Article Text

Download PDFPDF

Following the TRAIL from hepatitis C virus and alcohol to fatty liver
  1. S C Afford,
  2. D H Adams
  1. Liver Research Laboratories and MRC Centre for Immune Regulation, Institute for Biomedical Research, Queen Elizabeth Hospital, University of Birmingham, Birmingham, UK
  1. Correspondence to:
    Professor D Adams
    Liver Research Laboratories, Institute for Biomedical Research, University of Birmingham Medical School, Wolfson Drive, Birmingham B15 2TT, UK;

Statistics from

Tumour necrosis factor related apoptosis inducing ligand or TRAIL, is a novel mediator of fatty liver disease which may provide a mechanism to explain the development of steatosis in hepatitis C virus infection and in response to alcohol

Fatty liver or steatosis is a common finding in several liver diseases, most notably non-alcoholic fatty liver disease (NAFLD) and alcoholic liver disease (ALD) but is also seen more frequently in chronic hepatitis C virus (HCV) related liver disease than would be predicted by simple concurrence of the two diseases.1 Several mechanisms have been proposed to explain why steatosis might develop in chronic viral infection, including direct effects of the virus on lipid metabolism, but none of these is entirely convincing. Furthermore, steatosis is not a benign lesion but one that contributes to the progression of fibrosis, not only in NAFLD and ALD but also in HCV, in part by increasing the sensitivity of the liver to oxidative stress and cytokine mediated injury.2 The paper by Mundt and colleagues3 in this issue of Gut provides novel mechanistic insights into why steatosis develops in HCV and how it might accentuate liver injury (see page 1590). They have concentrated their efforts on a member of the tumour necrosis factor (TNF) superfamily called TRAIL or TNF related apoptosis inducing ligand. TRAIL is known to induce apoptosis in transformed cells and this group have previously shown TRAIL, acting through one of its receptors called TRAIL-DR5 or TRAIL-2, mediates hepatocyte apoptosis in viral hepatitis. What they now show is that this receptor-ligand pair can also mediate hepatic steatosis in both viral hepatitis and in response to alcohol.

The authors start with the observation that expression of TRAIL is increased in the livers of patients with HCV associated steatosis and they then proceed to investigate the functional relevance of these observations in animal models. They used adenoviral gene transfer to express TRAIL in mouse liver and found that overexpression of TRAIL itself had no effect in the healthy liver. However, if gene transfer of TRAIL was preceded by high dose adenoviral infection, this sensitised the liver to respond to TRAIL expression with steatosis and hepatocyte apoptosis. Furthermore, by looking at expression of TRAIL receptors in the infected livers, the authors were able to show that viral infection downregulates a TRAIL decoy receptor while increasing expression of the death domain associated TRAIL receptor TRAIL-R2, thereby providing a mechanism to explain how viral infection can amplify the effects of TRAIL. Furthermore, induction of steatosis was specific for TRAIL because overexpression of another TNFSF member Fas-L was associated with apoptosis but not steatosis.

Because nuclear factor κB (NFκB) activity is critical in determining the outcome of activation of many TNF receptors, the authors then investigated the effects of inhibiting NFκB and found that the loss of NFκB activation markedly sensitised hepatocytes to TRAIL mediated apoptosis and also resulted in the accumulation of larger amounts of fat in hepatocytes. Finally, because TNF-α plays an important role in alcoholic fatty liver disease, they investigated whether alcohol could sensitise mice to the effects of TRAIL. Intriguingly, feeding mice with 20% ethanol for four days did not lead to changes in TRAIL receptors but did sensitise animals to subsequent exposure to TRAIL but in contrast with virus infection ethanol led only to steatosis in response to TRAIL and not to apoptosis. TNF-α has been shown to mediate steatosis via effects on insulin resistance which can be inhibited by treatment with metformin leading the authors to investigate metformin treatment in their model. However, metformin had no protective effect on TRAIL mediated steatosis, suggesting that TRAIL induced steatosis is a direct consequence of TRAIL-R2 activation.

These observations are of major significance because they show that TRAIL is a novel mediator of fatty liver disease which may provide a mechanism to explain the development of steatosis in HCV infection. Furthermore, the fact that TRAIL expression is harmless in healthy livers whereas in the presence of viral infection it mediates hepatocyte apoptosis and steatosis and after alcohol mediates steatosis without apoptosis suggests that the outcome of TRAIL-2 activation will depend on the presence of local signals, including those mediated by other TNF family members. These data provide an exciting new insight into the pathogenesis of steatosis but also raise a note of caution for the clinical trials of TRAIL which, at present, show considerable promise for the treatment of malignant disease, including hepatocellular carcinoma.4–6 Thus the authors suggest that all trials involving TRAIL should consider the potential hepatotoxic side effects, particularly in patients with inflammatory, viral, and/or alcohol related liver disease.

There are at least 19 identified TNF superfamily ligands capable of binding to one or more of the 27 identified members of the TNF receptor (TNFR) superfamily.7 Several of these receptor ligand pairs are expressed in the liver, some constitutively and some at sites of active inflammation,8 making interpretation of the pathophysiological significance of interactions between any of these receptors and their ligands difficult. However, the study of Mundt and colleagues3 reinforces the potential benefits from studying the biological functions of this family of molecules in physiologically relevant model systems. TNFRs are seldom expressed in isolation on any particular cell type, and are usually found in combinations that regulate a wide range of functions, sometimes in an apparently cell specific manner.9,10 In general, expression of the ligands tends to be more restricted, which makes them a more attractive proposition to target therapeutically. However, under certain conditions, liver epithelial cells (hepatocytes and cholangiocytes) express both receptors and ligands allowing autocrine and paracrine interactions to regulate cellular responses.8 The situation is complicated further because several of the TNF ligands bind to more than one TNFR and moreover activation via one receptor can influence activation of another.9,10 An example of the latter mechanism is the ability of activation through CD40 on hepatocytes to induce increased expression of Fas and secretion of Fas ligand resulting in autocrine activation of the Fas pathway.11 Similar cooperative interactions have been described for other receptors, including TNFR2 and TNFR1.12 Because several of these receptors mediate the same effects (for example, induction of apoptosis) it must be presumed that a considerable degree of redundancy operates within this family of molecules.

Investigation of this molecular superfamily has come a long way since the initial reports of the immunomodulatory functions of TNF-α, Fas/FasL, and CD40/CD154.10 It is now apparent that TNFSF members influence a wide range of biological functions in many cell types. In the liver, they are involved in modulating cell survival and apoptosis, in epithelial, endothelial, and stromal cells. They are crucial for the control of liver growth and regeneration as well as mediating the consequences of both acute and chronic inflammation.13 TRAIL, which was originally identified by virtue of its sequence homology to Fas ligand, is a type II transmembrane protein with 281–291 amino acids and an extracellular region which is cleaved as a soluble molecule.14,15 TRAIL is upregulated on the membrane of activated lymphocytes, including hepatic NK and NK T cells, and secreted by neutrophils and monocytes on exposure to type 1 interferons.16 A major difference between TRAIL and CD95L or TNF-α is its ability to induce apoptosis of cell lines and tumour cells while displaying minimal toxicity on normal cells.17 This led to interest in TRAIL as a therapy for cancer and subsequent clinical trials with soluble TRAIL.18 There are currently four known membrane receptors for TRAIL and one dimeric soluble secreted receptor, osteoprotegerin.16 TRAIL-R1 and TRAIL-R2 are classical TNF receptors and contain a cytoplasmic death domain (which is also present in other family members, including TNFR1 and Fas19) and can activate both caspases and NFκB (fig 1). The other two membrane receptors, TRAIL-R3 (DcR1) and TRAIL-R4 (DcR2), have truncated death domains and are not capable of activating the caspase cascade. This led to the assumption that they act as decoy receptors but they may activate NFκB and block apoptosis16 and it is possible, given the precedent with other TNFRs including TNFR2, CD40, and CD30, that they might promote apoptosis via autocrine or paracrine activation of other TNFR1 and Fas.11

Figure 1

 Tumour necrosis factor related apoptosis inducing ligand (TRAIL) receptor signalling. Schematic diagram showing the major intracellular signalling pathways activated following TRAIL ligation of its membrane bound receptors. For TRAIL receptors 1 and 2 (TRAIL-R1 and TRAIL-R2) the dominant pathway is initiation of apoptosis. However, the ability to activate nuclear factor κB (NFκB) and c-Jun may modulate the outcome by inducing expression of both antiapoptotic and proapoptotic signals. Current evidence suggests that TRAIL-R4 may be able to activate NFκB, providing a mechanism by which it could antagonise the proapoptotic effects of TRAIL-R2 activation (adapted from Kimberley and Screaton16 and MacFarlane19).

The true biological function of TRAIL has been difficult to define but current evidence suggests that its main function is as a regulator of the innate immune response during immune surveillance against tumours and virus infected cells.15,18 The study by Mundt and colleagues3 provides further evidence that TRAIL is a critical determinant of outcome and tissue damage in inflammatory liver disease. Given the prominent role of TNF-α in the pathogenesis of steatohepatitis,20,21 it was logical to investigate other members of the TNF superfamily. TRAIL was particularly interesting because tumour cells were thought to be sensitive to TRAIL induced apoptosis whereas normal hepatocytes were not22 although such assumptions were probably too simplistic.23 TRAIL-R2 appears to be similar to other members of the TNFR family in having a potential role in the clearance of virally infected cells, including hepatitis B,24,25 but what is novel about the current study is the ability of TRAIL acting through TRAIL-R2 to induce steatosis as well as apoptosis. The authors show that the outcome of TRAIL-R2 activation is determined by the presence or absence of viral infection or toxic agents such as alcohol. Elucidating the mechanisms behind these effects will be important for understanding why particular disease triggers are associated with specific patterns of liver injury. The viral effect may be a consequence of the release of type 1 and type 2 interferons which can induce TRAIL expression on immune effector cells and prime cells for TNF mediated steatosis.24 Furthermore, transfection of Hep G2 cells with the HBV encoded X antigen increases their sensitivity to TRAIL mediated apoptosis, suggesting that some viral protein may have direct effects on this pathway.25 The mechanisms by which ethanol promotes TRAIL mediated steatosis is even less clear. Both apoptosis and steatosis mediated by TRAIL are accentuated if NFκB is inactivated, suggesting that NFκB plays a protective role.18,26 Thus some of increased sensitivity to the effects of TRAIL in inflammatory liver disease may be a consequence of alterations in NFκB activation, possibly as a result of signals provided by other TNF family members.24

It is not clear whether TRAIL in these models is presented as a membranous ligand by haematopoietic cells, as an autocrine/paracrine signal on hepatocytes, or as a soluble cytokine. Previous work in murine models of liver inflammation suggests that TRAIL expression on mononuclear cells is sufficient to drive liver damage27 but in models of adenoviral infection lymphocyte TRAIL is not necessary, suggesting a paracrine/autocrine mechanism involving hepatocyte TRAIL.24 In addition, TRAIL may be operating through other liver cell types. Apoptosis of stellate cells is an important component of the resolution phase of hepatic fibrosis28 and activated stellate cells show increased expression of TRAIL-R2 and increased sensitivity to apoptosis by TRAIL.29 The function of soluble TRAIL may prove to be important for the future of clinical trials with TRAIL. So far hepatotoxicity does not appear to have been a major side effect of TRAIL therapy but clearly from the study of Mundt and colleagues3 the chances of toxicity will be greater in patients with concomitant liver disease, many of whom will be screened out of phase 1/11 clinical trials and treated. Animal studies suggest that soluble TRAIL administered in vivo is much less likely to induce liver injury compared with membrane bound TRAIL. This may be because soluble TRAIL cannot engage death receptors as efficiently as membrane TRAIL or even because the soluble form inhibits activity of membrane TRAIL, as has been shown for soluble and membrane bound Fas-L.

Mundt and colleagues3 conclude that TRAIL may be a new therapeutic target for inhibition of hepatic steatosis. However, before this can become a reality, more needs to be known about the complex regulation and interplay of the different TNF family members in the context of liver inflammation. A further caveat also arises from the important differences between the murine and human TRAIL systems. Mice only express one death receptor, which is structurally related to human TRAIL-R2, and the mouse homologues of human TRAIL-R3 and TRAIL-R4 show distinct differences in structure with their human counterparts, suggesting independent evolutionary origins from the human receptors. Thus it is important to interpret results obtained with mouse models with caution before extrapolating them to humans.16,18

Tumour necrosis factor related apoptosis inducing ligand or TRAIL, is a novel mediator of fatty liver disease which may provide a mechanism to explain the development of steatosis in hepatitis C virus infection and in response to alcohol


View Abstract


  • Conflict of interest: None declared.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles