ReviewCell death

The Role of the Kinases RIP1 and RIP3 in TNF-Induced Necrosis

See allHide authors and affiliations

Science Signaling  30 Mar 2010:
Vol. 3, Issue 115, pp. re4
DOI: 10.1126/scisignal.3115re4


Tumor necrosis factor (TNF) is a pleiotropic molecule with a crucial role in cellular stress and inflammation during infection, tissue damage, and cancer. TNF signaling can lead to three distinct outcomes, each of which is initiated by different signaling complexes: the gene induction or survival mode, the apoptosis mode, and the necrosis mode. The kinases receptor-interacting protein 1 (RIP1) and RIP3 are key signaling molecules in necrosis and are regulated by caspases and ubiquitination. Moreover, TNF stimulation induces the formation of a necrosome in which RIP3 is activated and interacts with enzymes that control glycolytic flux and glutaminolysis. The necrosome induces mitochondrial complex I–mediated production of reactive oxygen species (ROS) and cytotoxicity, which suggest a functional link between increased bioenergetics and necrosis. In addition, other effector mechanisms also contribute to TNF-induced necrosis, such as recruitment of NADPH (the reduced form of nicotinamide adenine dinucleotide phosphate) oxidases and subsequent ROS production at the membrane-associated TNF receptor complex I; calcium mobilization; activation of phospholipase A2, lipoxygenases, and acid sphingomyelinases; and lysosomal destabilization. However, the link between RIP1 and RIP3 and these subcellular events remains to be established. The regulation of RIP1 and RIP3 and their downstream signaling cascades opens new therapeutic avenues for treatment of pathologies associated with cell loss, such as ischemia-reperfusion damage and neurodegeneration, and ways to stimulate alternative immunogenic cell death pathways in cancer.

TNF, a Pleiotropic Molecule

The name “tumor necrosis factor” emphasizes a histological feature, namely, the in vivo hemorrhagic necrosis of tumors after exposure to TNF, but does not refer to the type of cell death. TNF is a pleiotropic cytokine that plays a key role in inflammation that follows infection or tissue injury by prompting various cellular responses in different cell types, such as cell survival, proliferation, differentiation, induction of antimicrobial activity, and cell death (apoptosis or necrosis). Balkwill provides an excellent overview of the history of TNF (1). Here, we focus on recent insights into the molecular mechanisms of TNF-induced necrosis. To distinguish this programmed form of necrosis from accidental necrosis caused by physical or chemical injury, it has been renamed “necroptosis” (2). At the biochemical level, necroptosis is defined as a type of cell death that is inhibited by genetic or pharmacological ablation of the kinase activity of receptor-interacting protein 1 (RIP1) (3). However, the term “necroptosis” has not been generally adopted; therefore, we will use “necrosis” or “necrotic cell death.”

The first indirect description of the antitumor activity of TNF was in 1893, when Coley described the treatment of malignant tumors with a mixture of bacterial extracts called “Coley’s mixed toxins” (4). The ability of TNF to induce rapid hemorrhagic necrosis of experimentally induced tumors was formally identified in 1975 (5). This antitumor effect is partially caused by hemorrhagic necrosis resulting from the destruction of the tumor-associated vasculature and consequent lack of oxygen and nutrients (6). A major obstacle for the therapeutic exploitation of TNF in cancer treatment is its ability to elicit a systemic inflammatory response syndrome. Conversely, TNF-neutralizing strategies have been used to treat major inflammatory diseases, such as rheumatoid arthritis, Crohn’s disease, psoriasis, and sarcoidosis (7). Nearly 25 years ago, it was discovered that TNF and cachectin, a factor released by macrophages that reduces fat storage metabolism, were identical (8). TNF is involved in cachexia and wasting during chronic neoplastic and infectious diseases, which suggests a link between TNF-induced signaling and metabolism regulation; such a link could be central to the pathogenesis of diverse disease states (9, 10). A proposed connection between TNF and metabolism was supported by the demonstration in the 1990s that TNF-induced necrosis depends on complex I–mediated reactive oxygen species (ROS) production in the mitochondria (11, 12), at least in some cell lines.

Apoptosis and Necrosis, Two Sides of the Same Coin

Cells can die through apoptosis, necrosis, postapoptotic secondary necrosis, necroptosis, autophagic cell death, or mitotic catastrophe (3). Two extremes of cell death are apoptosis (which is caspase-dependent) and necrosis (which is caspase-independent). Hallmarks of apoptosis include activation of caspases, DNA fragmentation, and membrane blebbing (3, 13). Necrosis is characterized by swelling of the endoplasmic reticulum, mitochondria, and cytoplasm, with subsequent rupture of the plasma membrane and lysis of the cells (3, 14). However, there is no consensus on the biochemical changes that unequivocally identify necrosis. Therefore, necrotic cell death is still identified mainly by the absence of apoptotic or autophagic markers in the early stages of cell death (3). Necrosis should not be confused with secondary necrosis, which occurs in the late stages of apoptosis. But it is difficult to deduce whether in vivo necrotic tissue ensues from primary necrotic cell death or from secondary necrosis after apoptosis. Accumulating evidence shows that necrotic cell death is often as well controlled as caspase-dependent apoptosis (14, 15). The way cells die determines subsequent intercellular communication, immunological consequences, tissue repair, and regeneration (16). Initially, it was assumed that apoptosis is a tolerogenic cell death modality, whereas necrotic cell death is immunogenic. This simple dichotomous view has been challenged, and the immunological response, along with the tissue repair outcome, may be related not only to the type of cell death modality, but also to the intensity of the stimulus, the extent of cell death response, and the absence or presence of adequate phagocytic activity (16).

The crucial role of the kinase RIP1 in necrotic cell death has been shown in cellular models of necrosis induced by TNF receptor 1 (TNFR1) (17); TRAIL (tumor necrosis factor–related apoptosis-inducing ligand) receptor (18, 19); Fas (also known as CD95) (20); Toll-like receptor (TLR) 3 (21, 22); TLR4 (23); NALP3 (Nacht domain–, leucine-rich repeat–, and PYD-containing protein 3) (24, 102); or RIG-I (retinoic acid–inducible protein I) (25). RIP1 consists of an N-terminal kinase domain, an intermediate domain, a RIP homotypic interaction motif (RHIM), and a C-terminal death domain (DD) motif. RIP1 belongs to the RIP family of serine-threonine kinases that are involved in innate and adaptive immunity (26, 27) (Fig. 1). Degterev, Yuan, and others identified chemical compounds named necrostatins that allosterically inhibit the kinase activity of RIP1 and that can block necrosis without affecting RIP1-mediated activation of nuclear factor κB (NF-κB), p38 mitogen–activated protein kinase (MAPK), or c-Jun N-terminal kinase (JNK) (2, 15). This finding confirms the bifurcation between the kinase-independent proinflammatory function of RIP1 and its kinase-dependent function in necrosis. The in vivo protection provided by necrostatins in experimental models of ischemic brain injury (2), myocardial infarction (28), excitoxicity (29), and chemotherapy-induced cell death (30) further underscores the role of RIP1 kinase activity in these pathological conditions and its potential as a therapeutic target.

Fig. 1

Schematic representation of the structure of human RIP1 and RIP3. The number of amino acids (aa) corresponding to each domain and the modifications important for NF-κB activation (K63-linked polyubiquitin chains on K377 in RIP1) or necrosis (phosphorylation on RIP1 Ser161 and RIP3 Ser199) are indicated. The interaction between RIP1 and RIP3 is mediated by their RHIM domains. DD, death domain; KD, kinase domain.

Distinct TNF Complexes Determine the Signaling Mode

The multiple alternative outcomes of TNF-induced signaling—such as gene induction and survival, apoptosis, or necrosis—suggest the existence of several signaling complexes or pathways. In particular, regulation of caspase activity and the ubiquitin editing system play important roles in the biological outcome of TNF stimulation. TNFR1 exists in preformed receptor aggregates owing to the pre-ligand-binding assembly domain (PLAD) in the extracellular cysteine-rich domain 1 (CRD1) (31). Ligand binding leads to the formation of a membrane-associated complex I by causing a conformational change in TNFR1 that recruits TRADD (TNF receptor–associated death domain), which, in turn, recruits RIP1, cIAP1 (cellular inhibitor of apoptosis), cIAP2, TRAF2 (TNF receptor–associated factor 2), and TRAF5. TRAF2 stabilizes cIAP1 and cIAP2 by preventing their autoubiquitination and recruits them to the cytoplasmic domain of several members of the TNF receptor superfamily, including the TNFR1 (3235). RIP1 is K63-polyubiquitinated on K377 (36) by cIAP1 and cIAP2 (37), an event that enables it to act as a scaffold for the assembly of the TAK1 (transforming growth factor-β–activated kinase-1)−TAB2/3 (TAK-1 binding protein 2 or 3) complex, as well as that of the inhibitor κB kinase (IKK) complex, which consists of NEMO (NF-κB essential modulator), IKKα, and IKKβ. The assembly of the IKK complex activates the NF-κB activation.

It has been suggested that K63-polyubiquitinated RIP1 is crucial for cell survival and gene induction after TNF stimulation. However, the absolute requirement of RIP1 for TNF-induced NF-κB activation has been challenged by the observation that TNF-induced NF-κB is only partially inhibited in Ripk1-deficient cells (38), which implies the existence of alternative RIP1-independent pathways. This could be due to redundant signaling molecules, such as the components of the linear ubiquitin chain assembly complex (LUBAC) in the TNF signaling pathway (39). The ligase in the LUBAC is composed of the two RING (really interesting new gene) finger proteins HOIL-1L and HOIP, which conjugate a head-to-tail–linked linear polyubiquitin chain to NEMO, which stabilizes its interaction with IKKα and IKKβ. Indeed, in Rbck1 knockout mice (Rbck1encodes HOIL-1L) and cells derived from these mice, NF-κB signaling as induced by proinflammatory cytokines, such as TNF and IL-1β, was suppressed, which resulted in enhanced TNF-induced apoptosis in hepatocytes derived from Rbck1 knockout mice. Thus, TNF signaling research continues to explore new avenues that challenge widely accepted paradigms.

Depending on the cellular conditions and posttranslational regulatory mechanisms, the composition of complex I changes to form a cytosolic death-initiating signaling complex (DISC), the so-called complex II (40, 41). Caspases and ubiquitination control the composition and biological outcome of complex II (Fig. 2). The polyubiquitinated status of RIP1 prevents the transition of the membrane-associated complex I to the cytosolic complex II (42, 43). However, when K63-ubiquitination of RIP1 is removed by the deubiquitinases CYLD (cylindromatosis) or A20 (44, 45), RIP1 and RIP3 are recruited to a complex containing TRADD, Fas-associated death domain protein (FADD), and caspase-8, in which RIP1 and RIP3 are cleaved (42). In the presence of caspase inhibitors, the proteolytic inactivation of RIP1 and RIP3 is prevented, and that sensitizes some cell lines to TNF-induced necrosis. Smac mimetics are organic molecules resembling the IAP binding motif (IBM) of Smac (also known as DIABLO), a mitochondrial protein released during apoptosis that counteracts IAPs by inducing cIAP autoubiquitination and proteasomal degradation. cIAP ablation prevents RIP1 ubiquitination and favors the formation of an alternative complex II, containing RIP1, FADD, and caspase-8, which leads to RIP1 kinase–dependent activation of caspase-8 and subsequent apoptosis (37, 46). It remains unclear how RIP1 kinase activity contributes to apoptotic cell death, but it is conceivable that phosphorylation of the adapter molecule FADD or caspase-8 (or both) modulates their activities. Active caspase-8 blocks the necrosis mode under these conditions, probably because it cleaves RIP1 and RIP3 (43, 47). Indeed, reduced caspase activity (because of the presence of caspase inhibitors or the adaptor molecule FADD or the absence of caspase-8) prevents the apoptotic mode of complex II and favors the necrotic mode after TNF stimulation (18, 48). However, TNF-induced necrosis is inhibited in Fadd-deficient mouse embryonic fibroblasts (MEFs) (49). This suggests that different necrosis signaling complexes or necrosomes might exist in different cell types. Because both TNF-induced apoptosis and necrosis (in the presence of the pan-caspase inhibitor zVAD-fmk) are blocked in Tradd-deficient cells (50, 51), the TRADD-containing complex II can trigger TNF-induced apoptotic and necrotic signaling pathways (Fig. 2).

Fig. 2

The necrosome, a complex involved in TNF-induced necrosis. After TNFR1 stimulation, TRADD binds RIP1, TRAF2 (or TRAF5), cIAP1, and cIAP2 to form complex I, which activates the NF-κB and MAPK pathways. K63-linked polyubiquitination of RIP1 by cIAPs or of TRAF2 results in interaction of RIP1 with a complex containing TAK1 and TAB2/3. TAK1 activates the IKK complex, resulting in phosphorylation, K48 polyubiquitination, and proteasomal degradation of inhibitor of κB (IκB). NF-κB then translocates to the nucleus and induces transcription. These pathways constitute the gene induction and survival mode after TNF stimulation. In a negative-feedback loop, NF-κB–mediated up-regulation of A20 and CYLD targets RIP1 for K63 deubiquitination, which abolishes its ability to activate NF-κB. After receptor internalization, secondary cytosolic complexes are formed. When the TRADD-dependent complex II is formed, it involves FADD-mediated recruitment and activation of caspase-8 and leads to the apoptotic mode, which is associated with cleavage of RIP1 and RIP3. The alternative RIP1-dependent complex II, which is formed in the presence of Smac mimetics, proceeds independently of TRADD through a RIP1-FADD scaffold to activate caspase-8 in a RIP1-dependent manner. These secondary complexes are assumed to contain mostly nonubiquitinated RIP1. Smac mimetics facilitate cIAP autoubiquitination and proteasomal degradation and so favor the presence of nonubiquitinated RIP1 in complex II. When the activity of caspase-8 is blocked (indicated by the lighter color), RIP1 and RIP3 assemble in a complex with FADD and caspase-8 and, possibly, also TRADD. Mutual direct or indirect phosphorylation of RIP1 and RIP3 in the necrosome activates necrotic signaling. Dashed lines indicate putative steps or interactions.

In some cell types, treatment with zVAD-fmk or a Smac mimetic can induce necrotic cell death in a way that depends on endogenously produced TNF (46, 52). Hence, researchers should be careful when studying cell death pathways induced by chemicals that evoke cellular stress, which often may result in the endogenous production of TNF, which mediates the observed effects. TNF-induced apoptosis in the presence of Smac mimetics and TNF-induced necrosis in the presence of caspase inhibitors involve the CYLD deubiquitinating enzyme (46, 52), which confirms that deubiquitination is essential for the switch from the survival mode to the cell death mode. Indeed, ubiquitination of caspase-8 by a cullin-3–based E3 ligase in the TRAIL receptor 1 (TRAIL-R1) or TRAIL-R2 complex favors the aggregation and activation of caspase-8 (53). Furthermore, A20 can reverse caspase-8 ubiquitination, which suggests that activation of deubquitinating enzymes or the presence of inhibitors of ubiquitinating enzymes keeps caspase-8 inactive. In addition, phosphorylation at Tyr380 inhibits the proteolytic activation of caspase-8 (54). These findings highlight that regulation of caspase-8 recruitment and activation by phosphorylation or ubiquitination could affect the cell death response.

RIP1 and RIP3, the Danse Macabre of Two Kinases Leading to Necrosis

In most models of cell death, RIP1 kinase activity is specifically implicated in necrosis and not in apoptosis, as demonstrated by the use of necrostatins (15). However, RIP1 kinase activity has also been reported to be involved in TNF-mediated caspase-8–dependent apoptosis, at least in the presence of Smac mimetics (46). In addition to RIP1, RIP3 kinase activity is implicated in caspase-independent cell death (47). The observation that zVAD-fmk treatment protects many cell lines against TNF-induced apoptosis but causes others to switch to TNF-induced necrosis (14) raises the intriguing question of which factors determine whether the cells will undergo necrosis in response to TNF. It has become clear that RIP3 causes cells to undergo necrosis upon TNF treatment (42, 43, 55). RIP3 contains an N-terminal kinase domain and a C-terminal RHIM domain, which allows homotypic interaction with RIP1 (Fig. 1). In contrast to RIP1, RIP3 is not required for TNF-induced NF-κB activation (43, 56, 57). Experiments using cells lacking RIP3 (whether by RNA interference or genetic deficiency) show that RIP3 is crucial for TNF-induced necrosis and reveal the existence of a TNF-induced complex containing RIP1 and RIP3, which interact through homotypic RHIM domain interactions that are stabilized by Smac mimetics or caspase inhibitors.

After stimulation with TNF, Smac mimetic, and zVAD-fmk, deubiquitinated RIP1 dissociates from complex I and recruits RIP3, FADD, and caspase-8 (42). RIP3 is autophosphorylated on Ser199 and phosphorylates RIP1 (43). The presence of necrostatin-1 abolishes the formation of the RIP1−RIP3 complex, which suggests that RIP1 kinase activity is involved in necrosome formation. In agreement with this observation, RIP1 autophosphorylation on Ser161 is required for necrosis (15). In addition, RIP3-mediated phosphorylation of RIP1 is also inhibited by necrostatin-1, which suggests that the kinase activity of RIP1 is somehow involved in the regulation of RIP3 kinase activity, which illustrates an intricate relation between RIP1 and RIP3 that leads to necrosis (15, 42). The mechanism by which impaired caspase activity (due to presence of caspase inhibitors or to absence of caspase-8 or FADD) switches the cellular response from an apoptotic to a necrotic mode (20) is likely because caspase-8–mediated cleavage of RIP1 and RIP3 is blocked (19, 47).

Which Signaling Pathways are Downstream of the RIP1 and RIP3 Signaling Complex in TNF-Induced Necrosis?

During TNF-induced necrosis, many subcellular events occur (14); however, it is very difficult to distinguish early signaling events from cellular disintegration events (58). Since the 1990s, TNF-induced necrosis has been linked with energy metabolism. The Fiers group demonstrated that complex I–mediated production of ROS by the mitochondrial electron transfer system is crucial for TNF-induced necrosis (12, 59). This indicates that TNF signaling affects mitochondrial ROS production or ROS scavenging mechanisms (or both). In addition, these mitochondria-derived ROS are responsible for the ultrastructural changes that occur in the mitochondria and endoplasmic reticulum during cell death (14, 60). RIP1-dependent recruitment of membrane-associated NADPH (nicotinamide adenine dinucleotide phosphate, reduced form) oxidase has also been implicated in TNF-induced ROS production (61). After TNF stimulation, the regulatory NADPH oxidase organizer 1 (NOXO1) subunit associates with RIP1, TRADD, and Rac, and initiates ROS production by the cell membrane–associated NADPH oxidase NOX1 (61). These ROS provoke prolonged JNK activation, which contributes to TNF-induced necrotic cell death (61). The activation of NADPH oxidase is dependent on riboflavin kinase (RFK), which binds to both the TNFR1 death domain and to p22phox, the common subunit of NADPH oxidase isoforms (62). Whether RIP3 is also involved in this type of TNF-induced ROS production has not been investigated.

Although ROS production is not essential for TNF-induced necrosis in all cases (42, 63), RIP3 kinase activity links TNFR1-associated events, bioenergetics, and increased ROS production in several cell types. In cells treated with TNF and zVAD-fmk, and thus undergoing necrosis, the RIP3 complex contained seven metabolic enzymes (55), including the cytosolic enzymes glycogen phosphorylase (PYGL) and glutamate ammonia ligase (GLUL; also called glutamine synthase). PYGL catalyzes the degradation of glycogen to glucose-1-phosphate, which is subsequently converted into glucose-6-phosphate, which directly fuels glycolysis (64). GLUL catalyzes the condensation of glutamate and ammonia to form glutamine, which stimulates glutathione production, gluconeogenesis, and lipogenesis, though it is unknown whether these processes contribute to necrotic cell death. It is more likely that this amino acid undergoes glutaminolysis, because RIP3 also interacts with an enzyme of this pathway, namely the mitochondrial matrix–localized glutamate dehydrogenase 1 (GLUD1). Glutaminolysis increases the concentration of α-ketoglutarate, which then feeds into the tricarboxylic acid cycle (TCA) (65) (Fig. 3). In mitochondria, catabolism of glutamine also increases the ammonia load in these organelles, which in turn leads to increased ROS production (66). In addition, a direct link between glutaminolysis and the generation of cytotoxic ROS during TNF-induced necrosis has been established (67). Wild-type, but not kinase-dead, RIP3 enhances the activity of PYGL, GLUL, and GLUD1 in vitro and in cells (55), which suggests that these metabolic enzymes could be direct substrates of RIP3. Knockdown of PYGL, GLUL, or GLUD1 partially reduced the degree of TNF- and zVAD-fmk–mediated ROS production and necrosis. A role for glycolysis and glutaminolysis in TNF-induced necrosis under oxygen-rich conditions could explain why ischemia-reperfusion (IR) injury probably includes a necrotic component, as has been elegantly illustrated by the protective activity of necrostatin in models of myocardial or cerebral infarction (2, 28).

Fig. 3

Different intracellular signaling events contribute to TNF-induced necrosis. The kinase activities of RIP1 and RIP3 are needed for necrosis induced by death receptors and other stimuli. After they are recruited to complex II (Fig. 2), RIP1 and RIP3 undergo reciprocal auto- and trans-phosphorylation. RIP3 binds to and enhances the activity of three metabolic enzymes: GLUL, GLUD1, and PYGL. These metabolic enzymes eventually stimulate the TCA cycle and oxidative phosphorylation, which results in enhanced ROS production at complex I of the electron transport chain. Necrotic death in some cell types is associated with depletion of ATP, an event that could be due to RIP1-dependent inhibition of adenine nucleotide translocase (ANT) or activation of PARP1. RIP1-dependent activation of NOX1 also contributes to TNF-induced ROS production, which is dependent on RFK. α-KG, α-ketoglutarate; SDH, succinate dehydrogenase. Dashed lines indicate putative steps or interactions.

Taken together, these results are consistent with a model in which an excessive increase in energy metabolism is part of the execution mechanism of TNF-induced necrosis, at least in some cells. In an immunological context, the release of adenosine triphosphate (ATP) and other danger-associated molecular patterns (DAMPs) from cells dying by necrosis could serve as an alarm signal for cells or organs (16). Necrosis could also be a consequence of the function of TNF as a major antimicrobial cytokine during infection. In that capacity, TNF stimulates bioenergetics in order to fuel ATP-consuming enzymatic systems, such as NADPH oxidase, in the fight against infections (61). The protective effects of necrostatins in experimental models of neurodegenerative diseases, ischemic brain and heart injuries, and head trauma (2, 28, 68, 69) suggest that TNF produced as a consequence of cellular stress might contribute in an autocrine or a paracrine way to stimulate metabolism. It is possible that the primary goal of RIP3 signaling is to increase the cell’s energy stores to promote cell survival or repair following toxic insult, injury, or infection. This could imply that pathological necrosis is a by-product resulting from overstimulation of RIP1 or RIP3 (or both), similar to some models proposed for cell death occurring by excessive autophagy (3).

Glycolysis not only provides mitochondria with substrates to feed the respiratory chain, it is also the main biochemical pathway for the production of the cytotoxic dicarbonyl methylglyoxal (70). Production of methylglyoxal is not only inevitable but also proportional to glycolytic flux. The proposed role of RIP3 in stimulating glycogenolysis during necrotic cell death (55) is in line with the observed rise in methyglyoxal concentrations (71). Methylglyoxal binds covalently to proteins to form advanced glycation end products (AGEs), which create centers of sustained ROS production that alter the function of the protein involved (Fig. 3). Mitochondrial proteins are apparently especially prone to this methylglyoxal-mediated posttranslational modification (72). Methylglyoxal-derived AGEs are also formed during TNF-induced necrosis of L929 fibrosarcoma cells. Because inhibition of glycolysis inhibits cell death, whereas inhibition of methylglyoxal detoxifying pathways accelerates it, it is likely that these AGEs act as signaling molecules and might be sensed by intracellular receptors (71).

During TNF-induced necrosis, ATP-consuming mechanisms apparently remain intact, whereas in apoptosis they are immediately targeted by caspases. Examples of such processes are translation (73), proteasomal degradation (74), and poly(adenosine diphosphate–ribose) polymerase-1 (PARP-1) activity (75). PARP-1 is a nuclear enzyme involved in DNA repair, DNA stability, and transcriptional regulation (76, 77), and its activation is substantially increased in TNF-treated L929 cells (78), probably by mitochondrial ROS-induced DNA damage. But it has also been reported that PARP-1 hyperactivation causes mitochondrial dysfunction and prompts JNK activation, both of which can enhance necrotic cell death (79, 80). The activation of PARP-1 also leads to ATP depletion, a condition that is permissive for necrosis (81, 82). In agreement with this notion, PARP-1 inhibition blocks TNF-induced necrosis (78). In view of the dependence of TNF-induced ROS generation on RIP1, TNF-induced PARP-1 activation might be governed indirectly by RIP1 (15). It is surprising that both Ripk1-deficient and Traf2-deficient MEFs are resistant to PARP-1–induced cell death in response to the DNA alkylating agent N-methyl-N′-nitro-N-nitrosoguanidine (80), which indicates that RIP1 activation can also occur downstream of PARP-1 activation. A direct link has been postulated for RIP1 in the reduction of ATP concentrations and consequent necrotic cell death, because TNF induces RIP1-dependent inhibition of adenine nucleotide translocase (ANT) (83).

Another pathway involved in necrotic cell death is arachidonic acid (AA) metabolism. Phospholipase A2 (PLA2) is an esterase that produces AA from arachidonate-containing phospholipids (84). Treatment of L929 cells with TNF leads to the activation of PLA2, and overexpression of cytosolic PLA2 (cPLA2) sensitizes TNF-resistant L929 variants to TNF-induced necrosis (85). Resistance to TNF-induced cytotoxicity in MCF7 breast cancer cells correlates with cleavage of cPLA2 (86). AA is further converted by lipoxygenase (LOX) into free, cytotoxic lipid hydroperoxides (87), which further damage intact biological membranes (Fig. 3). In addition, lysosomes lose membrane integrity upon lipid oxidation, which results in lysosomal membrane permeability (LMP) (88). A role of this cPLA2-LOX pathway in necrosis was demonstrated in several in vitro and in vivo models (63, 85, 86, 89, 90).

Other mediators of LMP in TNF-induced necrotic signaling include acid sphingomyelinase (aSMase) and calcium. As a result of sphingomyelin catabolism, aSMase produces ceramide (a lipid second messenger) (91). RIP1 seems indispensable for activating aSMase and subsequent generation of ceramide (92). The specific role of ceramide in the necrotic cell death process remains unknown but might involve lipid peroxidation and ROS production (93). A connection between calcium homeostasis and LMP was indicated by the observation that TNF induces a moderate increase in intracellular Ca2+ concentrations, resulting in increased numbers and size of lysosomes (94). These oversized lysosomes experience LMP and induce collapse of the plasma membrane, a parameter often measured to monitor cell death (95). Cells depleted of plasma membrane calcium adenosine triphosphatase (ATPase) 4 (PMCA4), which extrudes Ca2+ from cells, have high intracellular concentrations of Ca2+. Sustained high intracellular Ca2+ load induces exocytosis of lysosomes, which prevents the TNF-induced build-up of oversized lysosomes and, thus, necrosis (94).

To better understand necrotic signaling, it is essential to distinguish the signaling events from events that are part of the destruction phase. The challenge for the necrosis research community will be to determine whether or how these TNF-induced signaling events are mechanistically linked to each other or to the signals emitted from the RIP1 and RIP3 necrosome signaling complex. Identification of substrates of RIP1 and RIP3 will be helpful in reaching these goals.

In Vivo Role of RIP1- and RIP3-Mediated Necrosis in Pathophysiology

Necrosis in pathological cell death might occur when activation of caspases is impossible—for example, because of changes in the redox status of the catalytic cysteine residue (96) or when caspase inhibitors are present. The latter is observed during infection with viruses that produce proteins that inhibit caspase-8 recruitment to the DISC (97). Necrosis occurs during activation-induced cell death (AICD) of T cells and in TNF-treated MEFs, both infected with vaccinia virus (VV) that encodes the viral caspase inhibitor B13R (also known as Spi2) (43). The requirement for RIP3 in induction of cell death in VV-infected cells suggests that, in vivo, the TNF-mediated protection against VV infection (98) could act through RIP3-dependent necrosis. Accordingly, Ripk3 knockout mice have high viral titers and succumb to VV infection, whereas wild-type mice are protected (43). This illustrates the dual role of necrosis in vivo: (i) as a backup mechanism to eliminate virus-infected cells if the apoptotic pathway is blocked and (ii) to release DAMPs, which, together with the viral antigens, act as endogenous adjuvants to boost the innate and adaptive immune system (16). These results also imply that viruses produce inhibitors of necrotic cell death during infection in order to evade the immune system. Indeed, the murine cytomegalovirus (CMV) expresses M36 protein, which inhibits death receptor–induced caspase-8 activation (99), and M45 protein. M45 is a RHIM-containing adaptor that interacts with RIP1 and RIP3 (100) and prevents TNF-induced activation of NF-κB and p38 MAPK, as well as caspase-independent cell death in CMV-infected cells (101). Thus, viruses have developed different strategies to block both apoptotic and necrotic cell death so they can propagate in host cells.

Ripk3 knockout mice also show decreased necrosis in a model of cerulein-induced pancreatitis (42, 55). Necrosis occurs in other medical conditions, such as ischemia-reperfusion damage during organ transplantation, cardiac infarction, stroke, and traumatic brain injury, and thus points to the possibility of developing kinase inhibitors of RIP1 and RIP3 to prevent pathological cell death. This has already been initiated with the development of the necrostatin RIP1 inhibitors (15). Moreover, stimulating the RIP1- and RIP3-dependent necrotic cascade could form an alternative cell death pathway that bypasses antiapoptotic mechanisms of cancer cells and stimulates the immune system by immunogenic DAMP signals (16).


In conclusion, the pleiotropic character of TNF has long challenged molecular cell biologists seeking to identify how all these different outcomes are signaled by the same molecule. A new model proposes that the TNF-induced biological response—gene induction and survival mode, apoptotic mode, or necrotic mode—is regulated by caspases and the ubiquitin editing system. Deubiquitination of RIP1 favors the switch from the gene induction and survival mode to a cell death mode. The next decision to be taken is the induction of apoptotic or necrotic cell death. Evidence supports the idea that caspase-8 activation or its absence is the main determinant of whether the cell will undergo apoptosis or necrosis. Furthermore, necrotic signaling is blocked during apoptosis by the caspase-mediated cleavage of RIP1 and RIP3. Rather than being a molecular switch between apoptosis and necrosis, the presence of RIP3 constitutes a permissive condition for the cell to undergo necrosis. RIP1- and RIP3-mediated necrosis is associated with increased alterations in carbohydrate metabolism and glutaminolysis, leading to increased ROS production mediated by AGEs and by complex I, which is part of the cytotoxicity mechanism in some cell types. However, it remains a challenge to link the RIP1-RIP3 signaling complex to downstream signaling events associated with or leading to necrotic cell death. The experimental use of necrostatins to target RIP1 kinase activity and the future development of RIP3 inhibitors might be applicable to various pathologies. Thus, TNF, as a multifaceted molecule, has challenged the scientific community to develop new concepts with important therapeutic implications.


Acknowledgments: We thank A. Bredan for editing and N. Takahashi for critically reading the manuscript. We apologize to the many authors whose original work could not be cited due to reference limitations. Funding: This research has been supported by VIB, Ghent University, Framework Programme 6 (FP6) ApopTrain, MRTN-CT-035624; FP6 Epistem, LSHB-CT-2005-019067; Apo-Sys FP7-200767, IAP 6/18, Research Foundation Flanders (FWO-Vlaanderen) (3G.0218.06 and G.0226.09) and Flemish Special Research Fund—Concerted Research Actions BOF-GOA 12.0505.02. T.V.B. holds a grant of the FWO and P.V. holds a Methusalem grant from the Flemish Government.

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.
View Abstract

Stay Connected to Science Signaling

Navigate This Article