Crosstalk between SUMOylation and other post-translational modifications in breast cancer

Abstract

Breast cancer represents the most prevalent tumor type and a foremost cause of mortality among women globally. The complex pathophysiological processes of breast cancer tumorigenesis and progression are regulated by protein post-translational modifications (PTMs), which are triggered by different carcinogenic factors and signaling pathways, with small ubiquitin-like modifier (SUMOylation) emerging as a particularly pivotal player in this context. Recent studies have demonstrated that SUMOylation does not act alone, but interacts with other PTMs, such as phosphorylation, ubiquitination, acetylation, and methylation, thereby leading to the regulation of various pathological activities in breast cancer. This review explores novel and existing mechanisms of crosstalk between SUMOylation and other PTMs. Typically, SUMOylation is regulated by phosphorylation to exert feedback control, while also modulates subsequent ubiquitination, acetylation, or methylation. The crosstalk pairs in promoting or inhibiting breast cancer are protein-specific and site-specific. In mechanism, alterations in amino acid side chain charges, protein conformations, or the occupation of specific sites at specific domains or sites underlie the complex crosstalk. In summary, this review centers on elucidating the crosstalk between SUMOylation and other PTMs in breast cancer oncogenesis and progression and discuss the molecular mechanisms contributing to these interactions, offering insights into their potential applications in facilitating novel treatments for breast cancer.

Introduction

Breast cancer is the most common cancer type worldwide, accounting for approximately 30% of cancers in women [1]. Annually, approximately 2 million women are newly diagnosed with breast cancer [2], and its global incidence has been increasing, with an annual increase of approximately 3.1%; what is worse is that this trend may continue [3, 4]. Despite notable advancements in high-quality prevention strategies, early detection, and treatment services that have led to a decline in breast cancer mortality rates, it still accounts for a substantial proportion of deaths, ranging from 15 to 30% among newly diagnosed cases [5,6,7]. Therefore, breast cancer remains a serious public health concern worldwide.

As a heterogeneous disease, breast cancer is commonly classified into three subtypes based on receptor expression in clinical settings: luminal estrogen receptor (ER) and progesterone receptor (PR)-positive breast cancer, human epidermal growth factor receptor 2 (HER2)-positive breast cancer, and triple-negative breast cancer (TNBC) (ER⁻, PR⁻, and HER2⁻) [8, 9]. Luminal ER and PR-positive breast cancer can be further divided into two subtypes based on the proliferation marker Ki-67: luminal A, which exhibits low Ki-67 levels, and luminal B, characterized by high Ki-67 levels [9,10,11]. TNBC can be divided into six categories: basal-like 1, basal-like 2, immunomodulatory, mesenchymal, mesenchymal stem cell-like, and luminal androgen receptor [12]. Notably, these breast cancer subtypes exhibit varying mortality rates, with HER2-positive breast cancer being associated with the highest mortality rate, followed by TNBC, Luminal A, and then Luminal B subtypes [13].

Breast cancer is often accompanied by two types of gene mutations: gain-of-function mutations in oncogenes and loss-of-function mutations in tumor suppressor genes. Approximately 10% of all cases are associated with genetic predisposition or family history [9]. Breast cancer susceptibility gene 1 (BRCA1) (located at 17q21) and BRCA2 (at 13q13) are two important and high-penetrance tumor suppressor genes whose mutations exhibit an autosomal dominant inheritance pattern [4, 14,15,16]. Germline mutations in BRCA1 or BRCA2 contribute to approximately 15%–20% of all TNBC cases and 10%–15% of HER2-negative, hormone receptor-positive breast cancers [17]. The development of next-generation sequencing has led to the identification of more mutated genes in a series of early breast cancers, including tumor protein p53 (TP53) (41% of tumors), phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) (30%), MYC (20%), phosphatase and tensin homolog (PTEN) (16%), cyclin D1 (CCND1) (16%), ERBB2 (13%), fibroblast growth factor receptor 1 (FGFR1) (11%), and GATA3 (10%) [18, 19]. This comprehensive genomic profiling has deepened insights into the molecular underpinnings of breast cancer development and potential therapeutic targets.

Although gene mutations affect specific protein sequences, the intricate functional regulation of proteins is primarily regulated by diverse post-translational modifications (PTMs). These PTMs involve chemical alterations to proteins that significantly modify their biochemical properties and are estimated to influence approximately 50% to 90% of all human proteins [20]. Each type of PTMs consists various enzymes to mediate the modification and de-modification to ensure it is in a dynamic balance. However, dysregulation of these enzymes contributes to a variety of pathologies to drive diseases. Owing to technological advancements over the past decade, several dysregulated enzymes [21,22,23,24,25,26,27,28] have been discovered contributing to imbalanced PTMs, including phosphorylation, ubiquitination, SUMOylation, neddylation, citrullination, acetylation, methylation, glycosylation, palmitoylation, succinylation, and S-Nitrosylation in breast cancer [24, 29,30,31,32,33,34,35,36,37,38,39,40,41]. These PTMs regulate DNA damage repair, signal transduction, immune responses, metabolic reprogramming, cell proliferation, cell cycle regulation, angiogenesis, malignant transformation, cell epithelial–mesenchymal transition (EMT) and invasion, and autophagy and apoptosis by effecting the stability, cellular localization, activity, interaction with other macromolecules, and cellular responses to different stimuli of the target substrates [20, 42,43,44,45,46,47,48] to play either a promotive role or a suppressing role in breast cancer.

SUMOylation, a type of ubiquitination-like modification, was described for the first time more than 25 years ago [49, 50]. It is characterized by the attachment of small ubiquitin-like modifier (SUMO) proteins to the lysine residues of target proteins [50, 51]. The SUMO family consists of three members: SUMO1–3, in which SUMO1 shares only 50% homology with SUMO2 and SUMO3, whereas SUMO2 and SUMO3 have > 97% sequence similarity [52,53,54]. SUMOylation is catalyzed by a cascade of three enzymes: activating enzyme (E1), conjugating enzyme (E2, UBC9), and ligating enzyme E3 [36, 55]. As one of the most dynamic modifications, six SUMO-specific protease (SENP) family proteins, namely, SENP1–3 and SENP5–7, can readily deconjugate SUMO molecules [56]. Among the SENPs, SENP1 plays a central role in deconjugating both SUMO1 and SUMO2/3 modifications in many target proteins and is therefore involved in many cellular processes [57]. Our studies, as well as other’s have demonstrated that SUMOylation plays important roles in multiple cellular processes, such as signaling transduction, gene regulation, DNA damage repair, cell death, and cell proliferation, primarily by affecting the cellular localization, stability, activity, protein–DNA, or protein–protein binding of substrates [57,58,59,60,61,62,63,64,65,66].

SUMOylation exerts critical functions in breast cancer progression. In general, SUMOylation promotes breast cancer by boosting tumor cell proliferation, migration and EMT. For example, SUMOylation of BRCA1 at K32 and K1690 has been shown to induced G0/G1 phase transition in the ER-positive breast cancer cells [67]. In addition, SUMOylation of talin at K2445 and K841 positively impacts the migration of MDA-MB-231 cells through the facilitation of focal adhesion disassembly [68]. Moreover, SUMOylation of transforming growth factor beta (TGF-β) receptor 1 (TβRI) at K389 promoted cancer cell metastasis by enhancing the interaction between TβRI and SMAD2/3, which in turn activates the TGF-β–SMAD signaling pathway and EMT [69]. However, SUMOylation of SMAD4 at K159 inhibited the TGF-β–SMAD4 signaling pathway by enhancing the interaction between SMAD4 and the transcriptional corepressor Daxx [70]. Additionally, the SUMOylation of PIN1 at K6 and K63 suppressed its oncogenic function [71]. These findings collectively indicate that SUMOylation can serve both pro-oncogenic and tumor-suppressive roles in breast cancer. Considering that SUMOylation is somewhat newly discovered, more functions and mechanisms by which SUMOylation is involved in breast cancer remains to be further investigated.

Nevertheless, breast cancer is regulated by a complex network of signaling pathways that are not controlled by only one PTM, but rather by the coordinated actions of PTM combinations. During breast cancer progression, multiple PTMs or the same PTM at different modification sites always occur on a substrate. Multiple PTMs may simultaneously or sequentially occur, which is necessary for the distinct outcomes of signaling cascades. Therefore, understanding the crosstalk between different PTMs is crucial for unraveling the molecular mechanisms and developing precise therapeutic strategies for breast cancer. Up to date, SUMOylation has been found to interact with other PTMs, in particular, with phosphorylation, ubiquitination, acetylation, and methylation in breast cancer. Therefore, in the present review, we discuss the crosstalk between SUMOylation and these four PTMs to detail the role and molecular mechanisms of each of these crosstalk pairs in regulating breast cancer oncogenesis, offering insights into their potential clinical applications in breast cancer treatment.

Phosphorylation and SUMOylation in breast cancer

Phosphorylation is possibly the most common PTM type and has a history of more than 60 years [72,73,74]. In breast cancer, phosphorylation is widely involved in multiple biological processes. Therefore, it frequently interacts with other types of PTM. Unsurprisingly, there is extensive crosstalk between phosphorylation and SUMOylation in breast cancer. To date, this interaction has mainly been observed in nuclear proteins and signaling transduction through phosphorylation-directed SUMOylation, although instances of SUMOylation-directed phosphorylation have also emerged. Thus, understanding the relationship between the two PTMs is vital for clarifying breast cancer pathogenesis, drug resistance mechanisms, or new therapeutic drug development.

Given that nuclear proteins are the predominant targets of SUMOylation [75], phosphorylation-directed SUMOylation mainly occurs on such proteins. In breast cancer, phosphorylation of Krüppel-like factor 8 (KLF8), a key oncogene regulating gene transcription and breast cancer-related cellular processes, at Ser-80 is needed for SUMOylation at K67 upon DNA damage; this may be a novel mechanism promoting DNA repair and cell survival in breast cancer due to the inhibitory role of KLF8 SUMOylation on its transcription activity, functioning as a negative feedback [76]. Interestingly, this feedback is broad in the phosphorylation-SUMOylation crosstalk. For instance, phosphorylation dependent GATA1 SUMOylation inhibits its DNA binding activity [77, 78], while signal transducer and activator of transcription (STAT) 1 phosphorylation at Y701 promotes SUMOylation at K703, which then suppresses further STAT1 phosphorylation to protect cells from interferon γ (IFNγ) hypersensitivity [79]. Therefore, phosphorylation dependent SUMOylation seems like a negative feedback mechanism to avoid substrate hyperactivity. However, SUMOylation can also positively reinforce substrate activity. The phosphorylation of ERRalpha1 at Ser19 enhances its SUMOylation at K14, further promoting the transcriptional activities of ERRalpha1 by affecting its response to coactivator [80]. However, in the case of ERβ, phosphorylation at Ser6, while it does enhance SUMOylation at K4, this subsequent SUMOylation suppresses the transcriptional activity of ERβ in breast cancer cells [81]. On the other hand, if SUMOylation is depressed by phosphorylation, the situation diverges. For example, the inhibited SUMOylation of tumor suppressor p53 mediated by its phosphorylation can enhance p53 transcription activity [82]. Another instance shows that the phosphorylation-mediated inhibition of SUMOylation on the pro-inflammatory factor inhibitor of kappa B alpha (IκBα) can boost IκBα ubiquitination, accelerating degradation and promoting p65/p50 translocation [83].

The biological process of tumorgenesis is driven by signaling transduction, of which, Rac-alpha serine/threonine-protein kinase (AKT) hyperactivation is one of the most commonly observed in breast cancer, typically stemming from PTMs rather than genetic mutations in the kinase. Among these PTMs, the crosstalk between phosphorylation and SUMOylation significantly affects AKT activity. AKT can be modified by SUMO1 and SUMO2; however, irrespective of SUMO1- or SUMO2-type modification, SUMOylation promotes AKT activity, thereby regulating MCF-7 cell proliferation [84]. This feedback affects not only cell proliferation but also macrophage polarization in tumors. Enhanced AKT1 SUMOylation upon SENP3 loss resulted in AKT1 hyperphosphorylation and activation, thereby facilitating M2 polarization, breast cancer cell proliferation and metastasis [85]. Similarly, other kinases exhibit crosstalk, as seen in DDX5 where phosphorylation-dependent SUMOylation stabilizes the protein and boosts the formation of the DDX5/Drosha/DGCR8 complex, promoting microRNA-10b processing and ultimately contributing to breast cancer cell proliferation, invasion, and metastasis [86].

What could be the molecular mechanism behind the phosphorylation-directed SUMOylation in breast cancer? The main mechanism may be owing to the presence of a phosphorylation-dependent SUMOylation motif (PDSM) characterized by ΨKx(D/E)xxSP, where ΨKx(D/E) represents a SUMO consensus site followed by any two residues and a serine and a proline-directed phosphorylation site [77, 87, 88]. Many nuclear proteins contain the PDSM, including KLF8 (Fig. 1A), ER, heat shock transcription factor 1 (HSF1), myocyte enhancer factor 2 (MEF2), GATA1, peroxisome proliferator activated receptor gamma (PPARγ), and nuclear receptor corepressor (NCoR) [77, 89]. The phosphorylation of the serine or proline residues of this motif provides the essential negative charge, enabling interaction with the basic residues of UBC9 or SUMOs, thereby enhancing SUMO conjugation [81, 90]. Similar to PDSM, the negative charge-dependent SUMOylation motif (NDSM) (ΨKXEXXEEEE) also contains the ΨKxE consensus motif, followed by at least two acidic residues localized < 10 residues away from the C-terminal end of the target lysine residue [90]. In addition to consensus covalent SUMOylation, the nonconsensus SUMO conjugation motif is also phosphorylation-dependent, where the consensus D/E residue is substituted for a serine residue, and whose phosphorylation provides the negative charge for nearby SUMOylation, such as ER [81]. This extended PDSM offers a valuable signature for predicting SUMO substrates that are regulated by protein kinases. Indeed, for PR, the PDSM is absent [87], and this may be why there is always controversial over whether there is phosphorylation dependent SUMOylation on PR. However, for proteins such as AKT, which also do not contain these motifs, its SUMOylation can also been modified by altering the characteristics of SUMO-related enzymes or molecules. Phosphorylation-dependent AKT SUMOylation could occur because AKT phosphorylation increases its own activity and directly phosphorylates UBC9 at Thr35 and SUMO1 at Thr76, fostering UBC9 thioester bond formation and SUMO1 stabilization, thereby amplifying AKT SUMOylation and creating a positive feedback loop. This heightened AKT-induced phosphorylation of UBC9 and SUMO1 also impacts the SUMOylation of other proteins, such as PTEN, further governing cellular processes in breast cancer in breast cancer (Fig. 1B) [91]. However, for cases where phosphorylation inhibits SUMOylation, the precise mechanism remains unclear. One possibility is that substrate phosphorylation inhibits its binding to SUMO ligase, supported by the p53 case where SUMOylation of p53 is inhibited by site-specific phosphorylation, which reduces the binding of p53 to UBC9 [92]. Although SUMOylation of c-Jun, ETS domain-containing protein Elk1 (ELK1), and promyelocytic leukemia (PML) are also repressed by phosphorylation, the crosstalk mechanism is still unknown, possibly due to a conformational change caused by phosphorylation that makes the SUMOylation site exposed to enable more rapid cleavage by SUMO proteases [93,94,95,96]. Taken together, phosphorylation-directed SUMOylation through PDSM or NDSM is the core mechanism driving the crosstalk between these two PTMs.

Fig. 1

Three potential mechanisms of crosstalk between SUMOylation and phosphorylation in breast cancer. A Phosphorylation dependent SUMOylation of KLF8. Upon DNA damage signals, phosphorylation of KLF8 at S80 promotes KLF8 SUMOylation at K67 because of the phosphorylation dependent SUMOylation motif at K67-S80, regulating DNA damage repair and breast cancer cell survival. B AKT SUMOylation enhances phosphorylation of UBC9 (the only SUMOylation E2) and SUMO1 to further promote global SUMOylation. Upon pro-tumorigenic stimuli, AKT undergoes phosphorylation at the T308 and S473 sites, and subsequently mediates SUMOylation at K276, leading to enhanced AKT activity. Upregulated AKT promotes mediated the phosphorylation of UBC9 at T35 and SUMO1 at T76, thereby further leading to the enhancement of SUMOylation of multiple proteins such as AKT, STAT1, PTEN, etc. Enhanced AKT SUMOylation further promotes AKT activity, thus forming a positive feedback loop to regulate the cellular function of tumor cells and the occurrence of tumors. C SUMOylation dependent phosphorylation of MZF1. SUMOylation of MZF1 at K23 promotes MZF1 phosphorylation at S27, thereby further regulating the ERBB2 signaling pathway and breast cancer cell invasion

In addition to phosphorylation-directed SUMOylation, there is also a new crosstalk called “SUMO-directed phosphorylation” in breast cancer. This process involves the poly-SUMOylation of myeloid zinc finger-1 (MZF1) at K23 directs MZF1 phosphorylation at S27 to further mediate invasive ERBB2 signaling in breast tumors (Fig. 1C) [97]. This crosstalk through a mechanism where SUMOylation at K23 opens up and exposes the S27, which otherwise is masked and not approachable for phosphorylation.

In conclusion, the phosphorylation of substrate proteins may have either positive or negative effects on SUMOylation in breast cancer, with the majority of studies suggesting a positive effect. Given that SUMOylation regulates the subcellular localization, protein stability, and protein–protein/DNA binding of substrate proteins, phosphorylation directed SUMOylation may play a feedback role through these ways to prevent substrate hyperactivity and cellular homeostasis or work synergistically with each other to enhance substrate activity. Through these mechanisms, phosphorylation-dependent SUMOylation can fundamentally alter the biological properties of substrate proteins, contributing further to breast cancer progression by influencing tumor cell proliferation, metastasis, and mitochondrial function.

SUMOylation and ubiquitination in breast cancer

Ubiquitination is a multi-step process catalyzed by ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2), and ubiquitin ligases (E3), which is similarly to SUMOylation [43]. Ubiquitination either promotes or suppress breast cancer. There is substantial evidence demonstrating that SUMOylation directly influences ubiquitination, which we will discuss by examining three aspects: SUMOylation-induced ubiquitination, SUMOylation-repressed ubiquitination, and SUMO type-specific effects on ubiquitination.

SUMOylation promotes ubiquitination to play a role in breast cancer suppression. For example, the SUMOylation of forkhead box M1 (FOXM1) at multiple sites—K201, K218, K460, K478, and K495—which facilitates ring finger protein (RNF) 168 recruitment, leading to FOXM1 ubiquitination and degradation, thereby inhibiting MCF-7 cell proliferation and mitotic progression [98, 99]; this may suppress breast cancer progression, metastasis, and genotoxic agent responses [100,101,102,103,104]. The proteasome degradation pathway of c-MYC, a frequently overexpressed oncogene in breast cancer, may also depend on SUMOylation; its SUMOylation at K326 results in its subsequent ubiquitylation and degradation by the proteasome (Fig. 2A) [105, 106]. This finding has been further confirmed by another study showing that SENP1, the major deSUMOylase often overexpressed in breast cancer tissues, leading to c-MYC deSUMOylation and the subsequent decrease in c-MYC polyubiquitination; this results in high c-MYC expression, leading to breast cancer cell proliferation and transformation [107].

Fig. 2

Main mechanisms of the crosstalk between SUMOylation and ubiquitination in breast cancer. A SUMOylation of c-MYC promotes its ubiquitination and degradation. SUMOylated c-MYC at K326 increases recruitment of RNF168 (ubiquitination E3 ligase) to enhance c-MYC ubiquitination. B Potential mechanisms of crosstalk between SUMOylation and ubiquitination at c-MYC. (1) SUMOylation and ubiquitination at different sites. (2) Ubiquitination occurs on SUMO molecules. The SUMOylation of c-MYC induces the binding of ubiquitin molecules on SUMO, leading to c-MYC proteasome degradation. (3) SUMOylation occurs on ubiquitin molecules, followed by further ubiquitination on SUMO. (4) A mixture of form (2), (3), and (4). C Progesterone receptor isoform B (PRB) SUMOylation competitively inhibits its ubiquitination. The ubiquitination site and the SUMOylation modification site are both located at the K388 of PRB. Once the K388 is mutated, neither SUMOylation nor ubiquitination of PRB can occur. D Different SUMOylation types of HDAC1 regulate different ubiquitination. In breast cancer tissue, SUMO2-type modification mediated by E3 ligase PIASy inhibits HDAC1 ubiquitination, thus improving HDAC1 stability. However, the SUMO1-type modification promotes HDAC1 ubiquitination, thereby reducing its protein stability

Several reasons may contribute to SUMOylation-dependent ubiquitination: 1) Similar to phosphorylation, these proteins may contain a SUMOylation-dependent ubiquitination motif to ensure SUMOylated proteins are better substrates than non-SUMOylated proteins (Fig. 2B, module 1). To support this, a study has reported that SUMOylation promotes an MYC mutant that cannot target FBW7 to be a FBW7 substrate [107]. 2) Substrates may be co-modified by both SUMO and ubiquitin to form a SUMO–ubiquitin chain because the ubiquitination of SUMO and SUMOylation at multiple lysine residues of ubiquitination have been identified [23, 108] (Fig. 2B, module 2–4). Evidence exists where MYC can be co-modified by both SUMO and ubiquitin, and SENP1 can stabilize MYC by removing ubiquitination via deSUMOylation [109]. Moreover, a single ubiquitin attached to MYC was identified by overexpressing SENP1 without proteasome inhibition; this indicates that SUMOylation occurs on a single ubiquitin molecule [109]. 3) SUMOylation may affect the stability or activity of the ubiquitin ligases, thereby promoting global ubiquitination [105, 109,110,111,112]. A similar mechanism has been observed during DNA double-strand break repair. Both SUMOylated RNF168 and HECT and RLD domain containing E3 ubiquitin protein ligase 2 (HERC2) enhance their association with RNF8, leading to the formation of an active UBC13-RNF8 complex that facilitates ubiquitin chain formation at the site of DNA damage [113]. 4) Another mechanism may involve the SUMO-targeted E3 ubiquitin ligase (STUbL) RNF4 [113]. The N-terminus of RNF4 contains four SUMO-interacting motifs (SIMs, SIM1-4) for recognizing poly-SUMOylated substrates. Among these motifs, SIM2 and SIM3 play a significant role in binding to SUMO-2 chains while SIM1 and SIM4 have a minor role [114, 115]. Binding to poly-SUMO chains induces the dimerization of the C-terminal RING domains of RNF4, which stabilizes the E2-ubiquitin thioester bond and subsequently catalyzes poly-ubiquitination of the substrates [113, 116,117,118,119]. Typically, this type of poly-ubiquitination results in proteasome-mediated degradation [120, 121]. In addition, RNF4 is also implicated in recruiting proteasome components to indirectly promote ubiquitin conjugation and proteasomal degradation [117]. However, additional studies are warranted to reveal the in-depth molecular mechanism of the crosstalk in breast cancer.

On the other hand, SUMOylation predominantly inhibits substrate ubiquitination and proteasomal degradation pathways by competing for the same lysine residues. Breast cancer-associated gene 2 (BCA2), an E3 SUMO ligase for IκBα in breast cancer cells, promotes IκBα SUMOylation, thereby preventing its ubiquitination for proteasomal degradation and boosting breast cancer cell proliferation and migration [122]; similarly, K379 of delta-lactoferrin (DLf), which can be either ubiquitinated or SUMOylated, is a key site for controlling DLf stability. SUMOylation competes with ubiquitination and protects DLf degradation by positively regulating its stability [123]. In progesterone receptor isoform B (PRB), both ubiquitination and SUMOylation occur at K388, and reduced SUMOylation accelerates PRB ubiquitination, leading to a decrease in T47D cell proliferation [124]. Mutations in the K388 SUMOylation site of PRB hinder progesterone-dependent PR degradation, indicating that K388 is a dual SUMOylation and ubiquitination site; when the conjugation site is mutated, neither modification can take place (Fig. 2C) [87]. However, there are cases where mutations in SUMOylation sites enhance ubiquitination. For example, SUMOylation at Lys-2806 of zinc finger homeobox 3 (ZFHX3) enhances the stability of ZFHX3 by interfering with its ubiquitination and proteasomal degradation, while the ZFHX3 K2806R mutant decreases its protein stability, further suppressing breast cancer growth [125]. The estrogen-induced SUMOylation of pescadillo ribosomal biogenesis factor 1 (PES1) stabilizes PES1 by inhibiting its ubiquitination, but mutation of K517R promotes the PES1 ubiquitin–proteasome pathway, thereby suppressing breast cancer cell proliferation and tumor growth [126]. These proteins may be ubiquitinated at residues other than the main SUMOylation sites.

In addition, the interaction between SUMOylation and ubiquitination depends on SUMO type. While SUMO1-type SUMOylation of histone deacetylase (HDAC) 1 promotes its ubiquitination and degradation, SUMO2-type SUMOylation of HDAC1 enhances its protein stability. This selective SUMOylation may be mediated by specific SUMO E3 ligases in specific cellular environments, further leading to ubiquitination regulation. Protein inhibitor of activated STAT 4 (PIASy), overexpressed in breast cancer cells, selectively promotes the conjugation of HDAC1 to SUMO2 (Fig. 2D) [127]. However, further investigation is necessary to elucidate the mechanisms underlying how different SUMOylation types differently affect ubiquitination.

Overall, SUMOylation plays an important role in regulating ubiquitination, either enhancing or repressing it. Regardless of the direction, SUMOylation-regulated ubiquitination constitutes a critical mechanism in breast progression. Targeting this regulatory mechanism presents a potential novel therapeutic strategy.

SUMOylation and acetylation in breast cancer

Acetylation is a reversible process mediated by lysine acetyltransferases and deacetylases for adding and removing the acetyl group from the side chain of lysine, respectively. Both histone and non-histone proteins are substrates of acetylation. Canonical acetylation occurs in histone proteins, where modifications play an essential role in breast cancer development and prognosis. Dysregulated deacetylation promotes cancer cell proliferation, cell cycle arrest, abnormal cell death, immune destruction, immune evasion, migration invasion, and metastasis [27, 128]. Interestingly, histone acetylation has crosstalk with non-histone protein SUMOylation. A notable example involves tripartite motif-containing protein 24 (TRIM24), a histone reader aberrantly expressed in breast cancer. In that study, researchers observed that the association of chromatin with TRIM24 leads to TRIM24 SUMOylation at lysine residues 723 and 741, which depends on the acetylated lysine 23 of histone H3, further promoting cell adhesion to extracellular matrix proteins (Fig. 3A) [129]. This interaction may be an important mechanism to explore the downstream functions that regulate specific genes implicated in breast cancer [129]. Besides it, histones themselves serve as substrates for SUMOylation, regulating multiple cellular process, such as gene regulation, chromatin condensation, p300-mediated transcription, double-strand break repair, and Set3-histone-deacetylase complex-mediated transcriptional regulation [130,131,132]. Therefore, it is not surprising that SUMOylation has crosstalk with acetylation on histone proteins. For example, SUMOylation of histone H4 at K12 inhibits H4 tail acetylation mediated by the acetyltransferase p300, indicating a negative crosstalk between histone SUMOylation and acetylation [131]. However, the role of this crosstalk in breast cancer progression remains unexplored.

Fig. 3

Crosstalk between SUMOylation and acetylation in breast cancer. A TRIM24 binds to chromatin and recognizes histone 3 with non-methylation at K4 and acetylation at K23 (H3K4me0/K23ac). Subsequently, TRIM24 undergoes SUMOylation at K723 and K741, leading to a decrease in TRIM24 stability. B SUMOylation of non-histone protein p53 inhibits its acetylation. p53 can be acetylated at multiple lysine sites under the mediation of P300. However, once the K386 of p53 is SUMOylated, it can inhibit p53 acetylation, thus affecting the breast cancer progression. Since the K386 is also one of the acetylation sites of P53, the SUMOylation at this site may have a competitive inhibitory effect on the acetylation at the same site. C Acetylation regulates SUMOylation by UBC9. Acetylation of UBC9 at K65 inhibits the binding of UBC9 to substrate proteins, thereby inhibiting systemic SUMOylation

Non-histone proteins are also subject to acetylation, a process referred to as non-canonical acetylation. In breast cancer, this type of acetylation effects metastasis, cancer cells proliferation, and the sensitivity of tumor cells to anti-tumor therapy, by regulating the functions of target proteins, such as Twist, RelA/P65, oncogene nuclear receptor coactivator amplified in breast cancer 1 (AIB1), homeobox B13 (HOXB13) [20]. In essence, the crosstalk between SUMOylation and acetylation mainly occurs competitively on non-histone proteins. For instance, SUMOylation of the tumor suppressor p53 at K386 inhibits its acetylation (Fig. 3B) [133], and a decrease in p53 acetylation is known to promote breast cancer development [134,135,136]. Another illustrative example involves DLf, when it is acetylated at K13, this decreases its SUMOylation and enhances the transcriptional activity of DLf, thereby may possess anti-breast cancer activity [123, 137]. A similar interplay can take place among transcription co-factors. When KRAB domain-associated protein 1 (KAP1) undergoes SUMOylation, it leads to decreased H3-K9 and H3-K14 acetylation and enhanced H3-K9 methylation at the p21 promoter, thereby regulating Dox-induced p21 expression, desensitizing MCF-7 cells to Dox-elicited cell death [138]. Moreover, acetylation extends its influence to global SUMOylation by regulating UBC9, the only E2 of SUMOylation. The acetylation of UBC9 at K65 attenuates the binding of UBC9 to substrates, decreasing overall SUMOylation (Fig. 3C) [139, 140]. The dynamic switch between deacetylation and SUMOylation may represent a novel mechanism underlying breast cancer progression.

The possible mechanism involves the crosstalk is that the conjugation site of SUMOylation and acetylation are the same. What can support it is that K386 is the competitive conjugation site for both acetylation and SUMOylation of p53 and K13 is the competitive conjugation site for both acetylation and SUMOylation of DLf [82, 123]. Another possible mechanism may involve the deacetylase, such as HDAC4, which has been considered as an E3 of SUMOylation of several proteins in breast cancer progression, such as silent information regulator 1 (SIRT1), IκBα, androgen receptor, and hypermethylated in cancer 1 (HIC1) [141,142,143,144]. This suggests that HDAC4 may mediate deacetylation/SUMOylation switch simultaneously. Additionally, it is intriguing to note that acetylation can neutralize the positive charge of lysine as well as SUMO surface, thereby preventing SUMO from binding to the negatively charged residues of SUMO-interaction motifs, which contain a hydrophobic residue core that binds to the hydrophobic pocket on the SUMO for non-covalent binding [145]. Therefore, it is plausible that acetylation might also inhibit SUMOylation in breast cancer through a comparable charge neutralization mechanism.

SUMOylation and protein methylation in breast cancer

Protein methylation is a process of transferring the methyl group from s-adenosyl methionine to the side chains of target protein amino acid residues, mediated by methyltransferases. The methylation at lysine and arginine residues of substrates is the prominent and universal types of methylation in breast cancer. In general, protein methylation promotes breast cancer progression by activating oncogenic signaling pathways, facilitates breast cancer cells EMT and migration by activating oncogenic genes or repress antitumor ability of tumor suppressor proteins, such as p53. Methylation at residue K370 on p53 represses its antitumor ability by inhibiting p53-mediated cancer cell apoptosis, while methylation at K382 represses its transcriptional activity [146, 147]. Interestingly, as mentioned above, p53 also can be SUMOylated. However, further study is required to analyze whether the methylation of p53 has a crosstalk with SUMOylation in breast cancer.

In fact, the crosstalk between protein methylation and SUMOylation is frequently observed in the regulation of methyltransferases or SUMO ligases themselves, rather than just oncogenes or tumor suppressors. For example, UBC9 can promote the SUMOylation of PR-Set7, a chromatin-modifying enzyme that specifically monomethylates lysine 20 of histone H4 (H4K20me1), to further decrease the expression of downstream genes mediated by PR-Set7, potentially in response to DNA damage in breast cancer [148]. Given that H4K20me1 has been reported to be associated with gene bodies, promoters, and enhancers, the repressive role mediated by SUMOylated PR-Set7 could be attributed to altered methylation at histone H4's K20 residue. On the other hand, methylation can also have an impact on SUMOylation. A case in point is the SUMO1 activating enzyme subunit 1 (SAE1), a SUMOylation enzyme; levels of DNA methylation at the SAE1 gene site (cg14042711) are negatively correlated with levels of SAE1 expression and global SUMOylation, indicating the regulatory role of methylation in SUMOylation in breast cancer [149]. Intriguingly, the AKT SUMOylation is mediated by SAE1 [150, 151], suggesting that SAE1 methylation may suppress cancer cell proliferation by reducing AKT SUMOylation.

In conclusion, unlike other types of protein modifications where crosstalk often takes place on the same substrate, the interaction between methylation and SUMOylation typically manifests through regulation of the enzymes involved in these processes, rather than direct modification of a shared substrate. This distinctive pattern may stem from the fact that protein methylation can take place at multiple amino acid residue side chains, whereas ubiquitination and acetylation, like SUMOylation, predominantly occur on lysine residues. This inherent difference allows for an indirect yet significant interplay between methylation and SUMOylation pathways.

Conclusion and perspectives

Breast cancer is the most prevalent malignant tumor and a leading cause of mortality among women worldwide. The progression of breast cancer is intricately governed by various protein PTMs, which are modulated by an array of cytokines, drugs, tumor microenvironments, genetic factors, and signaling pathways. However, clinical trials on PTM in breast cancer are primarily focus on protein acetylation, and most of which are only in their initial stages [20]. Recent advances in omics technologies, such as mass spectrometry, high-throughput sequencing, and bioinformatics, have facilitated the identification of new PTMs and helped reveal the mechanisms by which they regulate breast cancer progression [21, 23, 48]. SUMOylation, a recently identified modification, plays an important regulatory role in regulating cancer cell proliferation, migration, and metastasis. Notably, in the progression of breast cancer, SUMOylation frequently forms intricate networks with other PTMs, leading to complex interactions. In this review, we highlighted the crosstalk between SUMOylation and other PTMs, including phosphorylation, ubiquitination, acetylation, and methylation, in breast cancer. Generally, SUMOylation has a crosstalk with these PTMs, which together regulate the various stages of breast cancer development (Table 1). These pairs of crosstalk often serve as either redundant or negative feedback regulators to maintain cellular homeostasis; however, any dysregulation can disrupt this equilibrium and drive breast cancer progression. In the crosstalk network, SUMOylation plays a central role, providing feedback regulation to phosphorylation while also influencing subsequent ubiquitination, acetylation, and methylation. Mechanistically, changes in amino acid side chain charge distribution, conformational shifts, or the occupancy of specific sites due to post-translational modifications constitute the fundamental mechanisms driving crosstalk, with unique sequences like PDSM or specific modification sites serving as the mediators for such interactions. This comprehensive review may widen our current understanding of the relationship between SUMOylation and other PTMs in breast cancer and provide new perspectives for breast cancer treatment. However, future studies are required to address several important questions that are still unanswered.

Table 1 The role of crosstalk between SUMOylation and other PTMs in breast cancer

SUMOylation is a dynamic process, in which deSUMOylation is mediated by the SENP family proteins, while SUMOylation is primarily facilitated by various SUMOylation E3 ligases [36, 55, 56]. Notably, the crucial SUMO protease SENP1 can be upregulated during tumor development and progression [57, 107], possibly due to its transcriptional regulation by HIF1α, which is significantly upregulated by the tumor microenvironment [152,153,154,155,156,157]. In addition, SUMOylation E3 ligases like BCA2 also highly expressed in breast cancer [122, 158]. These alterations in the enzymes of deSUMOylation and SUMOylation collectively disrupt SUMOylation homeostasis during breast cancer progression. SUMOylation predominantly targets nuclear proteins, and thus, the interplay between SUMOylation and other PTMs is largely observed in transcription factors, transcriptional co-regulators, and nuclear receptors. However, our previous studies have found that SUMOylation also has dominant role in regulating membrane protein, such as FGFR1 [63], and mitochondria proteins like fission protein 1 (FIS1) [65]. Of significance, these two proteins are instrumental in breast cancer progression due to their critical roles in regulating EMT and mitochondrial function, respectively. Moreover, both FGFR1 and FIS1 exhibit additional PTMs beyond SUMOylation, including phosphorylation and ubiquitination [63, 159,160,161]. Therefore, it is worth further investigations to reveal the crosstalk of SUMOylation and other PTMs in these non-nuclear proteins in breast cancer. Such investigations could potentially uncover novel regulatory mechanisms and contribute significantly to our understanding of breast cancer development and progression.

In addition to PTMs mentioned above, some other rare PTMs have also been reported in breast cancer. Neddylation is another type of ubiquitination-like modification that involves covalent conjugation of neural precursor cell-expressed developmentally downregulated 8 (NEDD8) to a lysine residue in the target protein [162]. Studies have documented elevated levels of neddylation in breast cancer on various targets, such as p53, Smurfl, PTEN, murine double minute 2 (MDM2), BCA3, and TGF-β II [163,164,165,166,167,168,169]. Given the similarities between neddylation and SUMOylation, studies have reported the crosstalk between neddylation and SUMOylation. As an example, the SUMOylation of ribosomal protein L11 (RPL11) negatively modulates the conjugation of NEDD8 to RPL11 and promotes RPL11 translocation outside the nucleoli [170]. However, to date, no studies have specifically reported the crosstalk between SUMOylation and neddylation in breast cancer. Based on the interaction between SUMOylation and ubiquitination, it is plausible to hypothesize that SUMOylation might primarily suppress neddylation, thereby regulating substrate protein localization or activity and contributing to breast cancer tumorigenesis. Furthermore, protein glycosylation has been shown to play an oncogenic role in breast cancer by promoting proliferation and metastasis of cancer cells, inhibiting the sensitivity of tumor cells to anti-tumor therapy, and altering the immune microenvironment and antitumor immune response [20]. In addition, citrullination and palmitoylation have also been reported in breast cancer. Citrullination regulates epidermal growth factor (EGF)- phosphatidylinositol 3-kinase (PI3K) signaling, nuclear localization, and TGF-β signaling, further mediating gene transcription, cell proliferation, cell invasion and migration, and cancer cell EMT in breast cancer tumorigenesis and progression [30, 45]. Palmitoylation of cluster of differentiation (CD) 44 decreases its interaction with migratory binding partner ezrin, therefore inhibiting breast cancer cell migration [171]. Nevertheless, the crosstalk between SUMOylation and these three PTM type has not been revealed in breast cancer.

Because of the central role of SUMOylation in the crosstalk, some inhibitors targeting SUMOylation in breast cancer have been explored. However, these methods are primarily conducted by regulating enzymes that affect SUMOylation, including the SAE1/2 and the unique E2 UBC9 [172,173,174,175]. Although these inhibitors play crucial anticancer roles in breast cancer cell lines, including MDA-MB-231, MCF-7, and BT474, by accelerating autophagy-dependent cancer cell death or repressing cell migration and invasion [176, 177], it should be noted that targeting these enzymes can alter global SUMOylation patterns. The specific effects of such interventions therefore require further clarification. Indeed, although the majority of SUMOylation events may facilitate breast cancer tumorigenesis and progression through accelerating cell cycle transitions and promoting EMT and tumor cell migration, a subset of SUMOylation processes function as tumor suppressors. For example, the effects of SENP1- and SENP2-mediated deSUMOylation on tumor development are different, although both can suppress global SUMOylation [42]. Another example involves AKT and c-MYC, which are both deSUMOylated by SENP1 [107, 178], suggesting that using SENP1 inhibitors can simultaneously enhance the SUMOylation of the two substrates. Nevertheless, while c-MYC SUMOylation promotes its degradation and thereby exerts suppressive effects on breast cancer [105], AKT SUMOylation enhances AKT activity to drive breast cancer progression [84]. Consequently, broadly targeting SUMOylation enzymes to either augment or diminish global SUMOylation might compromise therapeutic efficacy due to potential off-target effects. To effectively inhibit breast cancer growth, intervention strategies need to be more precise and targeted, focusing on specific substrates and sites.

Previous discussion mentioned the crosstalk between SUMOylation and other PTMs, including phosphorylation, ubiquitination, acetylation, and methylation in breast cancer. However, a substrate protein always contains multiple types of PTM. These PTMs coordinately regulate the function of substrate. As mentioned above, IκBα has multiple sites for phosphorylation, SUMOylation, and ubiquitination; here, phosphorylation at certain sites depresses SUMOylation, facilitating ubiquitination because the same site is shared by both modifications. Similarly, the tumor suppressor p53 also contains multiple PTMs, such as phosphorylation, SUMOylation, ubiquitination, and acetylation. SUMOylation of p53 at K386 inhibits its acetylation by p300 and decreases DNA binding activity. These results suggest the central role SUMOylation plays in the PTM network. Thus, to achieve effective therapeutic outcomes without causing unintended consequences, the inhibition of SUMOylation or the crosstalk between SUMOylation and other PTMs must be targeted specifically rather than relying on enzymes that globally promote or remove SUMOylation. To gain it, specific PDSMs and SUMOylation modification sites for a certain substrate are needed to be characterized for specifically regulation.

Further studies are warranted to elucidate the balance between SUMOylation and other PTMs in cancers, particularly under infection–inflammation-associated events. To this end, additional studies are suggested for the following: (1) performing a global RNA sequencing or microarray analysis of SUMOylation E3 ligases in different primary cancer samples and using bioinformatics tools to provide clues and predict the targets of E3 ligase; (2) analyzing cytokine profiles using microarray to characterize the physiochemical properties of the tumor microenvironment to associate with the type of SUMOylation E3 ligase that is active. The insights gained from these studies will be vital for developing improved combinatorial therapeutic strategies with a well-balanced approach to control cancer cell death without affecting the survival of normal cells. (3) SUMOylation has strong heterogeneity and cannot be generalized. It has both enhancing and inhibitory effects on breast cancer and should be considered comprehensively and precisely. The process of drug development should focus on targeting specific sequences of specific target proteins, thereby improving accuracy and specificity and decreasing side effects by affecting this interaction. (4) Methods could be developed to construct SUMO chips because of the heterogeneity of SUMOylation; these chips can be combined with mass spectrometry to comprehensively analyze the relationship between SUMOylation and other PTMs. (5) In addition to SUMOylation, the crosstalk, among other modifications, can be explored to clarify the PTM network, facilitating the better development of drugs and therapeutic targets.

In conclusion, SUMOylation plays a vital role in breast cancer development; however, it does not regulate the biological characteristics of substrate proteins in a single manner but via crosstalk with various other PTMs. The crosstalk may potentially be used in breast cancer treatment. We believe that a deep understanding of the crosstalk between SUMOylation and other PTMs may facilitate a novel treatment for breast cancer.