Unraveling the regulatory role of endoplasmic-reticulum-associated degradation in tumor immunity

Abstract

During malignant transformation and cancer progression, tumor cells face both intrinsic and extrinsic stress, endoplasmic reticulum (ER) stress in particular. To survive and proliferate, tumor cells use multiple stress response pathways to mitigate ER stress, promoting disease aggression and treatment resistance. Among the stress response pathways is ER-associated degradation (ERAD), which consists of multiple components and steps working together to ensure protein quality and quantity. In addition to its established role in stress responses and tumor cell survival, ERAD has recently been shown to regulate tumor immunity. Here we summarize current knowledge on how ERAD promotes protein degradation, regulates immune cell development and function, participates in antigen presentation, exerts paradoxical roles on tumorigenesis and immunity, and thus impacts current cancer therapy. Collectively, ERAD is a critical protein homeostasis pathway intertwined with cancer development and tumor immunity. Of particular importance is the need to further unveil ERAD’s enigmatic roles in tumor immunity to develop effective targeted and combination therapy for successful treatment of cancer.

Introduction

In eukaryotes, approximately one-third of all proteins, including those important for immunity, are synthesized and folded in the endoplasmic reticulum (ER) (Jahn and Radford 2005; Oakes and Papa 2015). Physiological and pathological variations in the cellular environment often lead to the accumulation of misfolded and unfolded proteins in the ER, referred to as ER stress (Ma and Hendershot 2004). To restore ER homeostasis, cells have evolved two major protein quality control systems: the unfolded protein response (UPR) and ER-associated degradation (ERAD) (Chen et al. 2011). The UPR is triggered upstream of ERAD to restore protein homeostasis by transiently reducing protein translation, while selectively upregulating chaperone proteins to enhance protein folding and degradation (Janssens et al. 2014). The primary responsibility of ERAD is the elimination of misfolded and unfolded proteins to ensure protein quality. In addition, ERAD also regulates the quantity of properly folded proteins, including certain enzymes and lipid carriers (Fisher and Ginsberg 2002; Hampton 2002). Due to its crucial role in controlling both protein quality and quantity, altered ERAD function has been linked to numerous diseases, including neurodegenerative disorders, metabolic diseases, autoimmune diseases, and cancer (Guerriero and Brodsky 2012; Chen et al. 2018).

A critical component for cancer initiation and progression is the tumor microenvironment (TME), which is composed of tumor cells, immune cells, blood, and lymphatic vessels, as well as the many other types of cells (Junttila and de Sauvage 2013; Quail and Joyce 2013). The TME is a battlefield where immune cells struggle to gain control over the growth of tumor mass through antitumor immune responses. As the cycle of tumor regression and regrowth progresses, the TME becomes increasingly hypoxic and acidotic (Vaupel and Multhoff 2017). These adverse conditions thus heavily burden tumor and immune cells with ER stress (Cubillos-Ruiz et al. 2017). The increased ER stress impacts the proliferation and survival of tumor cells, as well as the activation and function of immune cells, thereby modulating tumor immunity, which is the process in which immune cells react to tumors (Janssens et al. 2014; Bettigole and Glimcher 2015; Grootjans et al. 2016; Cubillos-Ruiz et al. 2017).

To survive and proliferate in the harsh microenvironment, tumor cells activate the UPR to enhance ERAD for accelerated degradation of misfolded/unfolded proteins (Yadav et al. 2014). Due to the dependence of tumor cells on ERAD, a growing number of small-molecule inhibitors targeting ERAD have been developed as potential anticancer therapeutics (Tax et al. 2019). Interestingly, emerging evidence has established the role of ER stress and ERAD in regulating immunity, including ER stress induced by cancer treatment (Cubillos-Ruiz and Glimcher 2016; Cha et al. 2018). Hence, it is crucial to gain a comprehensive understanding of how ERAD impacts tumor immunity to guide both therapeutic development and cancer treatment. Here we present an overview of each individual step of the ERAD pathway and summarize the role of ERAD in immune cell development and function, as well as antigen presentation and antitumor immune responses. Finally, we discuss how small-molecule inhibitors, including standard cytotoxic chemotherapy and targeted therapy, may impact ERAD and tumor immunity.

ERAD, stepping down to degradation

ERAD research began in the 1980s when it demonstrated that unassembled membrane proteins in the ER were degraded (Lippincott-Schwartz et al. 1988; Bonifacino et al. 1990). Subsequent biochemical and genetic approaches led to the identification of specific ERAD components, spanning from yeast to metazoan (Table 1). In sequential steps, these components work together in this highly conserved pathway to recognize, retrotranslocate, and ubiquitinate proteins for proteasome-mediated degradation (Figure 1 and Table 1).

Table 1.

Components and substrates in the ERAD pathway.

Components/complexes Human (yeast)	Substrates	Function(s)	References
Substrate recognition ERMan1 (Mns1p)	ASGPR H2a	Glycan trimming	Avezov et al. (2008)
EDEM1, EDEM2, EDEM3 (Htm1p)	NHK AAT variant	Glycan trimming	Hosokawa et al. (2001), Hirao et al (2006), Hosokawa et al. (2010), Hotamisligil (2010), and Ninagawa et al. (2014)
GI/GII		Glycan trimming	D’Alessio and Dahms (2015)
OS-9 (Yos9p)	NHK AAT variant	Glycan binding	Bhamidipati et al. (2005), Szathmary et al. (2005), and Christianson et al. (2008, 2011)
XTP3-B		Glycan binding	Christianson et al. (2011) and Tyler et al. (2012)
GRP170	misfolded glycosylated client NHK	Nucleotide exchange factor	Inoue and Tsai (2016)
BiP (Kar2p)	nonsecreted Igκ light chain	Substrate recognition and recruitment	Okuda-Shimizu and Hendershot (2007)
PDI (Pdi1)	glyco-pro-α-factor; BACE457	Disulfide bonds rearrangement	Gillece et al. (1999), Molinari et al. (2002)
ERdj5-EDEM1-BiP		Substrate recognition	Ushioda et al. (2008)
UGGT		Glycoprotein glucosyltransferase	D’Alessio et al. (2010)
GRP94	NHK AAT variant	Chaperone	Christianson et al. (2008)
ERdj4	surfactant protein C	Co-chaperone	Dong et al. (2008)
ERdj5 (Scj1p)	NHK AAT variant	Co-chaperone	Dong et al. (2008) and Ushioda et al. (2008)
SEC22b (Sec22)		ER-phagosome traffic mediator	Alloatti et al. (2017)
ERp57	MHC-I heavy chain	Glycoprotein folding	Zhang, Baig et al. (2006)
Retrotranslocation and Ubiquitination Derlin-1,2,3 (Der1p)	US11-associated MHC-I heavy chain	Retrotranslocation channel	Lilley and Ploegh (2004) and Ye et al. (2004)
SEL1L-OS9-UBC6E-AUP1	US11-associated MHC-I heavy chain	Substrate recognition and dislocation	Mueller et al. (2008)
SEL1L (Hrd3p)	Igμ chain; US11-associated MHCI heavy chain	Substrate recognition and recruitment	Cattaneo et al. (2008) and Mueller et al. (2008)
SEC61 (Sec61p)		Retrotranslocation channel	Koopmann et al. (2000)
RMA1	mutant CFTR	E3 ubiquitin ligase	Morito et al. (2008)
TMEM129	US11-associated MHC-I heavy chain	E3 ubiquitin ligase	van de Weijer et al. (2014) and van den Boomen et al. (2014)
TMEM129-Derlin1	US11-associated MHC-I heavy chain	Substrate ubiquitination and translocation	van de Weijer et al. (2014) and van den Boomen et al. (2014)
HRD1-UBE2J1	non-β2m bound MHC I heavy chain	Substrate ubiquitination	Burr et al. (2011)
UBE2G2 (Ubc7p)	HMGCR	E2-conjugating enzyme	Miao et al. (2010)
SKP1-Cullin-F-box	Intergrin-β1	E3 ligase complex	Yoshida et al. (2005)
UBE2J1 (Ubc6p)	TCRα; mutant CFTR	E2-conjugating enzyme	Lenk et al. (2002) and Miao et al. (2010)
CHIP	mutant CFTR	U-box ubiquitin ligase	Younger et al. (2006)
TRC8	US2-associated MHC-I heavy chain	E3 ubiquitin ligase	Stagg et al. (2009)
HRD1 (Hrd1p)	unassembled Igμ; FAS; BLIMP-1; Nrf1	E3 ubiquitin ligase	Cattaneo et al. (2008), Tsuchiya et al. (2011), Yang et al. (2014), and Kong et al. (2016)
SEL1L-HRD1	IRE1α; P53; Pre-BCR; pro-AVP	Substrate ubiquitination and retrotranslocation	Yamasaki et al. (2007), Sun et al. (2015), Ji et al. (2016), and Shi et al. (2017)
HERP (Usa1p)	TCRα	Substrate ubiquitination	Schulze et al. (2005)
GP78	TCRα	E3 ubiquitin ligase	Zhang, Xu, et al. (2015)
TEB4 (Doa10)		E3 ubiquitin ligase	Hassink et al. (2005)
WDR20	TCRα	Adaptor Protein	Ju et al. (2018)
Dislocation and Degradation UBXD8 (Ubx2p)	ApoB100	P97 recruitment	Suzuki et al. (2012)
NGly (Png1)	glycosylated RTA	Deglycosylating enzyme	Kim et al. (2006)
RAD23 (Rad23)	glycosylated RTA	Ubiquitin binding protein	Kim et al. (2006)
AUP1 (Cue1p)	HMGCR	UBE2G2 recruitment	Jo et al. (2013)
Ataxin-3	TCRα; CD3δ	Deubiquitinating enzyme	Zhong and Pittman (2006)
P97-UFD1-NPL4 (Cdc48p-Ufd1p-Npl4p)		Substrates dislocation	Ye et al. (2001)
HERP (Usa1p)	CD3δ	Shuttle factors	Kim et al. (2008)
YOD1	TCRα; AAT	Deubiquitinating enzyme	Ernst et al. (2009)
BAG6-UBL4A-TRC35		Shuttle factors	Xu et al. (2012)
USP13	TCRα	Deubiquitinating enzyme	Sowa et al. (2009)

Figure 1.

Schematic illustration of the ERAD pathway. The first step of ERAD is substrate recognition. BiP recognizes and recruits nonglycosylated misfolded or unfolded proteins while ER resident lectins recognize glycosylated substrates. Substrate recognition is followed by retrotranslocation of substrates from the ER to the cytosol through translocon channels. The P97-UFD1-NPL4 complex facilitates the retrotranslocation of substrates that bind to P97. This complex serves as a platform for ubiquitination, which is carried out by E1 ubiquitin-activating enzymes, E2 ubiquitin-conjugating enzymes, and E3s such as HRD1. The polyubiquitinated protein is then degraded by the proteasome in the cytosol. A colored version of the figure is available online.

i. Substrate recognition

Substrate recognition, the first step in ERAD, must be precisely controlled to ensure accurate detection of target proteins (Figure 2) (Ruggiano et al. 2014; Moon et al. 2018). Initially, polysomes on the ER membrane synthesize proteins that enter the ER lumen in an unfolded state. Once inside the ER lumen, these polypeptides encounter an oxidizing and calcium-rich microenvironment, wherein N-linked glycosylation modification occurs and disulfide bonds form as proteins mature (Ma and Hendershot 2004). Properly folded polypeptides are subsequently shuttled to their respective destinations (Denic et al. 2006; Liu and Ye 2011; Walter and Ron 2011). If the protein fails to fold properly, it will bind to mannose-specific lectins that trim the terminal mannose residues from the core glycans, thereby marking this misfolded/unfolded protein for ERAD (Figure 2) (Xu and Ng 2015). Examples of these lectins include ER degradation-enhancing α-mannosidase-like lectin 2 (EDEM2) and ER mannosidase-I (ERmanI), which trim the first mannose, followed by further trimming of the glycan by EDEM1/EDEM3 (Hirao et al. 2006; Hosokawa et al. 2010; Ninagawa et al. 2014). The ER-resident lectin, Yeast OS-9 homolog (Yos9), or human osteosarcoma amplified 9 (OS-9) and XTP3-transactivated gene B (XTP3-B)/Erlectin, will then target misfolded/unfolded proteins for retrotranslocation (Szathmary et al. 2005; Christianson et al. 2011). Chaperone proteins, including binding immunoglobulin protein (BiP) and other members of the heat shock protein 70 (HSP70) family, can bind to hydrophobic patches of an unfolded substrate (Hammond and Helenius 1994). This prevents the aggregation of both glycosylated and nonglycosylated substrates (Hammond and Helenius 1994).

Figure 2.

Glycosylation-dependent recognition and targeting of ERAD substrates. Upon entry into the ER, the misfolded/unfolded proteins are often modified by core glycans, GlcNAc₂Man₉Glc₃, which is composed of three glucose (Glc₃), nine mannose (Man₉), and two N-Acetylglucosamine (GlcNAc₂). The addition of the glycans occurs at the Asn residue within the Asn-X-Ser/Thr motif. GI and GII remove two glucoses from the glycans. The resulting monoglucosylated proteins are then recognized by the ER chaperones, calnexin or CRT, which facilitate the folding of the immature proteins. Meanwhile, GI will remove the final glucose from the glycans. If proper folding occurs, the protein will then be modified by ERmanI that will remove the outer mannose. The mature protein achieving its native conformation will exit the ER. However, if the protein is improperly folded, UGGT will recognize and glucosylate this protein, allowing its reentry into the folding cycle again. If this protein fails to achieve its native conformation, it will undergo mannose trimming by EDEM and ERmanI, and subsequently targeted by OS-9 or XTP3-B for ERAD. GI: Glucosidase-I; GII: Glucosidase-I; CRT: calrecticulin; ERmanI: ER mannosidase-I; OST: oligosaccharyl transferase; UGGT: UDP-glucose: glycoprotein glucosyltransferase; and EDEM: ER degradation-enhancing α-mannosidase-like lectins.

In contrast to glycosylated proteins, the process of nonglycosylated substrate recognition and targeting is less clear. Nonglycosylated substrates are primarily recognized and recruited by BiP (Plemper et al. 1997; Skowronek et al. 1998). Additionally, mammalian non-glycosylated proteins can be recognized by EDEM2/OS-9 or EDEM1 (Cormier et al. 2009; Tang HY et al. 2014). Homocysteine-induced ER protein (HERP), a transmembrane protein, directs nonglycosylated substrates to the proteasome for degradation (Okuda-Shimizu and Hendershot 2007). In yeast, nonglycosylated substrates are often recognized by U1Snp1-asscociating 1 (Horn et al. 2009; Kanehara et al. 2010); however, some of them can also be directly recognized by Yos9 for degradation (Jaenicke et al. 2011).

ERAD utilizes the same general machinery to degrade misfolded/unfolded and properly folded proteins; however, the process of substrate recognition and targeting differs (Ruggiano et al. 2014). Each properly folded substrate tends to have its own specific adaptor although most of these adaptors are yet to be identified (Ruggiano et al. 2014). One interesting example is the identification of Derlin-related iRhom proteins as adaptors for the degradation of epidermal growth factor receptor (EGFR) ligands in regulation of sleep (Zettl et al. 2011). Additionally, cellular signaling also controls the recognition of certain enzymes, such as those in lipid metabolism and sterol biosynthesis (Ruggiano et al. 2014). One example is the degradation of 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMGR), a rate-limiting enzyme in cholesterol synthesis. Under low levels of sterol, HMGR does not bind to its adaptor, ER membrane protein insulin-induced gene 1(INSIG-1) or 2, and thus remains stable, serving as a precursor for sterol synthesis. Upon active sterol synthesis, the concentration of intermediate 24,25-dihydrolanosterol increases, enabling the binding of HMGR to INSIG-1 or -2 and the delivery of HMGR to its E3 ligase complex for ubiquitination and degradation (Song et al. 2005).

ii. Substrate retrotranslocation and ubiquitination

After recognition, substrates will be retrotranslocated (or dislocated) to the cytosol depending on their specific location in the ER, a step tightly coupled with substrate ubiquitination. If the substrate is in the lumen, the complete polypeptide will be retrotranslocated. For substrates located in the membrane, only certain domains will be retrotranslocated. Therefore, ubiquitination of the luminal substrates often occurs at later stages of retrotranslocation, whereas ubiquitination of most membrane substrates takes place simultaneously with retrotranslocation. Specific retrotranslocation channel proteins or certain E3 ligases can mediate substrate retrotranslocation. In yeast, Sec61 is a candidate channel protein for retrotranslocation, working within a multiprotein ensemble (Lilley and Ploegh 2004; Ravid et al. 2006; Kohlmann et al. 2008). The Derlin family of proteins serve as alternative candidate components for retrotranslocation (Jarosch et al. 2002; Richly et al. 2005), and there are three in mammalians, Derlin-1, 2, and 3 (Wahlman et al. 2007; Mehnert et al. 2014). Suppressor enhancer Lin12 1 like (SEL1L), an ortholog of yeast HMG-CoA reductase degradation 3 (Hrd3p), is essential for retrotranslocation of many ERAD substrates in human cells (Mueller et al. 2006). In chickens, the SEL1L-dependent substrate, ERAD-L, requires Derlin 2/3 and Herp 1/2 for its degradation (Sugimoto et al. 2017).

The P97 valosin-containing protein (VCP) is a type II AAA+protein family of ATPases, critical for substrate retrotranslocation from yeast to mammals (DeLaBarre and Brunger 2003; Huyton et al. 2003; Meyer 2012; Meyer et al. 2012). The N-terminal domain of P97 binds to substrates, while the C-terminal domain interacts with its adaptor proteins, mediated through the P97/VCP-interacting motif, P97/VCP-binding region, or SHP box (Meyer et al. 2012). Among P97 cofactors, the ubiquitin fusion degradation protein 1 (UFD1) and nuclear protein localization protein 4 (NPL4) form one of the major retrotranslocation complexes with P97 (i.e., the P97-UFD1-NPL4 complex) (Ye Y et al. 2001; Rabinovich et al. 2002). Because P97 cofactors also possess ubiquitin-binding domains, they can directly interact with ubiquitinated substrates. The P97 complexes not only retrotranslocate substrates through the ER membrane, but also provide a platform for various ERAD components, such as the E3 ligases, to interact with and subsequently ubiquitinate substrates (Ye et al. 2004; Liang et al. 2006; Greenblatt et al. 2011; Suzuki et al. 2012; Christianson and Ye 2014).

Ubiquitination is carried out through a series of sequential enzymatic reactions, facilitated by the ubiquitin-activating enzyme (E1), then the ubiquitin-conjugating enzyme (E2), and finally the ubiquitin ligase (E3). Ubiquitin is first covalently conjugated to a reactive cysteine residue of E1 and then transferred to a catalytic cysteine of E2. Finally, E3s facilitate the transfer of ubiquitin from E2 to a lysine residue or another ubiquitin moiety on ERAD substrates (Sommer and Jentsch 1993; Kikkert et al. 2004; Hassink et al. 2005). This reaction can be repeated several times to form polyubiquitin chains. Occasionally, an E4 ligase, a specialized ubiquitin ligase such as UFD2, works after E3s to elongate ubiquitin chains of oligoubiquitinated proteins to generate polyubiquitinated conjugates (Hoppe 2005). This suggests that the polyubiquitin appendage must reach a crucial length before a substrate can be retrotranslocated (Mayer et al. 1998).

A number of E3s involved in ERAD have been identified. Transmembrane E3s include HRD1, glycoprotein 78 (GP78), membrane-associated really interesting new gene (RING) finger protein 6, and RING finger protein 5 (Fang et al. 2001; Nadav et al. 2003; Kikkert et al. 2004; Hassink et al. 2005; Younger et al. 2006; Christianson and Ye 2014; Jeon et al. 2015; Tomati et al. 2015). Their functions are described in later sections. Cytosolic E3s include PARKIN, the C-terminus of HSP70-interacting protein (CHIP), ubiquitin protein ligase E3 component N-recognin 1/2 (UBR1/2), SMAD ubiquitination regulatory factor 1, E6-associated protein (E6-AP), and neuregulin receptor degradation pathway protein 1/fetal lever ring finger (Cummings et al. 1999; Meacham et al. 2001; Murata et al. 2001; Qiu and Goldberg 2002; Tsai et al. 2003; Nillegoda et al. 2010; Guo et al. 2011). Among them, CHIP, UBR1/2, PARKIN, and E6-AP can specifically recognize chaperone-bound misfolded proteins for proteasome-mediated degradation (Cummings et al. 1999; Murata et al. 2001; Tsai et al. 2003; Nillegoda et al. 2010). The SCF complexes, another major family of cytosolic E3s, are composed of Skp1, Cullin1, and specific F-box protein (Fbx, e.g., Fbx2, Fbx6, and β-transducin repeat containing proteins 1 and 2). This class of E3s can recognize and target certain glycoprotein substrates for degradation (Yoshida et al. 2003; 2005; Gong B et al. 2010; Magadan et al. 2010).

In yeast, E3s can function in cytosolic (ERAD-C), luminal (ERAD-L), and membrane (ERAD-M) pathways based on the location of their substrates (Deshaies et al. 1991; Wiertz et al. 1996; Vashist and Ng 2004). For instance, the degradation of alpha 10 (Doa10) is an E3 that targets substrates with lesions in the cytosolic domain, whereas the E3 Hrd1 can target the substrates with lesions in the transmembrane or luminal domain for degradation (Vashist and Ng 2004; Carvalho et al. 2006). Thus far, the three ERAD pathways mentioned above have only been identified in yeast. Mammals possess a more elaborate repertoire of ERAD components. Another unique property of yeast ubiquitin ligases is that they can form multispanning membrane protein complexes that perform a variety of functions including substrate recognition, targeting, and retrotranslocation. For instance, Hrd1 possesses retrotranslocation function (Schoebel et al. 2017). Autoubiquitination of Hrd1 on lysine residues of its RING-finger domain allows misfolded substrates to move across the membrane without the requirement of P97 ATPase (Baldridge and Rapoport 2016). Allowing a single E3 enzyme to harbor both ubiquitination and retrotranslocation activities may provide the most efficient way to target ERAD substrates for degradation in lower species (Bays et al. 2001). HRD1 and GP78 in mammals are homologous to the E3 Hrd1 in yeast. HRD1 ubiquitinates glycosylated substrates in a SEL1L-dependent manner, which can in turn regulate the stabilization of HRD1 (Gardner RG et al. 2000; Carvalho et al. 2006; Gauss et al. 2006; Christianson et al. 2008; Sun S et al. 2014). GP78 functions downstream of HRD1 through association with the human BCL2-associated athanogene 6 (BAG6) chaperone complex to prevent substrate aggregation, thus ensuring the efficient retrotranslocation of substrates for degradation (Zhang, Xu, et al. 2015).

iii. Substrate degradation

Once polyubiquitinated, substrates are extracted from the membrane prior to or during proteasome-mediated degradation. Interestingly, a substantial percentage of proteasome complexes reside at the surface of the ER membrane, ideally positioned to retrotranslocate, receive, and degrade substrates (Rivett 1993; Verma et al. 2000; Hollien and Weissman 2006). This begs the question: how do proteasome-ready substrates exist in ERAD? Quite often, P97 serves as a bridge to link the retrotranslocated substrates with cofactors in cytoplasm. For instance, the cytosolic N-glycanase 1 directly binds to P97 and cleaves N-linked glycans from retrotranslocated substrates, presenting these substrates to the proteasome for degradation (Kim et al. 2006; Bettigole and Glimcher 2015). P97 alone can recognize nonubiquitinated substrates, whereas polyubiquitinated substrates require both P97 and its cofactor UFD1-NPL4 for cytosolic dislocation, a process to export substrates from the ER to the cytosol (Ye et al. 2003). Deubiquitylation occurs during retrotranslocation or immediately prior to degradation, and is mediated by deubiquinating enzymes (DUBs) (Martinon 2012). Deubiquitination may be essential for substrates to enter the proteasome or escape from degradation. Example of DUBs, which associate with P97 directly or indirectly, include the ovarian tumor family of deubiquitinating enzyme 2, valosin-containing protein P97/P47 complex-interacting protein P135, ubiquitin-specific protease 13, and Ataxin-3 (Uchiyama et al. 2002; Wang et al. 2006; Ernst et al. 2009; Sowa et al. 2009).

In order to be degraded by the proteasome, misfolded substrates with hydrophobic motifs or transmembrane domains must remain soluble. The BAG6/ubiquitin-like protein 4 A (UBL4A)/transmembrane domain recognition complex 35 (TRC35) complex aids in this process. BAG6 maintains polypeptide solubility through its chaperone-like activity and subsequently transfers the polypeptides to the proteasome (Wang et al. 2011). The interaction of the BAG6-UBL4A-TRC35 complex with the co-chaperone small glutamine-rich TPR-containing protein further prevents the aggregation of ERAD substrates (Xu et al. 2012). Only properly folded proteins are transported to their destination through the secretory pathway, raising the question: how are they transported from the ER? It is likely that competition exists between protein transportation and ERAD selection machinery, suggesting that the decision to transport or degrade substrates may depend on cell type, stress, secretory-protein load, and/or signaling pathways.

ERAD impinges on immune cells

The adaptive and innate immune responses of our body must be tightly regulated, which is primarily carried out by different immune cells: T cells, B cells, natural killer (NK) cells, dendritic cells (DCs), macrophages, and neutrophils. ERAD plays a crucial role in regulating immune responses to avoid self-reactive (i.e., autoimmunity) or insufficient immune responses (i.e., immunodeficiency or immunosuppression) (Kim et al. 2011; Gong et al. 2014; Yang et al. 2014). ERAD regulates immune cell development by controlling the quantity of both T-cell receptor (TCR) and pre-B-cell receptor (pre-BCR) complexes (Ho et al. 2000; Ji et al. 2016). Additionally, ERAD is important for proper antigen assembly and cytokine secretion (Xu et al. 2016).

T cells

T cells are a type of lymphocytes that function in adaptive immunity and play a key role in the defense against pathogens and tumors. In mammals, the bone marrow is the primary site of hematopoiesis where both T and B lymphoid precursors originate, whereas the thymus is where early thymic progenitor cells seed and develop. A stepwise selection proceeds inside the thymus, starting as double negative (DN, CD4⁻CD8⁻) cells, then double positive (DP, CD4⁺CD8⁺) cells, and finally single positive (SP, CD4⁺ or CD8⁺) cells. During this long journey, the unique composition of the TCR complex identifies each T-cell clone and marks the self-responsive population for elimination. The TCR complex consists of heterodimer of TCRα/β together with homodimers or heterodimers of CD3γ, δ, ε, and ζ chains (Chan et al. 1994). The expression of the TCRβ chain on the DN thymocytes dictates their development into αβ T cells (Starr et al. 2003; Takahama 2006). The DN thymocytes first express the pre-TCR complex, which contains pre-α rather than mature α chains (Germain 2002). This complex is subsequently replaced by the mature TCR complex to facilitate positive and negative selection of thymocytes. Specifically, thymocytes expressing TCR complexes with the ability to bind to the major histocompatibility complexes (MHC) will survive (i.e., positive selection); conversely, thymocytes expressing TCR complexes that have strong affinity to self-antigens will be eliminated (i.e., negative selection). Mature SP T cells that survive the selection will exit the thymus and participate in the adaptive immune response (Germain 2002; von Boehmer and Melchers 2010).

Since the proper expression and assembly of the TCR complex on T cells is a prerequisite for their development in the thymus, the quality and quantity of TCR subunits must be properly controlled. TCR subunits that fail to assemble into the heteromeric complex will be degraded through the ERAD pathway (Bonifacino et al. 1989; Yang et al. 1998). Among these subunits, TCRα and CD3δ are particularly susceptible to ERAD. Recent studies have identified the specific components required for ERAD of TCR subunits. For instance, the mannose-trimming of TCRα is mediated by the ER lectin EDEM3 together with an ER-resident oxidoreductase ER-resident protein 46 (Yu S et al. 2018). Moreover, the ubiquitination of TCRα is essential for its retrotranslocation, which is a multi-step process regulated by P97 (Huppa and Ploegh 1997; Yu H and Kopito 1999; Moore et al. 2013). Interestingly, impaired UFD1 or NPL4 function does not impact TCRα levels, suggesting the P97-UFD1-NPL4 complex does not regulate its retrotranslocation (Nowis et al. 2006). The ER-localized E3 HRD1 mediates the ubiquitination of TCRα (Ishikura et al. 2010), whereas ubiquitin specific peptidase 19, a human DUB, can reverse this process (Hassink et al. 2009). Recently, WD repeat-containing protein 20 (WDR20), a novel regulator of ERAD, was found to facilitate TCRα ubiquitination by mediating the interaction of P97 and TCRα (Ju et al. 2018). Although TCRα and CD3δ are both TCR subunits, their ERAD components differ. The E3 GP78 facilitates the degradation of CD3δ through its physical interaction with P97 (Zhong et al. 2004); however, deletion of P97 only impacts the degradation of TCRα but not CD3δ (Wojcik et al. 2006; Lass et al. 2007). Instead, the degradation of CD3δ was mediated by the E3 Ring finger protein 2 (Lerner et al. 2007).

Despite the above knowledge, how ERAD specifically impacts T-cell development remains elusive. Using T-cell-specific Hrd1 knockout mice, a recent study showed that HRD1 depletion markedly reduced all T-cell populations, including CD4⁺ and CD8⁺ SP T cells (Xu et al. 2016). This observation is intriguing and suggests the contribution of the E3 HRD1 to the development of CD4⁺ and CD8⁺ T cells; however, the specific mechanism remains unclear. Additionally, the absolute number of T regulatory cells (Tregs) also decreased, likely due to fewer T cells in the spleen of Hrd1 knockout mice (Xu et al. 2016). Interestingly, HRD1 was shown to maintain the stability of Tregs through suppressing ER stress responses (Xu et al. 2019). Moreover, loss of HRD1 leads to a blockade in T-cell proliferation and differentiation, as well as interleukin-2 (IL-2) production (Xu et al. 2016). Despite these advances, the contribution of other ERAD components to T-cell immunity has yet to be elucidated. The use of animal models in which specific ERAD components are depleted could be particularly useful in defining their contribution to T-cell development and function.

B cells

In contrast to T cells, B cells develop in the bone marrow and primarily function in humoral adaptive immunity (LeBien and Tedder 2008). B cell precursors in the bone marrow undergo a multi-step developmental process, starting as B cell progenitors, then precursors, and finally immature B cells (Cambier et al. 2007). During early B-cell development, the SEL1L-HRD1 complex mediates ERAD of the pre-B-cell receptor (pre-BCR) complex, which is critical for the transition of large to small pre-B cells (Ji et al. 2016). Upon stimulation of thymus-independent antigens, such as lipopolysaccharides (LPS), B cells proliferate and differentiate into plasma cells that primarily secrete immunoglobulin M (IgM) antibodies (Shih et al. 2002; Nutt et al. 2015). To cope with this output, B cells must drastically improve their protein quality control system and secretory machinery (Gass et al. 2002); therefore, they enhance ERAD functions to reduce the accumulation of misfolded/unfolded Ig in the ER that results from escalating Ig synthesis. One such example is the upregulation of SEL1L and HRD1 in LPS-stimulated B cells, consistent with the abundant expression and secretion of IgM (Cattaneo et al. 2008).

IgM consists of two forms: a membrane-bound monomer and a secreted pentamer. The pentameric IgM consists of five monomers held together by J chains and disulfide bonds (Kubagawa et al. 2019). Similar to TCRα and CD3δ, unassembled IgM heavy chain (Igμ) is also degraded by ERAD (Ho et al. 2000). Igμ and J chains trimmed by ERManI are targeted to the cytosol for ERAD, as the use of Kifunensine, an inhibitor of ERManI, prevents degradation of Igμ and dislocation and degradation of the J chain (Fagioli and Sitia 2001). Other ERAD components, such as SEL1 and HRD1, are also involved in degradation of unassembled Igμ (Cattaneo et al. 2008). Additionally, unfolded/unassembled IgG is degraded by a Herp-dependent ERAD pathway (Lee J et al. 2012). In addition to impacting B-cell immunity through controlling the quality of Ig, HRD1 can also regulate other critical factors to affect B-cell activation. This is supported by a study that shows the death receptor FAS being degraded by HRD1 to sustain B-cell survival (Kong et al. 2016). The above evidence demonstrates the critical contribution of ERAD to B-cell development and activation. Further understanding of how ERAD controls plasma cell survival could guide the treatment of multiple myeloma, a disease with abnormal expansion of plasma cells.

Other immune cells

Recent studies also reveal the regulatory role of ERAD on other immune cells, such as DCs, macrophages, and NK cells. DCs are professional antigen-presenting cells (APCs), which play a critical role in initiating adaptive immune responses, as well as maintaining the tolerance of self-antigens (Heath and Carbone 2001; Mellman 2013). DCs can engulf pathogen-infected or malignantly transformed cells, a process that initiates exogenous antigen presentation through MHC class I molecules (MHC-I) to CD8⁺ T cells or MHC class II molecules (MHC-II) to CD4⁺ T cells (Joffre et al. 2012). B lymphocyte-induced maturation protein 1 (BLIMP-1), a transcriptional repressor, can suppress systemic lupus erythematosus in mice by maintaining tolerogenic function of DCs (Kim et al. 2011). Using mice with a DC-specific deletion of Blimp-1, a study showed that BLIMP-1 can transcriptionally repress the expression of MHC-II on the surface of DCs (Kim et al. 2013). Moreover, enhanced expression of HRD1 through Toll-like receptor stimulation promotes MHC-II expression by ubiquitinating and degrading BLIMP-1 (Yang et al. 2014). This regulatory role of HRD1 only impinges on MHC-II but not on MHC-I, impairing the priming of CD4⁺ T cells while sparing CD8⁺ T cells (Yang et al. 2014). Loss of HRD1 function in DCs protects mice from experimental autoimmune encephalomyelitis, a direct evidence demonstrating a physiological role of HRD1 in immunity (Yang et al. 2014). Additionally, ERAD has a vital role in antigen cross-presentation, which will be discussed below in the antigen presentation section.

Macrophages are derived from monocytes and exert the first line immune responses against infections (Labonte et al. 2014). Macrophages can differentiate into two subtypes: M1 that possesses antitumor activity and M2 that is detrimental to the host, contributing to tumor growth, angiogenesis, and metastasis (Dumont et al. 2008; Niino et al. 2010; Alfano et al. 2013; Labonte et al. 2014). A recent in vitro study showed that overexpression of HERP, a shuttling factor that delivers ubiquitinated substrates to the proteasome for degradation, promotes IL-4-induced polarization and migration of M2 macrophages (Kim et al. 2008; Li Y et al. 2018). Lipoprotein lipase (LPL) is a protein guarding the entry of fatty acids into tissues (Williams 2008). SEL1L mediates LPL secretion and systemic lipid metabolism in adipocytes, myocytes, and macrophages by preventing aggregation of LPL and facilitating its ER exit (Sha et al. 2014). Despite these studies, further studies are needed to understand how the ERAD pathway regulates macrophage differentiation and functions.

NK cells play a vital role in immunosurveillance by directly inducing cytotoxic responses to tumor cells (Degli-Esposti and Smyth 2005). The recognition of human tumor cells by NK cells is mediated by the binding of NK-cell activated receptor CD226 to its ligand CD155 (Carlsten et al. 2009). Although there is no direct evidence showing the regulation of ERAD in NK cells, activation of UPR and ERAD in hepatocellular carcinoma cell (HCC) can suppress NK-mediated cytotoxicity. Activation of two branches of the UPR pathways, transcription factor 6 (ATF6) and inositol-requiring enzyme 1α (IRE1α), can enhance HRD1 expression in HCC. Increased levels of HRD1 promote degradation of CD155 in HCC, which further perturbs the cytotoxic activity of NK cells (Gong et al. 2014). Since these experiments are only performed in vitro, additional in vivo work is required to establish the role of HRD1 in the development and activation of NK cells.

ERAD in antigen presentation

Antigen presentation is a prerequisite for initiating T-cell mediated antitumor responses. Dysfunctions in antigen processing and presentation are common causes of immune evasion in cancer, creating barriers to T-cell based immunotherapies (de Charette et al. 2016). To initiate antitumor immune responses, peptides derived from tumor antigens must be presented on the surface of APCs by MHCs to effector T cells (Burgdorf et al. 2008). MHC-I are expressed by all nucleated cells, including tumor cells and APCs, primarily presenting antigens to CD8⁺ T cells (Vyas et al. 2008). MHC-II are predominantly expressed by professional APCs, such as DCs, B cells, and macrophages, mainly presenting antigens to CD4⁺ T cells (Jensen 2007). Endogenous antigen presentation, which occurs in all nucleated cells, is the presentation of self-antigenic peptides to the cell surface through MHC-I (Neefjes et al. 2011). Endogenous antigen presentation can also occur through MHC-II in certain cell types, such as DCs, tumor cells, and thymic epithelial cells (Roche and Furuta 2015). In a similar manner, APCs present extracellular antigens through MHC-II to CD4⁺ T cells, a process referred to as exogenous antigen presentation (Watts 2004). Cross-presentation is the process in which APCs present exogenous antigens to naïve CD8⁺ T cells through MHC-I, thus activating them to cytotoxic CD8⁺ T cells (Kurts et al. 2010). Despite little known about how ERAD regulates antigen presentation, accumulating evidence demonstrates that ERAD regulates degradation of MHC-I and translocation of exogenous antigens from phagosomes to the cytosol, a critical step for cross-presentation (Grotzke and Cresswell 2015).

i. ERAD in endogenous antigen presentation

Cancer cells present tumor antigens through endogenous antigen presentation. One of the key players in this process is MHC-I, which is composed of a heavy chain and an invariant light chain known as β2-microglobulin (β2 m) (Vyas et al. 2008). The MHC-I is assembled in the ER and loaded with peptides derived from tumor antigens through the following sequential steps (Figure 3(A)). Tumor antigens, including proteins, peptides, and polysaccharides, are degraded by the proteasome into small peptides, which are subsequently translocated into the ER by the transporter associated with antigen presentation (TAP) (Howard 1995). These peptides are typically composed of eight to sixteen amino acids, which exceeds the peptide length required for MHC-I association. Hence, aminopeptidase, another ER luminal component, will trim these peptides into an approximate length of eight amino acids for MHC-I loading (York et al. 2002; Cruz et al. 2017). In mice, there is only one such enzyme called ER aminopeptidase that is associated with antigen processing, whereas in humans, two such enzymes have been identified: ER aminopeptidase-1 (ERAP1) and 2 (ERAP2) (Saric et al. 2002; Serwold et al. 2002; Saveanu et al. 2005). Before MHC-I can load the peptides, they must be stabilized by ER chaperone proteins, such as calreticulin (CRT), ER-resident protein 57 (ERp57), protein disulfide isomerase, and tapasin. Once stabilized, TAP and tapasin help translocate the peptides into the ER and deliver them to MHC-I (Sadasivan et al. 1996). The fully assembled peptide-MHC-I complexes will then be transported via the Golgi from the ER to the cell surface (Vyas et al. 2008).

Figure 3.

Endogenous and exogenous antigen presentation by MHC-I. (A) MHC-I presentation of endogenous antigenic peptides in cancer cells begins with degradation of the antigen by the proteasome. The degraded peptides are then translocated into the ER lumen through TAP. In the lumen, the peptides are trimmed by ERAP1/2 and loaded onto MHC-I. The fully assembled peptide-MHC-I complex is transported via the Golgi to the cell surface for presentation to CD8⁺ T cells. When β2m is absent or downregulated, the MHC-I heavy chain will be removed from the ER through ERAD. (B) Exogenous antigens can be presented to T cells by antigen presenting cells through two major pathways. The first pathway, the vacuolar pathway, begins with the exogenous antigen internalized into a phagosome, where it is degraded by cathepsin S and then loaded onto MHC-I for transportation to the cell surface. The second pathway, the endosome pathway, begins with the antigen exiting the phagosome through SEC61 and P97 to the cytosol for proteasome-mediated degradation. From this point, the peptides will be transported to the cell surface following a path similar to (A). Alternatively, the peptides will be transported back to the phagosome by TAP and loaded onto MHC-I for transportation to the cell surface. TAP: transporter associated with antigenic presentation; ERAP: ER aminopeptidase associated with antigen processing.

Cancer cells often escape immunosurveillance due to defective antigen presentation. Newly synthesized MHC-I heavy chain must bind to β2 m to incorporate into the peptide-loading complex (PLC), comprised of TAP, tapasin, CRT, and ERp57 (Chapman and Williams 2010). Upon increased ER stress, cancer cells often downregulate MHC-I, TAP, tapasin, CRT, and/or ERp57, impairing the formation of PLC and endogenous antigen presentation (Leone et al. 2013). Reduced presentation of the peptide-MHC-I complexes is observed in approximately 20–60% of solid tumors, such as melanoma, lung, and breast cancer (Campoli and Ferrone 2008). When β2 m is absent or downregulated, the MHC-I heavy chain cannot fold properly and will be removed from the ER through ERAD (Hughes et al. 1997). Misfolded MHC-I heavy chain is recognized in the ER lumen by EDEM1, OS-9, and XTP3-B in a glycan-dependent manner (Burr et al. 2013), and then targeted by HRD1/SEL11L/UBE2J1(ubiquitin conjugating enzyme E2 J1) for ERAD (Burr et al. 2011). The loss of UBE2J1 impairs the ERAD of misfolded MHC-I heavy chain (Burr et al. 2011). The extraction of the misfolded heavy chain from the ER membrane requires the UFD1-NPL4-P97 complex (Burr et al. 2013). Disruption of the complex through UFD1 deletion restores MHC-I levels in the β2 m-depleted cells (Burr et al. 2013); however, it is unknown whether the dysfunction of the UFD1-NPL4-P97 complex can increase the presentation of tumor antigens. Antigenic peptides that fail to bind to MHC-I will be transported from the ER lumen into the cytosol for degradation. One way to degrade these peptides is through cytosolic peptidase; however, this enzyme can only partially degrade the peptides (Roelse et al. 1994). TAP can then translocate these peptides into the ER, allowing them to re-associate with MHC-I (Roelse et al. 1994). ERAD serves as an additional way to eliminate antigen peptides that fail to associate with MHC-I (Koopmann et al. 2000). Hence, enhanced ERAD function in tumor cells may promote the degradation of MHC-I and antigen peptides, thereby preventing the reassociation of peptides with MHC-I, impairing antigen presentation and promoting immune evasion.

In addition to MHC-I, endogenous antigens can also be presented by MHC-II through autophagy-mediated antigen processing (Schmid and Munz 2007). This allows the recognition of intracellular antigens by CD4⁺ T cells, a process particularly important for positive and negative selection of thymocytes (Kasai et al. 2009). MHC-II is expressed by APCs, as well as by a variety of cancer cells, including melanoma, glioma, breast, colorectal, ovarian, prostate, and nonsmall cell lung cancer (NSCLC) cells (Axelrod et al. 2019). Interestingly, tumor-associated MHC-II expression predicts a favorable outcome in patients (Axelrod et al. 2019). Unlike MHC-I, MHC-II is composed of an α and a β chain, which are then associated with an invariant chain (Ii) in the ER. Binding of Ii to MHC-II prevents the loading of antigen peptides onto MHC-II through the blockade of the antigen-binding groove (Busch et al. 2000; Bryant and Ploegh 2004). The Ii-MHC-II complex is then transported to the late endosome, where part of Ii is cleaved, leaving the MHC-II-associated Ii peptide (CLIP) in the peptide binding groove. The late endosome is fused with an autophagosome that contains cytosolic antigens (van den Boorn et al. 2011). The nonclassical MHC-II molecule, human leukocyte antigen-DM, interacts with MHC-II and catalyzes the exchange of CLIP with the antigenic peptides. MHC-II loaded with antigenic peptide is then transported to the cell surface through endolysosomal tubules to present the antigen to CD4⁺ T cells (Vyas et al. 2008).

ii. Cross-presentation

In contrast to MHC II-restricted exogenous presentation, the process of cross-presentation is more complicated, and is primarily referred to as the presentation of exogenous antigens by APCs through MHC-I to CD8⁺ T cells (Figure 3(B)) (Vyas et al. 2008). Although other APCs, such as macrophages and B cells, can cross-present exogenous antigens (Ramirez and Sigal 2002; Marino et al. 2012), DCs are considered to be the most prominent cross-presenting APCs. DCs uptake exogenous antigens, including tumor cell debris, through endocytic pathways such as phagocytosis, macro-pinocytosis, and endocytosis (Guermonprez et al. 2002; Kamphorst et al. 2010). These extracellular antigens are then presented on the surface of APCs through two major pathways: TAP-dependent (endosome-to-cytosol) and independent (vacuolar) (Figure 3(B)). The endosome-to-cytosol pathway is unique to DCs, enabling the antigens to be internalized in the endosome and transported into the cytosol for proteasome-mediated degradation (Kovacsovics-Bankowski and Rock 1995; Rodriguez et al. 1999; Ackerman et al. 2003). The vacuolar pathway is utilized by APCs to cross-present virus and tumor antigens (Liu et al. 1997; Campbell et al. 2000; Ma et al. 2016). Exosomes secreted by tumor cells contain tumor antigens, which can be presented by DCs through their uptake of exosomes (Wolfers et al. 2001). Phagocytosis also results in efficient cross-presentation of tumor antigen (Berard et al. 2000; Nouri-Shirazi et al. 2000; Russo et al. 2000).

A key step in the endosome to cytosol pathway is the transfer of antigens from phagosomes to the cytosol (Ackerman et al. 2003; Embgenbroich and Burgdorf 2018). It is well accepted that ERAD components participate in the retrotranslocation of exogenous antigens, as demonstrated by the detection of ERAD components on the phagosomal membrane (Houde et al. 2003). One ERAD component involved in antigen presentation is the E3 CHIP. Downregulation of CHIP results in decreased cross-presentation of ovalbumin (OVA) and the accumulation of OVA in DCs (Imai et al. 2005). P97 is also involved in cross-presentation (Imai et al. 2005; Ackerman et al. 2006; Zehner et al. 2011; Menager et al. 2014). For instance, P97 can interact with exogenous OVA, and the knockdown of P97 reduces the degradation and cross-presentation of OVA (Imai et al. 2005). Mutant P97 that lacks the ATPase can inhibit cross-presentation, indicating that the ATPase activity of P97 is required for retrotranslocation of exogenous antigens (Ackerman et al. 2006). The recruitment of P97 toward the endosomal membrane is regulated by the polyubiquitination of the mannose receptor, C-type lectin (Zehner et al. 2011). Interestingly, P97, but not SEC61 or Derlin-1, can translocate the synthetic tumor antigen Melan-A peptides from the endosome into the cytosol (Menager et al. 2014). Although SEC61’s role in antigen presentation is controversial, several studies indicate that SEC61 participates in cross-presentation. Genetic inactivation of Sec61 decreases the degradation of exogenous OVA and its cross-presentation (Imai et al. 2005). Pharmacological inhibition of SEC61 by exotoxin A (ExoA) also impairs antigen cross-presentation through impairment of SEC61-dependent protein transport across the ER and phagosome membranes (Ackerman et al. 2006). An intracellular antibody, which can trap SEC61 inside the ER preventing its recruitment to the endosome, inhibits antigen translocation and cross-presentation (Zehner et al. 2015). Interestingly, a recent study shows that SEC61 blockade inhibits cross-presentation independently of the endosome to cytosol pathway (Grotzke et al. 2017). Together, these studies demonstrate an important role of ERAD components in antigen presentation, supporting its involvement in tumor immunity.

Paradoxical role of ERAD in tumor immunity

Tumor cells experience increased ER stress often resulting from oncogene activation, uncontrolled proliferation, hypoxia, nutrient deprivation, as well as anticancer therapy (Nakagawa et al. 2014; Cruz et al. 2017). Diverse strategies, such as the UPR and ERAD, are exploited by tumor cells to mitigate ER stress. However, the intrinsic stress response pathways in malignant cells exert paradoxical roles during tumor development. For instance, certain ERAD components can promote tumorigenesis, while others inhibit tumor cell survival and growth (Tax et al. 2019). In the TME, both the UPR and ERAD can impart extrinsic effects to promote immunosurveillance or immune evasion (Cubillos-Ruiz et al. 2017; Cha et al. 2018; Gong et al. 2014). When treated with chemotherapy or radiotherapy, ER stress can induce immunogenic cell death (ICD) of cancer cells, which activates antitumor immune responses (Garg et al. 2017). In DCs, ERAD facilitates the cross-presentation of tumor antigens (Li et al. 2011; Menager et al. 2014; Alloatti et al. 2017). However, increased ER stress in the TME can also dampen antitumor immune responses, leading to immune evasion of cancers.

i. The effect of ER stress on tumor immunity

In the TME, ER stress impacts tumor and immune cells simultaneously, exerting a paradoxical role in tumor immunity. When cells experience ER stress, BiP is dissociated from three stress sensors on the ER membrane: RNA-dependent protein kinase-like kinase (PERK), IRE1α, and ATF6 (Moenner et al. 2007; Yadav et al. 2014). BiP, as well as IRE1α in certain conditions, binds to misfolded/unfolded proteins to activate the UPR, the signaling pathways upstream of ERAD (Dorner et al. 1992; Bertolotti et al. 2000; Gardner BM and Walter 2011). When the UPR is triggered by ER stress, PERK phosphorylates eukaryotic initiation factor 2α (eIF2α), which in turn induces the expression of transcription factor C/EBP-homologous protein (CHOP), a proapoptotic protein (Harding et al. 1999; Ma et al. 2002; Marciniak et al. 2004). In various types of cancer, ER stress resulting from high hydrostatic pressure induces tumor cell death by activating the PERK-eIF2α pathway (McCullough et al. 2001; Moserova et al. 2017), allowing DCs to engulf tumor cell debris to activate antitumor immune responses. Oligomerized IRE1α, which possesses RNase activity, splices XBP1. As a highly active transcription factor, spliced XBP1 can enhance the expression of ER chaperones and ERAD components (Lee AH et al. 2003; Shaffer et al. 2004; Mohamed et al. 2017). In addition, the IRE1α-XBP1 axis has been shown to suppress the development of colorectal cancer (Niederreiter et al. 2013). However, at later stages of tumor development, the UPR transitions from being cytotoxic to cytoprotective, enabling tumor cells to thrive and proliferate (Vanacker et al. 2017; Sundaram et al. 2018).

Besides cell death induced by unresolved ER stress, cancer cells can also release “danger signals” known as damage-associated molecular patterns (DAMPs) to initiate ICD and protective immune responses (Ma et al. 2011; van Vliet et al. 2015). There are primarily three types of DAMPs: surface exposure of CRT, the secretion of ATP, and the release of high mobility group box 1 (HMGB1) (Krysko et al. 2012). ICD induced by DAMPs under ER stress promotes antitumor immunity (Casares et al. 2005; Garg et al. 2012). Exposure of CRT on the cell surface signals DCs to engulf dying tumor cells (Gardai et al. 2005). Secretion of ATP attracts DCs to promote activation of the pyrin domain containing-3 inflammasome and increased secretion of pro-inflammatory cytokines such as IL-1β and IL-18 (Ghiringhelli et al. 2009; Tschopp and Schroder 2010). HMGB1 can bind to Toll-like receptor 4 on DCs and subsequently enhance cross-presentation of tumor antigens to CD8⁺ T cells (Apetoh et al. 2007; Kepp et al. 2013). Since ICD induction elicits long-term antitumor immunity, it has been proposed as a potential novel strategy in the treatment of cancer.

ER stress is the ultimate driving force of ICD induction (Sagar et al. 2019). In refractory endometrial cancer cells, ER stress can increase membrane expression of CRT, the key signal in ICD induction (Xu et al. 2017). Blockade or knockdown of CRT can eliminate antitumor immune responses in mice (Obeid et al. 2007). Furthermore, in NSCLC, CRT exposure is correlated with the recruitment of antitumor immune cells to the TME (Zitvogel et al. 2010; Fucikova et al. 2016). ERp75, an ER chaperone, can promote phagocytosis by DCs through the enhancement of CRT exposure (Obeid 2008). CRT exposure on the cell surface is usually mediated through activation of the PERK-eIF2α pathway. Phosphorylation of eIF2α by PERK correlates with CRT exposure, allowing for the use of phospho-eIF2α as a biomarker of ICD (Bezu et al. 2018). In melanoma cells, however, eIF2α is phosphorylated by protein kinase R and general control nonderepressible 2 to facilitate the translocation of CRT to the surface of tumor cells (Giglio et al. 2018). Although it is clear that the UPR contributes to ICD induction, whether ERAD can stimulate ICD remains unknown.

While ER stress resulting from cancer therapy primarily enhances antitumor immunity through ICD induction, sustained ER stress in cancer cells can also contribute to the establishment of the immunosuppressive network within the TME. Cancer cells experiencing ER stress secrete unknown factors into the TME to activate and promote the UPR in tumor-infiltrating immune cells such as macrophages (Mahadevan et al. 2011). This process is referred to as “transmissible ER stress”, which induces a robust proinflammatory response in tumor cells and pro-tumor microenvironment (Mahadevan et al. 2011). Transmissible ER stress can suppress the antigen-presenting capacity of DCs, impairing the activation of CD8⁺ T cells (Mahadevan et al. 2012).

ER stress activation in immune cells can also inhibit antitumor immunity. Activation of the IRE1α-XBP1 axis in macrophage promotes the secretion of cathepsin to enhance tumor cell invasion (Yan et al. 2016). ER stress in myeloid-derived suppressor cells (MDSCs) increases their immunosuppressive capacity, accelerating cancer progression (Condamine et al. 2014; Lee BR et al. 2014; Condamine et al. 2016; Nan et al. 2018). Induction of ER stress in neutrophils from healthy donors can transform the cells into immunosuppressive MDSCs (Condamine et al. 2016). CHOP deficiency in mouse MDSCs leads to decreased immunosuppressive functions (Thevenot et al. 2014). Additionally, CHOP is upregulated in tumor-infiltrating CD8⁺ T cells to suppress their cytotoxic response (Cao et al. 2019). ER stress can also promote tumor immune escape through disruption of antigen presentation. In advanced ovarian cancer, activation of the IRE1α-XBP1 axis in tumor-associated DCs disrupts ER homeostasis, inhibiting their antigen-presentation capacity and ultimately suppressing T-cell antitumor immunity (Cubillos-Ruiz et al. 2015; Cubillos-Ruiz and Glimcher 2016). Increased ER stress in thymoma cells activates the PERK-eIF2α pathway, leading to the blockade of protein synthesis that impairs tumor antigen presentation mediated by MHC-I (Granados et al. 2009). ER stress also suppresses the expression of components of the MHC-I machinery, including tapasin and TAP, to decrease tumor antigen presentation (Pellicciotta et al. 2008; Bartoszewski et al. 2011).

While it is apparent that ER stress can impede antitumor immunity, it has also been conversely implicated in immunosurveillance. A low protein diet can increase IRE1α signaling and subsequently activate the regulated IRE1α-dependent decay of ER-localized RNAs to enhance the cytotoxic response of CD8⁺ T cells thus suppressing tumor growth (Rubio-Patino et al. 2018). Activation of eIF2α enhances antitumor immunity by producing novel tumor peptides associated with MHC-I (Schwab et al. 2004). In murine melanoma cell lines, accumulation of misfolded/unfolded mutant tyrosinase in the ER promotes the presentation of certain tyrosinase epitopes by MHC-I (Ostankovitch et al. 2005). Furthermore, autophagy induced by ER stress is another means to activate protective antitumor immunity. As demonstrated in an in-vivo murine model, autophagy activity is required to sustain CD8α-type DCs for efficient cross-presentation of tumor antigens (Parekh et al. 2017).

ii. The effect of enhanced ERAD on tumor cells

While ERAD alleviates ER stress through facilitating the degradation of misfolded/unfolded proteins, components of ERAD exert paradoxical roles in tumorigenesis (Table 2). It has been shown that certain components of ERAD, such as ERmanI, Derlin-1, Derlin-3, HRD1, P97, NPL4, and UFD1, exert pro-tumor effects (Table 2) (Tax et al. 2019). Reduced expression of ERmanI can inhibit N-glycosylation and impair survival of breast cancer cells (Legler et al. 2018). Derlin-1 and 3, retrotranslocation channel proteins, promote overall cancer growth of breast origin (Guiliano et al. 2014; Shibata et al. 2017). Additionally, overexpression of Derlin-1 has been found in several human cancers including NSCLC, bladder, and colorectal cancer, predicting poor prognosis (Tan et al. 2015; Dong et al. 2017; Mao et al. 2018). Moreover, upregulation of the E3 ligase HRD1 promotes invasion of colorectal cancer (Tan et al. 2019). Recently, UFD1, a component facilitating retrotranslocation of ERAD substrates, has been found to mediate MYC-driven leukemia aggression (Huiting et al. 2018). Inhibitors of P97, a UFD1 functional partner, can interfere with autophagy and induce cancer cell death (Magnaghi et al. 2013). Disulfiram, a drug that can target NPL4, another UFD1 functional partner, can inhibit cancer cell growth (Skrott et al. 2017). Collectively, the P97-NPL4-UFD1 complex in the ERAD pathway plays an essential role in the survival of cancer cells.

Table 2.

Paradoxical role of ERAD components in tumor development.

ERAD components	Effects on tumor	Tumor types	Experimental approaches	Mechanisms of action	References
Derlin-1	Tumor promoting	Breast, bladder, colorectal cancer, and NSCLC	Knockdown studies by shRNA in human cell culture	A retrotranslocation channel protein upregulated in colorectal cancer cells	Tan et al. (2015), Dong et al. (2017), and Mao et al. (2018)
Derlin-3	Tumor promoting	Breast cancer	SiRNA inhibition in breast cancer cell lines	A retrotranslocation channel protein, whose inhibition results in decreased invasion and proliferation of cancer cells	Guiliano et al. (2014) and Shibata et al. (2017)
ERmanI	Tumor promoting	Breast cancer	Gene expression profiling of patient samples and gene inactivation by shRNA in human cancer cell lines	A mannosidase, its reduced expression inhibits N-glycosylation of substrates, impairing cancer cell survival	Legler et al. (2018)
HRD1	Tumor promoting	Colorectal cancer	Knockdown studies by shRNA in human cancer cell line	An E3, promoting degradation of misfolded proteins overexpressed in colorectal cancer	Tan et al. (2019)
P97	Tumor promoting	Colorectal, cervical cancer, and osteosarcoma	Small molecule and RNAi inhibition in human cancer cell lines	In complex with UFD1-NPL4 to facilitate substrate retrotranslocation in ERAD	Magnaghi et al. (2013)
NPL4	Tumor promoting	Breast cancer, myeloma, and osteosarcoma	Human cancer cell line and a murine model treated with disulfiram	In complex with P97 and UFD1 to facilitate substrate retrotranslocation in ERAD	Skrott et al. (2017)
UFD1	Tumor promoting	T-cell acute lymphoblastic leukemia	Gene inactivation studies in zebrafish models of leukemia and human leukemic cell lines.	In complex with P97 and NPL4 to facilitate substrate retrotranslocation in ERAD and its inactivation resulting in cytotoxic UPR responses	Huiting et al. (2018)
SEC22b	Tumor suppressing	Lymphoma and fibrosarcoma	Conditional mouse knockout specific in DCs	An ER-phagosome traffic mediator, facilitating cross-presentation in DCs	Alloatti et al. (2017)
CHIP	Tumor suppressing	Breast cancer	Gene inactivation by shRNA in mouse xenograft models and human cancer cell lines	An E3 facilitating degradation of misfolded substrates	Kajiro et al. (2009)
OS-9	Tumor suppressing	Pancreatic ductal adenocarcinoma	Measuring mRNA expression levels in patient samples of pancreatic ductal adenocarcinoma	A substrate recognition factor facilitating ERAD	Sun YW et al. (2014)
SEL1L	Tumor promoting or suppressing	Pancreatic and breast cancer	Overexpression and downregulation studies in murine models	A substrate recognition and recruitment factor, facilitating substrate retrotranslocation; overexpression of SEL1L results in reduced pancreatic cancer invasion, while its downregulation predicts poor prognosis of breast cancer	Orlandi et al. (2002), Cattaneo et al. (2005), and Jeon et al. (2015)
GP78	Tumor promoting or suppressing	HCC, colorectal, esophageal, bladder, and breast cancer	Correlation studies in patients, knockout studies in mice, and gene inactivation studies by shRNA in murine models	An E3 participating in substrate degradation; high GP78 expression predicts poor prognosis of patients with esophageal squamous cell carcinoma, colorectal, and bladder cancer; Gp78−/− knockout increases HCC development; Gene inactivation shows tumor suppressive effect of GP78 in breast cancer	Nakamori et al. (1994), Otto et al. (1994), Maruyama et al. (1995), Silletti and Raz (1996), Zhang, Kho, et al. (2015), and Chang et al. (2016)

Conversely, some ERAD components engage in tumor suppression (Table 2). Expression of CHIP, a ubiquitin ligase, is inversely correlated with breast cancer progression (Kajiro et al. 2009). Depletion of CHIP correlates with an increase in growth of subcutaneous tumors in mice, indicating its role as a tumor suppressor (Kajiro et al. 2009). Similarly, enforced expression of OS-9, a substrate recognition factor, can suppress invasion of pancreatic ductal adenocarcinoma (Sun YW et al. 2014). The roles of SEL1L and GP78, two other ERAD components, differ based on the context of cancer (Table 2). Due to the complex roles of ERAD components, it is important to understand their respective function in ERAD and effect on tumor cell survival, which can impact tumor immunity.

iii. The effect of ERAD on tumor immunity

While the role of ERAD in regulating immune cell development and function is evident, its involvement in tumor-infiltrating immune cells is not well-defined. Nevertheless, ERAD participates in cross-presentation, indicating its importance in tumor immunity. In murine models, deficiency in SEC22b, an ERAD component required for effective cross-presentation, leads to impaired antitumor immune responses (Alloatti et al. 2017). Additionally, treatment of DCs with ExoA, an inhibitor of the SEC61 translocon protein, partially suppresses activation of CD8⁺ T cells (Li et al. 2011). Similarly, studies in DCs demonstrate a critical role of ERAD in processing tumor antigens (Xing et al. 2016). An alternative ERAD pathway was found in the cross-presentation of synthetic long peptides (SLP), which have been used as successful antitumor vaccines against multiple types of cancer (Menager et al. 2014). This study demonstrates that cross-presentation is dependent on P97, a critical ERAD component, to translocate the SLP16–40 antigen (Menager et al. 2014). Moreover, the ubiquitin proteasome system depends on P97 to translocate antigens to MHC-I for cross-presentation (Palmer and Dolan 2013). Glucose-regulated protein 170 (GRP170), an ER chaperone, can deliver tumor antigens to APCs for cross-presentation, resulting in antitumor immune responses by cytotoxic T cells (Wang et al. 2010; Wang et al. 2014). For instance, GRP170 in DCs facilitates the processing and presentation of melanoma antigen glycoprotein 100 by enhancing its interaction with ERAD components, such as SEC61, P97, CHIP and GRP78 (Wang et al. 2013; Wang et al., 2014). These findings suggest that upregulation of ERAD in DCs may potentially enhance cross-presentation of tumor antigens to elicit a stronger cytotoxic antitumor response.

To evade immunosurveillance, cancer cells upregulate programed death-ligand 1 (PD-L1), which binds to its receptor PD-1, a critical immune checkpoint protein on activated T cells (Pardoll 2012). Since ERAD can promote degradation of PD-L1, enhanced ERAD in tumor cells may be beneficial for antitumor immune responses (Cha et al. 2018). However, ERAD can be detrimental to antitumor responses in different contexts. In human HCC, IRE1α upregulates HRD1 to degrade the CD155 ligand for CD226, a major NK-cell activating receptor (Gong et al. 2014). Thus, ERAD upregulation in HCC has an adverse effect on immunosurveillance and results in a poor prognosis, demonstrating the paradoxical function of ERAD in tumor immunity (Gong et al. 2014). Hence, more studies on ERAD’s function in the TME and tumor immunity are needed to guide the design and efficacy of current immunotherapies.

Therapeutic influences on ER stress and tumor immunity

The key for successful immunosurveillance and effective cancer therapy is their ability to induce immune responses against tumor cells (Blankenstein et al. 2012). Most chemotherapy are immunosuppressive, as they are cytotoxic to immune cells (Table 3). Additionally, some of them can induce ER stress and treatment resistance (Table 3). However, targeted therapy and subsets of cytotoxic chemotherapy are able to promote antitumor immune responses (Table 3). This effect can occur through ER stress induced ICD, modulation of immune cells in the TME, and/or reduction of immunosuppressive cytokines (Garg et al. 2010). The initiation of antitumor immune responses by cancer therapy aids the immune system in eradicating cancer cells, allowing for sustained cancer remission (Hughes and Yong 2017). Given the diverse effects of anticancer therapy on tumor immunity, understanding their respective roles on cancer cells and immunity is critical to guide clinical application of immunotherapy and combination treatment.

Table 3.

Therapeutic effects on ER stress and tumor immunity.

Therapy	Tumor types	Mechanisms of action	Effects on ER stress/ERAD	References
Camptothecin	Leukemia	Cytotoxic therapy, targeting topoisomerase and inducing apoptosis but not ICD		Sauter et al. (2000), Obeid et al. (2007), and Tesniere et al. (2010)
Etoposide	Testicular cancer, lung, lymphoma, leukemia, neuroblastoma, and ovarian cancer	Cytotoxic therapy inducing apoptosis but not ICD	Promoting apoptosis in human hepatic stellate cells via ER stress induction	Sauter et al. (2000), Obeid et al. 2007), and Wang et al. (2016)
Mitomycin C	Adenocarcinoma of stomach and pancreas	Cytotoxic therapy inducing apoptosis but not ICD	Inducing ER stress	Sauter et al. (2000), Obeid et al. 2007), and Shi et al. (2013)
Cyclophosphamide	Lymphoma, leukemia, NSCLC, breast, and ovarian cancer	Cytotoxic therapy, high dose with immunosuppressive effects by inhibiting immune cell proliferation, while counteracting immunosuppression by Tregs in single low dose	Inducing ER stress to expose CRT and release HMGB1	Emadi et al. (2009), Schiavoni et al. (2011), Inoue and Tani (2014), and Ahlmann and Hempel (2016)
Cisplatin	Various types of cancers, such as breast, ovarian, testicular, brain, and lung cancer	Cytotoxic therapy, inducing apoptosis but not ICD unless combined with ER stress inducer		Martins et al. (2011)
Oxaliplatin	Colorectal cancer	Cytotoxic therapy, inducing ICD	Activation of the PERK-eIF2α pathway, resulting in CRT exposure	Tesniere et al. (2010) and Garg et al. (2012)
Mitoxantrone, idarubicin, epirubicin, and doxorubicin	Acute lymphoblastic leukemia, ovarian, and prostate cancer	Cytotoxic therapy, inducing ICD	Activation of the PERK-eIF2α pathway, resulting in CRT exposure	Fucikova et al. (2011) and Bezu et al. (2018)
Bleomycin	Testicular cancer and Hodgkin’s lymphoma	Cytotoxic therapy, inducing ICD and may also promoting Treg proliferation	Inducing ER stress through the generation of reactive oxygen species and exposing CRT though activation of the PERK-eIF2α pathway	Bugaut et al. (2013) and Garg et al. (2017)
Imatinib	Chronic myeloid leukemia, and gastric cancer	BCR-ABL tyrosine kinase inhibitor, promoting antitumor immune responses by inhibiting MDSCs and Treg function, while suppressing tumor immunity by inhibiting tumor antigens, DCs, and effector T cells	Inducing ER stress and resulting in apoptosis	Brauer et al. (2007), Larmonier et al. 2008), Heine et al. (2011), Hughes et al. (2017), and Kim et al. (2019)
Nilotinib and dasatinib	Chronic myeloid leukemia	BCR-ABL tyrosine kinase inhibitor, reducing MDSCs but not inducing ICD		Hughes et al. (2017)
Abemaciclib	Breast cancer	CDK4/6 inhibitor, enhancing tumor antigen presentation		Goel et al. (2017) and Schaer et al. (2018)
Palbociclib	Breast cancer	CDK4/6 inhibitor, activating effector T cells		Deng et al. (2018)
Gefitinib, erlontinib, afatinib, osimertinib, and olmutinib	NSCLC	EGFR inhibitors, decreasing PD-L1 expression and reducing Tregs		Yamaoka et al. (2017) and Li et al. (2018)
Cetuximab	Colorectal cancer	mAb competitively inhibiting EGFR, inducing ADCC and enhancing cross-presentation in DCs		Kimura et al. (2007), Srivastava et al. (2013), and Yamaoka et al. (2017)
Anti-GD2	Neuroblastoma	mAb targeting cell surface disialoganglioside GD2 and resulting in ADCC		Horta et al. (2016)
Ipilimumab	NSCLC and melanoma	mAb inhibiting CTLA-4		Pennock and Chow (2015) and Buchbinder and Desai (2016)
Nivolumab	NSCLC and melanoma	mAb inhibiting PD-1		Pennock and Chow (2015) and Buchbinder and Desai (2016)
Durvalumab	NSCLC	mAb targeting PD-L1		Antonia et al. (2017)
Metformin		Activating AMPK to phosphorylate PD-L1 and promote its degradation via ERAD	PD-L1 is degraded by ERAD but metformin does not target ERAD	Cha et al. (2018) and Dreher and Hoppe (2018)
Disulfram		Off-target effect on NPL4	Impairing ERAD by disrupting P97-NPL4-UFD1 complex and inhibiting proinflammatory macrophage responses	Skrott et al. (2017)

i. Effects of cytotoxic chemotherapy on ER stress and tumor immunity

Cytotoxic chemotherapy has historically been considered immunogenically silent or immunosuppressive (Harris et al. 1976). When killing cancer cells, most cytotoxic chemotherapy can induce ER stress to dampen antitumor immune responses (Bugaut et al. 2013). Additionally, this type of chemotherapy can interfere with hematopoiesis by eliminating hematopoietic stem cells, blood cells, and immune cells (Emadi et al. 2009). Cytotoxic drugs, for example, camptothecin, etoposide, and mitomycin C, induce apoptosis but not ICD, and thus are nonimmunogenic (Sauter et al. 2000; Obeid et al. 2007). High dose cyclophosphamide, used for treating hematological malignancies, has been shown to actively suppress the immune system by decreasing proliferation of immune cells (Emadi et al. 2009). Moreover, it can induce ER stress and activation of the antiapoptotic UPR, resulting in chemoresistance (Yarapureddy et al. 2019). In osteosarcoma, activation of ATF6 increases resistance to chemotherapy, leading to decreased overall and progression-free survival (Yarapureddy et al. 2019). A similar effect is seen upon BiP induction due to increased ER stress, which is correlated with increased tumor cell survival and chemoresistance (Lee AS 2007).

Although ER stress exerts a pro-tumor role, chemotherapy that results in ICD often utilizes the ER stress response pathways. For example, cisplatin alone is insufficient to induce ICD unless it is combined with ER stress inducers, such as thapsigargin or tunicamycin (Martins et al. 2011). Certain cytotoxic chemotherapy that do induce ER stress, such as oxaliplatin, mitoxantrone, idarubicin, epirubicin, doxorubicin, and cyclophosphamide, can cause ICD (Wu and Waxman 2018). Those chemotherapy can directly stimulate immunogenic responses primarily through the emission of one or multiple DAMPs (Krysko et al. 2012). Among them, the exposure of CRT on the cell surface is the primary determining factor for ICD induction by chemotherapy (Obeid et al. 2007; Krysko et al. 2012). Anthracycline chemotherapies, which include mitoaxantrone, idarubicin, epirubicin, and doxorubicin, have been shown to cause an antitumor immune response by inducing ICD in human prostate cancer, ovarian cancer, and acute lymphoblastic leukemia (Fucikova et al. 2011). To accomplish this function, this type of chemotherapy relies on the activation of the PERK-eIF2α pathway, which causes the translocation of CRT to the cell surface (Bezu et al. 2018). However, anthracyclines and oxaliplatin do not trigger the ATF6 or IRE1α branch of the UPR, suggesting that these UPR pathways do not contribute to ICD induction (Fucikova et al. 2011). Similarly, oxaliplatin, but not cisplatin, causes the translocation of CRT from the ER to the cell surface through the PERK-eIF2α pathway, and can induce ICD in colorectal cancer (Garg et al. 2012). Consistently, oxaliplatin treatment in colorectal cancer cells with CRT depletion cannot induce ICD (Tesniere et al. 2010).

Other ICD-inducing chemotherapy, such as cyclophosphamide and bleomycin, have a paradoxical role on antitumor immunity. Cyclophosphamide, one of the first anticancer agents, is immunosuppressive with high dose treatments depleting hematopoietic lineages, thus patients commonly receive bone marrow transplantation (Emadi et al. 2009). However, cyclophosphamide has also been shown to induce ICD in lymphoma cells by exposing CRT and releasing HMGB1 (Schiavoni et al. 2011; Inoue H and Tani 2014). When applied as a single and low dose, cyclophosphamide can mitigate the immunosuppressive effects of Tregs, aiding in cancer treatment (Ahlmann and Hempel 2016). Bleomycin, a glycopeptide antibiotic used in the treatment of testicular cancer and Hodgkin’s lymphoma, has also been shown to result in the induction of ICD (Bugaut et al. 2013; Garg et al. 2017). However, this effect is counteracted by bleomycin’s ability to stimulate Treg proliferation, making this drug both immunomodulatory and immunosuppressive, similar to cyclophosphamide (Bugaut et al. 2013).

ii. Effects of targeted therapy on ER stress and tumor immunity

Targeted cancer therapy can exert similar antitumor immune responses as some cytotoxic chemotherapy and are particularly important as they often minimize treatment side-effects. One of the first targeted cancer therapy, imatinib (Gleevec) and the more potent second-generation inhibitors nilotinib and dasatinib, can bolster the immune responses in chronic myeloid leukemia patients through reduction of MDSCs (Hughes et al. 2017). While imatinib, nilotinib, and dasatinib have not been shown to induce ICD, their ability to reduce immunosuppressive cells has lasting effects of promoting disease remission without the need for continuous treatment (Hughes and Yong 2017). Moreover, imatinib can impair the immunosuppressive functions of Tregs and enhance immunotherapy against lymphoma (Larmonier et al. 2008). However, imatinib downregulates tumor antigens in leukemic cells and inhibits DCs and effector T cells, suppressing antitumor immunity (Brauer et al. 2007; Heine et al. 2011). Inhibitors of cyclin-dependent kinase 4 and 6 (CDK4/6), such as ribociclib, abemaciclib, and palbociclib, recently gained traction in treating advanced breast cancer. In addition to inhibiting tumor cell proliferation, these inhibitors can also evoke antitumor immune responses. Abemaciclib enhances antigen presentation in tumor cells and suppresses Treg functions, while palbociclib activates effector T cells (Goel et al. 2017; Deng et al. 2018; Schaer et al. 2018).

Other targeted therapy inhibitors, such as EGFR inhibitors, have since been developed for the treatment of multiple cancers - such as NSCLC, colorectal, and breast cancer - and have been shown to promote antitumor immune responses (Li X et al. 2018). Gefitinib, erlotinib, afatinib, osimertinib, and olmutinib are approved treatments for metastatic NSCLC, while lapatinib is approved for HER2⁺ breast cancer (Yamaoka et al. 2017). Although EGFR inhibition has not been shown to induce ICD, it can increase antitumor immune responses. In NSCLC, EGFR inhibition with gefitinib results in decreased expression of PD-L1 (Li X et al. 2018). Additionally, gefitinib treatment reduces Tregs in mice engrafted with squamous cell carcinoma (Mascia et al. 2016). EGFR inhibitors, therefore, can affect both tumor cells and the TME. However, treatment resistance to EGFR inhibitors is common, demanding the use of other therapeutic strategies to overcome this clinical challenge (Stewart et al. 2015; Tang Z-H and Lu 2018).

Similar to cytotoxic chemotherapy, targeted therapy can also induce ER stress and alter ERAD, leading to treatment resistance. Melanomas with the BRAF^V600E mutation often become resistant to BRAF inhibitors, such as vemurafenib, due to the induction of ER stress and cytoprotective autophagy (Ma et al. 2014). This resistance can be overcome when combined with autophagy inhibition (Ma et al. 2014). Vemurafenib can also induce miR-211 to downregulate EDEM1, an ER chaperone critical for ERAD of tyrosinase, subsequently leading to tyrosinase accumulation and treatment resistance (Vitiello et al. 2017). Another targeted therapy is monoclonal antibodies (mAbs). Cetuximab, a mAb against EGFR and an effective competitive inhibitor, can result in antibody-dependent cell-mediated cytotoxicity (ADCC) in colorectal cancer (Kimura et al. 2007; Yamaoka et al. 2017). Additionally, cetuximab induces crosstalk between NK cells and DCs to promote the maturation of DCs, enhancing cross-presentation (Srivastava et al. 2013). A similar result is seen in the use of HER2 mAbs, which has led to better clinical outcomes in HER2⁺ patients (Costa and Czerniecki 2020). In neuroblastoma treatment, anti-GD2, the first mAb approved for treating childhood malignancies, has also been shown to result in ADCC (Horta et al. 2016). Due to their ability to specifically target cancer cells while eliciting immune responses, mAbs represent the most commonly used and approved immunotherapy in the treatment of cancer (Kimiz-Gebologlu et al. 2018).

Given the known role of ERAD in promoting tumor cell survival, it is reasonable to suspect that enhanced ERAD in tumor cells might result in resistance to ICD and decreased availability of tumor cell debris as antigens for DCs. As such, the benefits of ERAD inhibitors to cancer treatment could be two-fold: inducing cancer cell death and increasing antitumor immunity. The efficacy of such treatments is seen in proteasome inhibitors, such as bortezomib, which result in cancer cell death due to irresolvable ER stress as well as ICD induction (Obeng et al. 2006; Spisek et al. 2007; Strasser and Puthalakath 2008). The mechanism of ICD induction by bortezomib is through exposure of HSP90 (another DAMP) on the tumor cell surface (Spisek et al. 2007). Other components of the ERAD pathway may be similarly targetable. P97 is the AAA ATPase mainly associated with retrotranslocation of ERAD substrates (Carvalho et al. 2006). CB-5083, a selective and potent inhibitor for P97, results in irresolvable ER stress and subsequent cell death in multiple myeloma (Anderson et al. 2015). However, CB-5083 recently failed in phase I clinical trials due to high toxicity (NCT02243917). Interestingly, disulfram, an alcohol-aversion drug, can disrupt ERAD by inhibiting NPL4, a P97 partner. Disulfram’s use in cancer patients has led to lower mortality, due to its double-edged effects: promoting cancer cell death through ERAD impairment and inhibiting macrophage proinflammatory responses (Skrott et al. 2017; Terashima et al. 2020).

iii. ERAD’s impact on immunotherapy

Immunotherapy functions by promoting cytotoxic responses to tumor cells through cytokines and effector immune cells, leading to enduring antitumor immune responses. Advances in immunotherapy have improved cancer treatment, especially through the application of chimeric antigen receptor T-cells and immune checkpoint inhibitors, such as ipilimumab and nivolumab (Buchbinder and Desai 2016). However, current immunotherapy possesses significant side effects. A major adverse event, for example, is abnormal reactivity to normal tissues by hyperactivated T cells (Weber et al. 2015). Tumors in most patients respond to the treatment and are controllable for a long time, yet one-third of them relapse by some unknown mechanism (Ribas and Wolchok 2018).

Two immune checkpoints, cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) and PD-1, put the brakes on T-cell functions (Buchbinder and Desai 2016). CTLA-4 is expressed on Tregs, which ultimately inhibit T-cell activation and proliferation (Wing et al. 2008). PD-1 is expressed on activated T cells, suppressing their functions by binding to its ligand PD-L1 (Okazaki and Honjo 2007; Yokosuka et al. 2012). Ipilimumab and nivolumab, mAbs against CTLA-4 and PD-1 respectively, have achieved some success in treating melanoma and NSCLC (Pennock and Chow 2015; Buchbinder and Desai 2016). Since PD-L1 is often upregulated in malignant cells, the mAb durvalumab has been developed to target PD-L1, leading to increased progression-free survival in NSCLC (Antonia et al. 2017). Interestingly, metformin, a commonly used drug for type 2 diabetes, induces the degradation of PD-L1 through ERAD (Dreher and Hoppe 2018). Metformin can interact with the AMP-activated protein kinase (AMPK) and enhance AMPK-mediated phosphorylation of PD-L1. Although it has been observed that the phosphorylated form of PD-L1 accumulates in the ER and is subsequently degraded through ERAD, how ERAD targets PD-L1 remains unknown (Cha et al. 2018). In contrast, disruption of ERAD by disulfram upregulates PD-L1, leading to immunosuppression and treatment resistance in cancer patients (Zhou et al. 2019). A better understanding of ERAD in tumor immunity is needed for successful application of immunotherapy.

Future perspectives

Over the past four decades, significant insights have been acquired through in-vitro co-culture and in-vivo humanized animal model studies, establishing ERAD as a principle protein quality control mechanism, which often enables the survival and growth of tumor cells (Tax et al. 2019). Due to its ability to degrade key regulators of immune cells, such as TCR, BCR, and BLIMP-1, ERAD plays an indispensable role in immune cell development and function (Bonifacino et al. 1989; Yang et al. 1998; Yang et al. 2014; Ji et al. 2016). Through degrading MHC-I, processing tumor antigens, and translocating exogenous antigens for cross-presentation, components of ERAD participate in antitumor immune responses (Gong et al. 2014; Grotzke and Cresswell 2015). Despite some known effects of current anticancer therapy in evoking ER stress and their paradoxical roles in tumor immunity, our limited understanding of ERAD presents a significant barrier to applying current therapy for successful eradication of cancer.

While emerging evidence demonstrates ERAD’s regulatory role in immune cell development and function, little is known about how it impacts immune cells in the TME. In particular, several questions remain unanswered. How does altered ERAD in tumor cells affect functions of tumor-infiltrating immune cells and the induction of ICD by cancer therapy? How does altered ERAD in tumor-infiltrating immune cells impact their antitumor capacities? How can other forms of cancer therapy, particularly immunotherapy, be combined with ERAD inhibitors? Due to these unanswered questions, caution must be taken in combination treatment utilizing current immunotherapy with other therapy that are known to evoke ER stress and ERAD.

Therapy targeting ERAD is still in its infancy due to the complex and multifaceted roles of ERAD in tumorigenesis and immunity (Guerriero and Brodsky 2012). Although a number of small-molecule inhibitors targeting ERAD have been developed, none except proteasome inhibitors have progressed to clinical applications due to their toxicities. Therefore, a better understanding of ERAD is necessary for selection of therapeutic targets for development of minimally toxic treatments. Animal models with lineage-specific depletion of ERAD components will facilitate the dissection of ERAD’s effects on specific types of immune cells, as well as on tumor cells, in both physiological and pathological settings. With improved understanding of how ERAD regulates proteins relevant to immune checkpoints, such as PD-L1, a more effective combination therapy could be designed. Of course, the successful combination of ERAD inhibitors or other therapy with immunotherapy is likely to be context dependent as the ability for each therapy to evoke ER stress and impact tumor immunity differs in each cancer type. It is our hope that an improved understanding of ERAD’s role in tumor immunity will not only lead to development of novel and specific inhibitors but also guide rational application of currently available therapy in our battle against cancer.

Abbreviations:

AAT

alpha 1 antitrypsin

ADCC

antibody-dependent cell-mediated cytotoxicity

AMPK

AMP-activated protein kinase

APCs

antigen-presenting cells

ApoB100

apolipoprotein B

ASGPR

asialoglycoprotein receptor

ATF6

activating transcription factor 6

AUP1

ancient ubiquitous protein 1

BACE457

pancreatic isoform of human β-secretase

β2m

β2-microglobulin

BAG6

BCL2-associated athanogene 6

Pre-BCR

pre-B-cell receptor

BiP

binding immunoglobulin protein

BLIMP-1

B lymphocyte-induced maturation protein-1

CDK4/6

cyclin-dependent kinase 4 and 6

CFTR

mutant cystic fibrosis transmembrane regulator

CHIP

C-terminus of HSP70-interacting protein

CHOP

transcription factor C/EBP-homologous protein

CRT

calreticulin

CTLA-4

cytotoxic T-lymphocyte-associated antigen 4

DAMP

damage-associated molecular pattern

Doa10

degradation of alpha 10

DCs

dendritic cells

DUBs

deubiquinating enzymes

E1

ubiquitin-activating enzyme

E2

ubiquitin-conjugating enzyme

E3

ubiquitin ligase

E4

specialized ubiquitin ligase

E6-AP

E6-associated protein

EDEM 1/2/3

ER degradation-enhancing α-mannosidase-like lectin 1/2/3

EGFR

epidermal growth factor receptor

ER

endoplasmic reticulum

ERAD

endoplasmic reticulum-associated protein degradation

ERAP1/2

ER aminopeptidase-1/2

ERdj5

endoplasmic reticulum (ER)-resident protein-containing DnaJ and thioredoxin domains

ERManI

ER mannosidase-I

ERp57

ER-resident protein 57

eIF2α

eukaryotic initiation factor 2α

ExoA

exotoxin A

Fbx

F-box protein

Glc

glucose

GlcNAc₂

N-Acetylglucosamine

GP78

glycoprotein 78

GRP170

Glucose-regulated protein 170

HCC

hepatocellular carcinoma cell

HERP

Homocysteine-induced ER protein

HMGB1

high mobility group box 1

HMGR

hydroxy-3-methyl-glutaryl-CoA reductase

Hrd3

hmg CoA reductase degradation 3

HSP70

heat shock protein 70

IL-2

interleukin-2

Igμ

IgM heavy chain

IgG/M

immunoglobulin G/M

Ii

invariant chain

ICD

immunogenic cell death

INSIG-1

insulin-induced gene 1

IRE1α

inositol-requiring enzyme 1α

LPL

lipoprotein lipase

LPS

lipopolysaccharides

mAb

monoclonal antibody

MDSCs

myeloid-derived suppressor cells

MHC

major histocompatibility complexes

MHC-I

MHC class I molecules

MHC-II

MHC class II molecules

Man

mannose

NGly

N-glycanase

NK cells

natural killer cells

NPL4

nuclear protein localization protein 4

NSCLC

nonsmall cell lung cancer

OS-9

osteosarcoma 9

OST

oligosaccharyl transferase

OVA

ovalbumin

PERK

protein kinase RNA-like ER kinase

PDIs

protein disulfide isomerases

PD-L1

programed death ligand 1

PLC

peptide loading complex

RAD23

radiation-sensitive 23

RMA1

RING finger protein 5

RING

really interesting new gene

RTA

glycosylated protein-rich A chain

SEL1L

suppressor enhancer Lin12 1 like

SKP1

S-phase kinase-associated protein 1

SLP

synthetic long peptide

TAP

transporter associated with antigen processing

TCRα

T-cell receptor alpha chain

TCR

T-cell receptor

TME

tumor microenvironment

TMEM129

transmembrane protein 129

TRC8

translocation in renal carcinoma, chromosome 8 gene

TRC35

transmembrane domain recognition complex 35

Tregs

T regulatory cells

Ubc6p

ubiquitin-conjugating enzyme 6p

Ubx2p

UBX domain-containing protein 2

UBC6E

ubiquitin-conjugating enzyme 6e

UBE2G2

ubiquitin-conjugating enzyme E2 G2

UBE2J1

ubiquitin-conjugating enzyme E2 J1

UBL4A

ubiquitin-like protein 4A

UBR1/2

E3 component N-recognin 1/2

UBXD8

ubiquitinlike-domain-containing protein

UFD1

ubiquitin fusion degradation protein

UGGT

glycoprotein glycosyl-transferase

UPR

unfolded protein response

USP13

ubiquitin-specific protease 13

VCP

valosin-containing-protein

WDR20

WD repeat-containing protein 20

XTP3-B

XTP3-transactivated gene B

YOD1

ubiquitin thioesterase OTU1

Yos9

Yeast OS-9 homolog