This Article Abstract Full Text (PDF) All Versions of this Article: 28/4/365    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wullaert, A. Articles by Beyaert, R. Search for Related Content PubMed PubMed Citation Articles by Wullaert, A. Articles by Beyaert, R.

Hepatic Tumor Necrosis Factor Signaling and Nuclear Factor-B: Effects on Liver Homeostasis and Beyond

Unit of Molecular Signal Transduction in Inflammation, Department for Molecular Biomedical Research-VIB, and Department of Molecular Biology, Ghent University, Technologiepark 927, B-9052 Ghent (Zwijnaarde), Belgium

Correspondence: Address all correspondence and requests for reprints to: Rudi Beyaert, Department for Molecular Biomedical Research, VIB, Ghent University, Technologiepark 927, B-9052 Ghent (Zwijnaarde), Belgium. E-mail: rudi.beyaert{at}dmbr.ugent.be

	Abstract

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

 
The proinflammatory cytokine TNF has a pivotal role in liver pathophysiology because it holds the capacity to induce both hepatocyte cell death and hepatocyte proliferation. This dual effect of TNF on hepatocytes reflects its ability to induce both nuclear factor B (NF-B)-dependent gene expression and cell death. Multiple studies have demonstrated the crucial role of the transcription factor NF-B in the decision between life and death of a hepatocyte. Massive hepatocyte apoptosis preceding embryonic lethality in NF-B-deficient mice constituted the first indication of an essential antiapoptotic function of NF-B in the liver. Although many studies confirmed this crucial cytoprotective role of NF-B in adult liver, a number of genetic studies recently obtained conflicting results on the exact role of NF-B in different mouse models of TNF hepatotoxicity, demonstrating that caution should be taken when interpreting studies using different NF-B-deficient mice in distinct models of liver injury. Recent reports showing a role for hepatic NF-B activation in the proliferation of malignant cells during hepatocarcinogenesis, and in the progression of fatty liver diseases to insulin resistance and type 2 diabetes mellitus demonstrate that NF-B can also have more detrimental effects in the liver. Moreover, its role in the development of the metabolic syndrome emphasizes that hepatic NF-B activation might also have adverse effects on the endocrine system. Therefore, understanding the regulation of hepatic TNF signaling and NF-B activation is of critical therapeutic importance. In this review, we summarize how studies on the role of NF-B in different mouse models of liver pathologies have contributed to this understanding.

I. Introduction

II. Role of TNF in Liver Diseases

III. TNF-Induced Signaling Pathways in Hepatocytes

A. TNF signaling to the activation of NF-B and MAPKs

B. TNF-induced signaling pathways leading to hepatocyte apoptosis

IV. Crucial Role for NF-B in Regulating the Hepatocyte’s Response to TNF

V. Role of NF-B Activation in Mouse Models of TNF-Mediated Hepatic Diseases

A. Role of NF-B activation in mouse models of TNF-mediated liver toxicity

B. Role of NF-B activation in liver regeneration

C. Role of NF-B activation in mouse models of hepatocarcinogenesis

D. Role of hepatic NF-B activation in the metabolic syndrome

E. Role of NF-B-dependent inflammation in hereditary hemochromatosis

VI. Mechanisms of the Antiapoptotic Effect of NF-B

A. NF-B-dependent inhibition of caspase activation

B. NF-B-dependent inhibition of JNK activation

C. Mechanisms of crosstalk between TNF-induced activation of NF-B and JNK

VII. Concluding Remarks

	I. Introduction

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

 
LIVER CELL INJURY and cell death are prominent features in all liver disease processes. In healthy liver, constitutive production of cytokines is absent or very low. During liver inflammation, however, hepatocytes are exposed to increased levels of cytokines such as TNF, IL-1ß, and interferon , as well as oxidative stress and bile acids. Excessive exposure to these agents will eventually cause hepatocyte cell death, which can contribute to acute liver injuries, such as fulminant hepatitis and ischemia/reperfusion (I/R) damage, or even chronic sustained injuries, such as alcoholic liver disease, cholestatic liver disease, and viral hepatitis (1, 2). Conversely, insufficient hepatocyte apoptosis, associated with failure to remove mutated cells or with unrestrained proliferation within the context of a chronic inflammatory milieu, can promote the development of liver cancer. Paradoxically, hepatic carcinogenesis can also arise from sustained hepatocyte cell death due to the high rate of compensatory regeneration invoked in the tissue, thus elevating the risk of mitotic errors (3). These liver pathologies resulting either from excessive hepatocyte cell death or from a lack thereof emphasize the crucial role of the balance between hepatocyte survival and death in preserving liver homeostasis. The proinflammatory cytokine TNF is a key regulator of this vital balance because TNF can induce hepatocyte proliferation as well as hepatocyte cell death (Fig. 1) (2). Therefore, accurate knowledge of the TNF-induced signal transduction pathways implicated in the prevention as well as the execution of hepatocyte cell death is of major importance for understanding the development of liver diseases. Moreover, an improved understanding of the TNF-induced molecular pathways that determine the hepatocyte’s destiny, either survival or cell death, will undoubtedly contribute to the development of novel therapeutic strategies for preventing hepatocyte cell death in liver injury, or selectively killing malignant cells in liver tumors. In this review, we will focus on the intracellular signal transduction pathways that mediate the opposing effects of TNF in hepatocytes. We will discuss how the balance between these TNF-induced signaling pathways influences liver homeostasis. In addition, we will summarize how disturbing this balance promotes the development of liver pathologies, as well as how this affects the endocrine system and, as such, contributes to the metabolic syndrome and hemochromatosis.

View larger version (18K): [in this window] [in a new window]   FIG. 1. Role of TNF in liver homeostasis. TNF can induce both hepatocyte proliferation and hepatocyte cell death. Because excessive hepatocyte cell death contributes to liver injury and excessive hepatocyte proliferation leads to liver cancer, TNF has a crucial role in preserving liver homeostasis.

	II. Role of TNF in Liver Diseases

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

Although the above observations clearly demonstrate that TNF can contribute to hepatocyte cell death during liver diseases, the healthy liver has well-developed defense mechanisms that permit hepatocytes to adapt to TNF-initiated stress. Indeed, acute treatment with exogenous TNF will normally be well tolerated by the liver and will not lead to hepatocyte injury. In fact, systemic administration of TNF will induce healthy hepatocytes to proliferate rather than die (16). This proliferative response also occurs when the liver is confronted with partial hepatectomy, which acutely elicits hepatic production of TNF. Multiple studies have shown that this increased production of TNF is necessary for the remnant liver to regenerate, pointing to an essential role for TNF in the proliferation of hepatocytes after partial hepatectomy (17, 18, 19).

These studies show that depending on the context, exposure of hepatocytes to TNF induces signals that mediate cell death, or alternatively, survival pathways that allow hepatocytes to tolerate tissue damage or even recover from it. Although at first sight paradoxical, the activation of both apoptotic and survival pathways in response to the same stimulus ensures that neither aberrant cellular survival nor excessive cell death arises and, in doing so, preserves proper liver homeostasis. Moreover, next to its role in deciding between life and death of a hepatocyte, TNF also affects the metabolic functions of the liver. Indeed, TNF can cause insulin resistance and affects glucose metabolism in hepatocytes (20). Also hemochromatosis depends, at least partially, on the inflammatory activity of TNF (21). These observations add to the impressive activity range of TNF in the liver and emphasize the need of studying the intracellular signaling pathways that regulate cellular responses to TNF, because only this will help us to understand how a single cytokine, such as TNF, can exert such pleiotropic effects in the liver.

	III. TNF-Induced Signaling Pathways in Hepatocytes

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

 
TNF exerts its biological effects by binding two plasma membrane receptors, TNF-receptor 1 (TNF-R1) and TNF-R2. The majority of the biological effects of TNF are mediated by TNF-R1 triggering, which can initiate signaling pathways leading to the activation of the transcription factor nuclear factor-B (NF-B), the initiation of MAPK cascades, as well as cell death. This diversity of signaling pathways initiated by TNF-R1 is accomplished by the recruitment of different adapter proteins to its cytoplasmic part. According to a recent model (Fig. 2), TNF signaling starts when TNF-R-associated death domain (TRADD), TNF-R associated factor (TRAF)2, and receptor interacting protein (RIP)1 are recruited to TNF-R1 to form complex I, which signals to the activation of NF-B, as well as the MAPKs p38 and c-Jun N-terminal kinase (JNK). Once released from the receptor, the cytosolic complex II further recruits Fas-associated death domain (FADD) and pro-caspase-8, which will result in cell death (22). Alternatively, this cell death-inducing complex II has been suggested to result from internalization of the TNF-R1 receptosome (23).

View larger version (60K): [in this window] [in a new window]   FIG. 2. TNF-R1-induced signaling pathways via complex I and complex II. TNF binding to TNF-R1 leads to the recruitment of TRADD, TRAF2, and RIP1, forming complex I (1 ). TRAF2 catalyzes Lys-63-linked polyubiquitination of itself and RIP1, allowing RIP1 to recruit the TAK1 complex (2 ). This subsequently activates specific MKKs (3 ) as well as the IKK complex (4 ). Whereas MKKs activate the p38 and JNK MAPKs, the IKKß catalytic subunit of the IKK complex phosphorylates the NF-B-bound IBa (5 ), leading to its Lys-48-linked polyubiquitination and subsequent proteasomal degradation (6 ). This allows NF-B to translocate to the nucleus (7 ), where it induces the transcription of genes with an NF-B consensus binding site in their promoter. Binding of FADD and procaspase-8 to the TRADD/TRAF2/RIP1 complex results in the formation of the cytosolic complex II, which will signal toward apoptosis (8 ). See text for further details. DD, Death domain; Ub, ubiquitin; P, phosphate.

Before embarking on the complex regulation of interplay between the different TNF-induced intracellular signaling pathways, which will eventually determine the hepatocyte’s biological response to the TNF stimulus, we will first focus on some important mechanisms and molecules in each of the individual pathways. It should be stressed that our current knowledge of TNF signaling is still largely fragmentary and based on studies with different cell types and experimental set-ups. It can therefore be expected that the effect of specific adaptor proteins and kinases might sometimes be cell type-specific or depend on the TNF concentration that was used.

A. TNF signaling to the activation of NF-B and MAPKs
NF-B is a transcription factor that is normally held in the cytoplasm by members of the inhibitor of B (IB) protein family. The prototypical NF-B complex consists of a p65/p50 heterodimer and is bound to IB. TNF-induced signaling to NF-B involves the activation of the IB kinase (IKK) complex, which consists of two catalytic subunits, IKK (or IKK1) and IKKß (or IKK2), and a regulatory subunit, IKK (or NEMO). Once activated, the IKKß subunit phosphorylates IB, leading to its ubiquitination and subsequent proteasomal degradation. This allows the p65/p50 complex to translocate to the nucleus, where it can activate transcription of NF-B responsive genes by binding to B sites in their promoter regions. Although an "alternative" NF-B pathway that proceeds independent of IKKß and IKK can be induced by some other members of the TNF family, TNF itself is believed to be a specific activator of the "classical" IKKß/IKK-mediated NF-B pathway. We will therefore only focus on this pathway for our discussion of the more upstream signaling components (31).

TNF-induced signaling toward IKK activation is initiated by the sequential recruitment of the adapter proteins TRADD, TRAF2, and RIP1 to the cytoplasmic part of TNF-R1. Once this receptor complex I is formed, TRAF2 and RIP1 cooperate to recruit the TGF-ß activated kinase (TAK)1 complex. TRAF2 contains a ubiquitin ligase activity by which it catalyzes the attachment of ubiquitin Lys-63-linked polyubiquitin chains to itself as well as RIP1. In contrast to Lys-48-linked polyubiquitination, which labels proteins for proteasome-mediated degradation, Lys-63-linked polyubiquitination has emerged as an important signaling mechanism controlling various physiological processes such as the NF-B activation pathway (32, 33). The RIP1-associated Lys-63-linked polyubiquitin chains are recognized by TAK1-binding protein (TAB)2 and TAB3, both of which mediate the recruitment of TAK1 (34, 35). This event is crucial to TNF-induced IKK activation, because activated TAK1 will phosphorylate critical residues in IKKß, leading to its activation (36, 37, 38).

Because TAK1 is actually a member of the MAP3K family, recruitment of TAK1 to the TNF-R1 complex mediates not only TNF-induced activation of NF-B, but also TNF-induced activation of the p38 and JNK MAPKs (36, 37, 38). Although other members of the MAP3K family have also been suggested to play a role in TNF-induced activation of p38 and JNK, gene targeting studies could not confirm their role in TNF-induced MAPK activation. Possibly, different MAP3Ks make partial and additive contributions to TNF-induced JNK and p38 activation in a cell-type-specific manner. The activated MAP3Ks in turn activate the MAPK kinase (MKK)3/6 and MKK4/7 members of the MAP2K family that are responsible for the activation of p38 and JNK, respectively.

B. TNF-induced signaling pathways leading to hepatocyte apoptosis
Triggering TNF-R1 can lead to hepatocyte apoptosis by recruiting the adapter proteins TRADD and FADD to its cytoplasmic part. FADD contains a dead effector domain by which it subsequently recruits procaspase-8, thus constituting the death-inducing signaling complex (DISC) where clustering of pro-caspase-8 results in its autoactivation. The active caspase-8 is then able to proteolytically activate several effector caspases, which are responsible for many of the destructive cellular events that result in apoptosis. Whether recruitment of FADD and activation of caspase-8 really occur at the level of the TNF-R1 at the cell membrane or as part of a separate intracellular signaling complex is still a matter of debate (22, 23). In the case of hepatocytes, only a small amount of active caspase-8 is formed at the DISC, which has only a minor contribution to TNF-induced hepatocyte cell death. However, next to the direct activation of downstream executioner caspases by caspase-8 (the so-called type I cell death pathway), hepatocytes and some other cell types (all referred to as type II cells) require a second pathway that involves a mitochondrial amplification loop to undergo apoptotic cell death (39, 40, 41, 42). This mitochondrial pathway involves the release of numerous apoptogenic factors from the mitochondrial intermembrane space into the cytosol, including cytochrome c. Together with Apaf-1, deoxy-ATP, and procaspase-9, cytochrome c forms the apoptosome, a high-molecular weight protein complex that activates caspase-9. Activated caspase-9 then proteolytically activates effector caspases, which in turn amplify the apoptotic signal by activating procaspase-8 and -9. The full-blown caspase activity that emerges from this amplification loop ultimately kills the cell.

In hepatocytes, the mitochondrial pathway appears to be central in TNF-induced cell death because all signaling events directly or indirectly target the mitochondria (Fig. 3). Normally, a delicate balance between the antiapoptotic members of the Bcl-2 protein family, such as Bcl-2 itself and Bcl-xL, and its proapoptotic members, such as Bid, Bak, and Bax, preserves mitochondrial homeostasis. However, TNF-induced DISC formation leads to caspase-8 mediated cleavage of Bid, resulting in its active truncated form, tBid. This event establishes a molecular link between the DISC and the initiation of the mitochondrial pathway, because tBid will translocate to the mitochondria to induce the release of cytochrome c into the cytosol, thus shifting the Bcl-2 balance toward apoptosis. Several mechanisms have been proposed by which tBid disrupts the integrity of the mitochondrial membrane. For instance, tBid has been suggested to promote insertion and oligomerization of Bak and/or Bax into the outer mitochondrial membrane, thus leading to the formation of pores and resulting in the mitochondrial release of cytochrome c (43, 44, 45). In addition, tBid has been shown to mediate TNF-induced opening of the permeability transition pore in the outer mitochondrial membrane, a phenomenon that is called the mitochondrial permeability transition (MPT). MPT appears to be essential to TNF-induced hepatocyte apoptosis because specific inhibitors of MPT prevent the hepatotoxic effects of TNF in vitro as well as in vivo (40, 45, 46). Next to these direct mechanisms of inducing mitochondrial permeability, tBid also indirectly contributes to TNF-induced mitochondrial dysfunction via the release of cathepsin B from the lysosomes. Cathepsin B was previously shown to critically mediate TNF-induced mitochondrial release of cytochrome c in hepatocytes. Consistently, cathepsin B-deficient hepatocytes are more resistant to TNF-induced apoptosis than wild-type hepatocytes (47), and mice with a targeted deletion of the cathepsin B gene display less TNF-induced liver injury than their wild-type counterparts (48). The observation that TNF stimulation does not induce lysosomal permeabilization in Bid-deficient hepatocytes suggested that the reduced release of mitochondrial cytochrome c observed in these cells is due to the absence of cytosolic cathepsin B. Indeed, the release of cathepsin B into the cytosol leads to the activation of caspase-2, which subsequently facilitates efficient mitochondrial cytochrome c release. The mechanism by which active caspase-2 achieves this mitochondrial permeabilization is unclear (49). These studies demonstrate a central role for tBid in the induction of multiple pathways, all of which directly or indirectly converge on the mitochondrial release of cytochrome c and thus contribute to TNF-induced hepatocyte apoptosis.

View larger version (62K): [in this window] [in a new window]   FIG. 3. Signaling pathways in TNF-induced hepatocyte cell death: all roads lead to the mitochondria. After binding of TNF to TNF-R1, complex II is formed and caspase-8 is activated. Because hepatocytes are type II cells, the amount of active caspase-8 that is generated is not sufficient to directly activate caspase-3 (dashed line). Therefore, caspase-8 initiates the mitochondrial pathway by cleaving Bid to tBid, which induces mitochondrial permeabilization via three distinct mechanisms: 1) tBid mediates opening of the permeability transition pore (PTP), leading to MPT; 2) tBid mediates the insertion of Bak and Bax in the mitochondrial outer membrane, leading to the formation of the Bak/Bax pore; and 3) tBid mediates lysosomal permeabilization, which leads to mitochondrial permeabilization via a cathepsin B- and caspase-2-dependent mechanism. MPT can also be induced independently of tBid, via N-SMase/A-SMase-mediated generation of ceramide (4 ). These events result in the mitochondrial release of cytochrome c and ROS. Whereas cytochrome c contributes to apoptosis via activation of caspase-9 and subsequent activation of caspase-3, ROS mainly contribute to necrosis. See text for further details. DD, Death domain; pro, prodomain.

Although the studies above demonstrate a central role for Bid in the generation of ROS as well as the release of mitochondrial cytochrome c into the cytosol, Bid-deficient hepatocytes are not completely resistant to TNF-induced apoptosis. The absence of Bid significantly delays TNF-induced hepatocyte apoptosis but does not prevent it (45). Moreover, Bid-deficient mice are only partially protected against TNF-mediated liver injury (44). These observations suggest that Bid-independent pathways also contribute to TNF-induced hepatocyte apoptosis. In this respect, TNF-induced activation of acidic and neutral sphingomyelinase (A-SMase and N-SMase), both of which catalyze sphingomyelin degradation and formation of ceramide, has been implicated in hepatocyte cell death. A critical role for A-SMase in this process was demonstrated in A-SMase-deficient hepatocytes, which produce substantially less ceramide upon TNF treatment than wild-type hepatocytes, resulting in reduced cell death. Consistently, A-SMase-deficient mice are resistant to TNF-induced liver injury (53). Also N-SMase has been suggested to play a role in TNF-induced hepatotoxicity because mice with a targeted deletion of FAN, the adapter protein that mediates TNF-induced activation of N-SMase, are partially protected against TNF-mediated liver injury (54). These studies indicate that the intracellular level of ceramide is an important factor that determines the susceptibility of hepatocytes to TNF-induced apoptosis. Enhanced levels of ceramide seem to contribute to TNF-induced hepatocyte apoptosis by down-regulating expression of methionine adenosyltransferases, leading to depletion of intracellular S-adenosyl-L-methionine. This results in depletion of mitochondrial reduced glutathione (55), which sensitizes mitochondria to the harmful effects of ganglioside GD3, a glycosphingolipid that is synthesized from ceramide. GD3 traffics to the mitochondria and directly induces ROS production, followed by MPT, cytochrome c release, and caspase activation, thus contributing to hepatocyte cell death (56, 57, 58). Besides this ceramide-mediated pathway for ROS generation, other Bid-independent extramitochondrial sources of ROS might contribute to TNF-induced hepatocyte cell death, including ROS resulting from the 5-lipoxygenase-mediated metabolization of arachidonic acid that is produced upon triggering of cytosolic phospholipase A2 by TNF (59, 60). It should be mentioned, however, that this TNF-induced release of arachidonic acid, which also seems to involve cathepsin B (61), is thought to be a late process during TNF-induced cell death, occurring downstream of mitochondrial changes and effector caspase activation (62).

Considering the dense network of intracellular signaling pathways that TNF can induce in hepatocytes, it is hard to imagine the complexity that emerges if one also takes into account the signaling mechanisms that result from other factors that are indirectly induced by TNF. TNF undoubtedly induces the release of many cytokines and other factors that additively or synergistically induce cell injury, or instead try to limit the destructive activities of TNF. The most obvious example of the former is TNF-induced expression and release of TNF itself, which allows paracrine effects of TNF to gradually convert a localized inflammatory reaction into widespread tissue damage. However, TNF also induces many negative feedback mechanisms for preventing this exacerbation of the inflammatory response. In this context, TNF is an important inducer of acute phase proteins in the liver. These proteins together provoke a systemic antiinflammatory reaction known as the acute phase response, which has been shown to limit tissue damage in various models of liver injury. For instance, mice incapable of responding to IL-6, the most potent inducer of the acute phase response, are hypersensitive to lipopolysaccharide (LPS)-induced liver damage (63). Furthermore, induction of the acute phase response by IL-6 protects mice against I/R-induced liver injury and promotes the regeneration of hepatocytes (64). From the above, it is clear that increased expression of TNF in the liver induces benign systemic responses that protect the organism from excessive damage. However, although individual acute phase proteins such as 1-antitrypsin and 1-acid glycoprotein have been shown to exert hepatoprotective activities in TNF-mediated liver injury (65, 66), they fail to protect hepatocytes from the cytotoxic effects of TNF in vitro (67). This indicates that on their turn, these acute phase proteins employ indirect mechanisms for protecting hepatocytes against cell death. Although all these indirect and systemic effects of TNF undoubtedly contribute to the regulation of liver homeostasis, we have limited ourselves to review only the aforementioned TNF-induced intracellular, and thus cell autonomous, signaling pathways in hepatocytes that affect the overall response of the liver to TNF.

	IV. Crucial Role for NF-B in Regulating the Hepatocyte’s Response to TNF

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

 
Despite the fact that binding of TNF to TNF-R1 can simultaneously activate each of the above-mentioned signaling pathways, switching on an impressive cell death-inducing machinery, healthy hepatocytes are completely resistant to the cytotoxic effects of TNF. However, in the setting of global transcriptional arrest, such as by treatment with actinomycin D in vitro or D-(+)-galactosamine (GalN) in vivo, hepatocytes rapidly undergo cell death upon TNF stimulation. This indicates that hepatocytes are refractory to TNF-induced cell death due to the ability to up-regulate transcription of essential protective genes. Multiple studies have shown that NF-B activation is at least partially responsible for the induction of these cell death inhibitory factors.

Mice deficient in the p65 NF-B subunit provided the first indication that NF-B is important to protect hepatocytes against TNF-induced cell death. These mice, as well as mice without the essential IKKß or IKK subunits of the IKK complex, die during midgestation due to massive hepatocyte apoptosis (68, 69, 70, 71). Embryonic hepatocyte apoptosis was later shown to depend on TNF signaling because additional deletion of the TNF or the TNF-R1 gene rescued NF-B-deficient mice from embryonic lethality (72, 73, 74). The essential protective function of NF-B against TNF-induced hepatocyte apoptosis was confirmed by in vitro studies in which cultured hepatocytes are transfected with a mutant IB that lacks the normal phosphorylation sites (Ser 32 and Ser36) for IKKß so that it can no longer be marked for proteasomal degradation and thus acts as a superrepressor (IB^s) of NF-B activation. In this case, IB^s expression converts the hepatocyte’s response to TNF from proliferation to apoptosis (75, 76). In addition to this hepatoprotective role of NF-B in cultured hepatocytes and the developing embryo, several reports have shown that adult mice also depend on NF-B activation to avoid TNF-induced hepatocyte apoptosis (see Sections V.A and V.B) (77, 78, 79). Consequently, given the abundance of TNF during liver diseases together with the crucial role of NF-B in protecting hepatocytes against the cytotoxic effects of TNF, it is not surprising that the activation status of NF-B critically influences the pathogenesis of several TNF-mediated hepatic diseases. As such, the role of NF-B has been studied extensively in various mouse models of TNF-mediated liver injury.

	V. Role of NF-B Activation in Mouse Models of TNF-Mediated Hepatic Diseases

Top Abstract I. Introduction II. Role of TNF... III. TNF-Induced Signaling... IV. Crucial Role for... V. Role of NF-{kappa}B... VI. Mechanisms of the... VII. Concluding Remarks References

 
A. Role of NF-B activation in mouse models of TNF-mediated liver toxicity
Several mouse models have been developed to study TNF-mediated liver injury. Although TNF as such does not cause severe liver toxicity when injected in healthy mice, coadministration of GalN, a hepatotoxin that blocks transcription specifically in hepatocytes, drastically sensitizes mice to TNF hepatotoxicity and lethality (80, 81). Similarly, GalN sensitizes mice to LPS, which induces several cells to produce TNF, which then acts in a paracrine manner to induce liver injury. As such, administration of TNF/GalN or LPS/GalN causes an inflammatory hepatitis, characterized by infiltration of neutrophils and macrophages, and massive hepatocyte apoptosis and necrosis, which together result in lethality. TNF-R1-deficient mice are completely resistant against the lethal effects of TNF/GalN as well as LPS/GalN, demonstrating the essential role of TNF-R1 in these models (82, 83). It should be mentioned that the use of GalN may grossly exaggerate the effects of TNF, emphasizing the need for more advanced models of TNF-mediated liver injury to confirm the pathophysiological relevance of some of the findings made with the above-mentioned models. Nevertheless, the TNF/GalN-induced model of liver injury provides an elegant setting in which the direct effects of TNF signaling in hepatocytes can be studied in an in vivo context, without the interference of indirect effects of other cytokines or factors. Another model that is frequently used to study TNF-mediated liver injury is the administration of the T cell mitogenic plant lectin Con A (84). Con A induces a T cell-mediated hepatitis that, in contrast to the GalN-based models of liver injury, depends on both TNF-R1 and TNF-R2 (29).

A beneficial role for NF-B activation when facing TNF-mediated liver injury was already suggested when pretreatment of mice with NF-B activating stimuli, such as IL-1 or TNF itself, was found to confer protection against subsequent TNF/GalN-induced liver injury (85). Mice compromised in NF-B activation do not display this desensitization to the hepatotoxic effects of TNF (77), proving that NF-B activation induces potent defense mechanisms against TNF-induced liver injury. In the meantime, several mice deficient in NF-B activation, either by targeted deletion of essential NF-B signaling molecules or by transgenic expression of NF-B inhibitory proteins, have confirmed the hepatoprotective effects of NF-B. The responses of these various NF-B-deficient mice in different models of TNF-mediated liver injury are summarized in Table 1 and will be discussed in detail below.

View this table: [in this window] [in a new window]   TABLE 1. Response of NF-B-deficient mice in models of TNF-mediated liver injury

Remarkably, in some models of TNF-mediated hepatotoxicity, inhibition of NF-B activation has been reported to be protective rather than sensitizing (summarized in Table 2). Although at first sight this contrasts with the well-established role of NF-B in the liver as an antiapoptotic factor, a more detailed look reveals that in these models protection by NF-B inhibition is associated with decreased TNF levels, bypassing the danger of sensitizing hepatocytes to TNF cytotoxicity. Because TNF production in the liver is mainly derived from Kupffer cells, inhibition of hepatic NF-B activation can only be beneficial when it occurs in these cells. For instance, NF-B decoy oligonucleotides delivered by liposomes predominantly inhibit NF-B activation in Kupffer cells. Consequently, these NF-B decoys efficiently block LPS-induced production of TNF, thereby preventing cell death of hepatocytes sensitized by prior infection with Propionibacterium acnes (89). Because systemic administration of adenoviruses leads to infection of hepatocytes as well as Kupffer cells (90), adenoviral gene transfer of an IB^s is another opportunity to block hepatic TNF production. Indeed, adenoviral expression of an IB^s has been shown to protect mice against ethanol-induced liver injury. Chronic administration of ethanol leads to increased levels of gut-derived LPS in the portal circulation, thereby activating Kupffer cells to produce TNF. Adenoviral expression of an IB^s prevents this TNF synthesis by Kupffer cells and as such impairs subsequent liver injury (91). Also after I/R, Kupffer cell stimulation leading to TNF production is essential for the development of liver injury (15). As a result, adenoviral expression of an IB^s abrogated TNF production and liver injury after I/R (92), suggesting that in this model Kupffer cell NF-B activation needs to be blocked to prevent liver damage. However, also hepatocyte-specific NF-B-deficient mice display less hepatic TNF production after I/R of the liver, resulting in decreased liver damage (79). This indicates that after I/R, the initial trigger that activates Kupffer cells to secrete TNF might be produced by hepatocytes in an NF-B-dependent manner. Remarkably, the hepatocyte-specific IKKß-deficient mice that were used in this study show inhibition of NF-B activation after I/R, but not after TNF (79), indicating that the mechanisms by which TNF and I/R lead to NF-B activation in the liver are different. In support of this notion, mice in which an IBß transgene replaces the endogenous IB gene display reduced hepatic NF-B activation after I/R, leading to decreased TNF levels and diminished liver injury (93). In contrast, TNF-induced NF-B activation in these mice is not altered (94). Thus, while IB and IBß are redundant for regulating TNF-induced NF-B activation, IB seems to have a unique function in regulating I/R-induced NF-B activation. In this respect, tyrosine phosphorylation of IB has been proposed to mediate I/R-induced activation of NF-B (93, 95). Because IKKß-deficient mice display impaired NF-B activation after I/R, the IKK complex also seems to play a role in this pathway. However, the mechanism by which IKKß influences tyrosine phosphorylation of IB is not clear yet. Taken together, the above observations show that despite the well-established antiapoptotic role of NF-B in the liver, inhibition of NF-B activity can be beneficial in different conditions of TNF-mediated liver injury.

View this table: [in this window] [in a new window]   TABLE 2. Protective effect of inhibition of NF-B activation in models of TNF-mediated liver injury

B. Role of NF-B activation in liver regeneration

C. Role of NF-B activation in mouse models of hepatocarcinogenesis
The studies described above clearly demonstrate that NF-B activation in hepatocytes is essential for protection against apoptosis during TNF-mediated liver injury, whereas activation of NF-B in nonparenchymal cells of the liver is essential to stimulate hepatocyte proliferation during liver regeneration. Although both of these activities of NF-B are beneficial in these respective situations, they can also have detrimental effects. Indeed, considering the fact that both proliferation and evasion of apoptosis are important hallmarks of cancer (104), it is not surprising that NF-B activation in the liver was recently associated with an increased risk for hepatocarcinogenesis. Moreover, other hallmarks of cancer such as sustained angiogenesis, tissue invasion, and metastasis were also reported to be at least partially dependent on NF-B signaling (105, 106).

A major link between NF-B-dependent inflammation and hepatocarcinogenesis came from studies with genetically altered mice. In multidrug resistance gene 2-deficient mice, bile duct inflammation is followed by the appearance of hepatocarcinoma due to prolonged NF-B activation in hepatocytes (107). In this model of inflammation-associated cancer, TNF produced by the nonparenchymal cells of the liver activates NF-B in hepatocytes. Hepatocyte-specific expression of an IB^s as well as suppressing TNF activity resulted in apoptosis of transformed hepatocytes and failure to progress to carcinomas. This impaired tumorigenesis was only apparent when NF-B activity was inhibited between 7 and 14 months after birth, whereas its inactivation during the first 7 months caused no beneficial effects. These observations indicate that TNF-induced activation of NF-B in hepatocytes plays a detrimental role during late phases of tumor development by increasing survival of dysplastic hepatocytes.

In contrast, in the diethylnitrosamine (DEN)-induced model of hepatocarcinogenesis, the number and the size of hepatocarcinomas is significantly increased in hepatocyte-specific IKKß-deficient mice, pointing to a beneficial effect for hepatocyte NF-B activation (108). In these mice, DEN causes increased hepatocyte necrosis due to impaired hepatocyte NF-B activation, leading to more compensatory regeneration. The latter is enhanced by NF-B activation in Kupffer cells because additional ablation of IKKß in these cells resulted in less and smaller DEN-induced tumors. Thus, in this chemically induced model of hepatocarcinogenesis, hepatocyte NF-B activation avoids the initiation of tumorigenesis by preventing hepatocyte necrosis. However, once tumorigenesis has started, NF-B activation in the surrounding inflammatory cells acts as the driving force of further tumor development by promoting proliferation of the transformed hepatocytes.

Finally, a recent study using hepatocyte-specific IKK-deficient mice reveals a role for NF-B as a tumor suppressor in the liver because ablation of IKK causes the spontaneous development of chronic hepatitis resembling human nonalcoholic steatohepatitis that results in hepatocellular carcinoma. This carcinogenic process was shown to be initiated by hypersensitivity of the IKK-deficient hepatocytes to oxidative stress-dependent, FADD-mediated apoptosis, triggering compensatory hepatocyte proliferation, inflammation, and activation of liver progenitor cells (87). Thus, similar to hepatocyte-specific IKKß-deficient mice in the DEN-induced model of liver cancer, hepatocyte-specific IKK-deficient mice also indicate an essential role for the antiapoptotic effect of NF-B in the liver in avoiding compensatory proliferation leading to carcinogenesis. Moreover, in both of these models excessive JNK activation resulting from NF-B deficiency was suggested to play a causative role in hepatocyte apoptosis and consequent tumor development. DEN-induced carcinogenesis in mice with hepatocyte-specific deletion of both IKKß and JNK1 is completely abolished, showing that JNK activity is a principal mechanism by which IKKß deficiency in hepatocytes results in increased chemical hepatocarcinogenesis (109). In addition, IKK-deficient hepatocytes display a constitutive activation of JNK, suggesting that in these mice persistent activation of JNK also has a role in initiating hepatocyte apoptosis (87).

Taken together, the studies described above demonstrate a dual role for hepatocyte NF-B activation in hepatocarcinogenesis (Fig. 4). In early stages of tumorigenesis, its cytoprotective effect is beneficial by preventing hepatocyte cell death and thus avoiding compensatory proliferation, whereas in late stages of tumorigenesis it supports malignancy by promoting survival of the transformed hepatocytes. In nonparenchymal cells of the liver, NF-B activation during hepatocarcinogenesis is detrimental because it provides the tumor cells with essential growth factors, allowing them to keep on proliferating.

View larger version (67K): [in this window] [in a new window]   FIG. 4. Role of NF-B activation in hepatocarcinogenesis. NF-B activation in Kupffer cells leads to the production of TNF, which can cause either NF-B activation or cell death in hepatocytes, as well as IL-6 and other growth factors (GF), which can initiate hepatocyte proliferation. During early tumorigenesis, hepatocyte NF-B activation is beneficial because it protects against cell death and consequently prevents compensatory proliferation. If hepatocyte NF-B activation is impaired, TNF induces cell death, which leads to compensatory proliferation of other hepatocytes. This increases the risk of DNA mutations, which can give rise to tumor cells. During late tumorigenesis, NF-B activation protects the tumor cells against cell death. Moreover, the tumor cells are provided with essential growth factors due to NF-B activation in nonparenchymal cells of the liver, such as Kupffer cells. DD, Death domain.

D. Role of hepatic NF-B activation in the metabolic syndrome

NAFLDs are associated with a chronic subacute inflammatory state, and growing evidence links this inflammation to the development of the metabolic syndrome. The findings that TNF is overexpressed in adipose tissue of obese rodents and that TNF can cause insulin resistance constituted the first indications for this notion (112, 113). Further studies showed that antiinflammatory drugs can reverse insulin resistance because high doses of salicylates improved glycemia, insulin sensitivity, and hyperlipidemia in mice as well as in patients (114, 115, 116). Because salicylates are potent inhibitors of IKKß activity, these observations suggested an important role for NF-B activation in acquiring insulin resistance. Subsequent genetic evidence that IKK-dependent NF-B activation is implicated in the development of insulin resistance came from mice heterozygous for an IKKß null allele, which are partially protected from insulin resistance on an ob/ob background as well as after a high-fat diet (114). In contrast to this observation, another study could not detect an effect of IKKß heterozygosity in obesity-induced insulin resistance (117). Although a definite explanation for the conflicting results obtained in these studies is still lacking, it is possible that opposing contributions of NF-B activity in different organs obscure the effect of overall NF-B inhibition on insulin sensitivity.

Despite the controversy over the role of systemic NF-B activation in insulin resistance, two research groups have recently linked hepatic NF-B activation to the development of the metabolic syndrome. These studies applied transgenic approaches in genetic as well as high-fat diet-induced models of obesity to demonstrate a role for hepatic NF-B activation in the progression of fatty liver to steatohepatitis, insulin resistance, and type 2 diabetes mellitus (118, 119). Cai et al. (118) generated mice that express constitutively active IKKß in hepatocytes, which leads to hepatic production of proinflammatory cytokines such as TNF, IL-1ß, and IL-6. The subacute hepatic NF-B activation in these transgenic mice caused profound hepatic and more moderate systemic insulin resistance. Inhibition of NF-B activation by hepatocyte-specific expression of IB^s in these transgenic mice reversed this phenotype, reinforcing the notion that insulin resistance in these mice is a result of hepatic NF-B activation (118). Complementary to this study, Arkan et al. (119) used mice lacking IKKß specifically in hepatocytes. These mice retain insulin responsiveness and glucose tolerance in the liver, but develop peripheral insulin resistance in muscle and fat in response to high-fat diet, obesity, or aging (119). Taken together, these two studies show that NF-B activation in hepatocytes has a causative role in developing hepatic insulin resistance.

However, the observations that excessive NF-B activation in the liver leads to systemic insulin resistance and that lack of NF-B activation in hepatocytes could not prevent the development of insulin resistance in muscle and fat tissue suggest that different cell types could be involved in regulating systemic insulin sensitivity. In this context, it was shown that myeloid-specific deletion of IKKß leads to a global improvement of insulin sensitivity in mice on a high-fat diet (119). Because myeloid-specific deletion of IKKß prevents NF-B activation in macrophages, including Kupffer cells, this observation not only indicates a role for myeloid cells in regulating systemic insulin sensitivity but also suggests a role for Kupffer cells in the development of hepatic insulin resistance. Originally, adipose tissue was appreciated as the source for inflammatory cytokines that cause insulin resistance (20, 112). However, hepatocyte lipid accumulation represents a second potentially important site for inflammatory cytokine production (118). These proinflammatory cytokines, either produced by abdominal fat tissue and delivered via portal circulation, or produced by hepatocytes during steatosis, may then activate Kupffer cells. These activated Kupffer cells could be responsible for the subacute inflammation in the liver that causes progression to the metabolic syndrome (20). Overall, these findings suggest that IKKß-dependent NF-B activation in hepatocytes most likely acts in a paracrine manner to down-modulate insulin sensitivity in liver because production of inflammatory cytokines by hepatocytes that results from NF-B activation and leads to Kupffer cell activation causes hepatic insulin resistance. In contrast, systemic insulin resistance is caused by myeloid cells because IKKß-dependent NF-B activation in myeloid cells affects insulin action in all tissues (liver, fat, muscle).

It is important to mention that in line with the previously discussed models of TNF-mediated liver injury, during NAFLD NF-B activation is also essential to protect hepatocytes against TNF-induced cell death. In humans, obesity has been shown to increase the risk and severity of liver damage in alcoholics (120). Recently, inhibition of NF-B activation was shown to be responsible for this sensitizing effect of obesity in a mouse model for alcohol-induced liver injury. Ob/ob mice display reduced hepatic NF-B activation after chronic administration of ethanol, which unleashes the apoptotic effects of the simultaneously increased levels of hepatic TNF, resulting in more hepatocyte apoptosis in these obese mice in comparison with lean mice (121). Furthermore, a high-fat diet also sensitizes mice to TNF-mediated liver injury by diminishing NF-B activation. Indeed, feeding mice with a high-fat diet was shown to increase expression of IB in the liver, thus decreasing activation of NF-B and predisposing mice to increased hepatocyte apoptosis after partial hepatectomy (122). Although these studies are in support of a major role for NF-B inhibition in obesity-mediated sensitization to liver injury, one has to realize that leptin deficiency has a broad effect on innate as well as adaptive immunity, for instance by causing depletion of hepatic natural killer (NK) T cells (123). Moreover, depletion of these NKT cells not only occurs in genetically obese mice, but also after feeding wild-type mice a high-fat diet (124). Because hepatic NKT cell depletion promotes the production of proinflammatory cytokines in the liver, partially because of deregulated cytokine production by Kupffer cells (125), this phenomenon might be an important factor in the sensitization to TNF-mediated liver injury. Indeed, ob/ob mice as well as mice on a high-fat diet are extremely vulnerable to LPS-induced liver injury (126). In contrast to this observation, ob/ob mice are protected against Con A-induced T cell-mediated liver injury, which is associated with decreased hepatic production of TNF (127). Together, these data on the role of obesity in different models of TNF-mediated liver injury are quite puzzling. Given the systemic leptin deficiency of ob/ob mice, different effects of leptin on different cell types could account for some of these confusing results. Whereas leptin insufficiency could modulate the immune system such that, depending on the cell types involved, cytokine production in the liver is either more proinflammatory or more hepatoprotective, its effect on hepatocytes could decrease NF-B activity and sensitize these cells to second hits of liver injury.

Taken together, the above observations underscore the delicate balance of NF-B activation in hepatocytes during NAFLD, because excessive NF-B activation and subsequent inflammation may form the trigger for development and progression of some key features of the metabolic syndrome. On the other hand, NAFLD can predispose to certain second hits of TNF-mediated liver injury, which could partially be explained by deregulated cytokine production by nonparenchymal cells together with sudden drops in hepatocyte NF-B activity.

E. Role of NF-B-dependent inflammation in hereditary hemochromatosis
Hereditary hemochromatosis is a genetic disorder characterized by progressive iron overload in parenchymal tissue and is associated with a high risk of liver cirrhosis, development of hepatocellular carcinoma, cardiomyopathy, and diabetes (128, 129). In 90% of all patients, hereditary hemochromatosis is due to homozygosity for the point mutation C282Y in the HFE gene, which encodes for a major histocompatibility complex class I-like protein (130). Mice deficient in HFE or homozygous for the missense C282Y mutation develop iron overload that recapitulates hereditary hemochromatosis in humans, confirming that hereditary hemochromatosis arises from loss of HFE function (128, 131). In normal adults, storage iron is deposited in hepatocytes and tissue macrophages and mobilized in response to acute need. Erythrocytes are the major consumers of iron, and their demand normally exceeds the capacity of storage cells to mobilize iron for erythropoiesis, which enhances intestinal absorption of dietary iron and macrophage recycling of iron from hemoglobin. However, nonhematopoietic tissues easily assimilate iron, resulting in vigorous intestinal absorption and consequent deposition of excess iron in parenchymal cells (128). Tight regulation of iron levels is thus mandatory and is determined by both intestinal absorption and macrophage release of iron.

One key regulator of both intestinal iron absorption and macrophage iron release is hepcidin (132, 133, 134, 135), a peptide that is produced almost exclusively by hepatocytes in response to inflammatory stimuli and iron load (134, 136). Mice deficient in Usf2, a transcription factor essential for hepcidin production, lack hepcidin expression and demonstrate massive iron overload resembling the hemochromatosis phenotype in humans (137). On the other hand, mice with a hepcidin transgene develop severe iron deficiency anemia (138). These genetic studies in mice confirm the importance of regulating hepcidin synthesis for maintaining an appropriate iron balance. Induction of hepcidin expression by inflammatory stimuli seems to be mediated mainly by the inflammatory cytokine IL-6 produced by Kupffer cells, which triggers hepcidin transcription in nearby hepatocytes (139, 140). Kupffer cells are required for the activation of hepcidin synthesis during inflammation but are dispensable for the regulatory activity exerted by iron on hepatic hepcidin (141). Besides this crucial role for IL-6, TNF was also shown to regulate cellular iron distribution and as such modulates the severity of liver iron overload and damage (21, 142, 143, 144). The involvement of inflammatory cytokines and Kupffer cells in regulating the hepatic iron balance suggests a delicate role for NF-B activation in this process. Indeed, production of inflammatory cytokines resulting from NF-B activation in Kupffer cells causes in a paracrine way an increase of hepatocyte-derived hepcidin, which in turn augments the hepatic iron concentrations. Interestingly, chronic iron overload has been shown to trigger oxidative stress leading to activation of NF-B (145), suggesting that increased iron production in the liver can act as an amplifier of NF-B activation and hence of hepatic inflammation. Particularly important in this context, administration of iron to Kupffer cells was shown to activate NF-B and TNF promoter activity through enhancement of IKK activity (146). As such, uncontrolled NF-B activation and iron release form a feed-forward loop that can lead to excessive production of hepcidin and can result in serum anemia of inflammation. On the other hand, situations where NF-B activation is compromised might lead to insufficient hepcidin production and may therefore contribute to the pathology of hereditary hemochromatosis.

	VI. Mechanisms of the Antiapoptotic Effect of NF-B

Top