TNF signaling (Homo sapiens)

The caspase-8 zymogens are present in the cells as inactive monomers, which are recruited to the death-inducing signaling complex (DISC) by homophilic interactions with the DED domain of FADD. The monomeric zymogens undergo dimerization and the subsequent conformational changes at the receptor complex, which results in the formation of catalytically active form of procaspase-8.[Boatright KM et al 2003; Donepudi M et al 2003; Keller N et al 2010; Oberst A et al 2010].

The Reactome module describes necroptosis and pyroptosis.

BAG4, also known as silencer of death domain (SODD), belongs to the BAG family of anti-apoptotic proteins. Mammalian BAG4 was found to associate with TNFR1 preventing receptor signaling in the absence of ligand (Jiang Y et al. 1999; Miki K and Eddy EM 2002). Furthermore, crystallographic data and biochemical analysis showed that TNFR1 forms inactive homodimers or homotrimers in the absence of TNF by the N-terminal domain, the pre ligand assembly domain (PLAD) (Chan FK et al. 2000; Wang YL et al. 2011). Upon TNF-alpha binding BAG4 is quickly released from TNFR1 and three receptor molecules form a complex with the TNF trimer. The TNF-alpha homologue ligand, lymphotoxin-alpha (LTA, also known as TNF-beta), which as homotrimer only occurs as a soluble ligand, also interacts with TNFR1. LTA binds three receptor molecules and triggers the same effects as soluble TNF-alpha (Banner DW et al. 1993; Etemadi N et al. 2013).

The TNF-alpha:TNFR1 receptor complex then transmits the signal leading to cell death or survival. However, it remains unclear whether BAG4 binds to death domain of monomeric TNFR1 to prevent receptor oligomerization or recognizes receptor trimers to facilitate ATP-dependent TNFR1 trimer disassembly (Jiang Y et al. 1999; Miki K and Eddy EM 2002). Additionally, BAG4 is known to interact with HSP70, death receptor 3, and the anti-apoptotic protein Bcl-2 (Antoku et al. 2001; Brockmann et al. 2004; Jiang et al. 1999).

BAG4-overexpressing HeLa cells showed reduced cellular sensitivity to treatment with extracellular TNFalpha and CD95 ligand (Eichholtz-Wirth H et al. 2003). In addition, increased expression level of BAG4 in tumor cells leads to resistance of TNFalpha-induced cell death and is associated with pancreatic cancer, some types of melanoma, acute lymphoblastic leukemia etc.(Ozawa et al. 2000; Tao H et ql. 2007; Reuland SN et al. 2013). The physiological relevance of BAG4 for TNFR1 signaling, however, is difficult to judge because BAG4 knockout mice have no or only a mild effect on pro-inflammatory TNF signaling and give no evidence for an inhibitory role of BAG4 in TNFR1-induced cell death (Takada H et al. 2003; Endres R et al. 2003).

Several E3 ligases are involved in TNFalpha signaling to initiate an immediate and effective host response to infection or injury. Among them are anti-apoptotic regulators BIRC2 and BIRC3, also known as inhibitor of apoptosis proteins (cIAP1/2). BIRC2/3 were found to constitutively associate with TRAF2 and via TRAF2 they were recruited to the TNFR1 signaling complex (Samuel T et al. 2006; Bertrand et al, 2008; Varfolomeev E et al, 2008). BIRC2/3 can directly ubiquitinate RIPK1 within the TNFR1 receptor complex allowing it to bind to the TAB2:TAK1 complex, a process reversed by the deubiquitinase CYLD and A20 (Bertrand et al, 2008; Varfolomeev et al, 2008; Moquin DM et al. 2013; Shembade N et al. 2010; Wertz IE et al. 2004). In conjunction with the ubiquitin conjugating enzyme (E2) enzyme UbcH5a, BIRC2/3 was shown to mediate polymerization of both K63-linked and linear Met1-linked chains on RIPK1 (Varfolomeev E et al, 2008; Bertrand et al, 2008; Blackwell K et al. 2013). TRAF2 promotes BIRC-mediated linear and K63-linked ubiquitination of RIP1(Blackwell K et al. 2013). K11-linked polyubiquitination of RIPK1 may also depend on BIRC2 and BIRC3 (Dynek JN et al. 2010).

In addition, the linear polyubiquitination has been implicated in the NFkB activation. The linear ubiquitin chain assembly complex (LUBAC) ligase consisting of HOIL-1L, HOIP, and SHARPIN, specifically generates linear polyubiquitin chains (Kirisako T et al. 2006; Walczak H et al. 2012). IKBKG (NEMO), a regulatory component of the IκB kinase (IKK) complex, is a substrate of LUBAC. LUBAC-mediated IKBKG ubiquitination enhances IKBKG interaction with the TNF-alpha receptor signaling complex and stabilizes this protein complex to promote activation of NFkB (Haas TL et al. 2009).

Structural analysis revealed that NPL4 zinc finger 1 (NZF1) of HOIP can simultaneously bind both leucine zipper (CoZi) domains of NEMO (IKBKG) and ubiquitin and that both interactions are involved in the TNF alpha-mediated NFkappaB activation (Fujita H et al. 2014). In addition, NEMO (IKBKG) ubiquitination required RING-between-RING (RBR) domain of HOIL-1L (Smit JJ et al. 2013)

LUBAC consists of the chain-assembling E3 ligase HOIP as well as HOIL-1 and SHARPIN (Kirisako T et al. 2006; Walczak H et al. 2012). Importantly, deletion of the LUBAC component SHARPIN in mice or mutation of HOIL-1 in humans, lead to hyperinflammatory phenotypes, indicating key roles of LUBAC and linear Ub chains in the response to infection and inflammation (Gerlach B et al. 2011; Ikeda F et al. 2011; Tokunaga F et al. 2011; Boisson B et al. 2012).

HOIP belongs to the RING-between-RING (RBR) family of E3 ligases and is the catalytic component of LUBAC (Spratt DE et al. 2014). RBR E3 ligase domain and a conserved C-terminal extension of HOIP are responsible for assembling Met1-linked chains (Smit JJ et al., 2012; Stieglitz B et al. 2012 and 2013). HOIP was reported to act as RING/HECT hybrids, employing RING1 to recognize ubiquitin-loaded E2 while a conserved cysteine in RING2 domain subsequently forms a thioester intermediate with the transferred or “donor� ubiquitin (Stieglitz B et al. 2013). A Ub-associated (UBA) domain mediates interactions with HOIL-1L (Yagi H et al. 2012), while N-terminal PUB domain may be involved in interaction with regulatory proteins such as OTULIN to control NFkappaB signaling (Elliott PR et al. 2014). HOIP also comprises several NPL4 zinc finger (NZF) Ub binding domains (UBDs) that target it to ubiquitinated proteins (Haas et al. 2009; Fujita H et al. 2014).

Structural analysis revealed that NZF1 of HOIP can simultaneously bind both leucine zipper (CoZi) domains of NEMO (IKBKG) and ubiquitin and that both interactions are involved in TNF alpha-mediated NFkappaB activation (Fujita H et al. 2014). In addition, NEMO (IKBKG) ubiquitination required RBR domain of HOIL-1L (Smit JJ et al. 2013).

Factor associated with neutral sphingomyelinase activation (FAN or NSMAF) is an adaptor protein that constitutively binds to the neutral SMASE activation domain (NSD) of TNFR1 (Adam-Klages S et al. 1996). NSMAF (FAN) is thought to directly link TNFR1 to the activation of neutral sphingomyelinase (N-SMASE) such as sphingomyelin phosphodiesterase 2 or 3 (SMPD2, SMPD3).

The TRAFs (TRAF1 to TRAF6) are a group of structurally similar adaptor proteins, in most cases with E3 ligase activity, that are involved in downstream signaling of various cell surface receptors such as TNFR1, TNFR2, CD40, TLRs and TCR (Jang HD et al. 2001; Fotin-Mleczek M et al. 2004; Su X et al. 2006). The hallmark feature of all TRAFs is a C-terminal TRAF-domain of approximately 230 amino acids, which is responsible for homo- and heterooligomerization of TRAF molecules (Rothe M et al. 1994). The differences in amino acid sequences in TRAF-domains define the range of TRAF interaction partners. Another important structural element of TRAFs, with an exception of TRAF1, is the N-terminal RING finger domain that modulates induction of NFkappaB and MAPK activities. As TRAF1 has no RING finger domain, the effects of TRAF1 on NFkB activation are rather unclear. It is believed that TRAF1 regulates TNF receptor activity through its ability to interact with TRAF2 (Rothe M et al.1995; Zheng C et al. 2010; Fotin-Mleczek M et al. 2004). Structural and biochemical studies showed that TRAF1:TRAF2 heterotrimer (1:2) binds BIRC (cIAP2) more strongly than TRAF2 homotrimers, suggesting that TRAF1 may modulate the interaction between TRAF2 and BIRC (cIAP1/2) and thus suppress TNF-alpha induced apoptosis (Zheng C et al. 2010). Noteworthy, TRAF1:TRAF2 heterotrimers and TRAF2 homotrimers also differ in their capability with certain receptors but there seems to be no difference with respect to TNFR1 recruitment (Fotin-Mleczek M et al. 2004). On the contrary, TNF-induced caspase-mediated cleavage of TRAF1 generates a C-terminal fragment with NFkB-inhibitory, pro-apoptotic activity (Leo e et al. 2001; Jang HD et al. 2001; Henkler F et al. 2003). Thus, the current data suggest that depending on its cleavage status TRAF1 may exert either cytoprotective or cytotoxic effect in death domain-containing receptor signaling pathways.

The heteromeric complex TRAF1:TRAF2 has been also implicated in the cross-talk of TNFR1 and TNFR2 (Wicovsky A et al. 2009).

The Reactome pathway shows that upon TNF receptor 1 (TNFR1) stimulation, the E3 ubiquitin ligases such as baculoviral IAP repeat-containing protein (BIRC2/3 or cIAP1/2) mediate polyubiquitination of RIPK1 generating K63-linked chains. TNFAIP3 removes these K63-linked polyubiquitin chains, preventing the interaction of RIPK1 with NFkappaB essential modulator (NEMO) (Wertz IE et al. 2004). Subsequently, TNFAIP3 facilitates addition of K48-linked polyubiquitin chains to RIP1, targeting it for proteasomal degradation (Wertz IE et al. 2004). In this way, TNFAIP3 restricts TNF-induced NFkappaB signaling by sequential deubiquitination and ubiquitin-mediated degradation of RIPK1.

TNFAIP3 (A20) is thought to limit NFkappaB activation, however the inhibitory mechanisms for TNFAIP3 are not fully understood and are partially contradicting. A study with knockin mice expressing DUB-inactive Tnfaip3 C103A mutant reported that the deubiquitinase (DUB) activity was dispensable for LPS- or TNF-stimulated NFkappaB signaling (De A et al 2014). In contrast to Tnfaip3 knockout mice that develop perinatal lethality, the knockin Tnfaip3 C103A mice were normal and did not have an inflammatory phenotype (De A et al 2014). These finding are in agreement with an earlier study reporting that the Tnfaip3 C103A knockin mice were grossly normal for at least 4 months and contained a normal number of lymphocytes (Lu TT et al. 2013). In addition, bone marrow-derived macrophage cells (BMDM) from the knockin mice showed normal LPS- and TNF-induced NFkappaB activation and downstream gene expression, comparable to cells from wild-type mice (De A et al 2014). The study suggests that the deubiquitinase activity of TNFAIP3 in general is not obligate for most of the regulatory functions of TNFAIP3 and alternative mechanisms might be involved in TNFAIP3-mediated NFkappaB regulation (De A et al 2014). Indeed, there is biochemical evidence for different mechanisms that may contribute to TNFAIP3 inhibitory effects on TNFR1 pro-inflammatory signaling in a cell type-dependent manner:

• Deubiquitination of K63-linked ubiquitin chains from RIPK1 by the OTU domain and adding of K48-linked polyubiquitin chains via its forth zinc finger (ZnF4) domain (Wertz IE et al. 2004).
• Triggering degradation of E2 enzymes such as ubiquitin conjugating enzyme Ubc13 by adding K48-linked polyubiquitin chains and disruption of interactions between E2 and E3 (BIRC2,3 or TRAF2) enzymes in the TNFR1 pathway. This event was impaired with DUB-inactive TNFAIP3 (C103A, within the OTU domain) suggesting that there might be a crosstalk with DUB-dependent mechanisms (Shembade N et al. 2010).
• A non-catalytic blockage of TAK1-mediated IKK activation by TNFAIP3 binding to polyubiquitin chain via its seventh zinc finger (ZnF7) and forming a complex involving specific NEMO:TNFAIP3 interaction (Skaug B et al. 2011).
• A non-catalytic inhibition of TNF- and LUBAC-induced NFkappaB signalling by binding to linear polyubiquitin chains via its ZnF7, which prevents the LUBAC:NEMO interaction (Verhelst K et al. 2012; Tokunaga F et al. 2012).
• Attenuation of TNFR1 signaling complex formation (He KL & Ting AT 2002).

TAX1BP1 functions as an adaptor molecule for TNFAIP3 (A20) to block TNF-alpha-stimulated signaling to NFkappaB (Shembade N et al. 2007).

The Reactome pathway shows that upon TNF receptor 1 (TNFR1) stimulation, the E3 ubiquitin ligases such as baculoviral IAP repeat-containing protein (BIRC2/3 or cIAP1/2) mediate polyubiquitination of RIPK1 generating K63-linked chains. TNFAIP3 removes these K63-linked polyubiquitin chains, preventing the interaction of RIPK1 with NFkappaB essential modulator (NEMO) (Wertz IE et al. 2004). Subsequently, TNFAIP3 facilitates addition of K48-linked polyubiquitin chains to RIP1, targeting it for proteasomal degradation (Wertz IE et al. 2004). In this way, TNFAIP3 restricts TNF-induced NFkappaB signaling by sequential deubiquitination and ubiquitin-mediated degradation of RIPK1.

TNFAIP3 (A20) is thought to limit NFkappaB activation, however the inhibitory mechanisms for TNFAIP3 are not fully understood and are partially contradicting. A study with knockin mice expressing DUB-inactive Tnfaip3 C103A mutant reported that the deubiquitinase (DUB) activity was dispensable for LPS- or TNF-stimulated NFkappaB signaling (De A et al 2014). In contrast to Tnfaip3 knockout mice that develop perinatal lethality, the knockin Tnfaip3 C103A mice were normal and did not have an inflammatory phenotype (De A et al 2014). These finding are in agreement with an earlier study reporting that the Tnfaip3 C103A knockin mice were grossly normal for at least 4 months and contained a normal number of lymphocytes (Lu TT et al. 2013). In addition, bone marrow-derived macrophage cells (BMDM) from the knockin mice showed normal LPS- and TNF-induced NFkappaB activation and downstream gene expression, comparable to cells from wild-type mice (De A et al 2014). The study suggests that the deubiquitinase activity of TNFAIP3 in general is not obligate for most of the regulatory functions of TNFAIP3 and alternative mechanisms might be involved in TNFAIP3-mediated NFkappaB regulation (De A et al 2014). Indeed, there is biochemical evidence for different mechanisms that may contribute to TNFAIP3 inhibitory effects on TNFR1 pro-inflammatory signaling in a cell type-dependent manner:

• Deubiquitination of K63-linked ubiquitin chains from RIPK1 by the OTU domain and adding of K48-linked polyubiquitin chains via its forth zinc finger (ZnF4) domain (Wertz IE et al. 2004).
• Triggering degradation of E2 enzymes such as ubiquitin conjugating enzyme Ubc13 by adding K48-linked polyubiquitin chains and disruption of interactions between E2 and E3 (BIRC2,3 or TRAF2) enzymes in the TNFR1 pathway. This event was impaired with DUB-inactive TNFAIP3 (C103A, within the OTU domain) suggesting that there might be a crosstalk with DUB-dependent mechanisms (Shembade N et al. 2010).
• A non-catalytic blockage of TAK1-mediated IKK activation by TNFAIP3 binding to polyubiquitin chain via its seventh zinc finger (ZnF7) and forming a complex involving specific NEMO:TNFAIP3 interaction (Skaug B et al. 2011).
• A non-catalytic inhibition of TNF- and LUBAC-induced NFkappaB signalling by binding to linear polyubiquitin chains via its ZnF7, which prevents the LUBAC:NEMO interaction (Verhelst K et al. 2012; Tokunaga F et al. 2012).
• Attenuation of TNFR1 signaling complex formation (He KL & Ting AT 2002).

The TNF-alpha:TNFR1 receptor complex then transmits the signal leading to cell death or survival. However, it remains unclear whether BAG4 binds to death domain of monomeric TNFR1 to prevent receptor oligomerization or recognizes receptor trimers to facilitate ATP-dependent TNFR1 trimer disassembly (Jiang Y et al. 1999; Miki K and Eddy EM 2002). Additionally, BAG4 is known to interact with HSP70, death receptor 3, and the anti-apoptotic protein Bcl-2 (Antoku et al. 2001; Brockmann et al. 2004; Jiang et al. 1999).

BAG4-overexpressing HeLa cells showed reduced cellular sensitivity to treatment with extracellular TNFalpha and CD95 ligand (Eichholtz-Wirth H et al. 2003). In addition, increased expression level of BAG4 in tumor cells leads to resistance of TNFalpha-induced cell death and is associated with pancreatic cancer, some types of melanoma, acute lymphoblastic leukemia etc.(Ozawa et al. 2000; Tao H et ql. 2007; Reuland SN et al. 2013). The physiological relevance of BAG4 for TNFR1 signaling, however, is difficult to judge because BAG4 knockout mice have no or only a mild effect on pro-inflammatory TNF signaling and give no evidence for an inhibitory role of BAG4 in TNFR1-induced cell death (Takada H et al. 2003; Endres R et al. 2003).