E3 ubiquitin ligases in cancer stem cells: key regulators of cancer hallmarks and novel therapeutic opportunities

Abstract

Background

Human malignancies are composed of heterogeneous subpopulations of cancer cells with phenotypic and functional diversity. Among them, a unique subset of cancer stem cells (CSCs) has both the capacity for self-renewal and the potential to differentiate and contribute to multiple tumor properties. As such, CSCs are promising cellular targets for effective cancer therapy. At the molecular level, hyper-activation of multiple stemness regulatory signaling pathways and downstream transcription factors play critical roles in controlling CSCs establishment and maintenance. To regulate CSC properties, these stemness pathways are controlled by post-translational modifications including, but not limited to phosphorylation, acetylation, methylation, and ubiquitination.

Conclusion

In this review, we focus on E3 ubiquitin ligases and their roles and mechanisms in regulating essential hallmarks of CSCs, such as self-renewal, invasion and metastasis, metabolic reprogramming, immune evasion, and therapeutic resistance. Moreover, we discuss emerging therapeutic approaches to eliminate CSCs through targeting E3 ubiquitin ligases by chemical inhibitors and proteolysis-targeting chimera (PROTACs) which are currently under development at the discovery, preclinical, and clinical stages. Several outstanding issues such as roles for E3 ubiquitin ligases in heterogeneity and phenotypical/functional evolution of CSCs remain to be studied under pathologically and clinically relevant conditions. With the rapid application of functional genomic and proteomic approaches at single cell, spatiotemporal, and even single molecule levels, we anticipate that more specific and precise functions of E3 ubiquitin ligases will be delineated in dictating CSC properties. Rational design and proper translation of these mechanistic understandings may lead to novel therapeutic modalities for cancer procession medicine.

1. Introduction

In adult tissues, normal stem cells are endowed with the ability to differentiate into various cell types comprising multi-cellular organs [1]. Similarly, in human malignancies, a small subset of un-differentiated cancer cells, called cancer stem cells (CSCs), maintain tumorigenesis by infinite self-renewal and aberrant differentiation. Over the past few decades, CSCs have been intensively studied in the field of oncology. As a subset of tumor cells with unlimited proliferation potential, CSCs have increased expression of stem cell biomarkers. In addition, several stemness-related signaling pathways are activated to maintain unique transcriptomic signatures in CSCs, thereby driving metabolic reprograming, growth, and metastasis to fuel tumor progression and recurrence [2–4]. From a clinical perspective, CSCs are enriched in patient tumor samples undergoing chemo/radio-therapy and molecular targeted therapy. These CSCs are shown to be resilient to current therapeutic modalities in vitro and in vivo, leading to treatment failures in clinical settings. In this regard, treatment strategies tailored to eliminate CSCs could be promising anti-cancer strategies, thus warrant further investigation on the regulatory mechanisms that support CSCs self-renewal and fitness in the tumor microenvironment.

CSCs are maintained by multiple stemness regulatory signaling pathways and transcription factor networks, as well as the tumor microenvironment [5–7]. These signaling pathways include the Hippo, Notch, Wnt (Wingless/Integrated), Hedgehog (Hh), NF-κB (nuclear factor-κB), TGF-β (transforming growth factor)/SMAD, PI3K (Phosphatidylinositide 3-kinases)/AKT, JAK (Janus kinase)/STAT (signal transducers and activators of transcription) and PPAR (peroxisome proliferator-activated receptor) pathways [8, 9]. Aberrant activation of these pathways converge and activate several stemness transcription factors, e.g. Nanog, OCT4 (Octamer-Binding Transcription Factor 4) and SOX2 (SRY-Box Transcription Factor 2) for CSC maintenance and tumorigenesis [10]. Moreover, as a unique cellular subset residing in their niche, CSCs frequently interact with the tumor microenvironment through various soluble factors as well as physical contact, thereby dictating stemness regulatory signaling pathways for CSC survival and self-renewal [11].

Biochemically, these signaling pathways are regulated by various post-translational modifications (PTMs) such as phosphorylation, acetylation, methylation, and ubiquitination. PTMs on signaling components are controlled by corresponding writers and erasers, and recognized by reader proteins [12]. Physiologically, PTMs as well as their catalyzing machinery regulates normal stem cell functions, while aberrancies of PTMs on stemness signaling pathways and transcription factors are casually linked to CSC maintenance, thereby promoting cancer development [13]. While the roles of phosphorylation and acetylation have been extensively discussed in regulating signaling pathways and CSCs [14], a comprehensive narration of ubiquitination- and deubiquitination-mediated control of CSC properties is still lacking. To this end, we systematically review the versatile roles of E3 ubiquitin ligases in regulating biological properties of CSCs, with a focus on clinical translation strategies of targeting E3 ubiquitin ligases for anti-cancer therapy.

2. The ubiquitin proteasome system

2.1. Ubiquitination

The ubiquitin-proteosome system (UPS) is the most significant protein degradation pathway in human cells, which controls protein stability by degrading protein in two stages, including substrate protein ubiquitination followed by proteasome-mediated degradation of substrate proteins. Ubiquitination is an ATP-dependent PTM occurring on proteins. Ubiquitin (Ub), a 76-amino acid polypeptide, is covalently linked to substrate proteins which is catalyzed by a cascade of ubiquitin-activating enzyme E1, ubiquitin-binding enzyme E2, and ubiquitin ligase E3. The first enzyme involved in this process, an ubiquitin-activating E1 enzyme, forms a covalent intermediate with Ub via a thioester bond formed via a cysteine residue, driven by ATP. The E1 then transfers the activated Ub to the thiol group of a cysteine on the ubiquitin-binding E2 enzyme. Finally, an E3 ubiquitin ligase, a key factor in the selective degradation mechanism of the UPS, recognizes the degraded protein and ligates Ub to the substrate [6].

The classical ubiquitin modification types includes monoubiquitination, where a Ub is attached to a single lysine residue, and polyubiquitination as a result of either monoubiquitination at multiple lysine residues or the formation of a Ub chain in which the second, and subsequent Ubs, are attached to a lysine within the previously attached Ub protein [6]. Due to the presence of seven lysine residues in Ub, which can each serve as an acceptor site of the next Ub, the substrate protein has the ability to form a wide variety of chain linkages which serve as recognition signals for proteasome degradation. Each ubiquitin linkage can result in a different biological outcome: K6-linked chains play an increasingly prominent role in the tumor suppressor BRCA1, DNA repair, and mitochondrial regulation; K11 chains are a key regulator of mitogenic protein degradation to assists in the control of cell cycle progression and endoplasmic reticulum (ER)-mediated protein degradation; K27-linked chains inhibit caspase activation and TRAIL-induced apoptosis, thereby promoting cancer cell survival; K29-linked chains are involved in the regulation of proteasome function and affect epigenetics; K33-linked chains are one of the regulatory mechanisms of intracellular transport; K48-linked chains, the most common Ub linkage, mediates proteasome-dependent degradation; and K63-linked chains primarily regulates DNA damage response, intracellular sorting, autophagy of misfolded or aggregated proteins, and neurodegeneration [15, 16]. Therefore, ubiquitination regulates numerous biological processes, such as control of cell cycle progression, regulation of carcinogenesis, tumor immunity, transcriptional regulation and intracellular signaling pathways. Dysregulation of ubiquitination processes contributes to serious human diseases such as obesity, neurodegeneration, and cancer.

2.2. Deubiquitination

Deubiquitination is carried out by deubiquitinating enzymes (DUBs). Mechanistically, DUBs are equivalent to blocking E3 ubiquitin ligase function by disrupting the connection between the substrate and Ub proteins. Thus, DUB-mediated deubiquitination activity counteracts ubiquitination thereby reversing the effect on target protein function including protecting proteins from proteasomal degradation. The balance of ubiquitination and deubiquitination events helps regulate cellular protein homeostasis [7]. Approximately 100 deubiquitinating enzymes are split into seven distinct groups: (1) ovarian tumor proteases (OTUs), (2) ubiquitin-specific protease (USPs), (3) Monocyte chemotactic protein-induced protease (MCPIPs), (4) Josephin and JAB1/MPN⁺ (MJP), (5) JAB1/MPN/Mov34 metalloenzyme, (6) ubiquitin C-terminal hydrolases (UCHs) [17], and (7) MIU-containing novel DUB (MINDY) [18]. Importantly, DUBs have been shown to function in a variety of ubiquitin pathways and play central roles in several cell biological processes [7, 17, 19].

3. The classification and features of E3 ubiquitin ligases

E3 ubiquitin ligases are the most abundant enzyme in the UPS, with more than 600 enzymes, encoded in mammalian genomes, which are classified into four families according to molecular composition and activation mechanisms, including RING-type E3s, HECT-type E3s, U-box type E3s, and RBR-type E3s [20]. These different types of E3 ubiquitin ligase have low sequence homology and large variation in their composition. E3 ubiquitin ligases are responsible for binding to, and targeting, substrate proteins for ubiquitination, thus this high variability is integral in the ability of E3 ubiquitin ligases to each target specific proteins and regulate distinct processes within the cell. Finally, adding another layer of regulation, substrate proteins of E3 ubiquitin ligases typically contain specific amino acid recognition sequences, termed degron domains, which are often subjected to post-translational modifications, including but not limited to phosphorylation, acetylation, and hydroxylation, linking their degradation to upstream signal-dependent control mechanisms [21].

3.1. HECT‑type E3s

Some of the earliest studies in protein degradation have identified 28 HECT E3s in humans, which makes them some of the most well-studied E3 ubiquitin ligases. HECT E3 ubiquitin ligases have a variable N-terminal extension and a conserved catalytic HECT domain with an active site cysteine that controls substrate recognition selectivity. These E3s are further classified into three subgroups based on the varying N-terminal extensions, including Nedd4/Nedd4-like E3s contains WW domain (9 members), the HERC family E3s (6 members), and a subfamily without WW and RLD domains (13 members) [22, 23].

3.2. RING‑type E3s

In addition to HECT-type E3 ubiquitin ligases, many of the most well-known E3 ubiquitin ligases are found within the RING family and are distinguished by their RING or U-box domains. While the RING finger folds of the two domains are identical, the RING domain function is dependent on chelation of two zinc ions (Zn²⁺), whereas the U-box domains function is independent of Zn²⁺ [21]. RING E3s serve as a scaffold for the simultaneous binding of the target substrate and an active E2-Ub conjugate during ubiquitination, facilitating the Ub moiety translocation from the E2-Ub conjugate to a substrate lysine residue [20]. Notably, some Cullin-RING ligases (CRLs), which are RING-type E3s, form large multiprotein complexes with many subunits to facilitate ubiquitination. A minimum of four subunits are shared by all CRLs, including an E2-binding catalytic ring finger (Rbx), one of the seven Cullins (CUL1, CUL2, CUL3, CUL4A, CUL4B, CUL5, and CUL7), an adapter (Skp1) that connect the receptor and the Cullin scaffold, and a receptor for substrate recognition. The anaphase-promoting complex/cyclosome (APC/C) and Skp1/CUL1/F-box (SCF) E3 ubiquitin ligases are two well studied classes of CRLs [20].

3.3. U‑box‑type E3s

U-box E3 ubiquitin ligases are less abundant than other families of E3s and are more prevalent in plants, yet there are conserved members of this family from yeast to humans which utilize intramolecular interactions for its scaffold stabilization [24]. The C-terminus of U-box E3s contain a conserved U-box domain of approximately 70 amino acids and has a three-dimensional structure comparable to the RING finger domain [21]. Similar to CRLs, the U-box E3 catalyzes ubiquitination by interacting with a E2 and the U-box domain, which enables direct transfer of the Ub from the E2 to the target substrate lysine [24].

3.4. RBR‑type E3s

The RING-IBR-RING (RBR) type of E3s are a newly identified family, which are regarded as RING-HECT hybrid E3s. RBR E3s are characterized by a conserved catalytic region composed of three domains comprising two RING domains and a central in between-RINGs (IBR). The catalytic cysteine is located in the second RING domain, and the first RING recruits an E2 that is laden with ubiquitin [25]. In addition, several RBR E3s have unique domains that set them apart from one another. For instance, some RBR E3s can interact with other molecules to keep themselves in a state of autoinhibition. Furthermore, the RING2 domain has an active site cysteine that is absent from RING-type E3s but comparable to HECT-type E3s. As a result, the sequence and domain structure of RBR-type E3s suggest that they are RING-HECT hybrids. RBRs conjugate the Ub to the substrate via a two-step catalytic mechanism similar to the HECT-type E3s [25]. RBR E3s tend to ubiquitinate substrates by a distinct mechanism to produce linear ubiquitin chains, despite the fact that they generally resemble HECT E3 ubiquitin ligases. Therefore, the RBR E3 ubiquitin ligases selectivity point to a novel and intriguing mechanism that needs additional elucidation [26].

4. The roles of E3 ubiquitin ligases in regulating hallmarks of Cancer stem cells

Protein ubiquitination and subsequent degradation regulate biological functions of cells by maintaining protein stability and responding to stress stimuli. Therefore, disorder of the UPS contributes to a variety of human diseases, especially cancer. Ubiquitination and deubiquitination regulate CSC development, and dysregulation of this mechanism can alter stem cell properties to promote cancer. DUBs have also been implicated in tumor progression including the regulation of stem cell signaling pathways and transcription factors, and has recently been extensively discussed [17, 19]. E3 ubiquitin ligases have also been shown to be involved in CSC maintenance and differentiation, which in turn regulates cellular homeostasis and cancer development. It has been reported that depletion of E3 ubiquitin ligases in CSCs isolated from multiple patients promotes apoptosis or cell differentiation, suggesting targeting E3 ubiquitin ligase may serve as a novel mechanism to mitigate the impact of CSCs on tumor progression [27, 28]. In addition, E3 ubiquitin ligases regulate CSC self-renewal and cancer progression by reducing core transcription factor activity, activating specific signaling pathways, metabolic reprogramming, and activating the EMT program. Therefore, some E3 ligases may negatively affect the survival and proliferation of CSCs by degrading core transcription factors or inhibiting specific signaling pathways, while other E3 ubiquitin ligases may positively or negatively regulate stem cell phenotypes [29]. For greater clarity, specific E3 ubiquitin ligases and their mechanisms for regulating CSC hallmarks are described below.

4.1. Sustaining growth signaling to support CSC self‑renewal and proliferation

CSCs undergo asymmetric division to maintain a CSC population through self-renewal and undergo differentiation thereby promoting tumor formation [1]. That is to say, CSC cell division may lead to the formation of two CSCs, or a non-CSC and a CSC. Recent studies have also found that non-CSCs can dedifferentiation to increase the CSC population, thereby promoting cancer proliferation, tumor development, and treatment resistance. The self-renewal and proliferation potential of CSCs are controlled by various growth signaling pathways, which are extensively regulated by E3 ubiquitin ligase-mediated protein ubiquitination on their core components (Fig. 1).

Fig. 1.

The role of E3 ligases in regulating self-renewal and uncontrolled proliferation of CSCs. E3 ubiquitin ligases regulate self-renewal, proliferation and growth of CSCs through RTK, Notch, JAK-STAT, WNT, TGF-β, and SHH signaling pathways. Red and light green icons represent E3 ligases with tumor-promoting and tumor-suppressing activities, respectively

4.1.1. SKP2

Skp2 is a substrate recognition subunit within the SCF E3 ubiquitin ligase complex recognizing substrates through a leucine-rich repeats (LRR) motif. Skp2 can ubiquitinate and degrade multiple substrate proteins. For example, Skp2 regulates cell fate by promoting ubiquitination and degradation of p27, a G1/S cyclin dependent kinase inhibitor and a potent tumor suppressor [30]. As such, p27^−/− mice show multi-organ hyperplasia and the spontaneous growth of pituitary tumors, indicating that p27 is essential for attenuating proliferation of cancer cells, including CSCs [31]. As a core component of the PI3K/AKT/mTOR signaling, AKT is hyperactive and driving many types of cancers [32], which is partially due to Skp2-modified K63-linked ubiquitination to activate AKT [33]. In addition, TRAF6 also promotes K63-linked ubiquitination of AKT and is required for plasma membrane translocation of AKT, thereby promoting tumorigenicity [34]. Thus, CSCs may also be regulated by Skp2- and TRAF6-mediated activation of AKT signaling. Indeed, genetic depletion or pharmacological inhibition of Skp2 can significantly reduce the amount of ALDH⁺ CSCs in prostate cancer (PCa) [35]. Thus, Skp2 may be a novel cancer therapeutic target for eliminating CSCs.

4.1.2. FBXW7

FBXW7, another SCF complex member, targets various oncoproteins for ubiquitination and destruction, such as c-Myc and Notch. Notch is a powerful oncogene and a crucial regulator of CSC self-renewal, as observed in T cell acute lymphoblastic leukemia (T-ALL). Therefore, FBXW7 degradation of Notch through ubiquitination suppresses self-renewal of leukemia CSCs, thus inhibiting the progression of lymphocytic leukemia [36]. In chronic myelogenous leukemia (CML), FBXW7 functions to suppress initiation and progression of CML, through ubiquitination and degradation of c-Myc. Leukemic initiating cells (LICs) with FBXW7 mutations, or deletions overexpress c-Myc, which results in the development of CML [37].

4.1.3. SOCS6

The substrate recognition subunits of the CUL5 E3 ubiquitin ligases are well known and include the SOCS proteins. They perform ubiquitination function by using Src homology 2 (SH2) domains to recognize substrate proteins. Through the action of the SOCS box structure, SOCS6 facilitates the ubiquitination and degradation of receptor tyrosine kinases (RTKs) [38]. Previous research has shown that SOCS6 can bind and degrade the proteins YAP, JAK2, and Sin1 and functions as a tumor suppressor in esophageal carcinoma [39–41] where SOCS6 downregulation can increase cancer cell growth and the proportion of CSCs.

4.1.4. MULE

The HECT type E3 ubiquitin ligase Mule regulates the ubiquitination of multiple substrates, some of which have been found to have opposing roles in cell proliferation and tumorigenesis, such as p53 and c-Myc. At least in hyperproliferative settings, Mule can act as a “backup” E3 ubiquitin ligase for degradation of β-catenin, thereby helping to quench Wnt signaling in the intestine. Thus, Mule has indirect but important effects on intestinal stem cell maintenance [42]. Mule has been identified as a powerful tumor suppressor for colon cancer in mice, where its loss and APC mutation together induce c-Myc expression to enhance CSCs and tumorigenesis [43].

4.1.5. SPOP

Speckle-type POZ domain protein (SPOP) is a substrate recognition subunit of the CUL3 RING-type family of E3 ubiquitin ligases. Two studies demonstrate that SPOP regulates PCa CSCs traits by down-regulating the Nanog protein [44, 45]. Mechanistically, AMPK signaling induces phosphorylation of Nanog, thus priming Nanog for SPOP-mediated proteosomal destruction and regulating PCa CSC maintenance. Consistently, ectopic expression of SPOP leads to differentiation of mouse embryonic stem cells (mESCs) [45]. SPOP is also a tumor suppressor in pancreatic cancer, due to reduced expression and loss of function mutation, leading to Nanog overexpression to support CSC growth [46]. Moreover, the Bromodomain-containing proteins 4 (BRD4) and Cyclin E1 are also marked by SPOP for ubiquitination and proteosomal destruction to regulate PCa CSCs proliferation [47–50]. Thus, SPOP could be exploited to therapeutically eliminate CSCs and treat malignant cancers such as PCa.

4.1.6. TRIM8

As a member of the TRIM protein family, TRIM8 maintains glioblastoma stem cells (GSCs) stemness and self-renewal capacity by activating STAT3 signaling, which promotes the preservation and growth of GSCs [51]. In terms of mechanism, TRIM8 indirectly upregulates the expression of NESTIN, SOX2 and p-STAT3, thereby increasing the proliferation capacity of GSC, and knockdown of TRIM8 can promote glial cell differentiation. It has been reported that TRIM8 activates STAT3 by blocking the function of the STAT3 inhibitor PIAS3 through E3-mediated ubiquitination. Importantly treatment of cells with STX-0119, a small-molecule STAT3 inhibitor, led to the suppression of p-STAT3, SOX2, and c-Myc [52]. These findings imply that TRIM8 might serve as GBM treatment target.

4.1.7. TRIM16

TRIM16 inhibits tumor progression, migration, differentiation, and apoptosis through regulatory mechanisms of growth inhibition. In breast cancer, abnormal activation of TRIM16 dramatically reduced the development of spheres. TRIM16 depletion increases the proportion of CD44⁺/CD24⁻ cells, which suggests that TRIM16 expression level inversely correlate with spheroid formation ability and that TRIM16 may function to suppress multiple CSC phenotypes. Mechanistically, TRIM16 directly degrades the Gli-1 protein, a potent positive regulator of target genes downstream of the Sonic Hedgehog (SHH) pathway, through the ubiquitin-proteasome pathway, which in turn increases CSCs [53]. Thus, TRIM16 increases the number of CSCs in relation to the cells that compose the breast cancer tumor mass, which promotes tumorigenesis.

4.1.8. TRIM28

Another TRIM protein family member, TRIM28, downregulation reduces CSCs self-renewal, which significantly slows the progression of tumors. Meanwhile, inactivation of TRIM28 results in cell cycle disruption and a reduction in tumor cell metabolism [54]. CSCs are directly regulated by each of the methods outlined above. Furthermore, TRIM28, Cnot3, Zfx, c-Myc and other pluripotency markers formed a transcriptional pattern in the self-renewal transcriptional network that is different from the Nanog-Sox2-Oct4 module [55]. These findings reveal a role for TRIM28 in regulating CSCs through pluripotency factors, and may open the door for novel and more potent treatments that target CSCs.

4.1.9. NEDD4 and NEDD4L

It has been demonstrated that NEDD4 influences the expression of the stemness-related genes Sox9, Sox10, and FoxD3 to control the formation of craniofacial tumors [56]. Studies have demonstrated PTEN localization, and thus tumor suppressor function, is regulated by monoubiquitination potentially facilitated by NEDD4, thus suggesting it may serve as an oncoprotein [57]. In this context, expression of NEDD4 has been shown to target FOXA1 to promote tumorigenesis [58]. Contrary to these findings, deletion of NEDD4 enhanced colon cancer growth in the APC^min model [59]. Furthermore, crypt stem cells expresses NEDD4 and NEDD4L, which control intestinal stem cells (ISCs). Loss of Nedd4 and Nedd4L promotes the growth of ISCs and hastens the development of tumors. Mechanistically, NEDD4 and NEDD4L block Wnt/β-catenin signaling by promoting the ubiquitination and degradation of the LGR5 receptor and DVL2 [60]. Thus, loss of NEDD4 and NEDD4L increases ISC proliferation and self-renewal. Further study is necessary to fully explain the role of NEDD4 in tumorigenesis and in which contexts it may act as an oncogene versus a tumor suppressor gene, as well as whether it modulates CSC populations in either of these situations.

4.1.10. PJA1

PJA1 is a major TGF-β signaling regulator in hepatocellular carcinoma (HCC). PJA1 overexpression blocks SMAD3 and SPTBN1-mediated TGF-β tumor suppressor signaling and CSC-suppressing activities to promote HCC proliferation. In this context, the SMAD3/β2SP-mediated TGF-β pathway promotes tumor suppressor gene expression. However, PJA1 interacts with and mediates the ubiquitination of β2SP and SMAD3, thus reducing tumor suppressor gene expression and increasing cell proliferation [61, 62]. Furthermore, SMAD3 and β2SP regulate stem cell properties, which may contribute to the formation of CSCs, linking PJA1 to stemness capacity. Therefore, increased PJA1 may promote the progression of HCC by inhibiting TGF-β signaling to upregulate CSCs and tumor growth. Importantly, RTA405, an inhibitor of E3 ubiquitin ligases, increased SMAD3 activity and reduced the development of HCC tumors in mice [61]. Therefore, targeting PJA1 in the context of TGF-β signaling may be an effective treatment for HCC potentially through reducing CSC capacity.

4.1.11. UBE3C

In non-small cell lung cancer (NSCLC), the E3 ubiquitin ligase UBE3C ubiquitinates and degrades AHNAK to maintain stemness of tumor cells [63]. As a cofactor of p53, ANHAK facilitates the transcriptional repression of stemness-related genes via p53 binding to their promoters. Thus, stemness features were improved by ANHAK downregulation due to UBE3C overexpression, which eliminated p53-mediated repression of gene expression.

4.2. Promoting invasion and migration of cancer stem cells

Invasion and metastasis are two primary hallmarks of cancer and CSCs [2]. In-depth analysis of the roles of CSCs in maintaining tumor invasion and metastasis will lead to improved development of cancer therapies with greater efficacy. However, the specific mechanism by which CSCs initiate invasion and metastasis cascades remains largely unclear, and our limited knowledge primarily focuses on the EMT and the influence of the tumor microenvironment. Ubiquitination, and the E3 ubiquitin ligases mediating substrate selectivity, have been shown to regulate key signaling pathways that are involved in CSCs invasion and metastatic capacities (Fig. 2).

Fig. 2.

The role of E3 ligases in regulating invasion and metastasis of CSCs. E3 ubiquitin ligases regulate the invasion and metastasis of CSCs through RTK, WNT, and NF-kB signaling pathways. Red and light green icons represent E3 ligases with tumor-promoting and tumor-suppressing activities, respectively

4.2.1. NEDD4

NEDD4-mediated monoubiquitination of CD133, a well-known CSC marker, was recently shown to promote its incorporation into extracellular vesicles and through this enhance migration of CSCs [64]. As described above, NEDD4 has also been shown to promote the ubiquitination and degradation of FOXA1 which inhibits the EMT and cancer cell metastasis [58, 65]. Therefore, NEDD4 may promote CSC invasion and metastasis through this NEDD4/FOXA1/miR-340-5p/ATF1 axis. Furthermore, NEDD4 has also been shown to degrade the tumor suppressor PTEN to promote CSC invasion and metastasis [57, 66].

4.2.2. TRIM28

As described above, TRIM28 reduces self-renewal of breast CSCs, leading to a significant reduction in tumor growth. Furthermore, TRIM28 has been suggested to be involved in invasion and metastases in glioblastoma. TRIM28 is enriched in the tumor core, indicating that TRIM28 may regulate metastasis of CSCs. Importantly, the anti-TRIM28-selective nanobody NB237 can significantly inhibit the invasion and spread of CSCs in the zebrafish brain [67], suggesting a novel potential therapeutic for this cancer type.

4.2.3. FBXW7

A number of human malignancies have mutations in the tumor suppressor FBXW7. TWIST-1 and SNAIL-1, two EMT-regulated genes, are often overexpressed when FBXW7 is depleted [68]. As a target of FBXW7, mTOR is elevated in response to depletion of FBXW7, resulting in induction of the EMT, an increase in cell motility, invasiveness, and metastasis of colorectal cancer. The impact of FBXW7 depletion-induced invasion and metastasis is reversed by the mTOR inhibitor rapamycin [69, 70]. Additionally, it is probable that FBXW7 controls the EMT via Aurora A kinase, another one of its targets [71].

4.2.4. NRDP1

NRDP1 is a RING-type E3 that is downregulated in glioblastoma and promotes K63-linked ubiquitination of DVL2, which binds to Frizzled receptors leading to reduced non-canonical Wnt signaling thereby promoting CSCs invasion and migration [72].

4.2.5. RNF128

As a RING-type E3, RNF128 (also referred to as GRAIL) is downregulated in melanoma when compared to peritumoral tissue, and downregulation of RNF128 correlates with a poor prognosis. Mechanistically, Wnt signaling and cellular EMT are activated, and CSCs turnover is induced by the ubiquitination and degradation of CD44/CTTN when RNF128 is downregulated. Therefore, RNF128 is a reliable diagnostic and prognostic biomarker, and further understanding of RNF128 may lead to novel treatment strategies for melanoma [73]. However, more studies are necessary to unravel which pathways targeted by RNF128 may regulate invasion and metastasis through controlling CSC stemness and functionality.

4.2.6. TRIM21

Another TRIM family member, TRIM21, inhibits EMT progression through ubiquitination and degradation of Snail, a key EMT regulator, whose high expression correlates with tumor invasion and metastasis. Conversely, cancer-associated mutations of TRIM21 which have been shown to abolish TRIM21-mediated Snail ubiquitination and degradation, thereby increasing invasion and migration of breast CSCs [74].

4.2.7. SPOP

SPOP is often down-regulated in PCa tissues and predicts poor prognosis and shortened survival. Mechanistically, the SPOP promoter contains a SMAD-binding element (SBE) for SMAD3. Therefore, the TGF-β/SMAD pathway may reduce SPOP expression, thereby promoting migration and invasive properties of PCa CSCs. Notably, SB431542, a TGF-β pathway inhibitor, upregulates SPOP expression and suppresses invasiveness of CSCs [75]. In addition, loss of SPOP expression also promotes migration and invasion by maintaining stability of Nanog [44–46] and BRD4 [47–49] in CSCs.

4.2.8. HECTD3

HECTD3, a HECT E3 ubiquitin ligase family member, promotes tumor cell survival in many cancer types. Overexpression of HECTD3 promotes K63-linked ubiquitination and expression of MALT1. In breast cancer with angiotensin II receptor positivity, maintained MALT1 triggers the CARMA3/Bcl10/MALT1 signaling, contributing to growth and invasion of both bulk cancer cells and CSCs [76].

4.2.9. CUL1

The scaffold subunit CUL1 is a crucial part of SCF complexes which can interchange any one of 69 F-box proteins including many discussed above such as FBXW7 and β-TrCP1. Increased expression of CUL1 is positively correlated with poor prognosis of breast cancer patients, and CUL1 promotes migration, invasion, tube formation, angiogenesis, and metastasis of both breast cancer bulk tumor cells and CSCs in vivo. One contributing mechanism is that the PI3K/AKT/mTOR signaling pathway exacerbates breast cancer cell metastasis as a result of CUL1 controlling EZH2 to enhance production of autocrine cytokines CXCL8 and IL11. Therefore, CUL1 might offer a practical therapeutic target for preventing breast CSCs metastasis [77].

4.2.10. Parkin

The RBR type E3 ubiquitin ligase Parkin was initially identified due to it being mutated in Parkinson’s disease patients. Parkin has a well-established role in the regulation of mitophagy to manage ROS levels and mitochondrial homeostasis. In addition, Parkin serves as a tumor suppressor and is deactivated in many human malignancies. For instance, it has been confirmed that Parkin targets lysine 477 on HIF-1α for ubiquitination and degradation to prevent breast cancer cell migration and invasion [78]. Therefore, it is possible that Parkin may also assist in regulating the invasion and metastasis of CSCs. However, further studies are necessary to elucidate its role in these processes.

4.3. Facilitating metabolic reprogramming of cancer stem cells

Malignant transformation of normal cells is associated with nutrient uptake and metabolic reprogramming, to meet the needs of a perpetually proliferative state. Tumor cells modify their metabolism in also every aspect, such as changing glucose utilization from aerobic to anaerobic pathways, referred to as the Warburg effect [79]. In addition, tumor cells also change their lipid and amino acid metabolism. As a key cellular subset, CSCs display unique metabolic properties as compared to their relatively differentiated cancer cells. At the molecular level, metabolic pathways are controlled by rate-limiting metabolic enzymes and signaling pathways, which are regulated by E3 ubiquitin ligases to alter the metabolism of CSCs (Fig. 3).

Fig. 3.

The role of E3 ligases in regulating metabolic reprogramming of CSCs. E3 ubiquitin ligases regulate metabolism of CSCs through RTK, NF-κB and glucose metabolism signaling pathway, and hypoxic microenvironment. Red and light green icons represent E3 ligases with tumor-promoting and tumor-suppressing activities, respectively

4.3.1. CUL3

CUL3 binds to potassium channel tetramerization domain 2 (KCTD2), which interacts with c-Myc, promoting its ubiquitination and subsequent degradation. Interestingly, KCTD2 deficiency affects aerobic glycolysis through c-Myc protein regulation and expression changes of glycolysis-related genes, to promote the acquisition of CSCs characteristics [80].

4.3.2. SKP2

As a substrate recognition component of the SCF complex, Skp2 also mediates K63-linked ubiquitination of AKT which can promote aerobic glycolysis, and reduce senescence of CSCs [33, 81]. Highly selective pharmacological inactivation of Skp2 has strong anticancer properties and works with chemotherapeutics to lower cancer cell survival in a variety of animal models [35].

4.3.3. TRIM28

Overexpression of melanoma-associated antigen A3 and C2 (MAGE-A3 and MAGE-C2) in HCC enhances TRIM28-dependent degradation of FBP1 by forming a ubiquitin ligase complex with TRIM28, thereby increasing glucose consumption and lactate production, regulating tumor cells and CSCs metabolism and HCC tumorigenesis [82].

4.3.4. PPARγ

Peroxisome proliferator-activated receptors (PPARs) are key regulators of lipid metabolism and contain three major components, including PPARα, PPARγ and PPARβ/δ. Among these, PPARγ, which can function as an E3 ubiquitin ligase, inhibits NFκB-mediated inflammation and tumor growth by ubiquitinating and degrading NFκB/p65 [83]. PPARα can bind to the adipoQ receptor 3 (PAQR3) protein and progesterone, and PAQR3 enhances PPARα ubiquitination targeted by the E3 ubiquitin ligase HUWE1. However, PPARα degradation directly influences its function in lipid metabolism [84], and therefore targeting PPARr for degradation may be an important mechanism regulating lipid metabolism in tumor cells and CSCs.

4.3.5. CHIP

The U-box-containing E3 ubiquitin ligase CHIP can facilitate the ubiquitination and subsequent proteasome degradation of pyruvate kinase isoenzyme M2 (PKM2), a key tumor glycolysis regulator that is frequently dysregulated during tumorigenesis [85]. PKM2 is crucial for the metabolism and growth of tumors as it provides a boost in energy production and channels glycolytic intermediates for the synthesis of essential building blocks that cancer cells require [86]. CHIP targets PKM2 for proteasome degradation leading to inhibition of CSCs and ovarian development by inhibiting anaerobic glycolysis and reversing the Warburg effect [85].

4.3.6. VHL

The heterogeneous state of the tumor microenvironment is an integral component regulating metabolic reprogramming of solid tumors. One of the tumor microenvironmental factors closely linked to cancer metabolism is hypoxia. Hypoxia-inducible factor 1 (HIF-1), a key component of the cellular response to hypoxia, is a heterodimeric protein with two major components, HIF-1α and HIF-1β. HIF-1α is a transcription factor whose nuclear translocation upregulates multiple genes related to glucose metabolism in tumor cells. The tumor suppressor Von Hippel-Lindau (VHL) is the most well know E3 targeting HIF1-α under normoxic conditions. VHL recognizes and ubiquitinates HIF-1α through a proline hydroxylation modification, which suppresses the expression of genes involved in glucose metabolism in many malignancies and their CSCs, particularly renal cancer [87]. Therefore, the metabolic regulation of CSCs may depend on VHL, which is a potential target for cancer therapy.

4.3.7. TRAF6

As one of the key enzymes in the glycolytic pathway, Hexokinase 2 (HK2) regulates glucose metabolism and development of various cancers [88]. HK2 is also a substrate for selective autophagy and can undergo autophagic degradation. During liver tumorigenesis, HK2 is marked with K63-linked ubiquitination by the E3 ubiquitin ligase TRAF6, leading to autophagic disruption HK2 to downregulate glycolysis [89]. Thus, cells with high autophagy flux downregulate glycolysis by selective degradation of HK2 and slow cancer cell proliferation. On the other hand, TRAF6 also mediates K63-linked ubiquitination of AKT to enhance aerobic glycolysis and survival of CSCs [32, 34]. Therefore, TRAF6 may execute context-dependent functions in regulating glucose metabolism of CSCs.

4.3.8. RNF126

The essential metabolic enzyme Pyruvate Dehydrogenase Kinase 1 (PDK1) participates in several cancer signaling pathways such as the PI3K/AKT and Ras/MAPK pathways. It has been discovered that the E3 ubiquitin ligase RNF126 targets PDK1 for ubiquitination and degradation to support the growth of cancer cells [90]. Additionally, it has been suggested that the ubiquitin-like protein UFM1, which is attached to substrate proteins in a similar multi-step enzymatic process as found in ubiquitination, accelerates the degradation of PDK1 by ubiquitination, inhibiting PI3K/AKT signaling in the growth of gastric cancer. The small molecule DKM 2–93 covalently modifies the catalytic cysteine of ubiquitin-like modification activating enzyme 5 (UBA5) to inhibit the activity of UFM1 thus impairing the survival and growth of cancer cells in pancreatic cancer [91]. These studies demonstrate that DKM 2–93 may be a selective lead inhibitor and that UBA5 is a potential therapeutic target for pancreatic cancer. Therefore, two E3 ubiquitin ligases, RNF126 and UFM1, may regulate metabolic activity of CSCs and promote tumor progression by targeting PDK1 and inhibiting the PI3K/AKT signaling pathway.

4.4. Resisting death of cancer stem cells

Resistance to cell death is another major feature of CSCs. This phenotype is regulated by cell death pathways including apoptosis, iron death, and copper-related death. Considering that programmed cell death by apoptosis is a crucial cytoprotective mechanism against carcinogenic events, avoiding cell death may be a major factor in the survival of CSCs. E3 ubiquitin ligases also play key roles in enabling CSCs to escape death by targeting cell death-related proteins (Fig. 4).

Fig. 4.

The role of E3 ligases in in regulating resistance to cell death in CSCs. E3 ubiquitin ligases regulate CSCs death through TGF-β, NF-kB signaling pathways, and structural changes in VDAC. Red and light green icons represent E3 ligases with tumor-promoting and tumor-suppressing activities, respectively

4.4.1. PJA1

The E3 ubiquitin ligase PJA1 has been shown to inhibit apoptosis and promote growth of CSCs in HCC [61]. PJA1 is upregulated in liver cancers to interact with β2SP, thereby promoting the ubiquitination of β2SP and SMAD3, two essential components of the TGF-β/SMAD3 signaling pathway [61]. Thus, PJA1 could be potentially targeted to regulate TGF-β signaling and treating HCC [92].

4.4.2. NEDD4 and NEDD4L

NEDD4 promotes ion channel ubiquitination and proteasomal degradation following erastin treatment. Furthermore, cancer cells become more susceptible to erastin when NEDD4 is depleted because it restricts the breakdown of the VDAC2/3 protein. Thus, to overcome resistance to ferroptosis, a non-apoptotic form of cell death, erastin-induced resistance mediated by VDAC2/3 offers a novel route [93]. In addition, as a new ferroptosis inhibitor, NEDD4-like E3 ubiquitin ligase (NEDD4L) directly binds to lactotransferrin (LTF), an iron-binding transporter, ubiquitinates and degrades it, thereby preventing intracellular iron buildup and the ensuing ferroptosis brought on by oxidative damage in cancer cells [94]. These findings reveal a novel mechanistic relationship between the ubiquitination modification and ferroptosis, which may control the death of tumor cells and CSCs and hence be a potential target for anticancer therapy.

4.4.3. FBXW7

FBXW7 regulates lipid peroxidation and promotes ferroptosis. Mechanistically, FBXW7 inhibits the expression of stearoyl-CoA desaturase (SCD1), which has been reported to inhibit ferroptosis and apoptosis by targeting nuclear receptor subfamily 4 group A member 1 (NR4A1) [95]. Furthermore, the pro-survival Bcl-2 family member Mcl-1 can be ubiquitinated by FBXW7 and targeted for degradation to trigger cancer cell death in a GSK3 phosphorylation-dependent manner [96]. Therefore, inhibition of FBXW7 may block ferroptosis or apoptosis in cancer cells and CSCs, suggesting a mechanism for tumorigenesis mediated by deletion or mutation of FBXW7.

4.4.4. TRIM28

The E3 ubiquitin ligase TRIM28 can also target p53 and AMPK for ubiquitination and degradation, thereby inhibiting p53-mediated apoptosis, downregulating autophagy, and promoting tumor cell and CSC survival [55].

4.4.5. β‑TrCP1

In humans, excess copper can kill cells and cause damage to several organs in pathological conditions. This process relies on copper and mitochondrial respiration, differs from other well-known death mechanisms, thus it has been termed Cuprotosis [97]. The Drosophila Slmb (Slmb) gene encodes a protein homologous to β-TrCP1. By selecting several proteins for ubiquitination and subsequent proteasomal degradation, Slmb and β-TrCP1 regulate many biological activities. Depletion of Slmb resulted in a copper-deficient phenotype that might be reversed by raising the amount of copper present in cells by lowering efflux or boosting uptake. Further evidence that Slmb is necessary for the control of copper homeostasis comes from the fact that Slmb knockdown caused levels of the copper transporters Ctr1A and ATP7 to drop. Therefore, in humans, β-TrCP1 may be involved in regulating copper homeostasis of cancer cells and CSCs, rescue death caused by excessive copper, and promote tumor progression [98].

4.4.6. MDM2

MDM2, a RING-type E3 ubiquitin ligase with a single subunit, binds to the N-terminus of p53 to prevent it from carrying out its normal function as a transcriptional regulator. MDM2 inactivation leads to stabilization of p53, which impairs hematopoiesis by promoting cell cycle arrest, senescence, and ultimately cell death [99]. Additionally, MDM2 and MDMX inhibit p53 activity both individually and as a heteromeric complex. Furthermore, they hinder cells from developing a sufficient defense against lipid peroxidation, which may encourage CSC ferroptosis. Moreover, the promotion of ferroptosis by MDM2 and MDMX depends on PPAR activity, suggesting that the MDM2-MDMX complex controls lipids through modulating PPAR activity [100].

4.4.7. APC/C^Cdc20

An essential RING E3 that facilitates the metaphase to anaphase transition and is essential for regulating cell division and tumorigenesis is the anaphase promoting complex/cyclosome (APC/C). Two key APC/C activators Cdc20 and Cdh1 control substrate selectivity throughout mitotic progression. Among them, APC/C^Cdc20 regulates apoptosis by degrading the pro-survival protein Mcl-1 and the pro-apoptotic protein Bim [101]. Therefore, APC/C^Cdc20 promotes CSCs death and inhibits tumor development through its ubiquitin ligase activity.

4.4.8. CUL2

The protein p14^ARF has a short half-life that can induce ferroptosis through binding to the transcription factor NRF2 to inhibit the NRF2-mediated transcriptional activity of SLC7A11. The Cullin-RING E3 ubiquitin ligases can regulate the stability of p14^ARF, specifically Cul2^KLHDC3. p14^ARF is ubiquitinated and degraded by an E3 ubiquitin ligase complex composed of the proteins KLHDC3, CUL3, and RBX1. KLHDC3 is often up-regulated in ovarian cancer, leading to the decrease of p14^ARF protein level thus relieving the inhibitory effect of p14^ARF on NRF2 to promote SLC7A11 expression and inhibition of ferroptosis [102]. Therefore, the CUL2 RING E3 ubiquitin ligase can promote tumor by inhibiting ferroptosis of tumor cells and CSCs through the KLHDC3- p14^ARF-NRF2-SLC7A11 axis.

4.4.9. UBR2

Despite the fact that apoptosis is typically thought of as the primary pathway for cell death, caspase inhibition results in cell death being postponed rather than halted, resulting in caspase-independent cell death (CICD). As a regulator of CICD, the RING-type E3 ubiquitin ligase UBR2 overexpression prevents cell death while UBR2 downregulation makes cells more susceptible to CICD. UBR2-dependent protection against CICD is potentially mediated by the MAPK/ERK pathway. UBR2 overexpression contributes to CICD resistance and protects CSCs from cell death in many cancers [103]. Therefor targeting UBR2 may represent an innovative approach to inducing CSCs death.

4.5. Inducing therapy resistance of cancer stem cells

The CSC hypothesis can explain many clinical findings, such as the fact that many drugs initially kill cancer cells, but then the disease often relapses, which has attracted much attention. Compared with tumor cells, CSCs are more resistant to anticancer therapy. The therapeutic responsiveness of CSCs is cooperatively determined by the interaction between the tumor microenvironment, genetic abnormalities, and epigenetic alterations [1]. With this context, various E3 ubiquitin ligases have been identified to targeting the signaling network underlying therapeutic resistance (Fig. 5a).

Fig. 5.

The role of E3 ligases in regulating therapy resistance and immune evasion of CSCs. a) E3 ubiquitin ligases regulate cancer treatment sensitivity by regulating protein stability (such as p53, c-Myc, HIF1-α, etc.). b) E3 ubiquitin ligases affect anti-cancer immunity by regulating PD-1 or PD-L1 expression. Red and light green icons represent E3 ligases with tumor-promoting and tumor-suppressing activities, respectively

4.5.1. MDM2

MDM2, which as described above as a key regulator of p53, is upregulated in GSCs. Furthermore, MDM2 targets p53 for proteosomal degradation and maintains the expression of O6-methylguanine DNA methyltransferase (MGMT), thereby mediating the resistance of GSCs to temozolomide (TMZ) [104]. Therefore, identifying methods to target MDM2 could provide a novel therapeutic pathway to upregulate p53 and enhance efficacy of temozolomide.

4.5.2. FBXW7

Deletion or mutation of FBXW7 is often viewed as a tumor promoting mechanism due to the consequential upregulation of factors such as c-Jun, Mcl-1, Notch, c-Myc, and cyclin E. In glioblastoma CSCs, upregulation of FBXW7 is therefore considered a tumor suppressive mechanism. However, in colorectal CSCs, upregulation of FBXW7 and the resulting downregulation of c-Myc, appears to enhance resistance to anticancer agents and is a potential therapeutic strategy to eliminate colorectal CSCs [105]. However, more research is required to better understand the function of FBXW7 and the mechanisms by which it exerts its effects in different cancer cell types.

4.5.3. TRIM21

TRIM21 inhibits GSCs apoptosis via the p53-p21 pathway and boosts temozolomide resistance, which may serve as a mechanism by which patients with high TRIM21 expression have a poor prognosis [106].

4.5.4. VHL

Hypoxia is common within the tumor microenvironment where CSCs reside and have stabilized HIF-1α to promotes their proliferation and maintanance. Expansion of CSCs may promote tumor recurrence after radiotherapy or chemotherapy by preventing DNA damage as a result of reducing ROS and enhancing DNA checkpoint kinase activity [107]. Fortunately, VHL targets HIF-1α ubiquitination and inhibits tumor cell survival under hypoxic conditions [87]. Consequently, creating therapeutic strategies to increase VHL in cancer cells may be a promising way to lower therapeutic resistance.

4.5.5. HECTH9

CSCs are essential for maintaining cancer progression and drug resistance [1, 2, 108]. Recently, the E3 ubiquitin ligase HECTH9 was found to mediate K63-linked ubiquitination of HK2 to promote its mitochondrial localization, glucose metabolism, and CSC expansion. Disruption of the HECTH9-HK2 axis reduces CSCs and enhance sensitivity to chemotherapeutic agents such as doxorubicin and paclitaxel [109]. Therefore, HECTH9 might be a novel target to restore the susceptibility of CSCs to chemotherapy.

4.5.6. TRIM25

TRIM25 promotes stemness characteristics of CRC cells. In patients with CRC, high expression of TRIM25 is linked to oxaliplatin (OXA) resistance. Mechanistically, the E3 ubiquitin ligase TRAF6 binds to the histone methyltransferase EZH2 and TRIM25 prevents this interaction, stabilizing and upregulating EZH2 and boosting OXA resistance [110]. Thus, targeting TRIM25 expression may lead to the reduction of EZH2 to reduce tumor growth and increase response to OXA for CRC treatment.

4.5.7. SPOP

BET domain-containing epigenetic readers such as BRD4 represent promising molecular targets for cancer therapy. However, resistance to BET inhibitors JQ1 and I-BET, due to overexpression of BRD4, is observed for many types of cancers and their CSCs. In this regard, SPOP promotes ubiquitination and degradation of BET proteins to enhance sensitivity to BET inhibitors, while enzyme-deficient SPOP mutants in PCa lead to aberrant expression of BRD4 and drug resistance [47–49]. Thus, restoring SPOP expression and activity could be a potential approach to increase therapeutic efficacy of BET inhibitors.

4.5.8. FBXL2

FBXL2, a poorly understood SCF subunit, has been shown to ubiquitinate and degrade Cyclin D2, which may be a potential tumor therapeutic target by reducing CSCs and radio-resistance [111].

4.6. Avoiding immune destruction of cancer stem cells

Accumulating evidence suggests that CSCs may be the cellular origin of immune escape and poor response to immunotherapy. This unique property is mediated by crosstalk between CSCs and the tumor immune microenvironment (TIM) [112]. CSCs are known for their capability to shape the inhibitory TIM through intrinsic signaling pathways to modulate production of various cytokines and chemokines [11] .On the other hand, the TIM reprograms its cellular components and functionalities, thereby hindering antitumor immunity.

Furthermore, expression of immune checkpoints, such as programmed cell death protein 1 (PD-1) and its ligand PD-L1 reduces T cell effector function to facilitate immune escape. Therefore, therapeutic antibodies against these immune checkpoints are an effective immune strategy to treat human cancers [113] (Fig. 5b). E3 ubiquitin ligases can regulate both immunogenicity of CSCs as well as immune response by targeting these key immune signaling proteins. Unfortunately, immune checkpoint pathways can be disturbed in cancer cells through aberrant protein ubiquitination to circumvent immune surveillance and anti-cancer immune responses.

4.6.1. c‑CBL

The interaction between c-CBL and PD-1 facilitates the ubiquitination and degradation of PD-1 by c-CBLs E3 ubiquitin ligase function, which controls the PD-1/PD-L1 signaling pathway in colorectal cancer. Additionally, c-CBL suppresses PD-1 by deactivating PI3K/AKT, JAK/STAT, and MAPK-ERK signaling to increase the capacity of the immune system to combat cancer cells [114].

4.6.2. FBXW7

The oncogene c-Myc upregulates CD47 and PD-L1 in CSCs, thereby reducing macrophage and T cell infiltration into tumors in vivo [115]. However, FBXW7 has been well-established to target c-Myc for ubiquitination and degradation, thus overexpression of FBXW7 may lead to the downregulation of CD47 and PD-L1 to promote immune-mediated killing of CSCs.

4.6.3. PPARγ

Aberrant CDK5 activity is linked to the advancement of triple-negative breast cancer (TNBC). While CDK5 functions at multiple levels to control breast CSCs, CDK5 negatively regulates the activity of the E3 ubiquitin ligase PPARγ to protect ESRP1 from ubiquitin-dependent proteolysis. Therefore, in TNBC, inhibiting CDK5 activity may be a useful tactic to decrease stem cell transformation and boost response to PD-1 inhibition [116].

4.6.4. TRIM25

The TRIM protein family also participates in regulating innate immune responses. p53 not only acts as a tumor suppressor, but also essential for innate immunity by upregulating IRF9, a component of ISG factor 3 (ISGF3), and promoting interferon-stimulated genes (ISGs) expression. Previous studies have found that p53 can be ubiquitinated and degraded by TRIM25 in response to antitumor immune responses [117].

4.6.5. FBXO38

As a key molecule in tumor immunity, cell-surface PD-1 expression level can be specifically regulated by E3 ubiquitin ligases via Lys48-linked polyubiquitination. It has been reported that FBXO38 is necessary to maintain T cell activity against tumors. However, the tumor microenvironment inhibits its transcription. Fortunately, FBXO38 expression is restored by IL-2 in tumor-infiltrating T cells. Therefore, FBXO38 can regulate PD-1 expression to modulate T cells anti-tumor activity [118].

4.6.6. CUL3

As a PD-1-associated protein, KLHL22 bridges PD-1 to the CUL3 E3 ubiquitin ligase complex, thus mediating degradation of PD-1 prior to its transport to the cell surface. Therefore, loss or mutation of KLHL22 will promote PD-1 accumulation at the membrane and suppress T cell antitumor response. Interestingly, 5-fluorouracil (5-FU) treatment downregulates KLHL22 expression thereby increasing PD-1 levels suggesting therapeutic potential of combining 5-FU with anti-PD-1 in colorectal cancer patients [119]. In addition, CUL3 has been implicated in down regulating the PD-1/PD-L1 axis through the CUL3SPOP E3 ubiquitin ligase that was also shown to regulate CSCs through degrading the stemness transcription factor Nanog [120]. Therefore, CUL3 impinges on the PD-1/PD-L1 axis through regulating protein homeostasis of multiple target substrates, thereby conferring resistance to checkpoint inhibitors and maintaining stemness in CSCs.

4.6.7. ARIH1

While KLHL22 targets PD-1 for degradation, it has been demonstrated that PD-L1 is also targeted for ubiquitination and degradation by the E3 ubiquitin ligase ARIH1 following GSK3-mediated phosphorylation of PD-L1 in its degron domain. Thus, PD-1/PD-L1 axis inhibited by ARIH1-mediated PD-L1 degradation could promote anti-tumor immunity and may suggest that GSK3α and ARIH1 may be potential drug targets for promoting immunotherapy efficacy [121].

4.6.8. APC/C^Cdh1

APC/C^Cdh1 inhibits the tumor immune response by destabilizing SPOP which acts as a substrate recognition molecule for the E3 ubiquitin ligase CUL3, prompting CUL3 to recognize and ubiquitinate PD-L1. In this context, increased CDK4/6 activity in cancer cells promotes degradation of SPOP through APC/C^Cdh1, leading to increased PD-L1 and decreased tumor immunity, therefore APC/C^Cdh1 regulates the expression of PD-L1 indirectly to modulate the tumor immune response [101, 120].

5. E3 ubiquitin ligase and cancer stem cells targeted therapy

Mouse genetic studies have demonstrated that aberrancies in ubiquitin machinery are causally linked to tumor development and progression by targeting various oncogenes and tumor suppressors. A direct impact of E3 ubiquitin ligases on malignant properties of CSCs, as described above, is also revealed by regulating stemness signaling pathways and downstream transcription factors. Moreover, the importance of the UPS in CSCs is highlighted by the clinical success of proteasome inhibitors which has allowed for the development of several diagnostic and treatment methods against human cancers [21, 122]. In addition, proteolytic targeting chimeras (PROTACs), often referred to as bivalent chemical protein degraders, are a novel method for the degradation of disease-related proteins by commandeering E3 ubiquitin ligase pathways to target key cancer promoting proteins and have gained interest as potential treatment approaches for targeting CSCs.

5.1. E3 ubiquitin ligase inhibitors in CSCs targeted therapy

Bortezomib and MG132, two commonly used proteasome inhibitors, prevent the degradation of the entire proteome, but medications that target particular E3 ubiquitin ligases may be more selective, boost efficacy, and have lower toxicity against CSCs [123]. As described above, E3 ubiquitin ligases play critical roles in regulating the self-renewal, invasion and metastasis, immune evasion, and therapeutic resistance of CSCs. Therefore, targeting specific E3 ubiquitin ligases may provide promising strategies for cancer treatment. Inhibitors of E3 ubiquitin ligases currently under development are summarized in Table 1.

Table 1.

E3 ubiquitin ligases regulating CSCs properties and their inhibitors

E3 ligase	Substrates	Biological functions	Inhibitor	Reference
Skp2	p27, AKT	Self-renewal, proliferation, apoptosis, metabolism	SZL-P1-41, DT204, SKPin C1, Dioscin	[30, 33, 35, 81, 124–127]
TRAF6	AKT, HK2	Self-renewal, proliferation, metabolism	N/A	[34, 89]
TRIM8	PIAS3	Self-renewal, proliferation	STX-0119	[52]
TRIM28	p53, AMPK, FBP1	Self-renewal, invasion, migration, cell death, metabolism	NB237	[54, 55, 67, 82]
NEDD4	LGR5, DVL2, FOXA1, CD133, PTEN, VDAC2/3	Proliferation, self-renewal, invasion, migration, cell death	Diosgenin, I3C	[57–60, 64, 66, 93, 128, 129]
NEDD4L	LGD5, DVL2, LTF	Proliferation, cell death	N/A	[60, 94]
PJA1	β2SP, SMAD3	Proliferation, cell death	RTA405	[61]
UBE3C	AHNAK	Self-renewal, proliferation	N/A	[63]
NRDP1	DVL2	Invasion and migration	N/A	[72]
HECTD3	MALT1	Proliferation, invasion	PC3-15	[76, 130]
Cullin1	EZH2	Invasion, migration	N/A	[77]
RNF126	PDK1	Metabolism	N/A	[90]
UFM1	PDK1	Metabolism	DKM 2–93	[91]
MDM2	p53	Cell death, metabolism, therapy resistance	FC85, ISA27	[99, 100, 104, 131, 132]
APC/C	Bim, Mcl-1, SPOP	Cell death, immunity	N/A	[101, 120]
Cullin2	p14ARF	Cell death	N/A	[102]
UBR2	MAPK/Erk	Cell death	N/A	[103]
HECTH9	HK2	Therapy resistance	N/A	[109]
TRIM25	p53	Therapy resistance, immunity	N/A	[110, 117]
FBXO38	PD-1	Immunity	N/A	[118]
TRIM31	p52, Src	Proliferation, invasion, therapy resistance	MiR-551b	[133]

N/A, not available

The MDM2-p53 interaction is crucial for the development of tumors. Therapeutic inhibition of MDM2-p53 binding targets ALDH^highCD44^high CSCs in mucoepidermoid carcinoma for elimination by activating p53 and inhibiting CSC self-renewal. Additionally, mucoepidermoid carcinoma cells undergo cell cycle arrest and apoptosis when p53 is activated by preventing MDM2 binding, which results in cell death [131]. Novel AKT/mTOR inhibitors, FC85 and ISA27, reactivate p53 function to promote apoptosis, by blocking MDM2 activity towards p53 in GBM cells. Combination of FC85 with ISA27 synergistically suppressed cell viability via inducing the p53 pathway to promote differentiation of CSCs in GBM [132].

Many human cancers have elevated levels of Skp2 that is essential for cancer cell growth and dissemination. Treating lung adenocarcinoma cells with 5-FU reduces the CSC population by suppressing Skp2 leading to upregulation of p27, thereby inducing CSCs to enter a quiescent state [124]. Additionally, genetic and pharmacological Skp2 inactivation has been reported to effectively limit CSCs properties in prostate cancer cells increasing their sensitivity to chemotherapeutic treatments such as doxorubicin (Dox) and cyclophosphamide (CPA). The capacity of CSCs to produce spheres are decreased when cells are treated with Skp2 inhibitor SZL-P1-41, which disrupts the Skp2-Skp1 interaction with the SCF complex [35]. Recent studies has also shown that Skp2 inhibitor SKPin C1 may stabilize IDH1 and IDH2 expression in 22Rv1 and LNCaP cells in a dose-dependent manner, causing prostate cancer cells to switch from glycolysis to the TCA cycle, similar to Skp2 depletion [125]. Dioscin, a new Skp2 inhibitor derived from a natural steroidal saponin, may be used in the treatment of colorectal cancer in the future. Dioscin may have lower toxicity and fewer side effects compared to synthetic Skp2 chemical inhibitors [126]. Therefore, by eliminating CSC populations, Skp2 inhibitors may be able to overcome chemoresistance. In addition, chemical library screening identified DT204 as a compound able to prevent Skp2 from binding to CUL1 and Commd1, thus synergistically enhancing bortezomib (BTZ)-induced apoptosis. In a mouse model of myeloma, combining BTZ with DT204 reduces treatment resistance and inhibits tumor growth in vivo [127]. Therefore, potential exists for Skp2 targeted therapeutics to block CSCs maintenance and self-renewal.

MiR-551b, a proposed biomarker for ovarian cancer, inhibits FOXO3 and TRIM31, both of which have tumor suppressor functions in CSCs, suggesting miR-551b may be a potential target for cancer therapy [133]. The FBXW7/mTOR axis has emerged as a novel pathway to promote tumor invasion. The mTOR inhibitor rapamycin prevents FBXW7 depletion from promoting mTOR activity, which drives EMT, invasion, and stemness [66]. Furthermore, lysine-115 of SOX2 is ubiquitinated by the E3 ubiquitin ligase UBR5 and subsequently degraded. AKT-mediated phosphorylation of SOX2 at threonine-116 blocks its interaction of UBR5 leading to SOX2 stabilization. Therefore, AKT inhibitors may downregulate SOX2 to inhibit esophageal CSCs [134].

Through ubiquitin-dependent control of a variety of protein substrates, the E3 ubiquitin ligase NEDD4 is typically overexpressed in cancer and exhibits oncogenic characteristics. However, because NEDD4 is inhibited by indole-3-carbinol (I3C), a naturally occurring anticancer phytochemical, melanoma and other malignancies may be treated with I3C [128]. Similar to how diosgenin reduces the production of NEDD4 in prostate cancer cells, this finding raises the possibility of a novel prostate cancer treatment [129].

Studies have found that the ubiquitin-conjugating enzyme UbcH5b promotes HECTD3 autoubiquitination in vitro. On this basis, a fluorescence resonance energy transfer (FRET) assay revealed that PC3-15 could target HECTD3 thereby inhibiting lapatinib-induced autophagy and increasing lapatinib efficacy in vitro and in TNBC mouse xenografts. These results encourage further research into the combination of PC3-15 and lapatinib in TNBC therapy [130].

5.2. Proteolysis‑targeting chimera (PROTAC) for CSCs inhibition and cancer therapy

Traditional anti-cancer chemotherapy has the drawback of being poorly selective, broadly toxic, and susceptible to drug resistance. Thus, focusing on the discovery of novel targets in CSCs, small molecule inhibitors (SMIs) have become a critical alternative to develop precision therapeutic strategies. In recent years, proteolytic targeting chimeras (PROTACs), a novel approach to achieve disease-related protein degradation, has attracted attention (Fig. 6). PROTACs are composed of three parts; a linker molecule which acts as a bridge between the protein of interest (POI) binding ligand and the E3 ubiquitin ligase ligand. The ability of PROTACs to promote the degradation of target proteins is not restricted to binding sites within an active site of a POI; allowing degradation to be accomplished when a target protein’s catalytic activity is not its exclusive function or if a target protein lacks catalytic activity [135]. Therefore, PROTACs can be designed to target proteins that maintain CSC characteristics to be degraded through the UPS for the purpose of treating tumors. However, thus far only a limited subset of enzymes with small molecule ligands have been utilized in the development of PROTACs, such as SCF complexes, VHL, Cereblon (CRBN), inhibitor of apoptosis proteins (IAPs), and MDM2, etc. [136]. For better clarity, PROTACs for CSCs inhibition and cancer therapy currently under development are summarized in Table 2.

Fig. 6.

A Graphical representation of the components and process of PROTAC

Table 2.

PROTACs for CSCs inhibition and cancer therapy

PROTACS	Target	Indications	E3 Ligase	Reference
PROTAC 1/2	SMARCA 2/4	Acute myeloid leukemia	VHL	[137]
xStAx	β-catenin	Intestinal cancer	VHL	[138]
MZ1	BRD4	Triple negative breast cancer	VHL	[139]
ARV-825	BRD4	Triple negative breast cancer	CRBN	[139, 140]
A1874	BRD4	Colon cancer	MDM2	[141, 142]
WWL0245	BRD4	AR-positive prostate cancer	CRBN	[143]
A031	AR	Prostate cancer	VHL	[144]
MS9715	NSD3	Hematological cancer	VHL	[145]
MD-224	MDM2	Acute leukemia	Cullin 4A	[146]
MG-277	MDM2	Acute leukemia	CRBN	[147]
Degrader 32	MDM2	leukemia	CRBN	[148]
PRTC	CREPT	pancreatic cancer	VHL	[149]
L858R	EGFR	Non-small cell lung cancer	VHL	[150]
BP3	HSP90	Breast cancer	CRBN	[151]

As a novel structure-based PROTAC design technique, PROTAC degraders were constructed by linking a bromodomain ligand to a VHL ligand to target the BAF ATPase subunits SMARCA2 and SMARCA4. This novel PROTAC approach is under development for the treatment of tumors sensitive to loss of the BAF ATPase complex [137]. In addition, many types of cancers have been linked to abnormal stimulation of Wnt/β-catenin signaling, and thus β-catenin could be a therapeutic target for cancer [8, 152]. Therefore, in order to achieve effective β-catenin degradation, PROTACs have been developed that couple SAHPA1 or xStAx (which can improve or inhibit Wnt/β-catenin signaling, respectively) with a VHL ligand. This xStAx-VHL PROTAC led to enforced β-catenin degradation to inhibit Wnt signaling in cancer cells. Wnt/β-catenin are essential signaling that maintains CSCs in various types of cancers, underlining the potential of β-catenin PROTAC degraders as a promising new class of anti-cancer medicines and that xStAx-VHL may be able to eliminate CSCs [138].

Epigenetic readers, especially the acetylated-lysine reader BRD4, play important roles in maintaining hyperactive super enhancers to promote CSCs and tumorigenesis [153, 154]. Hence, BRD4 is an ideal target for designing PROTAC degraders. MZ1 and ARV-825, two BET-PROTACs, were demonstrated to decrease BRD4 protein levels in cellular models of breast cancer, establishing the groundwork for further development of these molecules for TNBC as well as CSCs [139]. In addition, ARV-825 efficiently suppresses c-Myc expression in pancreatic cancer models, providing a unique strategy to target Myc overexpressed cancers [140]. BRD4-degrader PROTACs (such as A1874) can inhibit colon cancer growth due to suppression of c-Myc, Bcl-2 and Cyclin D1. Surprisingly, A1874 induced more potent anti-colon cancer cell activity than the well characterized BRD4 inhibitors (JQ1, CPI203 and I-BET151) [141, 142]. WWL0245, an effective and isoform-selective BRD4-PROTAC, was found to elicit UPS-dependent degradation of BRD4 with sub-nanomolar concentrations and in vitro efficacy in AR-positive prostate cancer cell lines providing the basis to carry out further preclinical development of WWL0245 [143]. In addition, A031, a PROTAC targeting the androgen-receptor, significantly inhibited tumor growth in zebrafish transplanted with human prostate cancer cells [144]. Similarly, the nuclear receptor-binding SET domain protein 3 (NSD3) encodes a chromatin modulator and is an attractive tumor target. The NSD3-targeting PROTAC, MS9715 (designed by linking the NSD3 antagonist BI-9321 to the E3 ubiquitin ligase VHL ligand) specifically targets NSD3 and associated c-Myc in tumor cells, thus serving as a therapeutic strategy that jointly inhibits NSD3 and c-Myc-associated oncogenic nodes [145].

The tumor suppressive p53 pathway is frequently altered in human cancers but few strategies have been demonstrated to be effective in restoring p53 to suppress cancer. Prior studies have demonstrated that blocking the interaction between MDM2 and p53 may be a promising strategy for cancer treatment. For example, the MDM2 PROTAC degrader MD-224 has been designed to induce degradation of MDM2 in human leukemia cells and effectively restrain RS4-11 cell proliferation [146]. Further structural modifications to MD-224 led to the development of MG-277, which showed better effect on acute leukemia cell lines with variable p53 status. However, it does not have as significant an impact on MDM2 degradation. Mechanistically, it may be a true degrader of MDM2 or a molecular glue that attracts new substrate proteins to the CRBN E3 ubiquitin ligase for ubiquitination and subsequent degradation [147], for which further research is needed to determine the exact mechanism underlying its anti-cancer properties. Moreover, lenalidomide and the powerful MDM2 inhibitor RG7112 were combined to create a number of novel MDM2 PROTAC degraders [148]. Therefore, targeting the MDM2/p53 axis may serve as a potent anti-cancer mechanism based on increasing p53 activity which may suppress self-renewal of CSCs.

Application of new techniques or molecular targets in PROTAC is a field of intense investigation. In this regard, cell-permeable peptides could be utilized to enhance targeting of intracellular proteins. For example, in pancreatic cancer, a cell-permeable PROTAC was designed to degrade CREPT that is abundantly expressed and linked to uncontrolled proliferation and poor disease-free survival. When cancer cells were treated with the cell-permeable peptide-based PROTAC, it degraded CREPT in vivo and inhibited malignant growth of pancreatic cancer [149]. For NSCLC, epidermal growth factor receptor (EGFR) is a well-established target. Inhibitory immune checkpoints like PD-L1 and indoleamine-2,3-dioxygenase 1 (IDO1) are upregulated in response to EGFR activation through overexpression or mutation, which can promote the growth of NSCLC and enhance resistance to immunotherapy. To test the possibility that EGFR down-regulation can inhibit both PD-L1 and IDO1 and thus improve anti-tumor immunity in NSCLC, the PROTAC L858R, that degrades EGFR was developed. This EGFR L858R degrader drastically reduced the level of PD-L1 and IDO1 in NSCLC cells and inhibited tumor progression via enhancing the antitumor immune response towards in NSCLC cell xenografts [150]. Heat shock protein 90 (HSP90) is crucial for oncogenic transformation because it contributes to the activation and stability of oncogenic proteins. The PROTAC compound 16b (BP3) was developed to effectively degrade HSP90 and suppress the growth of human breast cancer cells, demonstrating the potential application of a HSP90-targeting PROTAC strategy in breast cancer therapy [151]. Therefore, the potential benefits of PROTAC technology might make up for the drawbacks of traditional chemo-therapy, which may encourage the development of new drugs to eradicate CSCs.

6. Conclusion and future perspectives

Recent studies have revealed CSCs as potentially the cells of origin in a variety of malignancies, and are responsible for tumor relapse, heterogeneity, metastasis, and therapy resistance. As with normal stem cells, CSCs are regulated by signaling pathways (Hhh, Wnt, Notch, TGF/BMP, Hippo, and other pro-survival pathways) and core transcription factors (Oct4, Nanog, Sox2, KLF4, c-Myc, etc). With this review, we briefly illustrate the principles of the protein ubiquitination system, and the critical function of E3 ubiquitin ligases in controlling CSCs characteristics. Various biological properties are regulated by E3 ubiquitin ligases, including CSC self-renewal, proliferation, Immunogenicity, and metabolism, which in turn influences cancer progression. Notably, E3 ubiquitin ligases have dual roles in the promotion and suppression of tumors. E3 ubiquitin ligases also specifically target particular substrates for proteasome degradation and alter the stemness characteristics of CSCs, in contrast to many other UPS components (such as E1 and E2 enzymes). Thus, they are more specific regulators of processes such as stemness signaling, which raises the possibility that they could be the focus of CSC-specific therapies.

Nonetheless, human cancer is notoriously heterogeneous, which hinders our complete understanding of its cellular origin and molecular basis of disease progression. As such, CSCs themselves are also cell mixtures that could be quiescent or fast cycling in term of growth kinetics, thereby maintaining everlasting growth of malignant cancers. Alternatively, CSCs may alter their phenotypes to adapt to environmental cues, e.g. CSCs leaving primary lesions will sense stromal signals and become migratory and resistant to anoikis. In clinical settings, CSCs will evolve with various treatment, such as chemo-therapy, radio-therapy, and immune therapy, leading to extensive genetic and epigenetics reprogramming. In all these above-mentioned situations, E3 ubiquitin ligases may undergo extensive regulation in their expression and activity, thereby functioning as important players in CSC biology. However, it remains largely unclear which E3 ubiquitin ligases regulate heterogeneity, plasticity, and evolution of CSCs during progression of a human cancer. Recently, high throughput muti-omics technologies have been developed and widely applied in biomedical research, offering valuable tools to address these open questions in CSC biology. Specifically, various genomic, transcriptomic, and proteomic approaches at single cell and single molecule levels will better define that heterogeneous expression profiles of E3 ubiquitin ligases in CSCs, while emerging spatiotemporal techniques will inform the roles of E3 ubiquitin ligases in CSCs in situ and in a time-dependent manner [155, 156]. Systemic integration of large datasets, as well as conventional molecular, cellular, and in vivo genetic studies, will no doubt reveal precise roles of E3 ubiquitin ligases in dictating CSC properties under various pathological conditions.

Moreover, these cutting-edge findings could provide further insights into which E3 ubiquitin ligases should be exploited to abrogate CSCs. Through these approaches, E3 ubiquitin ligases may be targeted to inhibit CSCs or to sensitize these cells to apoptosis induced by conventional anti-cancer therapy. Furthermore, the promising results of recently developed PROTACs indicate they may be more effective in many cases than inhibition of specific E3 ubiquitin ligases. Both PROTACs and small molecule inhibitors have significant limitations, despite extensive testing in both preclinical and clinical development. Therefore, while the application of new anticancer drugs based on the ubiquitin-proteosome system still has a way to go, the discovery and development of more efficient therapies based on E3 ubiquitin ligase biology remains a promising avenue in cancer treatment.

Data availability

Not applicable.