p53 and pRB in these mice blocked tumorigenesis due to cell cycle arrest mediated by accumulated p27, suggesting that SKP2 could be a promising target for p53 mutant cancer therapy.¹⁰⁷ Likewise, APC/C co-activator CDC20 has been identified as an oncoprotein that is highly expressed in several human cancers.^108-112 Tumorigenic role of CDC20 is linked to its involvement in diverse cellular pathways where it targets tumour suppressors. CDC20 aberrant expression or the disruption of SAC-mediated inhibition of CDC20 leads to tumorigenesis due to aneuploidy.¹¹³ Conversely, APC/C^CDH1 has been associated with a tumour suppressive role as majority of its substrates including SKP2 are known oncoproteins. APC/C^CDH1-mediated regulation of SKP2 turnover in cell cycle significantly contributes to its control of tumorigenesis in cells. Studies have shown that inhibition of CDH1 resulted in high cellular proliferation while an induced overexpression in solid tumours is associated with patients’ survival and low histological tumour grade in solid tumours.^{114, 115} Nonetheless, SCF^β-TrCP performs dual roles in the cell cycle, promoting both cell cycle progression and cell cycle arrest. This suggests both oncogenic and tumour-suppressive roles. Hence, the expression of β-TrCP in cancers is not clearly defined; however, the expression could be linked to a context-dependent role. The Cul4 E3 ligase complex is also a key regulator of cell cycle that targets CDK inhibitors p21 and p27 as well as cyclin E for degradation. Aberrant expression of Cul4 has been indicated in breast cancer, lung cancer, ovarian cancer and squamous cell cancer.¹¹⁶ Studies show that Cul4 depletion reduces the proliferation of lung squamous cell carcinoma and small cell lung cancer.¹¹⁷ Moreover, conditional knockdown of Cul4A in mice reduces sensitivity to UV-induced skin cancer in a mechanism linked to its regulation of p21.¹¹⁸ MDM2 is the primary inhibitor of p53 and is amplified in most cancers, especially p53^WT cancer types.^{119, 120} FBXW7, on the other hand, is an established tumour suppressor that functions as SCF substrate recognition protein targeting several oncogenic proteins including cyclin E and c-MYC for proteasome degradation. Concomitant with its tumour-suppressive role, FBXW7 is frequently underexpressed or inactivated in most human cancers.¹²¹ Taken together, these studies show that E3 ligases are critical regulators of cell cycles, and their cell cycle-dependent oncogenic function could be considered in the development of therapies that selectively target E3 ligases to induce the cell cycle arrest and corresponding death of cancer cells with considerably lesser toxicity.

3.2 Roles of E3 ubiquitin ligases in immune response and inflammation

Evasion of immune response and stimulation of tumour-promoting inflammation are among the several strategies adopted by cancer cells to sustain proliferation and progression. Host cells through the innate and adaptive immune response, release proinflammatory cytokines that activate macrophages, dendritic cells and natural killer (NK) cells for the destruction of tumour cells and infectious agent that threatens cell immune homeostasis. During immune surveillance, growing cancer cells escape from immune destruction and colonise the infected tissue. They achieve this by promoting the activation of tumour suppressive cells such as the regulatory T cell (Treg), myeloid-derived suppressor cell and regulatory B cell which suppress anti-tumour immunity within the tumour microenvironment (TME).^122-125 Strikingly, cancer cells promote uncontrollable inflammation which becomes chronic, causing a damaging effect that increases cancer progression.¹²⁵ Evidence of chronic inflammation progression into cancer is exemplified in the study where Mdr2-knockout mice developed liver inflammation that subsequently progressed to liver cancer.^{126, 127} E3 ligases are actively involved in immune response and inflammatory signalling either as promoters or suppressors and could thus be considered novel therapeutic targets for improving anti-tumour immunity and blocking tumour-promoting inflammation.

To achieve immune tolerance, cancer cells increase the expression of E3 ligases that promote the activity of Treg and decrease anti-tumour immunity. Among these E3 ligases, the TRAF6 ligase promotes the activity of Treg within TME by mediating K63-linked polyubiquitination and activation of FoxP3, a transcription factor for Treg. TRAF6-deficient Treg cells resist growth of implanted tumours, abolish immune tolerance in cancer cells and enhance anti-tumour immunity.¹²⁸ The effect of TRAF6 inhibition for immunotherapy was investigated in Hela 1–6 tumour model and the result showed that TRAF6 inhibitors accelerated T cell-mediated anti-tumour immunity and blocked Treg infiltration to reduce the Treg tumour population.¹²⁹ Therefore, TRAF6 represents a promising targeting candidate for cancer immunotherapy. Elsewhere, it has also been reported that the E3 ligase ITCH positively regulates Treg by catalysing the monoubiquitination of the transcription factor TIEG1. Monoubiquitination promotes the nuclear translocation of TIEG1 necessary for FoxP3 expression.⁵⁸ Interestingly, ITCH can also mediate IL-6-dependent K27-linked polyubiquitination of TIEG1 which in opposite to monoubiquitination abrogates TIEG1 nuclear translocation thereby preventing FoxP3 expression.¹³⁰ Inhibition of Tregs in this case is found to correlate with an increase in Th17 response and enhance anti-tumour immunity. Given these contrasting functions of ITCH-mediated mono- and poly-ubiquitination, targeting ITCH might not be a best option for increasing anti-tumour immunity but the driving mechanism involved in each case could be considered. Another E3 ligase casitas B-lineage lymphoma proto-oncogene-b (Cbl-b) has been identified as an important immune checkpoint regulator of CD8⁺ T-cell and NK cells.^{131, 132} Findings from a recent study showed that depletion of Cbl-b in tumours restores the effector activity of CD8⁺ tumour-infiltrating lymphocytes and prevents chimeric antigen receptor T-cell exhaustion.¹³³ Given this immune suppressive role, APN401, a small interfering RNA-based cellular immunotherapy that specifically targets and silences Cbl-b is currently in clinical studies in patients with advanced solid tumours.¹³⁴

Tumour-promoting inflammation is mainly regulated via nuclear factor-kappa B (NF-κB) signalling. NF-κB is the key mediator of downstream inflammatory response and is composed of five members (RelA/p65, RelB, c-Rel, NF-κB1/p50 and NF-κB2/p52), from which p52 and p50 are generated from p105 and p100 precursors, respectively.⁷⁰ Both canonical and non-canonical NF-κB signalling pathways are regulated by several E3 ligases. For instance, SCF^β-TrCP via a non-canonical pathway mediates ubiquitination and partial degradation of NF-κB precursor p100 to generate p52 which then translocates to the nucleus for target gene expression.¹³⁵ In the physiological state, the non-canonical pathway is inhibited by the E3 ligases cIAP, TRAF2 and TRAF3. These E3 ligases form a destruction complex with NF-κB-inducing kinase (NIK) where cIAP conjugates K48-linked polyubiquitin on NIK to induce its proteasome degradation.^136-140 When released from the destruction complex, NIK activates downstream IκB kinase alpha (IKKα), which in turn phosphorylates p100 for its partial degradation to p52.^{137, 141} Under stimulation, TRAF2 induces cIAP to catalyse proteasome degradation of TRAF3 leading to destruction complex disassembly and consequent release of NIK.^{142, 143} Consistent with negative regulation, TRAF2, TRAF3 and cIAPs inactivation mutation has been detected in B-cell malignancy and their tumour suppressive function is mainly associated with non-canonical NF-κB inhibition.^{137, 144–147}

Activation of the canonical NF-κB signalling pathway promotes nuclear translocation of NF-κB dimer (RelA-p50 heterodimer), where they activate target genes. Under normal physiological conditions, NF-κB inhibitor alpha (IκBα) arrests NF-κB dimer in the cytoplasm to prevent their nuclear translocation and consequent expression of target genes.¹⁴⁸ However, with inflammation signalling, IκBα is phosphorylated by IκB kinase (IKK) complex (containing NEMO, IKKα and IKKβ), to promote its ubiquitination and degradation by SCF^β-TrCP E3 ligase.^149-151

Upstream signalling activators of canonical NF-κB such as tumour necrosis factor receptor 1 (TNFR1) and pattern recognition receptors (PRR) including TLR, RIG-I-like receptor (RLR) and NOD-like receptor (NLR) are regulated by E3 ligases. Upon binding of inflammatory effector ligands (TNFα and cytokines) to their cell surface receptors, TRADD (TNFR-associated death domain protein) and RIP1 (receptor-interacting protein 1) are recruited via their death domain to the TNFR1 death domain.^{152, 153} TRADD then recruits TRAF2 E3 ligase which further recruits another E3 ligase cIAP to form complex I, undergoing self-polyubiquitination and K63, K11 and K48 ubiquitination of RIP1 (Figure 2).^{154, 155} These polyubiquitin chains act as scaffolds for the recruitment of another ubiquitin ligase LUBAC for M1 linear polyubiquitination of RIP1 and NEMO.^156-160 TGF-β activated Kinase 1 (TAK1) is also recruited by the K63 polyubiquitin scaffold, promoting downstream signalling by inducing the phosphorylation and activation of the IKK complex as shown in Figure 2.^{161, 162} Key E3 ligases involved in canonical NF-κB signalling are listed in Table 2. Other E3 ligases regulating NF-κB signalling include Makorin ring finger 2 (MKRN2) E3 ligase which suppresses NF-κB signalling via proteasome degradation of p65 and the E3 ligase ITCH which cooperates with deubiquitinating enzyme cylindromatosis tumour suppressor protein (CYLD) to mediate deubiquitination and degradation of TAK1.¹⁶³ Although TRAF2 exhibits anti-tumour activity via noncanonical NF-κB, it has also been shown to elicit protooncogenic function by activating canonic NF-κB signalling.¹⁶⁴ As TRAF2 is required to interact with cIAP1 or cIAP2 to promote TNFR1-induced NF-κB signalling and suppress TNF-induced cell death, inhibition of TRAF2 or cIAPs is necessary to sensitise tumour cells to TNF-induced cell death. To this end, several IAP antagonists have been developed for inducing anti-tumour immunity and have proven effective as a monotherapy or a combination therapy with immune checkpoint blockers such as programmed death 1.^{165, 166}

TLRs play essential roles in the innate immune system, which include both cell surface and intracellular receptor types. TLRs contain TIR domain through which they recruit adaptor proteins MyD88 (myeloid differentiating factor 88) and TRIF (TIR domain-containing adaptor-inducing IFN-β) for both NF-κB and/or interferon (IFN) induction.¹⁶⁷ TRAF6 and Pellino 1 (Pel1) are the key E3 ligases regulating TLR-induced NF-κB signalling. They mediate K63 polyubiquitination for induction and activation of TAK1 complex and IKK complex necessary for downstream NF-κB activation.^{168, 169} K63 polyubiquitin is conjugated on TAB2 and TAB3 which complexes with TAK1 (TAK1–TAB1–TAB2 or TAB3 complexes) to aid TRAF6–TAK1 interaction and activation of TAK1.^{170, 171} Besides, a ring finger containing E3 ligase BICP0 (bovine herpes virus-encoded protein, infected cell protein 0), functions as a negative regulator of NF-κB signalling by mediating K48 polyubiquitination and degradation of TRAF6.¹⁷² Besides, c-Cbl has also been reported to inhibit NF-κB signalling by promoting K48-linked ubiquitination of TRAF6.¹⁷³ TRAF members particularly TRAF6 are known key regulators of tumour-promoting inflammation and immune response along TLR–NF-κB pathway, therefore, small molecule inhibitors of TRAF6-induced NF-κB activating inflammation have been identified. Such inhibitor like the TRAF–CD40 inhibitor has been shown to suppress breast cancer metastasis effectively as either a single agent or in combination therapy with conventional chemotherapy.¹⁷⁴ NLRs and RLRs are cytosolic PRR activated following recognition by bacteria cell walls and cytosolic RNA virus, respectively. These two receptors are master regulators of the innate immune response against pathogens and are considered relevant for inducing immunogenic cell death and anti-tumour immunity. However, their constitute activation due to dysregulation may elicit proinflammatory signals and chronic inflammation that predisposes to malignancies as such they could be considered a double-edged sword in inflammation and cancer. NOD1 and NOD2 are the major NLR family that drives NF-κB signalling. They contain caspase activating and recruiting domain (CARD) for interaction with RIP2.¹⁷⁵ This complex promotes the recruitment of E3 ligases such as XIAP, cIAPs and LUBAC for recruiting downstream NF-κB activating cascade.^176-178

Polyubiquitination of RIG-I by the E3 ligases TRIM25, TRIM14, MEX3c and Riplet (RNF135) is essential for activation and initiation of RIG-I-mediated signalling.¹⁷⁹ RLRs on the other hand include RIG-I, MDA5 and LGP2 are known to mediate anti-viral immunity by binding the mitochondria anti-viral proteins (MAVs). This promotes MAVs interactions with TRAFs (TRAF2/3/5/6) for K63 polyubiquitination and subsequent IKK complex induction.^{180, 181} The regulation of NLR and RLR-driven NF-κB signalling by E3 ligases has not been clearly described in cancer; however, a few studies have demonstrated the cancer implication of E3 ligase regulation of NOD1/2-driven NF-κB signalling. For instance, overexpression of TRIM22 has been shown to decrease the proliferation and migration of endometrial cancer cells as knockdown of TRIM22 was found to accelerate cancer progression via NOD-NF-κB pathway.¹⁸² More so, NOD1 can attenuate Helicobacter pylori infection-induced caudal-related homeobox 2 (Cdx2) expression and intestinal metaplasia via induction of TRAF3. Since, H. pylori infection-induced Cdx2 expression is dependent on NF-κB activation, NOD1 induction of TRAF3 inhibits NF-κB signalling and protects the intestinal cells from precancerous changes.¹⁸³

Wnt/β-catenin is a significant pathway in carcinogenesis and embryonic development, which has been shown to regulate inflammatory response via cross-talking with the NF-κB pathway.^{184, 185} SCF^β-TrCP is a key regulator of both Wnt/β-catenin and NF-κB signalling pathways. SCF^β-TrCP negatively regulates Wnt/β-catenin by catalysing K48-polyubiquitin-induced proteasome degradation of β-catenin, subsequently inhibiting its nuclear translocation and target gene expression. It is shown that Wnt/β-catenin signalling induces high β-TrCP expression which in turn activates NF-κB signalling.¹⁸⁶ Strikingly, an integrated association of β-TrCP, β-catenin and NF-κB is detected in colorectal cancer and is considered important for tumour metastasis and apoptosis inhibition.¹⁸⁷ In addition, β-TrCP promotes lymphocytic leukaemia cell proliferation through concomitant activation of NF-κB and β-catenin/TCF signalling pathways, suggesting that β-TrCP–NF-κB–β-catenin pathway could be considered a potential target for cancer therapy.¹⁸⁸

Active Rel/NF-κB act as a transcription factor for hundreds of genes involved in various biological process including cytokines/chemokines, immunoreceptors, apoptotic regulators, growth factors and transcription factors.¹⁸⁹ It is reported that increasing NF-κB signalling is central to driving cancer cell proliferation. As a result, genetic alterations of E3 ligases regulating this pathway have been identified in several human cancers. Oncogenic functions of immunomodulating E3 ligases are mostly attributed to the activation of NF-κB signalling. For instance, TRAF6 is overexpressed in colon cancer and lung cancer and investigation of its oncogenic function shows the activation of NF-κB signalling pathway.^190-192 TRAF6-induced NF-κB activation promotes multiple myeloma cell adhesion to bone marrow stroma.¹⁹³ Similarly, TRAF2 functions as a NF-κB activating oncogene that promotes epithelial cancers and skeletal tumour growth in osteotropic breast cancer.^{164, 194} Furthermore, mutational activation of BRAF in melanoma promotes β-TrCP expression and induces NF-κB activation and melanoma cell growth.¹⁹⁵ Constitutively, enhanced expression of β-TrCP in pancreatic carcinoma cells positively correlates with NF-κB expression and chemoresistance in pancreatic carcinoma.¹⁹⁶ On the other hand, the U-box containing E3 ligase CHIP (carboxyl terminus of Hsc70-interacting protein) as a tumour suppressor weakly expressed in colon cancer and gastric cancer and negatively regulates NF-κB signalling.^{197, 198} CHIP was found to repress the growth of colorectal cancer cells by mediating ubiquitination and degradation of p65 and its expression correlated with TNM stages with the lowest expression in stage 3 and 4 patients.¹⁹⁸ Further a similar result revealed that TRIM7 E3 ligase negatively regulates NF-κB signalling by promoting the ubiquitination and proteasomal degradation of p65 in lung cancer.¹⁹⁹ In this work, Jin and colleagues not only detected low expression levels of TRIM7 in tumours compared to normal cells, but also found that tumour size decreases with stable expression of TRIM7.¹⁹⁹ With an increasing understanding of the target substrates and mechanisms of action, RNF40 could be considered a target for colorectal cancer treatment. Conversely, TRIM37, another member of the TRIM family, is reported to activate NF-κB signalling and promote non-small cell lung cancer aggressiveness due to its K63-mediated polyubiquitination and activation of TRAF2.²⁰⁰ These studies verify the active involvement of E3 ligases in cancer-promoting inflammatory signalling presenting them as excellent targets for inducing tumour regression via inhibition of NF-κB tumour-promotion inflammation.

3.3 Evasion of apoptosis by E3 ubiquitin ligases

Cell death is a natural suicidal event for destroying malignant and potentially harmful cells. Of all modes of cell death, programmed cell death (apoptosis) is considered the most significantly inhibited type by cancer cells. This is primarily because it does not elicit adverse effects, thus, diverse therapies targeting apoptotic pathways have been developed for cancer treatment.^{201, 202} Table 3 lists significant E3 ubiquitin ligases associated with apoptotic regulation.

3.3.2 Extrinsic apoptotic pathway

TNFR family containing death receptors (DR) interact with effector ligands such as Fas ligand (Fasl), TNFα and TNF-related apoptosis-inducing ligand (TRAIL) and recruit death domain-containing adaptor molecules (TRADD and FADD) to effectuate intracellular signalling that led to cell death.²²⁴ Unlike other ligands which trigger only death response, TNFα associates with the receptor TNFR1 and induces either pro-survival/inflammatory signalling via complex I or apoptotic signalling via complex II.^{225, 226} E3 ligases are important regulators of TNFα–TNFR1-mediated switch from pro-survival/inflammatory signalling to apoptosis signalling. To induce apoptosis signalling under TNFα stimulation, TNFR1 adaptor TRADD recruits FADD which in turn recruits pro-caspase 8 to form complex II for activation of downstream apoptosis effector.²²⁷ TRADD recruitment of the E3 ligases TRAF2 and cIAP on the other hand inhibits apoptosis and promotes complex I formation for pro-survival/inflammatory signalling.^{153, 155} RIP1 can also recruit FADD to form another complex II that comprises RIP1, FADD and caspase 8, so RIP1 interplays between pro-survival/inflammation and apoptosis signalling.^228-230 Ubiquitination of RIP1 by cIAPs and TRAF2 inhibits RIP1 induction of apoptotic signalling and favours pro-survival signalling.^{229, 231, 232} In the absence of cIAP and TRAF2 or in deubiquitinated form, RIP1 dissociated from membrane-bound complex I to recruit FADD for the formation of apoptotic complex II signalling (Figure 3).²²⁴ LUBAC is another E3 ligase that impedes RIP1-dependent apoptosis signalling by mediating the linear ubiquitination of RIP1. This ubiquitination is necessary for recruiting a hybrid protein A20 that blocks CYLD from deubiquitinating RIP1thereby stabilising RIP1 in pro-survival response.²³³ Cellular FLICE inhibitory protein (c-FLIP) which interferes with pro-caspase 8 activation in complex II is inhibited by ITCH that degrades c-FLIP to promote apoptosis.²³⁴ In addition, the apoptotic inhibitor, TRAF2 promotes polyubiquitination and degradation of caspase 8 and this contributes to insensitivity to TRAIL ligand-induced apoptosis in gastric cancer.^{235, 236} Likewise, two E3 ligases, Siah2 (Seven in absentia homolog) and POSH (plenty of SH3s), interact together to abrogate TRAIL or FAS ligand-induced apoptosis by inhibiting downstream caspase 8 activity.²³⁷ The stability of FADD is impaired in the presence of the E3 ligase MKRN1 which promotes FADD proteasomal degradation.²³⁸

The expression levels of E3 ligases that negatively affect apoptosis are high in most cancers. An example is seen in XIAP, which is abundantly expressed in anaplastic thyroid carcinoma (ATC). Most notably, this overexpression is associated with proliferation, migration, invasion and chemoresistance in ATC cells.²³⁹ Additionally, XIAP overexpression has also been reported in breast cancer, renal cancer, colon cancer and leukaemia and correlated with poor overall survival, thus, positioning it as a potential therapeutic for cancer treatment.^240-244 In a related study, overexpression of TRIM9 is opined to correlate with the upregulation of Bcl-2 and downregulation of caspases 9/7 leading to apoptosis inhibition in lung cancer.²⁴⁵ Similarly, Yang et al.²⁴⁶ reported a correlation between TRIM9 overexpression and decreased caspase 3 activity in uterine leiomyoma. TRAF2 shows high expression in prostate cancer cells which is associated with inhibition of TRAIL-induced apoptosis. In vitro study in prostate DU-145 cells shows that TRAF2 affects caspase-8, cFLIP and SIRT1 expression.²⁴⁷ Altogether, E3 ligases represent an attractive target that could be considered in clinical application for promoting apoptosis in tumour cells.

4.1 Small molecules targeting cell cycle E3 ligases

To promote cell cycle arrest and inhibit cellular proliferation in cancer, several small molecule compounds have been developed to target cell cycle E3 ligases such as MDM2, SKP2, FBXW7, APC/C and Cullin4.

4.2 Small molecules targeting E3 ligase along inflammatory NF-κB pathway and apoptotic pathway

Among the targeted pathways for cancer treatment are the tumour promoting inflammation pathway and apoptotic pathway. The strong involvement of E3 ligases in NF-κB and apoptotic pathways, present them as an important target for inducing apoptosis and inhibiting inflammation in cancer cells. Hence, small compounds targeting the concerned E3 ligases such as SCF^β-TrCP, TRAF6, LUBAC and IAPs have been developed for cancer treatment.

β-TrCP is targeted by the small molecule GS143 for the inhibition of NF-κB-mediated inflammation in cancer. GS143 inhibits β-TrCP-mediated ubiquitination of IκBα to prevent activation of NF-κB.²⁸⁷

TRAF6 in association with the E2 ubiquitin-conjugating enzyme Ubc13 mediates K63 polyubiquitination for NF-κB inflammation signalling.³⁵ The small molecule inhibitor of TRAF6, C25-140, inhibits TRAF6–Ubc13 interaction to reduce TRAF6 E3 ligase activity.²⁸⁸ The interaction of CD40 with TRAFs particularly TRAF6 has been shown to activate inflammatory signalling via the NF-κB pathway. This finding prompted the identification of a small molecule inhibitor of TRAF6–CD40, 6877002 that binds TRAF6 and blocks TRAF6 interaction with CD40 thereby suppressing NF-κB inflammation signalling.^{174, 289}

Most small molecule inhibitors of LUBAC, target HOIP, the catalytic subunit in LUBAC. The small molecule, Gliotoxin, inhibits LUBAC by binding to the Ring-IBR-Ring domain of HOIP to prevent linear ubiquitin chain formation.²⁹⁰ Johansson et al.²⁹¹ designed α, β-unsaturated methyl ester-linked compound (Compound 5) which binds the active cysteine to inhibit HOIP E3-UB thioester bond formation thereby preventing NF-κB activation. Another similar LUBAC antagonist, HOIPINs bind active Cys885 and other associated residues in HOIP to suppress the RING–HECT-hybrid reaction of HOIP, thus hindering LUBAC-mediated linear ubiquitination.²⁹²

IAP contains a BIR domain that binds caspases in the cell to inhibit cell apoptosis.²⁰⁸ Natural IAP antagonist mimetics (SMAC mimetics) are currently developed and under clinical evaluation in cancer subjects.²¹⁷ Small compounds AT-406 (DEBIO1143), GDC-0152, LCL161 and Birinapant (TL32711), are SMAC mimetics under clinical evaluation (Table 4), but the clinical trial for GDC-0152 was terminated after only one phase I study of safety and pharmacokinetics in patients with advanced malignancy although the reason for the termination is not related to patients’ safety or anti-tumour activity (ClinicalTrials.gov: NCT00977067).^293-295 NF-κB inflammation signalling and the extrinsic apoptotic pathway crosstalk between complex I and complex II via TNFR1 signalling, and this point could be considered a potential target for cancer therapy.²²⁶ For instance, TRAF2 recruits cIAP to complex I and prevents the formation of apoptotic promoting complex II. In this case, Birinapant with a high affinity for the BIR3 domain of cIAP binds cIAPs that are associated with TRAF2 to induce their autoubiquitination and degradation.²⁹⁶ By this means, Birinapant abrogates complex I-induced NF-κB activation and promotes RIP1-mediated complex II formation for apoptosis of tumour cells.²⁹⁶

Besides the polyubiquitinating E3 ligase, small molecule inhibitors of monoubiquitinating E3 ligase such as PRC1 responsible for H2A monoubiquitination have been developed. The catalytic components of PRC1 such as RING1 A or B/BMI1 complex are associated with several cancers and correlate with poor prognosis, thus PRC1 represents an attractive therapeutic target for cancer. PRT4165 was initially designed as small molecule inhibitor of BMI1/Ring1A-mediated ubiquitination and degradation of topoisomerase II α (Top2α)/drug complex and was later found to inhibit PRC1-induced H2A and H2AX ubiquitination at DNA damage site. PRT4165 abrogates DNA double-strand break repair thereby promoting cell cycle arrest and death of DNA-damaged cells.^{297, 298} Su and co-workers²⁹⁹ designed GW-516, an inhibitor of H2A ubiquitination that catalytically inhibits the RNF2 component of PRC1 to truncate immunosuppression and prevent metastasis of prostate cancer cells. Recently, Grembecka and Cierpicki's laboratory developed the small molecule RB-3 that inhibits RING1B/BMI1-mediated H2A ubiquitination in cancer cells. RB-3 binds RING1B/BMI1 complex and induces conformational changes that disrupt their interaction with nucleosomes. RB-3 was also shown to induce differentiation in leukaemia cell lines and AML, establishing that RB-3 may act as a promising therapeutic compound for leukemic cells.³⁰⁰

In summary, these studies are convincing evidence that E3 ligases are credible and attractive targets for cancer therapy. The studies also show that small molecule compounds can be directed at E3 ligases, E3 ligase–substrate interaction point or E3 ligase substrates as in the case of RITA to inhibit E3 ligase-mediated substrate degradation. With a complete understanding of the E3 ligase mechanism within ubiquitin–proteasome system, more efficient and potent E3 ligase inhibitors will be developed for clinical assessment.