The SUMOylation and ubiquitination crosstalk in cancer

Abstract

Background

The cancer occurrence and progression are largely affected by the post-translational modifications (PTMs) of proteins. Currently, it has been shown that the relationship between ubiquitination and SUMOylation is highly complex and interactive. SUMOylation affects the process of ubiquitination and degradation of substrates. Contrarily, SUMOylation-related proteins are also regulated by the ubiquitination process thus altering their protein levels or activity. Emerging evidence suggests that the abnormal regulation between this crosstalk may lead to tumorigenesis.

Purpose

In this review, we have discussed the study of the relationship between ubiquitination and SUMOylation, as well as the possibility of a corresponding application in tumor therapy.

Methods

The relevant literatures from PubMed have been reviewed for this article.

Conclusion

The interaction between ubiquitination and SUMOylation is crucial for the occurrence and development of cancer. A greater understanding of the crosstalk of SUMOylation and ubiquitination may be more conducive to the development of more selective and effective SUMOylation inhibitors, as well as a promotion of synergy with other tumor treatment strategies.

Introduction

PTMs of proteins control the activity and stability of proteins by cutting precursor proteins or covalently adding modified groups to proteins, thus responding to stimuli in and out of the cell (Venne et al. 2014). At present, many different PTMs have been discovered, such as phosphorylation, ubiquitination, SUMOylation, acetylation and so on (Venne et al. 2014). Ubiquitination of proteins is controlled by three steps: ubiquitin-activating enzymes (E1s), ubiquitin-binding enzymes (E2s) and ubiquitin ligases (E3) (Venne et al. 2014; Capili and Lima 2007). Mechanistically, ubiquitin (Ub) is covalently bound to the lysine residues of the substrate protein via E1s activation, E2s coupling, and E3s connection (Venne et al. 2014; Capili and Lima 2007; Zhao et al. 2022) (Fig. 1A). Due to the presence of seven lysine residues (K6, K11, K27, K29, K33, K48, and K63) and one methionine residue (M1) in the ubiquitin molecule, substrates have different forms of ubiquitination chains, which typically determine the fate of substrates (Ikeda and Dikic 2008). Commonly, substrates with K48- and K11- linkage ubiquitin chains are for protein degradation via 26s of proteasomes. However, K6-linkage chains are associated with DNA damage repair, and K27-linkage chain may change the substrate sub-location to regulate the function of mitochondrial maintenance and autophagy. As for K29-linkage chains, itis usually associated with lysosomal degradation, while the K33 linker plays a dominant role in regulating somatic signal transduction in T cells. In addition, M1 and linear ubiquitin chains activate nuclear factors κB (NF-κB) (Zhai et al. 2022a; Mansour 2018). Ubiquitination interference usually leads to functional disorders of proteins in cells, which is closely related to the occurrence and development of cancer (Zhao et al. 2022; Zhai et al. 2022b). In addition, the ubiquitination modification of proteins can also be reversed by deubiquitinases (DUBs) (Yang et al. 2013). DUBs are also crucial for almost cell signaling pathways (such as cell cycle, apoptosis, receptor downregulation and gene transcription) by removing Ub from substrates (Yang et al. 2013).

Fig. 1.

The process and the crosstalk of ubiquitination and Sumoylation. A Ubiquitin conjugation cycle. First, the ubiquitin E1 enzyme activates by ATP and binds to Ub. Then, the Ub–E1 intermediate transfers this activated Ub to the ubiquitin E2 enzyme, after which it is transferred Ub to the specific substrates via E3 ubiquitin ligases. E3 ligases and transfer Ub to substrates. Ubiquitination may lead to degradation of substrate protein, or nondegradable modification. Finally, DUB allows deubiquitination to maintain a steady state ubiquitination cycle. B SUMO conjugation cycle. During each conjugation cycle, the SUMO E1 activating enzyme, which is a heterodimer of SAE1/SAE2, activates SUMO proteins in an ATP-dependent manner. Ubc9 catalyzes the transfer of SUMO to the target protein. This transfer step always requires an E3 enzyme; however, it is not necessary. Finally, SENP allows deSUMOylation to maintain a steady state SUMOylation cycle. C Crosstalk between Ubiquitin and SUMO: SUMO and Ub can form a hybrid chain. SUMO and ubiquitin can interact in many ways: 1. Antagonistic. SUMO can antagonize ubiquitination and subsequent degradation of target proteins. 2. Promotional. SUMO can change the subcellular localization of the substrate protein, thus promoting the ubiquitination of the substrate protein. 3: Dependent. The SUMOylation of the substrate protein can recruit STUbL, which can ubiquitin the existing SUMO chain or other lysine residues of the substrate protein. STUP can reverse the action of STUbL and remove ubiquitin from SUMO targets, thus stabilizing proteins. 4: SUMO-SIM. Ubiquitin E3 ligase can combine with SUMOylated substrate protein through its SIM domain to promote ubiquitination of substrate protein. At the same time, the SUMOylated ubiquitin E3 ligase can promote the ubiquitination of the substrate protein by binding to the SIM domain of the substrate protein

Additionally, SUMOylation itself is critical for cell growth, proliferation, and maintenance of cell homeostasis (Han et al. 2018; Chang and Yeh 2020). SUMO is a ubiquitin-like protein that affects the function of a target protein by binding to specific lysine residues of the target protein (Gill 2004). The SUMOylation machinery comprises the SUMO family of proteins, the SUMO E1 activation enzymes, the E2 conjugation enzyme, the E3 ligases, and the SUMO proteases (SENPs) (Capili and Lima 2007; Chang and Yeh 2020). Generally, SUMO molecule is also covalently bound to the substrates through three steps of enzymatic reaction: activation, coupling, and connection (Fig. 1B) (Capili and Lima 2007). In humans, there are five distinct SUMO isoforms (Vertegaal 2022). However, SUMO4 and SUMO5 have been little studied so far (Vertegaal 2022). SUMO2 and SUMO3 have extremely identical amino acid sequences and are commonly written as SUMO2/3 (Chang and Yeh 2020). SUMO2/3 can be SUMOylated and form a poly-SUMO chain (Chang and Yeh 2020). However, SUMO1, which lacks internal lysine residues, cannot be SUMOylated, attached to a single or multiple lysine site of the target protein as a monomeric molecule, or attached at the end of a poly-SUMO chain (Fig. 1C) (Han et al. 2018). Unlike ubiquitination, SUMOylation has only one E2 enzyme (UBC9), and SUMO E3 ligase is not necessary because UBC9 can also bind to substrates (Gong et al. 1997). SUMOylation alters the properties of its substrates by altering their activity, localization, stabilization, and interaction with proteins (Chang and Yeh 2020).

Currently, studies have found that SUMOylation plays an important role in various human diseases, including cancer, neurological disorders, cardiovascular diseases, and liver diseases(Vertegaal 2022; Zeng et al. 2020). Moreover, the level of SUMOylation of intracellular proteins is usually related to the external pressure on the cell, and SUMO can also be ubiquitinated or phosphorylated (Vertegaal 2022). Recently, the interaction between SUMOylation and ubiquitination has been extensively studied. It has been proved that SUMOylation affects the ubiquitination of a variety of substrates, including androgen receptor (AR) (Yang et al. 2019), phosphatase and tensin homolog (PTEN) (Bawa-Khalfe et al. 2017), p53 (Ashikari et al. 2017), and the aberrations of these modifications in the substrates may lead to the occurrence and progression of cancer. Intriguingly, since SUMOylation and ubiquitination act simultaneously on the lysine residues of the substrate, it has been shown that SUMOylation on the same site of the substrate can suppress its ubiquitination. However, in some conditions, if the substrate is SUMOylated, SUMOylation can promote its subsequent ubiquitination. STUbLs (SUMO-targeted ubiquitin ligases), such as RNF4 (Ring finger protein 4), can promote the ubiquitination of SUMOylated substrates (Kumar and Sabapathy 2019). These results suggest that SUMOylation has a profound effect on the ubiquitin–proteasome system (Fig. 1C). Understanding what the crosstalk between SUMOylation and ubiquitination is, and how the dysregulation of this crosstalk contributes to tumorigenesis and malignancy in cancer, is essential, and may provide new directions for applications, as well as for the development of related drugs based on SUMOylation and ubiquitination.

Ubiquitination in controlling the SUMOylation of substrates in cancer

Ubiquitination also affects the SUMOylation of substrates. However, unlike SUMOylation, the ubiquitination of the substrates usually promotes the degradation of the substrates. Although the research on the non-degradable ubiquitination is gradually deepened (Zhai et al. 2022a; Shi et al. 2022), there is little evidence that the non-degradable ubiquitination of the substrate affects its SUMOylation. Herein, we focus on the effect of ubiquitination on SUMO E3 ligases and SUMO proteases.

PIAS3 (protein inhibitors of activated STAT protein 3)

PIAS3, as a SUMO E3 ligase, affects the occurrence and progression of cancer by promoting SUMOylation of various substrates, such as Smurf2 (SMAD specific E3 ubiquitin protein ligase 2), IRF-1 (Chandhoke et al. 2017; Nakagawa and Yokosawa 2002). In addition, it has been found that PIAS3 binds to activated STAT3 and prevents it from binding to DNA, thereby inhibiting tumor growth (Pan et al. 2018). Jiao et al. reported that Smurf1 was recruited to promote the ubiquitination and degradation of PIAS3, thereby promoting the growth of glioma cells and the initiation of stem cell-like cells (Jiao et al. 2018).

SENP3

SENP3 is characterized by a sentrin/SUMO-specific protease, which is overexpressed in a variety of cancer cells (Yang et al. 2020; Ren et al. 2014). In addition, SENP3 can prevent substrate, such as Sp1 (Specificity protein 1), from being degraded by RNF4-mediated ubiquitination by removing SUMOylation of substrate, thus promoting the progress of cancer (Wang et al. 2011). Yan et al. found that, under normal conditions, CHIP (Carboxy-terminus of Hsp70-interacting protein) can promote its proteasome degradation by combining with SENP3; Under stress, Hsp90 (Heat shock protein 90) binds to SENP3 to protect SENP3 from CHIP-mediated ubiquitination and subsequent degradation (Yan et al. 2010). Moreover, the interaction between Hsp90 and SENP3 was enhanced in hepatoma cells, and thus the ubiquitination modification of SENP3 was reduced (Yan et al. 2010).

SENP7

Similarly, SENP7 is also a SUMO-specific protease, which can promote tumor progression by removing SUMO from the substrates (Gallardo-Chamizo et al. 2022). The SUMOylation of HP1α (Heterochromatin protein 1α) inhibits its transcriptional activity, and then promotes the aging of cancer cells, while SENP7 can down-regulate the SUMOylation of HP1α (Maison et al. 2012; Zhu et al. 2015). Zhu et al. found that SPOP (Speckle-type pox virus and zinc finger protein) promotes aging and inhibits tumor cell proliferation and metastasis through SPOP-mediated SENP7 degradation (Zhu et al. 2015; Ji et al. 2018).

SUMOylation controlling the ubiquitination of substrates in cancer

As previously mentioned, ubiquitination is an acute enzyme cascade pathway, which is usually affected by other types of PTMs, such as SUMOylation and phosphorylation (Venne et al. 2014). It has been found that SUMOylation of substrate usually affects the interaction between the substrate and E3 ubiquitin ligase. However, the SUMOylation of the substrate has different effects on its ubiquitination and may depend on the SUMO types and properties of the substrates (Table 1). Here, we discussed how SUMOylation participates and affects the process of ubiquitination.

Table 1.

Summary of the crosstalk between SUMOylation and ubiquitination and its biological function when SUMOylation exits in substrates

Substrate	Sumoylated site	Modified by SUMO	SUMO E3 ligase	Ubiquitinated site	Ubiquitin E3 ligase	Method	How SUMOylation affects ubiquitination of substrate	Biological function	References
AR	K386	SUMO3	PIAS1	–	MDM2	co-IP	Promotion	Inhibit PCa tumorigenesis	Yang et al. (2019)
PTEN	K254/266	SUMO1	PIASxα	–	WWP2	co-IP	Promotion	Promote PCa progression	Bawa-Khalfe et al. (2017)
p53	K386	SUMO1	RanBP2	K370/372/373/381/382/386	MDM2	co-IP	Promotion	Promote cancer progression	Ashikari et al. (2017), Li et al. (2003)
CDK6	K216	SUMO1	–	K147	–	IP	Inhibition	Promoting the progression of glioblastoma	Bellail et al. (2014)
IRF-1	K275/K299	SUMO1	PIAS3	K275/K299	–	co-IP	Inhibition	Increase the stability of IRF-1 and decrease the anti-tumor effect of IRF-1	Park et al. (2007)
PRB	K7/K388/K531	SUMO1	PIAS3	K388	CUEDC2	co-IP	Inhibition	Promote BC progression	Zhang et al. (2007)
ZFHX3	K2349/2806/3258	SUMO1	PIAS2	K2806	EFP	co-IP	Inhibition	Promote BC progression	Wu et al. (2020), Dong et al. (2012)
HnRNP-K	K422	SUMO1/2/3	PIAS3/Pc2	–	MDM2	co-IP	Inhibition	Promote p53 transcription activation and inhibit tumor progression	Lee et al. (2012), Pelisch et al. (2012)
PES1	K517	SUMO1/2/3	–	–	TRIM23	IP	Inhibition	Promote BC progression	Li et al. 2016)

SUMOylation of substrates: promoting ubiquitination degradation

The subcellular localization of the protein affects its ubiquitination-dependent degradation, and E3 ubiquitin ligases such as NEDD4-1 (Neural precursor cell expressed developmentally downregulated 4-1) and WWP2 (HECT E3 ligases WW domain-containing protein 2) may be more prone to mediate ubiquitination and degradation of PTEN in cytoplasm, while FBXO22 (F-box only protein 22) selectively degrades PTEN in nucleus (Ge et al. 2020). Moreover, the current research found that SUMOylation can affect the ubiquitination degradation of the substrate by changing its subcellular location.

AR

AR, which belongs to the nuclear receptor superfamily of steroid hormones, is overexpressed during all stages of prostate cancer (PCa) including castration-resistant PCa (CRPC) (Dai et al. 2017). Several ubiquitin E3 ligases have been found to participate in ubiquitination and degradation of AR, such as MDM2 (Mouse double minute 2) (Lin et al. 2002), SPOP (An et al. 2014), Siah2 (Seven in absentia homolog 2) (Li et al. 2022), CHIP (Sarkar et al. 2017). Moreover, the current study confirms that AR could be modified by SUMO at K386 and K520 sites (Nishida and Yasuda 2002). Interestingly, PIAS1 (Protein inhibitors of activated STAT protein 1) enhances the SUMO3 modification of AR at the K386 site and then SUMO3-modified AR and PIAS1 complexes exited the nucleus together, thus PIAS1 recruiting MDM2 to ubiquitinated and degraded AR in cytoplasm (Yang et al. 2019).

PTEN

It is recognized that PTEN can exert tumor suppressive effects, but these functions are affected by its subcellular localization (Song et al. 2008). Furthermore, both ubiquitin and SUMO modifications can regulate the subcellular localization of PTEN. Cytosolic PTEN is poly-ubiquitinated and targeted for proteasomal degradation, while mono-ubiquitination of PTEN leads to its translocation to the nucleus (Trotman et al. 2007). While SUMOylation contributes to nuclear export of PTEN, especially to the cell membrane (Matunis and Guzzo 2012). Wang et al. found that PIASxα/PIAS2α (Protein inhibitors of activated STAT protein 2α) could promote the SUMO1 modification of PTEN at K254 and K266 sites (Wang et al. 2014). Also importantly, Bawa-Khalfe et al. found that the SUMO1-modified PTEN could be excluded from the nucleus, and then interact with WWP2 for ubiquitination and subsequent degradation (Bawa-Khalfe et al. 2017).

p53

P53 is a transcription factor that acts as a tumor suppressor by regulating various biological behaviors such as cell cycle arrest, apoptosis, senescence, DNA repair and cell metabolism (Kruiswijk et al. 2015; Devine and Dai 2013). In PCa, p53 in the nucleus often plays an anti-tumor role, while p53 in the cytoplasm is related to tumor progression (Ashikari et al. 2017). It has been found that cytosolic p53 may bind to MDM2 and thus be poly-ubiquitinated and degraded. But MDM2 frequently promotes the mono-ubiquitination of nucleus localized p53, and thus promoting the nuclear export of p53, which is then poly-ubiquitinated and degraded in the cytoplasm (Li et al. 2003). In addition, the SUMOylation of p53 contributes to its nuclear export (Ashikari et al. 2017). Takayama et al. found that RanBP2 (Ran Binding Protein 2) can bind TRIM25 (Tripartite motif protein 25) and androgen-induced G3BP2 (Ras GTPase-activating protein-binding protein 2) to increase SUMO1 modification of p53 and subsequent nuclear p53 export (Ashikari et al. 2017), resulting in MDM2-mediated ubiquitination and degradation.

SUMOylation of substrates: reduce proteasomal degradation

SUMOylation of the substrate can also inhibit its ubiquitination by targeting different lysine residues, competing for the same lysine site, and inhibiting the binding of the substrate with E3 ubiquitin ligase.

Target different lysine residues

CDK6 (cyclin-dependent kinase 6)

CDK6 is elevated in glioblastoma and promotes tumor progression by promoting cell cycle transition (Bellail et al. 2012). UBC9 contributes to the SUMOylation of CDK6 at the K216 site and impairs its ubiquitination and degradation, which preserves its protein level, thus promoting the progression of glioblastoma (Bellail et al. 2014).

Compete for the same lysine site

IRF-1 (interferon regulatory factor-1)

IRF-1 was originally identified as a regulator of IFNα/β (Miyamoto et al. 1988). The accumulating evidence supports the function of IRF-1 as a tumor suppressor (Alsamman and El-Masry 2018; Armstrong et al. 2015; Yan et al. 2021). In human tumors, IRF-1 is inactivated to prevent apoptosis and cell cycle arrest (Alsamman and El-Masry 2018). However, the stability of IRF-1 can be mediated by the crosstalk of SUMOylation and ubiquitylation. Park et al. found that PIAS3 promoted the SUMOylation of IRF-1 at K275 and K299 sites (Nakagawa and Yokosawa 2002; Park et al. 2007). Moreover, in vivo ubiquitination test showed that compared with wild-type IRF-1, the level of ubiquitination of IRF-1 with SUMOylation sites mutation was significantly reduced (Park et al. 2007). Co-expression of wild-type IRF-1 and SUMO1 also decreased the ubiquitination level of IRF-1 (Park et al. 2007). This indicates that the PIAS3-mediated SUMOylation of IRF-1 inhibits the ubiquitination modification and degradation of IRF-1 by competing for the same lysine site. Interestingly, although SUMOylation increases the stability of IRF-1, it decreases the anti-tumor effect of IRF-1, leading to the eventual occurrence of ovarian cancer (OCa) (Park et al. 2007, 2010).

PRB (progesterone receptor B)

PR is an important member of the nuclear receptor family of ligand-activated transcription factors, which promotes the proliferation of breast cancer (BC) cells and is associated with the progression of BC (Hilton et al. 2018). In addition, PR is one of the markers of hormone-dependent BC and disease prognosis (Hilton et al. 2018). Moreover, PR can undergo a variety of post-translational modifications, including phosphorylation, ubiquitination, and SUMOylation, which have been shown to modulate the function and stability of PR (Weigel 1996; Abdel-Hafiz et al. 2002; Lange et al. 2000). PR exists as two isoforms in BC, namely PRA and PRB. Zhang et al. found that CUEDC2 (containing CUE domain 2) interacts with PRB to promote the ubiquitination and degradation of PRB (Zhang et al. 2007). Moreover, the K388 site of PRB is the key site to regulate the degradation of PRB by CUEDC2 (Zhang et al. 2007). Man et al. found that PIAS3 promoted the SUMO1 modification of PRB at K7, K388, and K531 sites (Man et al. 2006). At the same time, in vitro experiments showed that PRB did not undergo ubiquitination and ubiquitination when the K388 mutation of PRB occurred (Zhang et al. 2007). This suggests that SUMOylation of PRB may inhibit its ubiquitination and degradation by competing for the same lysine site.

ZFHX3 (zinc finger homeobox 3)

ZFHX3 promotes the occurrence and progression of BC, and its protein level in BC cell was significantly higher than that in the normal cell (Dong et al. 2020). Notably, SUMOylation can regulate the stability of ZFHX3. It has been found that the PIAS2 could SUMOylate ZFHX3 at K2349, K2806 and K3258 sites and subsequently prevent the degradation and ubiquitination mediated by E3 ligase EFP (Estrogen-responsive finger protein) (Wu et al. 2020; Dong et al. 2012). K2806 is the principal SUMO site of ZFHX3 (Wu et al. 2020). Moreover, when MDA-MB-231 cells expressing wild-type or mutated ZFHX3 were subcutaneously injected into nude mice, wild-type ZFHX3 significantly promoted tumor growth, while K2806R mutant significantly inhibited tumor growth (Wu et al. 2020). These results indicate that SUMOylation-mediated inhibition of the ubiquitination of ZFHX3 can result in the proliferation of BC cancers.

Inhibit the binding of substrate and E3 ubiquitin ligases

HnRNP-K (heterogeneous nuclear ribonucleoprotein K)

HnRNP-K can hardly be considered a tumor suppressor or an oncoprotein (Gallardo et al. 2016). Clinical studies have shown that overexpression of HnRNP-K is associated with poor prognosis in a variety of malignant tumors, including lung cancer (LC), PCa, BC (Gallardo et al. 2016). However, HnRNP-K plays its anti-tumor role by activating the p53 pathway (Lee et al. 2012). Under UV irradiation, PIAS3 and Pc2/CBX4 could promote SUMOylation of HnRNP-K at K422 site and increase the stability of HnRNP-K (Lee et al. 2012; Pelisch et al. 2012). Further mechanism studies showed that the SUMOylation of HnRNP-K could inhibit its interaction with MDM2 and escape from MDM2-mediated ubiquitination and degradation (Lee et al. 2012). Interestingly, SUMOylation of HnRNP-K increases its binding with p53 and promotes p53 activation (Lee et al. 2012; Pelisch et al. 2012). Moreover, Peng et al. found that Circ-GALNT16 inhibits the progression of colorectal cancer (CRC) by upregulating SUMOylation of HnRNP-K, thereby activating p53 signal (Peng et al. 2021).

PES1 (pescadillo ribosomal biogenesis factor 1)

Several studies have shown that PES1 is closely related to tumorigenesis by promoting cell proliferation and participating in cell cycle progression (Qiu et al. 2019; Jin et al. 2019; Cheng et al. 2012). Moreover, PES1 is highly expressed in several cancers, including CRC and BC (Cheng et al. 2012; Bian et al. 2021). The stability of PES1 was negatively regulated by TRIM23 (Tripartite motif protein 23) through proteasome degradation, and SUMOylation of PES1 at K517 site inhibits its interaction with TRIM23 (Li et al. 2016). Furthermore, although PES1 can be modified by SUMO1 and SUMO2/3, SUMO1 modification is the major type(Li et al. 2016). Moreover, in vitro and in vivo, the SUMOylation of PES1 promotes the proliferation of BC cells(Li et al. 2016).

SUMOylation of E3 ubiquitin ligase: enhance its interaction with a substrate

Recently, many proteins have found that they contain SIM (SUMO interaction motif) domains, therefore the SUMOylated ubiquitin E3 ligases may bind to the SIM domain of the substrate through interaction between SUMO and SIM, thereby further promoting the ubiquitination of the substrate (Fig. 1C). Once the ubiquitin E3 ligases are SUMOylated, the interaction between substrates and the ubiquitin E3 ligases may sometimes be enhanced (Table 2).

Table 2.

Summary of the SUMOylation of ubiquitin E3 ligase

Ubiquitin E3 ligase	Sumoylated site	Modified by SUMO	SUMO E3 ligase	Method	Ubiquitination activity	Biological function	References
WWP2	K473	SUMO1/2/3	–	co-IP	Upregulate	Promote cancer tumorigenesis	Bawa-Khalfe et al. (2017)
Smurf2	K26/369	SUMO1/2/3	PIAS3	co-IP	Upregulate	Inhibit tumor progression	Chandhoke et al. (2016)
BRCA1	–	SUMO1/2/3	PIAS1/4	co-IP	Upregulate	Inhibit tumor progression	Morris et al. (2009)
pVHL	K171	SUMO1/2/3	PIAS4	co-IP	Downregulate	Inhibit tumor progression	Núñez-O’Mara and Berra (2013)
MDM2	K446	SUMO1/2/3	PIAS1/3	co-IP	Upregulate	Promote cancer tumorigenesis	Buschmann et al. (2001)

WWP2

WWP2 is abnormally expressed in a variety of cancers, including LC, Hepatocellular carcinoma (HCC), PCa, and Oca (Soond et al. 2013; Yang et al. 2016; Qin et al. 2016; Jung et al. 2014). In LC, PCa, and HCC, WWP2 is highly expressed, usually promotes the occurrence and progression of cancer, but it exhibits the reversed effect in Oca (Soond et al. 2013; Yang et al. 2016; Qin et al. 2016; Jung et al. 2014). Maddika et al. reported that WWP2 degrades PTEN through ubiquitination, thereby activating PI3K/AKT pathway and promoting the survival of PCa cells (Maddika et al. 2011). At the same time, in vitro experiments showed that down-regulation of WWP2 led to increased PTEN expression and decreased Akt activity, leading to delayed cell proliferation (Maddika et al. 2011). Moreover, Bawa-Khalfe et al. reported that WWP2 could be SUMOylated at the K473 site, which enhanced its interaction with PTEN for further degradation (Bawa-Khalfe et al. 2017).

Smurf2

Smurf2, a HECT-type E3 ubiquitin ligase, plays a dual role as a tumor promoter and inhibitor by regulating protein stability during tumorigenesis and progression (Fu et al. 2020a). Current studies have found that Smurf2 inhibits TGF-β-mediated EMT (Epithelial–mesenchymal transformation), thereby inhibiting the progression of a variety of cancers, such as pancreatic cancer, BC and HCC (Chandhoke et al. 2017; Song et al. 2022; Zhang et al. 2022a). Mechanistically, Smurf2 can interact with TβRI (TGFβ type I receptor) and promote the ubiquitin-mediated degradation of TβRI, thus blocking the TGF-β-induced EMT (Kim et al. 2019). Chandhoke et al. reported that PIAS3 interacts with Smurf2 and promotes its SUMOylation at the K26 and K369 sites (Chandhoke et al. 2016). The SUMOylation of Smurf2 enhances its ability to induce the degradation of the TβRI and thereby suppresses EMT (Chandhoke et al. 2016). Therefore, SUMOylation promotes the tumor suppressive function of Smurf2 by enhancing its ubiquitination activity by degrading the corresponding substrates.

BRCA1 (breast cancer 1 protein)

BRCA1 is a tumor suppressor that participates in the basic cellular functions necessary for cell replication and DNA synthesis (Morris et al. 2009). However, BRCA1 expression is reduced due to mutation or epigenetic inactivation, thus increasing the risk of BC and OCa (Romagnolo et al. 2015). Current research shows that BRCA1 participates in DNA damage response (DDR) by acting as ubiquitination E3 ligase, thus maintaining genomic integrity (Morris et al. 2009). PIAS1 and PIAS4 (Protein inhibitors of activated STAT protein 4) promote the SUMOylation of BRCA1, thus dramatically improving its E3 ligase activity, and contributing to the DDR process (Morris et al. 2009).

pVHL (von Hippel–Lindau protein)

pVHL is a tumor suppressor and the main regulator of HIF (hypoxia-inducible factor) activity (Minervini et al. 2020). The functional inactivation of pVHL is the cause of von Hippel–Lindau disease, which is a genetic predisposition to different cancers (Minervini et al. 2020). Moreover, pVHL needs to combine with Elongin C, Elongin B and Cullin-2 to form an active E3 ligase to degrade substrates (Minervini et al. 2020). PIAS4 promotes the SUMOylation of pVHL at K171 site, thus inhibiting the E3 ligase activity of pVHL (Núñez-O’Mara and Berra 2013). Mechanistically, the SUMOylation of pVHL inhibits the combination of pVHL with Elongin C and Elongin B, thus inhibiting HIF-1α ubiquitination (Núñez-O’Mara and Berra 2013). Therefore, SUMOylation upregulates HIF-1α by suppressing the E3 ligase activity of pVHL to promote the progress of the tumor.

SUMOylation of DUBs

DUBs remove the Ub molecule from the substrates to antagonize ubiquitination induced by E3 ligase (Fig. 1A), thus ensuring the balance of ubiquitin protein in cells. Moreover, some DUBs can be regulated by SUMOylation, suppressing the activity of DUBs (Table 3).

Table 3.

Summary of the SUMOylation of DUBs

DUB	Sumoylated site	Modified by SUMO	SUMO E3 ligase	Method	DUB activity	Biological function	References
CYLD	K40	SUMO1/2/3	–	IP	Downregulate	Activate NF-κB signaling pathway and Promote cancer tumorigenesis	Masoumi and Massoumi (2016)
USP25	K99/141	SUMO1/2/3	–	IP	Downregulate	Inhibit tumor progression	Denuc et al. (2009)
USP39	K6/16/29/51/73	SUMO1/2/3	–	IP	Downregulate	Inhibit tumor progression	Wen et al. (2014)
USP28	K99	SUMO1/2/3	–	IP	Downregulate	Inhibit tumor progression	Zhen et al. (2014)

CYLD (cylindromatosis)

CYLD is a tumor suppressor and its tumor suppressor function has been reported in several cancers, including CRC, LC, melanoma, and multiple myeloma (Hellerbrand et al. 2007; Massoumi et al. 2009; Annunziata et al. 2007). Mechanistically, CYLD negatively regulates NF-κB pathway, thereby promoting cancer cell apoptosis and inhibiting tumorigenesis (Kovalenko et al. 2003). TRAF2 (TNF Receptor-Associated Factor 2), TRAF6 (TNF Receptor-Associated Factor 6) and NEMO (NF-κB essential modulator, also known as IκB kinase γ) are key protein members of NF-κB pathway, and the K63-linked ubiquitination of TRAF2, TRAF6 and NEMO is essential for the activation of NF-κB pathway (Kovalenko et al. 2003). However, CYLD combines with TRAF2, TRAF6 and NEMO to remove their K63-linked ubiquitination, thereby negatively regulating NF-κB signaling (Trompouki et al. 2003; Zhao et al. 2015). In neuroblastoma, all-trans retinoic acid (ATRA) can induce the SUMOylation of CYLD at K40 site (Masoumi and Massoumi 2016). SUMOylation of CYLD can inhibit its DUB activity, thereby activating NF-κB signal and inhibiting the apoptosis of tumor cells (Masoumi and Massoumi 2016). However, prolonged ATRA treatment for more than 6 days led to reduce the SUMOylation of CYLD, which, in turn, prevented NF-κB signal and promoted cell death (Masoumi and Massoumi 2016). In conclusion, SUMOylation of CYLD interferes with its DUB activity, which is essential for its tumor suppressive effect.

USP25 (ubiquitin specific peptidase 25)

USP25 a DUB, has long been associated with immune response, inflammation, and cancer (Wang et al. 2020). It has a SIM domain and multiple ubiquitin-binding domains at its N-terminal (Wang et al. 2020). USP25 can recognize and bind SUMO and ubiquitin chain through its SIM and ubiquitin-binding domain (Denuc et al. 2009). In addition, the catalytic activity of USP25 is regulated by SUMOylation or ubiquitination. In USP25, the same lysine residue K99 can be either SUMOylated or ubiquitinated (Denuc et al. 2009). These mutually exclusive modifications have the opposite effect on USP25. The SUMOylation of K99 and K141 sites inhibits the DUB activity of USP25 by reducing the binding of USP25 to the ubiquitin chain (Denuc et al. 2009). Conversely, the ubiquitination of K99 may activate the DUB activity by preventing SUMOylation (Denuc et al. 2009). At present, USP25 has been found to be highly expressed in a variety of cancers, such as BC and LC (Deng et al. 2007; Li et al. 2014). Moreover, in vitro experiments showed that overexpression of USP25 was beneficial to tumor cell migration, invasion, and metastasis (Deng et al. 2007; Li et al. 2014). These results strongly suggest that overexpression of USP25 contributes to carcinogenesis and metastasis, while SUMOylation of USP25 inhibits tumor progression by inhibiting the DUB activity of USP25, which may be one of the hypothetical targets of new drugs.

USP39 (ubiquitin specific peptidase 39)

USP39 is a member of the ubiquitination family, which contains a central zinc finger ubiquitin binding domain and a ubiquitin C-terminal hydrolase domain (Wang et al. 2021). More and more data show that USP39 plays a crucial role in the development of cancer (Huang et al. 2016). USP39 is overexpressed in BC and HCC cells, and silencing USP39 can significantly inhibit cell proliferation and increase the number of apoptotic cells (Wang et al. 2013; Dong et al. 2021). At the same time, overexpression of USP39 can enhance the proliferation of androgen-dependent PC-3 cells and androgen-dependent LNCaP cells (Huang et al. 2016). In addition, USP39 can be SUMOylated at K6, K16, K29, K51 and K73 sites. The mutation of the SUMOylation sites of USP39 further strengthened its ability to enhance the proliferation of PCa cells (Wen et al. 2014).

USP28 (ubiquitin specific peptidase 28)

Like USP25, USP28 contains the SIM domain and ubiquitin-binding domains (Masoumi and Massoumi 2016). USP28 can modify and stabilize various oncoproteins such as c-Myc, cyclin E1 and Claspin by deubiquitylation to promote the proliferation and metastasis of tumor cells (Popov et al. 2007; Aziz et al. 2022; Ito et al. 2018). Moreover, USP28 expression is upregulated in several tumor tissues and is associated with poor prognosis (Wang et al. 2018). Zhen et al. found that USP28 is SUMOylated by SUMO2 at K99 site, and the SUMOylation of USP28 can inhibit its DUB activity (Zhen et al. 2014).

STUbLs

In addition, Ub and SUMO molecules can form hybrid chains. Endogenous SUMO can also be modified by Ub, especially at K11 of the SUMO molecule (Fig. 1C). SUMO-Ub chain establishment is catalyzed by a specific set of STUbLs. STUbLs directly connect SUMO and ubiquitin pathways: STUbLs binds to the poly-SUMO chain of target proteins through their SIM, and then promote the ubiquitination and degradation of these proteins (Chang and Yeh 2020). Currently, the most studied STUbLs is RNF4 (Table 4).

Table 4.

Summary of STUbLs and their downstream substrates

SUMO-targeted ubiquitin ligase (STUbL)	Substrate	SUMO	Sumoylated site	SUMO E3 ligase	Method	degradation or non-degradation	Biological function	References
RNF4	PML(PML-RARα)	SUMO1/2/3	–	PIAS1	co-IP	Degradation	Inhibit APL progression	Lallemand-Breitenbach et al. (2008)
	c-Myc	SUMO1/2/3	K52/148/157/317/323/326/389/392/398/430	–	co-IP	Degradation	Inhibit tumor progression	González-Prieto et al. (2015)
	PIM1	SUMO1/2/3	K169	PIAS1, PIAS3	co-IP	Degradation	Inhibit tumor progression	Iyer et al. (2017)
	KDM5B	SUMO2/3	K242/278	Pc2	co-IP	Degradation	Inhibit tumor progression	Bueno and Richard (2013)
	Sp1	SUMO1/2/3	K16	–	co-IP	Degradation	Inhibit tumor progression	Wang et al. (2011)
	NDRG2	SUMO1	K333	–	co-IP	Degradation	Inhibit tumor progression	Tantai et al. (2016)
	HIF-2α	SUMO1/2/3	K394	–	co-IP	Degradation	Inhibit tumor progression	Hagen et al. (2010)

PML (promyelocytic leukemia)

PML is a tumor suppressor, which plays a key role in a variety of tumor inhibition, including growth inhibition, apoptosis, replicative aging, and inhibition of migration and angiogenesis (Hsu and Kao 2018). It has been found that PML can be SUMOylated at K160 site (Liu et al. 2017). The SUMOylated PML binds to RNF4 and is then degraded by RNF4-mediated ubiquitination (Liu et al. 2017). PML-RARα caused by gene fusion plays a vital role in the occurrence and progression of APL (Acute promyelocytic leukemia) (Gong et al. 2022). At present, arsenic trioxide (ATO) is the first choice of anti-APL therapeutic agents (Gong et al. 2022). Mechanistically, ATO promotes the SUMOylation of PML, the part of the whole PML-RARα, thereby promoting RNF4-mediated ubiquitination degradation of PML-RARα (Lallemand-Breitenbach et al. 2008). In addition, PIAS1 may participate in the SUMOylation of PML triggered by ATO (Rabellino et al. 2012). And patients with the PML-RARα K160 mutant were resistant to ATO treatment (Liu et al. 2017).

c-Myc

The c-Myc is the most frequently overexpressed oncogene in tumors, including CRC, BC, and LC (Beroukhim et al. 2010). González-Prieto et al. found that c-Myc can be modified by SUMO (González-Prieto et al. 2015). At the same time, 10 SUMOylation sites of c-Myc were identified by mass spectrometry: K52, K148, K157, K317, K323, K326, K389, K392, K398 and K430 (González-Prieto et al. 2015). Interestingly, the mutation of all 10 SUMOylation sites did not reduce the SUMOylation level of c-Myc (González-Prieto et al. 2015), indicating other SUMOylation sites may exist. Moreover, PIAS1 knockout led to a decrease in the SUMOylation of c-Myc, suggesting that PIAS1 may induce the SUMOylation of c-Myc (González-Prieto et al. 2015). In addition, González-Prieto et al. also found that RNF4 can promote the ubiquitination and degradation of SUMOylated c-Myc (González-Prieto et al. 2015). Therefore, the association of c-Myc and SUMO may provide a new therapeutic option for c-Myc-driven tumors.

HIF (hypoxia-inducible factor)

HIF is a key transcription factor that mediates cell survival under hypoxia (Wiesener et al. 1998). Under normal conditions, HIF is continuously synthesized in cells and degraded by the proteasome pathway (Ivan et al. 2001). In tumors, HIF is often overexpressed and over activates the transcription of target genes, thus enabling tumor cells to survive under hypoxia (Albadari et al. 2019).

The best characterized HIF isoforms, HIF-1α and HIF-2α. Although there are many studies on SUMOylation of HIF-1α, the effect of SUMOylation on its stability is a controversial issue. Cheng et al. found that the SUMOylation of HIF-1α in SENP1^−/− cells increased but the stability of HIF-1α decreased during hypoxia (Cheng et al. 2007). Moreover, the SUMOylation of HIF-1α promoted the interaction between HIF-1α and pVHL, resulting in the ubiquitination and degradation of HIF-1α (Cheng et al. 2007). At the same time, PIAS4 is also considered to be a specific SUMO E3 ligase that promotes the SUMO1 modification of HIF-1α (Seeler and Dejean 2017). At the same time, Berta et al. found that SUMO E3 ligase RanBP2 promoted the modification of SUMO1 and SUMO2/3 at K391 and K477 residues of HIF-1α (Berta et al. 2007). However, Berta et al. showed that SUMO modification did not affect the stability of HIF-1α protein but downregulated its transcription activity (Berta et al. 2007).

The stability of HIF-2α protein is strictly regulated by SUMOylation and ubiquitination. Van et al. also found the K394 site of HIF-2α is the main SUMO site, and SUMOylation reduces the transcriptional activity of HIF-2α (Hagen et al. 2010). Moreover, RNF4 can promote the ubiquitination and degradation of the SUMOylated HIF-2α (Hagen et al. 2010).

PIM1 (proviral integration site for Moloney murine leukemia virus 1)

The Ser/Thr protein kinase PIM1 has been reported to participate in multiple biological processes such as cell cycle progression (Blanco-Aparicio and Carnero 2013). The PIM1 protein is highly expressed in malignancies, and its expression correlates with high tumor grade (Nawijn et al. 2011; Warfel and Kraft 2015). In addition, PIM1 inhibitors are considered promising targets for cancer therapy and have been an important focus of drug development (Warfel and Kraft 2015). Iyer et. al found that the stability of PIM1 is negatively regulated by RNF4-mediated ubiquitination and degradation (Iyer et al. 2017). Both PIAS1 and PIAS3 promote SUMOylation modification of PIM1 at the K169 site, and then RNF4 combines with SUMOylated PIM1 and promotes its ubiquitination and degradation (Iyer et al. 2017). Therefore, the regulation of PIM1 ubiquitination and SUMOylation may provide an opportunity to enhance the drug targeting of PIM1 kinase activity in cancer treatment.

KDM5B (lysine-specific demethylase 5B)

KDM5B has been found to be overexpressed in many cancers (Xhabija and Kidder 2019; Zheng et al. 2019). In addition, it is also regarded as an anti-cancer target, hinting that the protein level of KDM5B may be crucial for drug discovery (Fu et al. 2020b). Moreover, KDM5B protein may be differentially regulated by ubiquitination and SUMOylation. Pc2-mediated SUMOylation of KDM5B at the K242 site competitively inhibits TRAF6-mediated K63-linked ubiquitination at the same site of KDM5B, thereby inhibiting nuclear import of KDM5B (Bueno and Richard 2013). In addition, RNF4 could promote the degradation of SUMOylated KDM5B, thereby promoting the aging of cancer cells and inhibiting the progression of cancer (Bueno and Richard 2013).

Sp1

Sp1 is a common transcription factor that is highly expressed in a variety of cancers (Beishline and Azizkhan-Clifford 2015). In gastric cancer (GC), Sp1 expression increases with tumor progression and is associated with poor prognosis (Xu et al. 2018). Wang et al. found that the SUMOylation of Sp1 at K16 promoted ubiquitination and degradation of Sp1 by recruiting RNF4 as ubiquitin E3 ligase (Wang et al. 2011). In addition, during tumor development, the SUMOylation of Sp1 is attenuated and thus the stability of Sp1 is increased (Wang et al. 2011).

NDRG2 (N-myc downstream-regulated gene 2)

NDRG2 is frequently downregulated in various cancers and acts as a candidate tumor suppressor (Chen et al. 2018). The NDRG2 could be modified by SUMO1 at the K333 site, and the SUMOylation of NDRG2 increases the ability of NDRG2 to inhibit LC tumorigenesis (Tantai et al. 2016). Moreover, Tantai et al. found that RNF4 promotes the proteasome degradation of SUMOylated NDRG2, thereby further inhibiting the occurrence and progression of LC (Tantai et al. 2016).

SUMO-targeted ubiquitin proteases (STUPs)

As previously mentioned, SUMO2/3 and Ub can form a hybrid chain, including via the K32 site of SUMO2 or the K33 site of SUMO3 (Fig. 1C) (Schimmel et al. 2008). The substrate is usually degraded after being modified by SUMO-Ub chain, which can be counteracted by STUPs (Fig. 1C). At present, the most studied STUPs are USP11 (Ubiquitin specific peptidase 11) and USP7 (Ubiquitin specific peptidase 7).

USP11

The USP11 was identified as a binding partner of RNF4 through mass spectrometry analysis (Hendriks et al. 2015). In vitro experiments showed that USP11 was able to deubiquitinate hybrid SUMO2-Ub chains produced by RNF4. USP11 could counteract RNF4 under normal growth conditions and DDR process (Hendriks et al. 2015). The involvement of USP11 in the DNA damage response reflects the importance of reversible ubiquitination of SUMOylated protein (Meng et al. 2021). In vitro experiments showed that USP11 combined with PML to remove Ub from PML, and then stabilized PML to inhibit the proliferation, migration and invasion of glioblastoma cells (Wu et al. 2014).

USP7

The identification of USP7 as a STUP involved in DNA replication further proved the concept of deubiquitinase-specific target hybridization SUMO-Ub chain (Smits and Freire 2016). USP7 has been proved to continuously block the clearance of SUMO-related proteins from DNA replication sites by limiting their ubiquitination modification, which promotes fork progression and stimulates new replication origins, thereby promoting DNA replication (Lecona et al. 2016). Downregulation of USP7 can inhibit DNA replication (Lecona et al. 2016), which should be considered when using USP7 inhibitors as anti-cancer therapy.

Ligases exhibiting both ubiquitination and SUMOylation activity

Currently, it is found that many proteins have exhibited both ubiquitin E3 activity and SUMO E3 activity. They often combine with ubiquitin E2 ligase or UBC9 through different domains (Table 5). However, it is not clear how they choose the different types of PTM for the substrate.

Table 5.

Summary of the ligases exhibiting both ubiquitination and SUMOylation activity and their downstream substrates

Ligase	Substrates	Ubiquitination or SUMOylation	Modification site	Method	Biological function	References
MDM2	p53	Ubiquitination	K370/372/373/381/382/386	co-IP	Promote tumor progression by promoting the degradation of p53	Li et al. (2003)
	p53	SUMOylation	K386	co-IP	Suppress the transcriptional activity of p53 and promote its nuclear output	Stindt et al. (2011), Chen and Chen (2003)
TRIM28	RLIM	Ubiquitination	–	co-IP	Promote the ubiquitination degradation of RLIM to stabilize MDM2 and promote the development of LC	Jin et al. (2021)
	AMPK	Ubiquitination	–	co-IP	Suppress autophagy and promote tumor progression	Pineda and Potts (2015)
	p27	Ubiquitination	–	co-IP	Accelerate cell cycle and promote the development of HCC	Pineda and Potts (2015)
	HDAC6	Ubiquitination	–	co-IP	Promote the development of HCC	Pineda and Potts (2015)
	CSDE1	Ubiquitination	–	co-IP	Promote the development of HCC	Pineda and Potts (2015)
	Vps34	SUMOylation	–	IP	Promote the progress of osteosarcoma	Tsang et al. (2022)
	PCNA	SUMOylation	K164	co-IP	Promote DDR	Li et al. (2018)

MDM2

MDM2 is a key negative regulator of the tumor suppressor p53 and plays a key role in controlling transcriptional activity, protein stability, and nuclear localization of p53 (Wang et al. 2017). MDM2 expression is upregulated in many cancers, leading to the loss of p53-dependent activities such as apoptosis and cell cycle arrest, which promotes the occurrence and progress of various cancers (Wang et al. 2017). On the one hand, MDM2 degrades p53 through polyubiquitination, and on the other hand, low-level MDM2 can also catalyze monoubiquitination of p53, resulting in a nuclear output of p53 (Li et al. 2003). However, the regulation of p53 by MDM2 is far more complex than simple ubiquitination modification. Several studies have found that MDM2 can promote SUMO1 and SUMO2/3 modifications of p53 (Stindt et al. 2011; Chen and Chen 2003).

However, the SUMO 1 or SUMO 2/3 modifications of p53 may have different results. The SUMO 1 modification of p53 promotes the nuclear output of p53, but the SUMO 2/3 modification of p53 does not affect the subcellular localization of p53 but does inhibit the p53 transcription activation (Stindt et al. 2011; Chen and Chen 2003). Notably, inhibition of ubiquitin E3 activity of MDM2 would promote the SUMO modification of p53. While the effect of SUMO2/3 modification on p53 activity is limited, there are many examples showing that relatively small changes in p53 activity can also affect the biological behavior of cells.

The E3 ligase activity of MDM2 is also affected by PIAS1 and PIAS3-mediated SUMOylation (Buschmann et al. 2001; Ding et al. 2012). The SUMOylation of MDM2 occurs at the K446 site, which is responsible for the ubiquitination of MDM2 (Buschmann et al. 2001). Therefore, SUMOylated MDM2 is more relatively high stable to mediate the ubiquitnation and degradation of p53 (Buschmann et al. 2001; Ding et al. 2012; Lee et al. 2006). Moreover, PIAS1 and PIAS3 have been proved to promote SUMOylation of MDM2.

TRIM28 (tripartite motif protein 28)

TRIM28 also known as KAP1 (Krüppel-Associated Box (KRAB)-Associated Protein 1) (Czerwińska et al. 2017). Immunohistochemical analysis showed significantly increased levels of the TRIM28 protein in BC, HCC, GC, and PCa, suggesting that TRIM28 may play a role in promoting cancer in multiple cancers (Czerwińska et al. 2017). At the same time, several studies have shown that the expression level of TRIM28 is associated with poor prognosis of BC, HCC, and GC (Czerwińska et al. 2017; Wang et al. 2016; Wei et al. 2016). However, the inversed conclusion has also been reported. High expression of the TRIM28 was associated with a better prognosis in early LC (Chen et al. 2012), suggesting that TRIM28 may also display a tumor-suppressive role in LC.

TRIM28 acts to promote tumorigenesis through ubiquitination and degradation of multiple tumor suppressors. TRIM28 can act as an MDM2-binding protein to promote MDM2-mediated p53 degradation, thus playing its oncogenic role in tumorigenesis (Liu et al. 2022a). However, Jin et al. showed that TRIM28 can stabilize the level of MDM2 in cells by promoting the degradation of RLIM, an E3 ligase of MDM2, thereby promoting the proliferation, migration, and invasion of LC cells (Jin et al. 2021). Moreover, TRIM28 can significantly reduce the autophagy level of cancer cells by promoting the ubiquitination and degradation of AMPK, a major cell energy sensor and regulator, thereby promoting the progress of cancer (Pineda and Potts 2015). In HCC, TRIM28 promotes the progression of HCC by degrading several tumor suppressors such as p27, HDAC6 (Histone deacetylase 6), and CSDE1 (Cold shock domain-containing E1) (Zhang et al. 2021; Li et al. 2021; Liu et al. 2022b).

Besides, TRIM28 also has SUMO E3 enzyme activity. Tsang et al. found that TRIM28 promoted the SUMOylation of Vps34 (phosphatidylinositol 3-kinase catalytic subunit type 3) with the help of PVT-1 (Plasmacytoma variant translocation-1), which led to the ubiquitination and degradation of tumor suppressor complex 2 (TSC2), thus promoting the progress of osteosarcoma (Tsang et al. 2022). Moreover, TRIM28 is associated with the DDR process. TRIM28 binds to PCNA (Proliferating cell nuclear antigen) through its PIP motif to promote SUMO2 modification at the K164 site of PCNA, thereby promoting DDR (Li et al. 2020). Interestingly, TRIM28 haploid dysfunction has been linked to nephroblastoma (Diets et al. 2019). The inhibition of SUMO2 modification of PCNA has also been shown to be relevant in nephroblastoma(Li et al. 2018). Therefore, it is necessary to further determine whether TRIM28 can inhibit the progression of nephroblastoma by promoting the SUMO modification of PCNA.

Tumor therapies based on SUMOylation

The multiple regulatory modalities of SUMOylation toward ubiquitination shed light on SUMO-based therapies for those diseases caused by dysregulation of ubiquitination. Targeted SUMOylation of substrates may have additional therapeutic effects on tumor treatment strategy, as they can modulate the stability of certain oncoproteins and tumor suppressors by altering their SUMOylation and ubiquitination properties. Although there is a lack of SUMOylated drugs targeting specific substrates, SUMOylated inhibitors have proved their safety and effectiveness in tumor therapy in vitro and in vivo (Table 6).

Table 6.

Summary of the SUMO Inhibitors

Target	Inhibitor	Structure	IC50 (μM)	Type	References
SUMO E1	Ginkgolic acid		3.0	Natural	Fukuda et al. (2009a)
	Anacardic acid		2.2	Natural	Fukuda et al. (2009a)
	Kerriamycin B		11.7	Natural	Fukuda et al. (2009b)
	Davidiin		0.15	Natural	Takemoto et al. (2014)
	Tannic acid		12.8	Natural	Suzawa et al. (2015)
	ML-792		0.003 (SUMO1), 0.011(SUMO2)	Synthetic	He et al. (2017)
	COH000		0.2	Synthetic	Lv et al. (2018)
	ML-93		0.037	Synthetic	Biederstädt et al. (2020)
	TAK-981		nM range	Synthetic	Langston et al. (2021)
SUMO E2	2-D08		6.0	Synthetic	Kim et al. (2013)
	GSK145A		12.5	Synthetic	Brandt et al. (2013)
	Spectomycin B		4.4	Natural	Hirohama et al. (2013)
	Compound 2		74	Synthetic	Zlotkowski et al. (2017)
SENP	Triptolide		0.071–0.076	Natural	Huang et al. (2012)
	Momordine Ic		15.37	Natural	Wu et al. (2016)
	Compound 3		3.55 (SENP1), 2.98 (SENP2)	Synthetic	Zhao et al. (2016)
	Compound 7		0.99	Synthetic	Lindenmann et al. (2020)
	Compound J5		2.385	Synthetic	Lindenmann et al. (2020)
	Compound 13m		3.5	Synthetic	Lindenmann et al. (2020)
	Compound 10		7.5	Synthetic	Uno et al. (2012)
	Compound 38		9.2	Synthetic	Qiao et al. (2011)
	Compound 69		5.9	Synthetic	Kumar et al. (2014)
	Compound 117		3.7	Synthetic	Kumar et al. (2014)

SUMO E1 inhibitors

The SUMO E1 enzyme, which is responsible for the first step of SUMOylation, is composed of SAE1 and UBA2/SAE2 subunits and is overexpressed in many cancers (Seeler and Dejean 2017; Liu et al. 2015). It has been found that knockdown of SUMO E1 enzyme inhibits the proliferation of cancer cell in various tumor cells, such as HCT116 colon cancer cells, U87 and U251 human glioma cells (He et al. 2017; Langston et al. 2021).

Several SUMO E1 inhibitors have been identified, including ginkgolic acid, anacardic acid, kerriamycin B Davidin, tannic acid, ML-792, COH000, ML-93 and TAK-981 (He et al. 2017; Langston et al. 2021; Fukuda et al. 2009a, 2009b; Takemoto et al. 2014; Suzawa et al. 2015; Lv et al. 2018; Biederstädt et al. 2020). These drugs inhibit the formation of SUMO E1-SUMO, thereby blocking the coupling of SUMO to the target protein, and thus acting as an anti-cancer agent. ML-981 is currently the most potent SUMO E1 inhibitor (Langston et al. 2021). Although several DDR-associated proteins are SUMO substrates, ML-981 does not induce DNA damage and does not synergize with known DNA damage agents such as cisplatin or hydroxyurea (Langston et al. 2021). However, ML-981 induced apoptosis and inhibited cell proliferation in cancer cells (Langston et al. 2021). Besides, TAK-981 is currently in Phase I clinical trials for the treatment of lymphoma and metastatic solid tumors.

SUMO E2 inhibitors

Unlike ubiquitination, only one SUMO E2 ligase UBC9 has been found to be responsible for coupling in the SUMOylation pathway. Therefore, there is an opportunity to inhibit the SUMO pathway by targeting UBC9. Currently, UBC9 has been found to play a cancer-promoting role in a variety of cancers, such as BC, PCa and Oca (Xu et al. 2016; Moschos et al. 2010). In addition, inhibition of UBC9 increases the sensitivity of melanoma cells and hepatocellular carcinoma cells to chemotherapy drugs and inhibits the proliferation of colon cancer cells (Moschos et al. 2007; Fang et al. 2017).

SUMO E2 inhibitors that have been discovered include GSK145A, 2-D08, Spectomycin B, compound 2 (Brandt et al. 2013; Kim et al. 2013; Hirohama et al. 2013; Zlotkowski et al. 2017). Among them, 2-D08 can reduce the growth of non-APL and acute myeloid leukemia (AML) cells, induce apoptosis, and inhibit cell migration in K-Ras mutant pancreatic cancer cells (Baik et al. 2018), suggesting that SUMO E2 inhibitors can target cancer cells by blocking the SUMO cascade reaction.

SENPs inhibitors

As previously mentioned, SENPs can remove the coupling between SUMO and the substrates. Currently, six types of SENPs have been identified, including SENP1, SENP2, SENP3, SENP5, SENP6, and SENP7 (Seeler and Dejean 2017). Moreover, SENP1-3 is up-regulated in multiple tumors (Seeler and Dejean 2017). Therefore, targeting SENPs may be a new direction of tumor therapy.

At present, natural inhibitors of SENP have been found, such as triptolide and Momordine Ic (Huang et al. 2012; Wu et al. 2016). HIF-2α is an essential oncoprotein in PCa (Palayoor et al. 2003). SUMOylated HIF-2α would be degraded by RNF4-mediated ubiquitination; SENP1 can inhibit the SUMOylation of HIF-2α to improve its stability (Hagen et al. 2010). Therefore, SENPs inhibitors may promote HIF-2α degradation to induce cell apoptosis. Meanwhile, in vitro experiments showed that triptolide and Momordin Ic could reduce the proliferation of LNCaP and PC-3 cells (Wu et al. 2016). In addition, other SENP inhibitors have been developed through artificial synthesis of computer screening agents, including compounds 3, 6, 7, J5, 10, 13m, 38, 69, and 117 (Lindenmann et al. 2020; Kumar et al. 2014; Uno et al. 2012; Zhao et al. 2016; Qiao et al. 2011). The role of these SENPs inhibitors in cancer treatment needs further study.

Other related drugs

SUMO1 is highly expressed in various types of human cancer, and knockout of SUMO1 inhibits the growth of cancer cells (Bellail et al. 2014; Guo et al. 2011). Bellail et al. found that compound CPD1 and its derivative HB007 can facilitate the ubiquitination and degradation of SUMO1 (Bellail et al. 2021). Mechanistically, HB007 binds to CAPRIN1 (cytoplasmic activation/proliferation-associated protein 1), inducing an interaction between CAPRIN1 and FBXO42 (F-box protein 42). FBXO42 then recruits SUMO1 into the CAPRIN1-CUL1-FBXO42 ubiquitin ligase complex, resulting in ubiquitination and degradation of SUMO1 (Bellail et al. 2021). At the same time, CPD1 showed extensive anti-cancer activity on cell lines from the brain, BC, CRC and LC (Bellail et al. 2021).

Discussion

Different PTMs can be affected by each other. Ubiquitination primarily controls the homeostasis of proteins in the cell by promoting the degradation of substrates. Ubiquitination disorder is closely related to the occurrence and progression of cancer, such as SPOP in PCa (Zhang et al. 2022b). The SUMOylation of the substrates can either promote or inhibit its ubiquitination modification. Therefore, disorders in SUMOylation may contribute to cancer by affecting the ubiquitination process. These substrates may participate in the same signal pathway, such as p53 pathway, HIF pathway (Fig. 2), and the effects of SUMOylation and ubiquitination on different substrates may be different, so SUMOylation and ubiquitination may jointly promote or inhibit the activation of these pathways.

Fig. 2.

The SUMOylation and ubiquitination crosstalk in the pathway. p53 pathway: In general, MDM2 degrades p53 through ubiquitination, thus stabilizing its abundance in cells. However, MDM2 can not only promote the ubiquitination degradation of p53, but also promote the SUM modification of p53. Moreover, the SUMO 1 or SUMO 2/3 modifications of p53 may have different results. The SUMO 1 modification of p53 promotes the nuclear output of p53, but the SUMO 2/3 modification of p53 does not affect the subcellular localization of p53 but does regulate the activation and inhibition of many p53 target genes. Moreover, PIAS3 and Pc2 inhibit the ubiquitination degradation of HnRNP-K by promoting its SUMOylation. The SUMOylated HnRNP-K enters the nucleus and promotes the transcriptional activity of p53. HIF-1α pathway: In general, pVHL degrades HIF-1α through ubiquitination to maintain low intracellular level. PIAS4 promotes the SUMOylation of pVHL, inhibits the ubiquitination E3 activity of pVHL, and down-regulates the Ubiquitination level of HIF-1α; Interestingly, PIAS4 can also promote the SUMOylation of HIF-1α. SUMOylated HIF-1α is easier to be degraded by pVHL ubiquitination. Moreover, in the nucleus, RanBP2 promotes the SUMylation of HIF-1α to inhibit its gene transcription activity

As mentioned earlier, SUMO1 normally forms a single SUMO modification or is attached at the end of the poly-SUMO chain. However, SUMO2/3 mostly forms the poly-SUMO chain and can form the SUMO-Ub chain with ubiquitin. Due to differences in the SUMO proteins themselves, different types of SUMOylation of certain substrates have different effects on their ubiquitination. For instance, the SUMO1 modification of SHMT1 (Serine hydroxymethyltransferase 1) and IκBα inhibit their ubiquitination, while SUMO2/3 modification promote the ubiquitination (Anderson et al. 2012; Mabb and Miyamoto 2007). In contrast, the SUMO1 modification of HDAC1 promotes ubiquitination, while the SUMO2 modification inhibits this process (Citro and Chiocca 2017). Alternatively, the SUMOylation type of a substrate may be determined by its SUMO E3 enzyme. For example, PIAS4 selectively promotes the SUMO2 modification of HDAC1, while PIAS1 is not selective (Citro and Chiocca 2017). Further research in this area is currently needed, which will expand knowledge of ubiquitination and establish a framework for SUMO–Ub interactions that may lead to new potential therapies for targeting these proteins.

Even though SUMO inhibitors have achieved some success in basic research and clinical trials, the underlying mechanism by which different SUMO subtypes modify substrates is still unclear. Moreover, no drug targeting SUMO E3 has been developed yet, which remains to be studied in the future. Considering that substrates are diverse, consisting of oncoproteins and tumor suppressor proteins, SUMO inhibitors may play a more effectively anti-cancer role by modulating the ubiquitin modification of substrate proteins. Therefore, a greater understanding the crosstalk of SUMOylation and ubiquitination may be more conducive to the development of more selective and effective SUMOylation inhibitors, as well as a promotion of synergy with other tumor treatment strategies.

Author contributions

Conceptualization, KL and JH; methodology, MY and XJ; writing—original draft preparation, KL and MY; writing—review and editing, XJ, MY, YX, and KL; visualization, KL; supervision, XJ; funding acquisition, XJ. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by The National Natural Science Foundation of China (32270821), The Natural Science Foundation of Ningbo (Grant No. 2021J065), The Natural Science Foundation of Ningbo (Grant No. 2022J040), The Natural Science Foundation of Ningbo (Grant No. 2022J230), The Fundamental Research Funds for the Provincial Universities of Zhejiang (Grant No. SJLZ2022004), and The K.C. Wong Magna Fund in Ningbo University.

Data availability

Not applicable.

Declarations

Conflict of interest

The authors declare no conflict of interest.

Institutional review board statement

Not applicable.

Informed consent

Not applicable.