TNFα (tumour necrosis factor α) is an extensively studied pleiotropic cytokine associated with the pathogenesis of a variety of inflammatory diseases. It elicits a wide spectrum of cellular responses which mediates and regulates inflammation, immune response, cell survival, proliferation and apoptosis. TNFα initiates its responses by binding to its receptors. TNFα-induced effector responses are mediated by the actions and interactions among the various intracellular signalling mediators in the cell. TNFα induces both survival and apoptotic signal in a TRADD (TNF receptor-associated DD)-dependent and -independent way. The signals are further transduced via a variety of signalling mediators, including caspases, MAPKs (mitogen-activated protein kinases), phospholipid mediators and miRNA/miR (microRNA), whose roles in specific functional responses is not fully understood. Elucidating the complexity and cross talks among signalling mediators involved in the TNFα-mediated responses will certainly aid in the identification of molecular targets, which can potentially lead to the development of novel therapeutics to treat TNFα-associated disorders and in dampening inflammation.
TNFα (tumour necrosis factor α) is a pleiotropic cytokine extensively studied for its role in the pathogenesis of a variety of disease conditions, and is known to have a wide range of beneficial and deleterious effects in humans [1,2]. TNFα is produced by a variety of cells, which include: macrophages, monocytes, lymphocytes, NK cells (natural killer cells), eosinophils, keratinocytes, Langerhan cells, Kupffer cells, glial cells, adipocytes and fibroblasts [1–3]. This cytokine is known to be produced in response to a wide range of stimuli such as, bacterial toxins [LPS (lipopolysaccharide)], infections (bacterial, viral, fungal, mycobacterial and parasitic), antigen–antibody complexes, injury, host inflammatory agents (products of the complement activation, auto-antibodies and cytokines), as well as toxic and non-toxic environmental challenges [1,3]. TNFα elicits a wide spectrum of cellular responses, which mediates inflammation, regulates immune response and also induces apoptosis in certain types of cancer cells [4,5]. It brings about its immuno-modulatory and inflammatory functions by inducing the production of various cytokines, activation of leucocytes and lymphocytes, enhancing the expression of adhesion molecules, initiating adherence of neutrophils and monocytes to the endothelium, promoting inflammatory cell migration, and it is also known to stimulate cellular proliferation and differentiation [1–3,5,6]. Appropriate levels of TNFα are necessary for homoeostatic functions such as protection from infection, haematopoiesis, immune response regulation, tumour regression and immune surveillance . Dysregulation in TNFα production or signalling has been associated with a range of inflammation-based disorders [1,2,4,7–9].
The observations by Dr P. Bruns on spontaneous regression of tumours in patients with acute bacterial infections  and the follow-up study by William B. Coley on tumour reduction by the use of bacteria-free filtrate preparation from bacterial extracts , more than a century ago, have been the real building blocks that led to the discovery of TNF. This factor found in the serum, which is capable of causing regression of certain tumours, was termed tumour necrosis factor (TNF) in 1975 [12,13]. A decade later, successful attempts to isolate and characterize this factor led to the identification of two structurally related factors, which were termed TNF or TNFα and LT (lymphotoxin) [14–16]. Here, TNFα will be discussed in detail.
TNFα brings about its varied functions and responses through the ligand–receptor complex formation mechanism, where TNFα is a ligand. It is a homotrimer of 157 amino acid subunits  and is a type II transmembrane protein . It is present as a biologically active mTNF (membrane bound TNF) and sTNF (soluble bound TNF) forms. Soluble TNFα is originally expressed as a 26 kDa membrane bound pre-protein, which can be cleaved by a metalloproteinase [TACE (TNFα converting enzyme] to be released as a 17 kDa mature form of TNFα [17–21]. The secreted, mature TNFα subunits form non-covalently bound homotrimers: sTNF [5,6]. This trimerization is necessary for the TNFα to be biologically potent or activate its receptors . The membrane-bound form of a TNFα ligand is useful in the transfer of signals by cell-to-cell contact and the effect is said to be localized, whereas the soluble form of TNFα is said to have a more dispersed or systemic effect . TNF will be used to represent TNFα in the following sections.
TNFRs (TNF RECEPTORS)
TNFRs are a family of transmembrane proteins characterized by their cysteine-rich extracellular domain sequence homology. There are more than 27 members in the TNFR family, which includes TNFR1 (TNFR type 1), TNFR2 (TNFR type 2), LTβ receptor, Fas, CD40, OX40, TRAIL (TNF-related apoptosis-inducing ligand) receptors, RANK [receptor activator of NF-κB (nuclear factor κB)], DRs (death receptors), decoy receptors, BAFFR (B-cell-activating factor receptor) and many more . TNFR1 and TNFR2 are the two main receptors that specifically interact with the TNFα ligand and will be the focus of attention in the present review. It should also be noted that the ligand–receptor interaction between the other TNF family members also contributes towards inflammatory and immune responses [23,24]. TNFα ligand induces its different cellular responses by binding to TNFRs TNFR1 (TNFR1/p55TNFR/CD120a) and/or TNFR2 (TNFR2/p75TNFR/CD120b) [4,5,25–27]. The binding process induces receptor trimerization and/or recruitment of signalling proteins to the cytoplasmic domains of the TNFRs and the subsequent cellular responses depends on the type of receptor bound to the ligand. Both receptors contain a PLAD (pre-ligand binding assembly domain) that pre-complexes the receptors and encourages them to trimerize, particularly upon activation by the TNF ligand . TNFR1 can equally be activated by both sTNF and mTNF, whereas mTNF is considered to be superior to sTNF in activating TNFR2 [29,30]. TNFα activates TNFR1 with higher efficiency due to the fact that it binds to TNFR1 with increased affinity (Kd of 19 pM) and exhibits a slow rate of dissociation (t1/2=33 min) from the receptor, subsequent to binding . These kinetics are different for TNFR2, where the affinity of TNF to TNFR2 is lower and the dissociation rate is faster compared to that of TNFR1 . TNFR1 is said to be constitutively expressed in a vast majority of cell types and tissues including immune cells, whereas TNFR2 expression is mainly found in immune cells [4,32]. TNFR1 is considered to mediate most of the biological effects and responses induced by TNF , and TNFR1 is the most widely activated and studied TNFR in an experimental set-up, due to the use of soluble recombinant TNF, which predominantly activates TNFR1. A novel isoform of TNFR2 (icTNFR2) reported recently is thought to have originated from TNFR2 gene as a new transcriptional start followed by alternative splicing of the transcript and it was found to be expressed only within the cell . Although overexpression studies have suggested that icTNFR2 signalling is similar to that of TNFR2 , more work has to be done to evaluate the relevance of icTNFR2 compared with that of the well-studied TNFRs. TNFR lacks intrinsic enzymatic activity and it brings about its functions by the recruiting different adaptor molecules, which in turn transduces signals for a range of responses, such as apoptotic cell death, inflammation, cellular survival and proliferation.
The binding of the ligand, TNF to TNFR1 leads to conformational changes in the intracellular domain of the TNFR, which initiates the recruitment and subsequent interaction of several adapter proteins to the receptors . TNFR1 contains a DD (death domain) motif towards the carboxy-end of the receptor and is critical in the death-inducing and survival or inflammatory activity of TNFR1 . The DD of a protein can associate with the DD of other proteins. SODD (silencer of DD), an adaptor protein, is bound to TNFR1 through its DD, thereby preventing the binding of other molecules with DD to the TNFR . In this manner, SODD acts as an endogenous inhibitor of TNF signalling. Receptor ligation of TNFR1 initiates internalization of the ligand–receptor complex, which is another protective mechanism present naturally to disrupt any sustained signalling subsequent to receptor activation [37,38]. Receptor ligation leads to the disassociation of SODD from TNFR1, thereby allowing other DD-containing proteins to interact with the DD of TNFR1. This event results in the recruitment of a DD-containing signal transduction adapter molecule, TRADD (TNF receptor associated DD), which is said to mediate different signalling pathways by interacting with other related adaptor molecules. A schematic representation of the TNFα-induced intracellular signalling pathway is illustrated in Figure 1.
TNFα-induced intracellular signalling events
TRADD-dependent signalling is unique to TNFR1 and not TNFR2. This is due to the absence of DD in TNFR2. TRADD plays a unique role in distinct signalling pathways such as the pro-apoptotic or survival pathways by forming a signalling complex with FADD (Fas associated DD) and TRAF2 (TNF receptor-associated factor 2), respectively [39,40].
Intracellular signalling mediators regulating TNF-induced pro-apoptotic signalling discussed here are primarily FADD and JNK (c-Jun N-terminal kinase). The role of RIP (receptor interacting protein) and cPLA2 (cytosolic phospholipase A2) in this process is also discussed.
TNFRs also known as DRs are known to induce death signals through the extrinsic apoptotic pathway using an initiator caspase (caspase 8) [41–44] and TRADD. Subsequent to TNFR1 receptor activation, the TNFR1–TRADD complex recruits another DD containing protein FADD, which in turn activates the procaspase 8 and 10, which eventually leads to DNA damage and PCD (programmed cell death) [45,46]. FADD recruitment of procaspase 8 is facilitated by protein–protein interaction through their DEDs (death effector domains), thus leading to the formation of a DISC (death-inducing signalling complex) along with TNFR1 and TRADD. Procaspase 8 is then autolytically cleaved to release an active form of caspase 8 from the DISC, which then acts on downstream effector caspases, especially caspase 3. These effector caspases cleave or lead to the degradation of a number of substrates necessary for normal cell homoeostasis, thus resulting in morphologic and biochemical characteristics of apoptotic cell death. Caspase 8 also cleaves Bid (BH3-interaction DD agonist, a member of Bcl-2 family) to form t-Bit, which translocates to mitochondria causing the release of cytochrome c . This is followed by the formation of apoptosome—a complex made up of cytochrome c, Apaf-1 (apoptotic protease-activating factor 1) and procaspase 9, which is capable of activating effector caspases such as caspase 3 to bring about PCD . It was also found that activated caspase 3 can in turn activate the cleavage of procaspase 8 to caspase 8 and thereby creating a positive feedback system .
JNK, an important member of the MAPK (mitogen-activated protein kinase) family, influences a variety of cellular functions including apoptosis . JNK is known to be activated by a variety of stimuli-like cytokines, UV irradiation, growth factors, serum and DNA damaging agents. TNFα-induced prolonged activation of JNK has long been associated with the induction of apoptosis or necrosis. It is stated that TNFR2 activates JNK and not other MAPKs . TRAF2 was identified to be essential in the activation of JNK [52,53]. Efforts to elucidate the mechanism of TNFα-induced JNK activation has led to the identification of the role of ASK1 (apoptosis signal-regulating kinase 1) in the process . TRAF2 brings about JNK activation by influencing ASK1, a MAPKKK (MAPK kinase kinase)  and GCK (germinal centre kinase) . ASK1 activates or phosphorylates MKK4 (MAPK kinase 4) and MKK7 leading to the activation of JNK . TNFα activation of JNK is brought about in two distinct phases. There is TNFα-induced TRAF2 mediated early and transient JNK activation , which is followed by a delayed and persistent activation of JNK mediated by ROS (reactive oxygen species) induced by TNF . It was also reported that ROS-mediated JNK activation is followed by ROS production mediated by JNK, which exhibits a positive feedback of ROS production and JNK activation , which would contribute to the induction of apoptosis. TNF-induced ROS was found to facilitate the activation of ASK1 by oxidizing thioredoxin and its dissociation from ASK1. This is followed by oligomerization and phosphorylation of ASK1 and subsequent JNK activation, leading to apoptosis [60,61]. A more recent study has shown that, in addition to the activation of ASK1, TNF-induced increased ROS production inhibited the function of MAPK phosphatase thereby leading to a prolonged or persistent JNK activation resulting in cell death . An elegant report from Shen and Pervaiz  can be referred to for a detailed discussion on the role of ROS in TNF-induced JNK activation and associated cell death. All these reports indicate the role of ROS as a co-activator in TNF-induced JNK-mediated cell death. JNK is well known to induce c-Jun-dependent transcription leading to AP-1 (activator protein 1) activation [57,64] and it also phosphorylates transcription factors such as ATF2 (activating transcription factor 2), NF-AT (nuclear factor of activated T-cells), HSF (heat-shock factor) and STAT3 (signal transducer and activator of transcription 3) to bring about it various cellular responses .
RIP was found to be associated with TNF-induced apoptosis . It was found that caspase 8 cleaves RIP to form RIPc and nRIP. It has been reported that RIPc blocks TNF-induced survival signalling and the association of RIPc with TRADD was associated with TNF-induced apoptosis . Subsequent to cleavage, RIP was found to dissociate from the TRADD, thus enabling the recruitment of FADD and thereby indicating its role in TNF-induced apoptosis .
cPLA2 has been associated with TNFα-induced apoptosis [67–70]. On the contrary, some earlier studies reported that cPLA2 and sPLA2 mediates TNF-induced NF-κB activation in human keratinocytes [71,72]. However, other studies have clearly shown and emphasized that cPLA2 protein and its activity is vital for TNF-induced cell death in a variety of cell types [73–76]. cPLA2 activated by TNFRs [77–79] leads to the generation of arachidonic acid , which is further converted to form eicosanoids (prostaglandins and leucotrienes). cPLA2-based eicosanoids induces the generation of ROS, which in turn plays a role in determining cell fate . TNFR1 is thought to play a role in the activation of cPLA2, whereas TNFR2 is said to be involved in regulating the expression of cPLA2 [5,78].
Survival or inflammatory
TNFR1 facilitates a survival or an anti-apoptotic signal through the recruitment of TRAF to the receptor signalling complex through TRADD. The TRAF family of proteins are involved in a range of cellular responses such as survival, inflammation and anti-apoptosis . Of the seven members of the mammalian TRAF family, TRAF2 is well studied and was found to be associated with TNFR1 and TNFR2 signalling . TRADD was found to be specific for TRAF1 and TRAF2. It has also been reported that TRAF2 has a higher affinity for TRADD than the TNFRs  and this indicates the predominant role of TRAF2 in TNFR1 signalling. Along with TRAF2, RIP has also been found to be recruited by TRADD to the receptor signalling complex. RIP was found to be necessary in TNFα-mediated survival signals [39,84] and is also involved in TNF-induced activation of NF-κB, an important transcription factor . TRAF2 and TRAF5 together are also necessary in TNFα-induced NF-κB activation . Intracellular signalling mediators other than NF-κB playing a crucial role in pro-survival and/or anti pro-apoptotic signalling discussed here are IAPs (inhibitor of apoptosis proteins), PLC (phospholipase C), PLD (phospholipase D), PKC (protein kinase C), sphingolipids, p38 MAPK and PKB (protein kinase B). The other survival mediators like ERK1/2 (extracellular-signal-regulated kinase 1/2), TNFR2 and FAN [factor associated with NSmase (neutral sphingomyelinase)] are discussed in the following TRADD-independent signalling section.
TNF brings about survival responses predominantly by the activation of NF-κB, an important anti-apoptotic and pro-inflammatory transcription factor. NF-κB is known to play an important role in regulating the expression of genes responsible for mediating immune and inflammatory responses [87,88]. Activation of cell adhesion molecules on endothelial cells by TNF, an important event in inflammatory response is mediated specifically by the TNFR1-induced NF-κB pathway rather than the various MAPK pathways . If NF-κB signalling is blocked, TNFα can bring about rapid PCD in a variety of cells , which indicates the importance of NF-κB in the cellular response specificity, when induced by TNF. The canonical or the classical NF-κB pathway  is considered to be the predominantly triggered pathway in most cells and stimuli. Binding of TNF to TNFR1 is known as the potent activator of the classical NF-κB pathway. In this pathway, NF-κB is present in the form of a dimer made up of p50 and p65 subunits. In unstimulated cells or in cells prior to activation, NF-κB is sequestered in the cytoplasm in an inactive form and its DNA-binding ability is prevented by IκB (inhibitor of nuclear factor κB) family of inhibitor proteins . On TNF stimulation, there is a rapid activation of the IKK (IκB kinase) complex . The IKK complex consists of two catalytic subunits, IKKα and IKKβ, and a regulatory subunit, NEMO (NF-κB essential modulator) . Subsequent to TNFα stimulation, TRAF2 is suggested to interact with IKKα and IKKβ recruits the IKK complex to the receptor complex . RIP is said to help in the shuttling of the IKK complex to the receptor complex by binding to NEMO  and this role of RIP independent of TRAF2 is necessary for the activation of the IKK complex . Activation of the IKK complex is followed by the ubiquitination and proteosomal degradation of IκBα within ~10 min . TRAF2 was also found to be capable of interacting with NIK (NF-κB-inducing kinase), which is reported to phosphorylate IκB and which leads to the latter's degradation [96,97]. TNF-induced degradation of IκBα triggers the activation of NF-κB by the release of the dimer to translocate from the cytoplasm to the nucleus and to initiate the activation of its target genes by binding to the DNA [92,98].
It was originally thought that NF-κB was a potent inhibitor of JNK activation induced cell death and this was mediated by the activation of NF-κB target genes like XIAP (X-chromosome linked inhibitor of apoptosis), GADD45β (growth-arrest and DNA-damage-inducible protein 45β) and cFLIP (cellular FLICE inhibitory protein) [99–101]. This effect was challenged when reports showed that JNK activation was not affected despite the disruption of XIAP and GADD45β [102,103]. This fuelled the search for a mediator or mediators to enable the cross-talk between NF-κB and JNK following TNFα stimulation. This led to the identification of ROS in the interplay between TNFα-induced cell survival and cell death [104,105]. It is now known that NF-κB brings about the regulation of ROS through Mn-SOD (manganese superoxide dismutase) and FHC (ferritin heavy chain) [106,107]. The antioxidant properties of Mn-SOD and FHC aid in the elimination and prevention of ROS build up in TNF-treated cells [106,108]. This event prevents the ROS-mediated JNK activation. For further information on the role of ROS in TNF-induced NF-κB activation and pro-survival signalling, one can refer to a well-discussed review by Shen and Pervaiz . PKB activated via PI3K (phosphoinositide 3-kinase) signalling and GSK3β (glycogen synthase kinase 3β) was also found to be necessary in the transcriptional activity of NF-κB, thereby preventing TNF-induced apoptosis [109–111]. Therefore, NF-κB brings about its anti-apoptotic, pro-survival signalling by activating key apoptosis inhibitory proteins and also by inhibiting ROS-mediated JNK activation and subsequent cell death.
TNFR1 also mediates survival by regulating or controlling its pro-apoptotic signalling by the recruitment of IAPs, cIAP1 and cIAP2 by TRAF2 to the TNFR1 signalling complex [112,113]. cIAP1 and cIAP2 are proteins that inhibit apoptosis by inhibiting caspase 3 and caspase 7  and proteosomal degradation of caspase 3 and caspase 7 . Therefore, in TNFR1 signalling, TRAF2-recruited cIAPs can effectively inhibit caspase 8, which is recruited to the TNFR1-signalling complex by the TRADD–FADD complex . This would impede the signalling for PCD and facilitate cell survival.
Lipases are involved in the transduction and regulation of TNF-mediated signals and responses. PLC is a well-known phospholipase that acts on PIP2 (phosphatidylinositol 4,5 bisphosphate) to generate DAG (diacylglycerol) and IP3 (inositol 1,4,5-trisphosphate). The activation of PC-PLC (phosphatidyl-choline-specific PLC) leads to the production of choline and DAG, a potent secondary messenger involved in signal transduction. PLC is known to be activated on TNFR stimulation [116–118] and the resulting second messengers such as DAG are extremely relevant in the downstream signal transduction events, like the activation of PKC .
TNF signalling was also found to be associated with the activity of another lipase, PLD (phospholipase D). However, conflicting roles were reported about the role of apoptotic and survival role PLD in TNF signalling [118,120–124]. PLD is a phosphodiesterase that hydrolyses PC to form PA (phosphatidic acid), a potent second messenger and a relatively inert choline. PA can be further converted to DAG by PA phosphodehydrolase. PLD has two major isoforms (PLD1 and PLD2) and a few splice variants of the isoforms have been identified so far. The role of PLD as a key intracellular signalling molecule was extensively discussed in various reviews [125–128]. PLD was shown to be involved in a range of immune cell signalling and responses. A detailed review by Melendez and Allen  on the role of PLD in immune functions can be referred to for further reading. TNF-induced PLD was found to influence ERK1/2 phosphorylation and p38 MAPK in neutrophil-like HL 60 cells . Our results indicate TNF-induced PLD activity and PLD1 translocation from the cytosol to the cell membrane in monocytic cells . Our results also indicate the importance of PLD in selective activation of ERK1/2 rather than p38 MAPK and the role of PLD in TNF-induced ERK activation by influencing Raf translocation [130,131], is in concordance with the established ERK activation pathway discussed by Andresen et al. . We have also found that PLD is necessary in TNF-induced NF-κB activation in monocytic cells [130,131]. This is in acceptance of the report that implicates PLD in TNF-induced NF-κB activation in immature acute myeloid leukaemia cells . The mentioned data set also indicates that PLD is necessary for TNF-induced IκBα degradation, a key event in the NF-κB activation pathway . Our in vitro and in vivo experiments clearly indicate that TNF-induced pro-inflammatory responses are dependent on PLD1 [130,131]. All these findings indicate an anti-apoptotic and inflammatory role of PLD in TNF-induced signalling and responses. Recent studies were more conclusive and detailed about the role of PLD in TNF signalling. The apoptotic and survival role of PLD in general including those induced by TNF are well discussed by Nozawa  and a recent report has indicated a protective role of PLD against TNF-induced apoptosis . ROS, which was found to be a key regulator in TNF-induced apoptosis and cell survival, was also found to play a role in the activation of PLD in endothelial cells . All these facts strongly urge that future studies should address the need to understand the role and interaction of PLD with other TNF signalling mediators.
The role of PKC in TNF-induced signalling and responses in a variety of cells was well summarized by Schutz et al. . TNF-induced PKC activation is considered to be via DAG produced by PLC rather than by PLA2 or DAG produced by PLD. Since earlier kinetic studies have shown that TNF-induced PC-PLC activation precedes PKC activation and that TNF-induced PLA2 and PC-PLD activation is a delayed response occurring much later than the PKC activation . A variety of PKC isoforms exist with distinct activation patterns and distributions. Some of the isoforms such as PKCα, β, δ and atypical isoforms like PKC ζ, λ, τ have been implicated in TNF-induced responses and signalling events such as the activation of MAPKs and NF-κB [137–142]. An atypical PKC-associated protein, p62, interacts exclusively with RIP to facilitate the involvement of a PKC to the TNF-induced NF-κB pathway .
The sphingolipid-based sphingomyelin pathway is known to be utilized in TNF-mediated signal transduction. TNF stimulates both NSmase (neutral sphingomyelinase) and acidic SMase (sphingomyelinase) [143–146]. TNF-induced NF-κB activation is said to be via the endosomal acidic SMase . Studies have proved that PC-PLC-derived DAG is involved in the activation of acidic SMase, a C-type phospholipases hydrolyses sphingomyelin to produce ceramide [145,147,148]. Ceramide is hydrolysed to form sphingosine, which is then converted into S1P (sphingosine-1-phosphate) by SphK (sphingosine kinase). Evidence has shown that the sphingomyelin pathway is involved in TNF-induced NF-κB activation in neutrophil-like cells . However, the mechanism by which SphK mediates TNF-induced NF-κB activation is still under investigation by our group and others.
TNFR1 was found to stimulate membrane-associated NSmase [117,146] either directly or through the activation of another adaptor protein called FAN (factor associated with NSmose). This pathway involves the hydrolysis of sphingomyelin, a major membrane-associated sphingolipid to ceramide by action of an NSmase and subsequent stimulation of CAPK (ceramide-activated protein kinase) . A study on alveolar epithelial cells has reported that ceramide production by sphingomyelin hydrolysis induced by TNF activation can activate a survival pathway rather than initiating PCD . Ceramide functions as a selective mediator of the cytotoxic or cytostatic effects of TNF by playing a positive feedback role in the activation of NF-κB . Ceramide can also be converted to form sphingosine. The latter is phosphorylated by SphK to form S1P, a potent second messenger that can act both intracellularly and extracellularly through specific receptors. S1P through its receptor can influence a range of responses such as proliferation, differentiation, migration, inflammation and apoptosis. It is also known to play a role in the mobilization of intracellular Ca2+ . There are two isoforms of SphK (SphK1 and SphK2). SphK and its activity were associated with a variety of immune cell signalling and immune responses, including its potential pro-inflammatory role. The role of SphK in immune cell signalling and responses was extensively reviewed by Kee et al.  and Melendez , and can be referred for further reading. SphK activity can be induced by a variety of stimuli. We have shown that Fc receptor ligation, C5a, TNF, LPS and fMLP (N-formylmethionyl-leucyl-phenylalanine) stimulate SphK activity. TNF-induced signalling and effector responses are mediated by SphK1 in human monocytes . We have also shown that SphK is necessary for TNF-induced NF-κB activation, which is essential for the survival and pro-inflammatory signalling . Based on our results, we can now suggest that TNF-induced SphK is mediated by PLD and the role of SphK1 in TNF-mediated inflammatory response in vivo [130,131]. The main effect of SphK is brought about by S1P, which induces mitogenesis, stimulates ERK and survival transcription factors and hinders ceramide-mediated JNK activation. Thus, the fate of the cell depends on the tight balance between SIP and ceramide. However, the effects of S1P in the above-mentioned downstream targets are yet to be determined in TNFα-induced signalling.
A stress-activated protein kinase, p38 MAPK, is also well known to be activated by environmental stress and pro-inflammatory cytokines including TNF. TNF-induced cellular responses were found to be mediated by p38 MAPK and is considered predominantly to mediate a pro-inflammatory response [156,157]. p38 MAPK activates downstream targets like cPLA2, microtubule-associated protein Tau, transcription factors such as ATF1, ATF2, MEFs (myeloid Elf-1-like factors), Elk-1 [ETS (E twenty-six)-like kinase 1] , AP-1  and several MAPK-activated protein kinases . The activation of p38 MAPK by TNF is mediated by its upstream MAPKKKs (ASK1)  and MEKs (MAPK/ERK kinases) (MEK3 and MEK6)  is said to be transduced through TRAF2 . p38 MAPK is activated only by TNFR1 and not TNFR2 . Our report showed that unlike ERK1/2, TNF-induced p38 MAPK activation is not dependent on PLD in monocytes . However, fMLP-activated p38 MAPK was indeed dependent on PLD in neutrophil-like cells . It is known that p38 MAPK is involved in TNF-induced IL-6 production in HeLa cells  and from our studies in human monocytic cells [130,131]. Our studies in monocytes is in concordance with the study in HeLa cells  that p38 MAPK does not influence the phosphorylation of IκBα subunits and the translocation of NF-κB components to the nucleus. However, it is considered that p38 MAPK exhibits its indirect influence on TNF-induced NF-κB activation by affecting the expression of a reporter gene, which is driven by an NF-κB element containing promoter .
The importance of PKB/Akt in the regulation of cell survival responses is by the modulation of signalling molecules like IKK, Raf, ERK, Bad and caspase 9 . PI3Ks generate specific inositol lipids like PIP3 (phosphatidylinositol 3,4,5 triphosphate), which were found to be relevant in the regulation of cell proliferation, survival and differentiation. PIP3 recruits PKB to the cellular membrane on stimulation and the latter is subsequently phosphorylated by PDK1 (3′-phosphoinositide-dependent kinase-1) . PKB is known to play an important role in TNFα-induced stimulation of NF-κB activity, since NIK activation is PKB dependent [163,164]. Recently, TNF-induced PKB and NF-κB activation in endothelial cell survival was reported along with the role of integrin-mediated adhesion contributing to the mentioned response . The destruction of PKB by caspases and the inhibition of PKB by ceramide result in the inhibition of survival signals and favours cell death [166–168]. The anti-apoptotic activity of PKB partly contributes to cyclic nucleotide (cAMP and cGMP)-mediated suppression of apoptosis . PKB activated via PI3K signalling and GSK3β was also found to be necessary in the transcriptional activity of NF-κB, thereby preventing TNF-induced apoptosis [109–111].
It is known that there are proteins which interact with the TNFRs without dependence or association with the DD. TNFR1 also exhibits TRADD-independent signalling, as there are proteins which associate directly with TNFR1. The nature of signalling and resulting responses depends on the type of the associating protein and its influence in the signalling pathways.
ERK1/2 is a classical member of the MAPK family. ERK1/2 activation pathway, like p38 MAPK, is considered as a potential target for cancer therapeutics due to its regulatory effect in cellular proliferation . ERK activation-induced cellular responses are brought about by the phosphorylation of membrane proteins (including Syk, CD120a), nuclear substrates [including NF-AT, Elk-1, c-Fos (protein which is the product of the c-fos gene), c-Myc and STAT3], cytoskeletal proteins and MAPK-activated protein kinases [158,171]. It is known that MAPK are activated by TNF [65,158] and ERK1/2 (p44/p42) were shown to be activated by TNF through their respective MEK (MEK1 and MEK2) phosphorylations . Like p38 MAPK, ERK activation was found to be triggered by TNFR1 only and not TNFR2 . Studies have shown the relevance of ERK in modulating TNF-induced survival and apoptotic signals [173,174]. It was also shown that TNF-induced ERK activation was absent or lower than the stress-activated protein kinases in most of the cells . TNFR1 is thought to bring about its ERK activation through an adaptor protein, MADD (MAPK-activating DD). This protein with low homology to DD was found to interact directly with the DD of TNFR1 and activate the MAPK signalling pathways including ERK . TNFR1 induces the activation of ERK also via Grb2 (growth-factor-receptor-bound protein 2), an adaptor protein that interacts with the PLAP motif of TNFR1 . Grb2 is known to interact with SOS (Son of sevenless), which in turn leads to the activation of Ras, followed by c-Raf which leads to the activation of ERK . However, FAN-derived CAPK is necessary along with Grb2 for the efficient activation of ERK . In our study, we have shown the role of PLD in TNF-induced ERK phosphorylation and the relevance of ERK in TNF-mediated inflammatory response . Besides aiding in ERK activation, MADD was also reported to play a role in TNFR1-induced JNK activation and cPLA2 , thus contributing to the pleiotropic effects induced by TNF, which includes cell survival, inflammatory responses and cell death.
FAN is an adaptor protein found to have the ability to bind to the membrane-proximal region of TNFR1 , and it has been found to aid in the stimulation of NSmase [177–179]. SMase catalyses the sphingolipids to form ceramide, which in turn can influence caspase 3 activation and bring about cell death. It is also capable of activating CAPK . The signalling events following the activation of SMase and the resulting possibility of transducing apoptotic and survival responses have been discussed earlier.
TNFR2-mediated signalling can be categorized as TRADD-independent signalling, since TNFR2 is devoid of DD. TRAF2 is directly associated with TNFR2 to initiate its pro-survival signalling and responses. TNFR2 is also known to induce cell death  and this is thought to be brought about by the induction of endogenous mTNF, which activates TNFR1 [181,182] along with the depletion of TRAF2 by cIAPs. It is also thought that TNFR2 induces apoptotic signals directly by the ligand-passing mechanism, where it increases the local TNF concentration in the TNFR1 region enabling the latter to initiate apoptosis .
TRAPs (TNFR-associated proteins) were found to be associated directly with the regions proximal to the DD in TNFRs to contribute their influence in TNF signalling and responses . TRAP1 [Hsp75 (heat-shock protein 75)] and TRAP2 (p97) were found to regulate the TNFR-associated function by exhibiting DD-independent interactions with TNFRs [184–187]. TRAP1 (Hsp75) is localized in the mitochondria  and its interaction with TNFR1 takes place in the N-terminal half of cytoplasmic domain of the TNFR1 . In relevance to its intracellular localization, it was reported that TRAP1 is either cytosolic  or in the mitochondria . The mechanism by which TRAP1 in the mitochondria interacts with TNFR is yet to be elucidated. TRAP2 (p97) which influences the regulation of proteosomal functions was found to exhibit DD independent association with TNFRs [184,186,187]. p97 is 26S proteasome non-ATPase regulatory subunit 2. 26S proteasome is a multi-subunit protein complex that promotes the non-lysosomal degradation of cellular proteins, particularly those conjugated with ubiquitin. It is involved in the degradation of constitutively short-lived proteins, regulatory, abnormal and misfolded proteins that promote normal cell growth and metabolism. p97 was found to interact with TNFRs and it was proposed to play a role in the latter's degradation by 26S proteasome complex [184,186].
BRE, a stress response protein was found to interact directly with TNFR1 and it is thought to regulate the TNFR-mediated response by inhibiting and dampening TNFR1-induced signalling . An alternative pathway in TNFα-induced phosphatidy linositol signalling via PIP5K-IIβ (phosphoinositol 4 phosphate 5 kinase IIβ) was reported. TNFR1 but not TNFR2 was found to interact in its juxta-membrane region and activates PIP5K-IIβ . PIP5K-IIβ is necessary for the generation of PIP2, a substrate for PI-PLC-mediated IP3 and DAG .
In general, the binding of TNF to the TNFRs (TNFR1 and TNFR2) can bring about its varied effects such as cellular apoptosis, proliferation or inflammatory responses due to the interplay of signalling mediators, in a TRADD-dependent and -independent pathway. The differences in the site of action, cell or tissue type, expression patterns and interactions of these signalling molecules play a pivotal role in determining the effect of the wide range of cellular function and responses. In other words, the outcome depends on the micro-environment within and around the cell. This includes the influence of sphingolipid rheostat and the level of intracellular ROS, which influences or determines cell fate, which is an active area of research. All these factors play a pivotal role in determining the effect of the wide range of cellular functions and responses.
More recent advances have led to the identification of miRNAs (microRNAs) as potent regulators of gene expression. miRNAs are 21–23 nt long, non-coding RNA oligonucleotides and are evolutionarily conserved . miRNAs like miR146a/b, miR155, miR125b were recently reported to be involved in immune responses, including inflammatory responses induced by LPS/TNFα [192–195]. It was also found that miR155 regulate the expression of vital TNFα signalling molecules like RIP, FADD, IKK . TNF-induced miRNAs miR-31 and miR-17–3p were reported to regulate TNF-induced endothelial-cell-adhesion molecules activation . TNF-associated aberrant inflammatory response can be reversed by targeting specific miRNA by using miRNA mimics. Despite the advances achieved in the understanding of the various signalling events and cross talks during TNF signalling, there is a necessity to understand and elucidate further the various interactions and cross-talk among the signalling mediators in this signalling process.
TNFα SIGNALLING MEDIATORS AS THERAPEUTIC TARGETS
Inflammation is involved in the maintenance of tissue homoeostasis, defence against infection and mediating immune responses. However, a dysregulated or prolonged inflammatory process contributes to tissue injury and morbidity, especially in certain chronic diseases and autoimmune conditions. This leads to the necessity of dampening the inflammatory response. TNF is well known for its role in host defence to bacterial, viral and parasitic infections by mediating and amplifying the inflammatory response. Therefore, aberrant TNF response was associated with a spectrum of inflammatory disorders. Biopharmaceutical agents such as antibodies and soluble receptors that target TNF production are being increasingly used in the management of TNF-related disorders. A range of them are currently licensed as TNF blocking agents and are being used in the management of inflammatory diseases like rheumatoid arthritis, ankylosing spondylitis and Crohn's disease. TNF antagonists are also being used in the management of endometriosis based on the interaction of TNF with GnRH (gonadotropin-releasing hormone) signalling . The mechanism of action and therapeutic effects in a variety of conditions for most of the currently available TNF antagonists was extensively discussed in some recent reviews [9,198–201]. There are several pitfalls in anti-TNF therapy which includes a TNF antagonist exhibiting a varying degree of efficacy, increased susceptibility to infections and cost . More specifically, TNF blockade was associated with increase in susceptibility to bacterial, viral and parasitic infections including Listeria, Mycobacteria and granulomatous infections [203–206]. It has also been found to be associated with the incidence of opportunistic infection, demyelinating syndromes and autoimmune conditions like lupus. A recent report by Jan Lin et al.  has discussed in detail the adverse effects induced by TNF blockade, which clearly indicates the limitations of the use of such biopharmaceuticals. The lack of responsiveness to certain disorders, susceptibility to infections and resistance on long-term use has increased the lookout for alternative therapeutic agents.
The search led to the consideration of signalling molecules and mediators regulating inflammatory signals and responses as potential targets. Mediators and molecules involved in various pathways triggered by TNFRs including that of the MAPK and NF-κB have gained attention in the recent past. TNF blockade by inhibiting its production and molecules involved in the TNFα-triggered signalling pathways such as phosphodiesterase 4, p38 MAPK and NF-κB inhibitors are currently being explored . Compounds targeting specific molecules of these pathways such as p38 MAPK inhibitors and IKK inhibitors are in various stages of drug development including clinical trials. The current status and future in the treatment of inflammatory diseases by targeting signal transduction have been well discussed by O'Neill . The advancements in the elucidation of signalling pathways have increased the choice of such targets.
The identification of suitable and predictable targets for intervention in TNF-associated inflammatory disorders requires a more detailed understanding of the signalling action and interaction profile of the various signalling mediators. Based on the facts discussed earlier about our work, it can be suggested that signalling molecules like PLD and SphK can be therapeutically targeted to dampen the pathological effects of inflammation in autoimmune diseases and in some inflammatory disorders. In vivo experiments are being carried out to further validate the effectiveness of targeting PLD and SphK in an isoform-specific way to dampen inflammation and its feasibility to be used in the whole organism. Some recent studies have paved the way for more work, to elucidate the role of miRNAs in regulating the various signalling mediators present in the TNFα signalling pathway and there is potential for the use of antagomirs against specific miRNA in the therapeutic management of inflammatory disorders. Therefore, an in-depth understanding of the signalling events and signal transduction process is imperative and it will certainly aid in identifying novel molecular targets and in planning strategies in the therapeutic intervention of immune-mediated inflammatory and TNF-associated diseases.
activator protein 1
apoptosis signal-regulating kinase 1
activating transcription factor
ceramide-activated protein kinase
cytosolic phospholipase A2
death-inducing signalling complex
ETS (E twenty-six)-like kinase 1
extracellular-signal-regulated kinase 1/2
ferritin heavy chain
growth arrest and DNA-damage-inducible protein 45β
growth-factor-receptor-bound protein 2
glycogen synthase kinase 3β
heat-shock protein 75
inhibitor of apoptosis protein
inhibitor of nuclear factor κB
c-Jun N-terminal kinase
mitogen-activated protein kinase
manganese superoxide dismutase
nuclear factor of activated T-cells
nuclear factor κB
NFκB essential modulator
factor associated with NSmase
programmed cell death
phosphatidylinositol 4,5 bisphosphate
phosphatidylinositol 3,4,5 triphosphate
phosphoinositol 4 phosphate 5 kinase IIβ
protein kinase B
protein kinase C
receptor interacting protein
reactive oxygen species
sphingosine 1 phosphate
silencer of DD
signal transducer and activator of transcription
tumour necrosis factor
soluble bound TNF
TNFR-associated factor 2
X-chromosome linked inhibitor of apoptosis
We thank A.-K. Fraser-Andrews for proofreading the manuscript prior to submission.
This work was supported by the UK Medical Research Council [grant number G0700794].