Article Text

Download PDFPDF

BASL abstracts

Statistics from

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.

This abstract book has been produced by the BMJ Publishing Group from electronic files supplied by the authors. The abstracts have been formatted for consistency but not edited for content.

Every effort has been made to reproduce faithfully the abstracts as submitted. However, no responsibility is assumed by the publishers or organisers for any injury and/or damage to persons or property as a matter of product liability, negligence or otherwise, or from any use or operation of any methods, products, instruments, or ideas contained in the material herein. We recommend independent verification of diagnoses and drug dosages.


D. Schuppan.Department of Medicine I, University of Erlangen-Nuerberg, Germany

Antifibrotic therapies should preferentially be targeted to the activated hepatic mesenchymal cells. Those cells resemble wound healing myofibroblasts and synthesise an excess of matrix proteins. They derive from quiescent hepatic stellate cells and (myo-) fibroblasts. Several fibrogenic cytokines and other mediators trigger or mitigate their activation. Although various agents have been shown to inhibit hepatic stellate cell/myofibroblast proliferation and collagen synthesis in vitro, only few of them are effective in suitable animal models in vivo, and finally in man. Useful animal models are rat secondary biliary cirrhosis or the speed of reversion of fibrosis after withdrawal of a hepatotoxin like thioacetamide.

The interferons (IFNγ>αβ) have proven antiproliferative and fibrosuppressive activity on mesenchymal cells in culture. Retrospective data suggest that IFNα therapy for hepatitis C can halt or even reverse fibrosis. However, this has to be confirmed by randomised prospective studies. Strategies to inhibit the key profibrogenic cytokine TGF-β and the associated connective tissue growth factor, eg by soluble decoy receptors, are evolving, but results are not yet convincing. IL-10 has little if any antifibrotic effect and hepatocyte growth factor, an epithelial mitogen, enhances hepatocyte regeneration with the inherent risk of promoting the development of hepatocellular carcinoma in cirrhotic livers.

Combination therapy of several potential antifibrotic agents appears promising. Such agents are silymarin, a defined mixture of flavonoids, sho-saiko-to, which contains related compounds like baicalein, halofuginone, another plant-derived drug, the phosphodiesterase inhibitor pentoxifylline, oral inhibitors of the endothelin-A-receptor, or inhibitors of the renin-angiotensin system.

Drug targeting to the fibrogenic liver cells is now possible by use of cyclic peptides that bind to receptors that are specifically upregulated on activated stellate cells, eg those for platelet derived growth factor or collagen type VI. Together with the evolving validation of serological markers of fibrogenesis and fibrolysis an effective and individualised treatment of liver fibrosis is anticipated.


















C. Trautwein.

Department of Gastroenterology, Hepatology and Endocrinology, Medizinische Hochschule Hannover, 30625 Hannover, Germany

Earlier results indicated that interleukin 6 (IL–6) is involved in the pathogenesis of acute and chronic liver diseases. IL–6 belongs to a family comprising of IL–6, IL–11, leukemia inhibitory factor (LIF), oncostatin M (OSM), ciliary neurotropic factor (CNTF), and cardiotropin 1 (CT-1). All family members need the gp130 molecule for signal transduction. In contrast to knockout animals for individual IL–6 family members null mice for the gp130 receptor are embryonal lethal.

We first studied the role of IL–6 in acute and chronic liver diseases. IL–6 expression in the liver of patients with acute and chronic liver diseases was significantly enhanced compared with normal livers. Additionally IL–6 serum levels in the patients with chronic liver disease were significantly increased compared to healthy controls and correlated with the degree of fibrosis according to Child-Pugh. Moreover we found in these patients an inverse correlation of IL–6 with serum markers of liver function and a positive correlation with inflammatory markers and signs of portal hypertension.

In order to study the relevance of these clinical findings we used the Cre/lox system to generate conditional gp130 knockout animals. Animals were generated where gp130 is deleted specifically in hepatocytes or in all cells of the liver (parenchymal and non-parenchymal liver cells). Using this approach we studied the relevance of gp130 dependent pathways in hepatocytes or non-parenchymal liver cells during different pathophysiological conditions. The animals were used to perform partial hepatectomies, LPS-stimulation experiments, and chronic CCL4 injection studies. These results collectively demonstrated that gp130 dependent pathways in parenchymal and non-parenchymal liver cells activate protective pathways in order to maintain liver physiology while no direct impact on hepatocyte proliferation was found.

As these results indicated a direct role of the IL–6/gp130 system in the progression of liver disease, we now developed approaches to potentially treat patients with acute and chronic liver diseases. Two molecules were investigated. ME3738 has been shown before that it attenuates liver disease in several models of acute and chronic liver injury. We used the Concanavalin A model of acute liver failure to study its molecular mechanism. By using IL–6 and gp130 hepatocyte specific knockout animals we show that ME3738 induces IL–6. Consecutively by activating gp130 dependent pathways specifically in hepatocytes the animals are protected from Con A induced liver failure. The second compound is a designer protein (hyper-IL–6) where the IL–6 has been fused to its soluble receptor gp80. This molecule is 10-fold more potent to activate gp130 dependent pathways in hepatocytes and also this drug is able to confer protection in different models of liver injury.

In summary, our results indicate an essential role of the IL–6/gp130 system in the progression of acute and chronic liver disease in humans and in animal models. Therefore, new approaches to activate this pathway in vivo could be of potential therapeutic interest.


N. Naoumov.

Institute of Hepatology, University College London and UCL Hospitals, UK

The hepatitis B virus (HBV) is a non-cytopathic virus and the outcome of HBV infection depends primarily on the host immune response to the virus. It is established that self-limited acute hepatitis B is associated with strong T cell responses against viral antigens, while these responses are weak or undetectable in patients with chronic HBV infection.1 Furthermore, studies using HBV transgenic mice revealed that virus-specific cytotoxic T cells can abolish HBV gene expression and replication without killing infected hepatocytes.2,3 This potent antiviral effect is mediated by IFN-γ and tumour necrosis factor alpha (TNF-α), which are secreted by activated T cells. Based on this concept, the strategy for treating patients with chronic hepatitis B should aim to enhance T cell reactivity to HBV and to establish a long term, host immune control over viral replication.

One potential approach is by using interleukin-12 (IL-12), as the most potent inducer of IFN-γ production from T and NKT lymphocytes. In the HBV transgenic mouse model, the administration of recombinant IL-12 for only 3 days led to disappearance of viraemia and complete inhibition of viral replication in the liver without biochemical or histological evidence of hepatitis.4 We investigated the antiviral activity of recombinant human IL-12 using an in vitro model where lymphocytes obtained from patients with chronic hepatitis B were co-cultured with transfected hepatocytes supporting HBV replication.5 The effector and target cells were separated by a membrane, thus only allowing a transfer of soluble factors. There was a good correlation between the level of IFN-γ produced by patients’ lymphocytes and the degree of HBV DNA reduction in the target cells. We then progressed to a randomised trial to investigate in vivo the antiviral effect of human recombinant IL-12 given in combination with lamivudine in comparison to treatment with lamivudine alone.6 Although the combination regimen had a significantly greater antiviral activity, it did not abolish HBV replication in vivo.

Another approach to alter the balance in virus—host interactions is by directly suppressing HBV replication with antiviral agents. In this context we investigated the HBV-specific T cell reactivity during treatment with an antiviral agent—adefovir dipivoxil.7 One important aspect of these analyses is that they allowed demonstrating the heterogeneity of the patients and their response to antiviral treatment. Restoration of T cell reactivity was observed only in a proportion of patients treated with adefovir dipivoxil and these were the cases with a sustained virological response.

A further possibility is to directly correct the inefficient T cell responses to HBV. We have recently demonstrated this in chronic HBsAg carriers, who received bone marrow from an HLA-identical donor with natural immunity to HBV.8 The replacement of host’s lymphocytes with T cells from the donor led to resolution of chronic hepatitis B with seroconversion to anti-HBs. Furthermore, this study provided evidence that transfer of hepatitis B core specific CD4+ and CD8+ T cells is associated with resolution of chronic HBV infection. The data suggest that therapeutic immunisation with HBV core antigen deserves further in vivo investigations in patients with chronic hepatitis B.










D. Rockey.

Duke University Medical Center, USA

The response to recurrent injury, in liver and in other organs, is one of wound healing. Chronic injury and wound healing in the liver ultimately lead to hepatic fibrosis and cirrhosis. One of the most important clinical sequelae of cirrhosis is portal hypertension, which is associated with significant morbidity and mortality. Although the pathophysiologic basis of portal hypertension is multifactorial, the most prominent aspect is an increased intrahepatic resistance to blood flow, which in turn is the result of both fixed and dynamically modulable vascular elements. Recent data point to several key factors in the altered vascular response that occurs in cirrhosis. For example, during liver injury and cirrhosis, resident smooth muscle-like perisinusoidal cells termed hepatic stellate cells within the hepatic sinusoid (analogous to the capillary in the systemic circulation) undergo transformation from a quiescent to an “activated” state. This process is characterised by de novo expression of smooth muscle proteins including smooth muscle isoforms of each actin and myosin. Moreover, these abundant contractile proteins impart on stellate cells enhanced contractility, elicited in particular by the endothelins, a group of 21 amino acid peptides known primarily for their vasoactive properties. Moreover, in injured liver, endothelin-1 (ET-1) is produced in increased quantities by stellate cells themselves; this paracrine loop involving ET-1 in the context of highly contractile (vasoregulatory) stellate cells results in increased resistance to blood flow typical of portal hypertension.

The dynamic nature of blood flow regulation emphasises the importance of interplay between vasoconstrictive and vasodilatory compounds. Since ET-1, a potent vasoactive peptide, is overproduced during liver injury and contributes to increased intrahepatic resistance via contraction of stellate cells, we postulated that nitric oxide (NO), a key molecule in normal vascular homeostasis, could balance the effect of ET-1 in liver, serving as an important relaxing factor for stellate cells. Although this indeed appears to be the case in the normal liver, in the injured liver, we and others have found that endothelium-derived NO (ie that produced by endothelial cell NO synthase) is not produced in increased quantities such that it counterbalances ET-1, but rather that endothelium-derived NO production is reduced. These data suggest that the increased intrahepatic resistance to blood flow typical of cirrhosis is the result of an endothelialopathy in which reduced intrahepatic NO plays an important role.

Given the clear dysregulation of ET-1 and NO in the hepatic microvascular unit after injury, we have proposed to address these defects by inhibiting ET-1 production and/or replenishing NO in the cirrhotic liver in specific cellular compartments.


S. Gupta.

Albert Einstein Institute, New York

The ability to repopulate the liver with transplanted cells will help develop novel therapies. A variety of recent animal studies have begun to advance insights into mechanisms regulating liver repopulation. Although transplanted hepatocytes integrate in the liver parenchyma and function normally throughout the life of rodents, transplanted cells do not proliferate significantly in the normal, unperturbed liver. Transplanted cell proliferation is activated when survival of native hepatocytes is perturbed. Under suitable situations, the liver can be repopulated extensively with transplanted hepatocytes. Moreover, transplanted hepatocytes can repopulate the liver of animals with chronic liver disease. Hepatocyte transplantation will thus apply to chronic liver disease, where structural and cellular perturbations alter hepatic function, inherited genetic disorders with extra-hepatic target organs, such as kernicterus due to congenital jaundice, encephalopathy due to hyperammonemia, and additional conditions, such as familial hypercholesterolemia, which is associated with severe atherosclerosis. Replacement of deficient liver function after orthotopic liver transplantation (OLT) in these conditions suggested that hepatocyte transplantation will be successful. Furthermore, if transplanted cells proliferated in the host liver extensively, people could be treated with relatively small numbers of hepatocytes from a single liver. This implies that cells from a single donor liver could be used to treat many patients. Proliferation of transplanted cells could potentially prolong survival in acute liver failure and offer a bridge toward OLT in less difficult circumstances; perhaps, OLT could be avoided altogether if the native liver regenerated in the additional time provided by hepatocyte transplantation. Therefore, an ability to repopulate the liver with cells offers exciting new opportunities for correcting genetic disorders and for treating liver failure. Use of several animal models has begun to establish the framework for therapeutic liver repopulation. Among presently unresolved issues are ways to repopulate the normal liver in an optimal fashion, as well as development of alternative sources for human cells in view of the scarcity of donor livers. Significant efforts are being devoted to these issues. Use of liver stem cells could be one potential way to alleviate organ shortages and advance applications of cell therapy. It is now established that the liver contains progenitor cells capable of differentiating into hepatocytes. Similarly, stem cells derived from other organs, such as bone marrow, can differentiate into mature hepatocytes. Finally, pluripotential human embryonic stem cells have also been isolated. Progenitor cells are capable of proliferating extensively in culture and are often amenable to highly efficient gene transfer. These advances indicate that liver directed cell therapy offers exciting possibilities for novel therapies.

Animal studies


















Clinical studies






Recent review



I.N. Crispe.

The David H Smith Center for Vaccine Biology and Immunology, The University of Rochester, Rochester NY, USA

The activation of CD8+ T cells in a diverse set of experimental models results in the preferential accumulation and apoptosis of these cells in the liver. This results, at least in part, from the expression of adhesion molecules on activated T cells, and of receptors for these adhesion molecules on the endothelium of the liver. However, the capacity of the liver to present antigen also influences T cell accumulation and apoptosis. Current experiments are designed to determine whether the apoptosis of CD8+ T cells in the liver is an example of activation induced cell death, mediated through death receptors, or of passive cell death, due to the lack of survival signals.

The death of T cells in the liver is accompanied by damage to hepatocytes. Some of this may be direct cytotoxic damage due to CTL recognition of antigenic peptide presented by hepatocytes, but some is independent of antigen recognition on hepatocytes. This mechanism we term “collateral damage”.

We have proposed that pathogens may exploit the capacity of the liver to promote T cell apoptosis, and a prime candidate is hepatitis C virus. To investigate this, we have built a series of baculovirus vectors encoding individual HCV proteins linked to a defined T cell target epitope. The HCV core protein exerts effects on CD8+ T cells, when these cells recognise antigen expressed by HCV core-expressing hepatocytes. This may be one of the many mechanisms through which HCV evades the immune system.


P.A. Knolle.

Institute of Molecular Medicine and Experimental Immunology, University of Bonn, Germany

Among the different cell populations in the liver, the sinusoidal endothelial cells (LSEC) are strategically positioned in the hepatic microvessels to interact with T cells passing with the blood stream through the liver. LSEC constitutively express MHC class I and II molecules as well as co-stimulatory molecules such as CD80, CD86, and CD40 suggesting an APC function. Furthermore, LSEC are endowed with pattern recognition receptors such as the mannose receptor, scavenger receptors, and L-SIGN allowing very efficient receptor mediated uptake of antigens.

Similar to dendritic cells, LSEC have the capacity to present soluble exogenous antigens on MHC class I molecules to CD8+ T cells, a process termed cross-presentation. Cross-presentation by LSEC involves receptor-mediated endocytosis, endosomal maturation, proteasomal processing, and TAP-mediated loading onto de novo synthesised MHC class I molecules. On a cell to cell basis, LSEC are almost 100 times more efficient in cross-presentation than dendritic cells. Encounter of naive CD8+ T cells with LSEC cross-presenting the cognate antigen results in T cell stimulation as evidenced by proliferation and cytokine release. In contrast to dendritic cells, LSEC do not induce effector function of CD8+ T cells. Contact with cross-presenting LSEC leads to loss of cytokine expression and specific cytotoxicity during subsequent antigen-specific (re)stimulation or encounter of target cells presenting the relevant antigen. Thus, LSEC appear to render CD8+ T cells nonfunctional in an antigen-specific manner. Using adoptive transfer of LSEC, we can demonstrate that specific immune tolerance is induced by LSEC in vivo.

These observations have implications for control of local immune reactions in the liver as well as immune reactions to those antigens distributed via the blood stream, which finally end up in the liver. Oral antigens can be found within 1–2 hours after ingestion in LSEC leading to T cell proliferation, cytokine release but finally rendering CD8± T cells tolerant to the orally applied antigen. We propose a new role of LSEC in modulation of immune reactions towards antigens expressed in the liver (by hepatocytes or other hepatic cell populations) and antigens distributed via the blood stream, thus rendering the liver an important immunological organ.


J.L. Boyer.

Liver Center, Yale University School of Medicine, New Haven, CT 06520-8019, US

Bile formation is a vital function of the liver that facilitates intestinal lipid absorption and enables the excretion of toxic endogenous substances and xenobiotics. Advances in cellular and molecular biology have rapidly increased our knowledge of the physiology and pathophysiology of this important process. This lecture will briefly summarise the historical developments in this field and then discuss current concepts of the molecular mechanisms of cholestasis.

A number of ABC transporters in liver and kidney regulate the transport of bile salts and other organic solutes. Recent studies indicate that several of these transport proteins undergo adaptive regulation in response to cholestatic liver injury that serves to protect the liver from the accumulation of bile salts and other toxicants. Our laboratory has been interested in how these transporters respond to cholestatic liver injury and how these adaptive responses are regulated.

Using common bile duct ligation (CBDL) and other models of cholestasis in the rat, we and others have determined that the expression of hepatic Mrp2 (Abcc2) is down regulated,1,2 whereas Mrp3 (Abcc3) is substantially up regulated in hepatocytes and cholangiocytes3,4 The P-glycoprotein Mdr1 (Abcb1), and the sister of P-glycoprotein (Abcb11), now known as the bile salt export pump, Bsep, are upregulated or minimally down regulated respectively in the liver.5 Following cholestasis, urinary bile salt excretion is substantially increased and the ileal sodium dependent bile salt transporter, Isbt (Slc10a2), that is expressed on the luminal membrane of the proximal tubule (and normally reabsorbs bile salts from the glomerular filtrate) is down regulated.

In contrast to the liver, Mrp2, which is also expressed at the luminal membrane of the renal proximal tubule, is up regulated.6 Mrp3 is expressed on the basolateral membrane of the proximal tubule of rat kidney and does not change after CBDL. Mrp2 and Mrp3 are also expressed on the luminal and basolateral membranes, respectively, of the intestine. However, the levels of expression vary reciprocally with Mrp2 expression highest in the duodenum while the highest expression of Mrp3 is in the ileum and colon. This pattern of Mrp expression in intestine is not altered after CBDL (Soroka et al, submitted). Mdr1 is also expressed predominantly in the distal intestine on the luminal membrane and is down regulated after CBDL.

Recent studies indicate that the Mrp2 promoter contains a RARα:RXRα cis element and that Il-1β may suppress Mrp2 promotor induction in-vitro.7 Cytokines, bile acids, and other substances that accumulate in the liver during cholestasis thus might alter the expression of Mrp2 by acting as specific ligands for nuclear hormone receptors such as RARα:RXRα, CAR, PXR, and FXR.7,8 We therefore examined the effects of CBDL on the nuclear expression and Mrp2 promoter binding of RARα and RXRα. Results indicate that CBDL down regulates liver Mrp2 RNA and protein in association with a loss of RARα and RXRα nuclear proteins and diminishes RNA expression. Binding of RARα:RXRα to the Mrp2 promoter is diminished. In contrast, renal Mrp2 protein is upregulated, RNA is unchanged and there is no change in renal RARα and RXRα nuclear protein or RNA. Cytokine treatment of primary hepatocytes reduces RXRα nuclear protein levels.9 These studies indicate that CBDL induced cholestasis leads to differences in expression of the same ABC transporter in liver and kidney and that these differences may relate to organ specific effects of ligand mediated nuclear receptor regulation of gene expression. Preservation of Mrp2 expression in kidney may permit urinary excretion of toxic organic anions and xenobiotics when biliary excretion is impaired.











L.J. Elrick, V. Leel, M.C. Wright.

Department of Molecular and Cell Biology, University of Aberdeen, Aberdeen, UK

Cytochrome P450s are drug metabolising enzymes that are expressed at high levels in the liver. An important aspect of cytochrome P450 regulation is their modulated expression in response to administered drugs. One of the earliest identified and most potent inducers of cytochrome P450 expression is phenobarbitone. Phenobarbitone activates the transcription of certain cytochrome P450 genes—notably CYP2B and CYP3A sub-families—through activation of the nuclear constitutive androstane receptor (CAR). However, phenobarbitone does not directly bind to the CAR. The mechanism by which phenobarbitone activates the CAR is unknown. Previous work in this lab has shown that the synthetic glucocorticoid dexamethasone binds to membrane bound fractions isolated from rat liver. Binding activity was referred to as “low affinity glucocorticoid binding site” (LAGS) activity and previous work has shown that LAGS also bound phenobarbitone. The identity of this binding site was therefore examined.

The interaction of radiolabelled dexamethasone in rat liver microsomes was determined by equilibrium analysis and shown to be specific and saturable (with a KD of 50 nM). Ligand selectivity was examined by competition analysis. Other steroids were able to compete with dexamethasone for binding in a concentration-dependent manner—in particular prednisolone and progesterone (IC50 ∼ 100 nM), whereas non-steroidal glucocorticoid receptor antagonists RU486 and triamcinolone acetonid did not (IC50 > 100 μM). The cytochrome P450 2B and 3A sub-family inducers metyrapone and phenobarbitone, at concentrations required for the induction of cytochrome P450 expression in hepatocytes, also competed with dexamethasone for binding to this site. A cDNA sequence encoding a putative membrane-associated progesterone binding protein termed ratp28 has recently been identified in rat liver. The ratp28 cDNA was cloned and an additional sequence-related variant termed HC5 identified. Ratp28 was expressed in COS-7 cells and reconstituted dexamethasone binding activity in cell extracts. Other LAGS drug/xenobiotic ligands (metyrapone, phenobarbital) also competed with dexamethasone for binding to the expressed ratp28. These data suggest that the ratp28 cDNA encodes LAGS binding activity. The identity of the phenobarbitone receptor in rat liver may therefore be ratp28. These results suggest the existence of a novel steroid/xenobiotic signalling pathway in the regulation of orphan nuclear receptor function and cytochrome P450 gene expression.


SC Ryder on behalf of The Trent Hepatitis C Study Group.

Queen’s Medical Centre, Nottingham

Studies of the rate and risk factors for progression of fibrosis in hepatitis C mainly rely on single liver biopsies with estimated dates of infection. This has inherent potential bias. We have studied 221 patients with untreated hepatitis C who have been prospectively followed and had a repeat liver biopsy at a mean interval of 2.95 years. Fibrosis and necroinflammatory change (Ishak) were assessed by three pathologists unaware of the date of biopsy. Each pair was scored by the same pathologist. 67/221 (30%) showed progression of one point in the Ishak fibrosis scale and 12% showed progression of 2 points or more. If a linear rate of progression is assumed, mean rates are 0.33 Ishak points per year, giving an infection to cirrhosis average duration of 18 years. Risk factors for progression >2 Ishak points are shown in table (excludes patients with Ishak 5 on index biopsy).

Abstract 10

Alcohol intake, ferritin, mean and peak ALT, steatosis score, gender, route of transmission, prior hepatitis B infection, and HCV genotype were not significantly associated with fibrosis progression. Identical results were seen with 1 point progression. We conclude, even relatively mild HCV related liver disease has a significant risk of progression. The strong relationship of progression with fibrosis on the index biopsy and the much weaker relationship with age at infection or duration of infection strongly suggests that fibrosis in HCV is a non-linear process, accelerating once fibrosis is initiated. Increased hepatic inflammation predicts fibrosis progression, focal necrosis being the most significant component of the Ishak NI score.


M. Bertrand, M.P.J. Arthur, D.A. Mann.

Liver Group, Southampton General Hospital, Southampton SO16 6YD, UK

The tissue inhibitor of metalloproteases 1 (TIMP-1) has been shown to play a major role in the pathogenesis of liver fibrosis and is overexpressed in activated hepatic stellate cells (HSC). Therefore, identifying the transcription factors involved in the regulation of TIMP-1 gene expression is a critical step towards the understanding of the disease. A regulator element in the TIMP-1 promoter region, called UTE-1, has been described to be important for the transcriptional regulation as the deletion of this site decreases the activity of the promoter. The aim of this study was to identify the proteins that bind to the UTE-1 site and to characterise how UTE-1 binding proteins regulate transcription.

Using the yeast one hybrid system, the UTE-1 binding protein was identified as belonging to the Runx protein family, which included three proteins (Runx1, 2, and 3). These proteins are known to bind to the same DNA sequence within different gene promoters. EMSA data confirmed the binding properties of Runx1 and 2 to the UTE-1 binding site. RT-PCR and western blot confirmed the presence of Runx1 and 2 at mRNA and protein level in human and rat HSC. The function of Runx1 and 2 was analysed in activated rat HSC using a reporter gene system where TIMP-1 minimum promoter was cloned upstream the chloramphenicol acetyltransferase (CAT) gene and co-transfected with expression vector for Runx1 or Runx2. The level of the gene activation was measured and it showed that overexpression of Runx1 in vitro inhibited the transcription of the reporter gene whereas Runx2 activated it. However, when overexpressed in HeLa cells, both Runx1 and 2 activated the TIMP-1 promoter.

We have shown for the first time that TIMP-1 gene transcription is controlled by the Runx family of transcription factors. Runx1 and 2 are both present in activated HSC but only Runx2 activates the TIMP-1 promoter and Runx1 has a cell-specific inhibitory effect. We propose that in HSC the promoter activation depends on a competition between the two Runx proteins for the DNA binding site.


S. Saksena, J.B. Leathart, A.D. Burt, O.F.W. James, A.K. Daly, C.P. Day.

Centre for Liver Research, Medical School, University of Newcastle-upon-Tyne, Framlington Place, Newcastle-upon-Tyne NE2 4HH, UK

Non alcoholic fatty liver disease (NAFLD) is an increasingly common condition that can progress to cirrhosis. The risk factors for NAFLD are obesity and insulin resistance/type 2 diabetes mellitus (T2DM), however, < 20% of individuals with these risk factors have advanced disease—NASH and/or fibrosis, the majority have fatty liver only. Circumstantial evidence suggests that hepatic steatosis is the first “hit” in the pathogenesis of advanced NAFLD. This implies that, along with “inflammatory” or “fibrotic” genes, genetic factors influencing the severity of steatosis may play a role in determining individual susceptibility to advanced NAFLD. Microsomal triglyceride transfer protein (MTP) is crucial for the assembly and secretion of hepatic triglyceride as VLDL. The MTP gene promoter contains a functional SNP (-493 G/T) and the G allele has been associated with decreased gene transcription, increased ALT in T2DM, and fibrotic alcoholic liver disease. The aims of this study were first, to determine whether the severity of steatosis correlates with the presence of advanced disease in NAFLD and, second to determine whether the MTP polymorphism influences the severity of steatosis and/or fibrosis in patients with NAFLD.

The study included 65 consecutive patients referred with non-cirrhotic NAFLD. Alternative diagnoses were excluded by serology/history. Liver biopsies were scored for severity of steatosis, fibrosis (metavir) and necroinflammation.

The severity of steatosis correlated with the presence of advanced fibrosis (36% of grade 3 steatosis had >F1 fibrosis v only 12% of grade 1–2, OR: 3.94 (1.14–13.7)). GG homozygotes had an increased risk of steatosis and fibrosis compared to other genotypes. 50% of GG had grade 3 steatosis v 15% of T/*, OR: 5.5 (1.5–20); 41% of GG had > F1 fibrosis v 15% of T/*, OR: 4 (1.1–14.3). These results show that genetic polymorphisms influencing the severity of steatosis in NAFLD are also associated with advanced fibrosis strongly support a role for steatosis—the first-hit, in the pathogenesis of fibrotic NAFLD.


J.A. Byrne1, S.S. Strautnieks1, G. Mieli-Vergani1, C.F. Higgins2, K.J. Linton2, R.J. Thompson1.

1 Institute of Liver Studies, King’s College Hospital, Guy’s, King’s and St Thomas’ School of Medicine, Bessemer Road, London, UK; 2 MRC Clinical Sciences Centre, Imperial College School of Medicine, Hammersmith Hospital Campus, Du Cane Road, London, UK

The human bile salt export pump BSEP is situated in the canalicular membrane of hepatocytes. The gene encoding BSEP, ABCB11, is mutated in a form of progressive familial intrahepatic cholestasis. It is believed that BSEP is the major bile salt export pump of human liver. However, this has not been proven. Drug- and hormone-induced cholestases are a frequent problem in clinical medicine. Inhibition of BSEP is a potential mechanism for the development of such acquired cholestatic liver diseases. At present, inhibitors of human BSEP are unknown. The aim of this study is to express the human bile salt export pump in vitro, determine its substrate specificity, and identify inhibitors of function.

The human ABCB11 coding cDNA was amplified by reverse transcription PCR and subsequent nested PCR from human liver total RNA. A histidine tag was introduced at the 3’ end of the cDNA. The Bac-N-Blue baculovirus expression system was used to generate a recombinant ABCB11 baculovirus. Membranes prepared from High Five™ insect cells were shown to express a 140 kDa protein using an anti-histidine tag antibody, which was absent in mock-infected and uninfected cells. An ATPase assay showed BSEP to have a high basal, vanadate-sensitive, ATPase activity; indicative of the presence of an ABC transporter. Transport assays were performed by measuring the initial rates of ATP-dependent uptake of increasing concentrations of [3H]-taurocholate by inside-out membrane vesicles prepared from High Five™ cells infected with the ABCB11 baculovirus. The Michaelis constant (Km) for taurocholate was defined as 4.25 μM with a maximum velocity (Vmax) of 200 pmol min-1 mg-1 protein. The specificity of BSEP for a range of other bile salts was determined by their ability to complete with [3H]-taurocholate uptake. Furthermore, taurocholate transport by BSEP was inhibited by therapeutic drugs.

This study has provided the first opportunity to analyse the function of human BSEP in isolation outside the canalicular membrane of hepatocytes. It is now clear that BSEP is the main transporter of bile salts in liver and thus inhibition of this transporter is likely to be highly significant in acquired cholestatic liver disease.


S.M. Rushbrook, E. Unitt, L. Morris, I. Scott, N. Coleman, G.J. Alexander.

University of Cambridge School of Clinical Medicine, Level 5, Box 157, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK; MRC Cancer Cell Unit, Hutchison/MRC Research Centre, Hills Road, Cambridge CB2 2XZ, UK

Hypothesis: The impaired T cell responses characteristic of chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infection are mediated by the generation of viral peptide specific T regulatory lymphocytes.

Background: CD4 T helper cell responses are characterised currently by a Th1 or Th2 cytokine profile. Recently in humans T regulatory lymphocytes (T regs) have been characterised that have the immune phenotype CD4+CD25+CD45RO+, are anergic, antigen specific, require contact inhibition for suppression, prevent the proliferation and secretion of interleukin-2 in CD4+CD25- cells, prevent activation of Th0/Th1/Th2/CD8 responses, and require IL-10 and TGF-β for their function.

Methods: The proportion of T reg cells in portal tracts and parenchyma was assessed in paraffin embedded liver tissue from patients with HBV (n = 20), HCV (n = 32), and normal liver (n = 10) stained for CD4, CD25, CD45RO, and TGF-β and analysed with conventional confocal microscopy to assess co localisation of these markers. The number of T regs in human blood from controls (n = 10), HBV (n = 7), and HCV (n = 19) was determined by flow cytometry.

Results: In the portal tracts of patients with HBV and HCV we have identified cells with a phenotype of CD4+CD25+CD45RO+. We have also identified CD4 positive T lymphocytes with TGF-β on their cell membrane. The proportion of T regs in portal tracts ranged from 25–75% in active HBV and HCV infection but never exceeded 10% in control tissue. The proportion of such cells rose from 25–70% in early and late fibrosis, respectively. T reg expression was less than 10% in portal tracts from patients with HCV who were negative for HCV RNA in serum. In human peripheral blood there was no significant difference between the groups in the number of T regs identified by triple staining flow cytometry (CD4+/C25+/CD45R0+) (p = 0.14 and 0.25 in HBV and HCV, respectively).

Conclusions: T regulatory lymphocytes are the predominant T cell in the liver tissue of chronic HBV and HCV cirrhosis. The proportion of such cells is invariant in peripheral blood. T reg cells correlate with viral replication and tissue damage, supporting the hypothesis that these cells are crucial to chronic carriage and liver damage.


P.N. Newsome1, I. Johannessen2, S. Boyle3, K. Samuel4, F. Rae5, M. Turner4, L. Forrester4, P.C. Hayes1, D.J. Harrison5, W.A. Bickmore3, J.N. Plevris1.

1Liver Cell Biology Laboratories, Department of Medicine, Royal Infirmary of Edinburgh, Edinburgh, EH3 9YW; 2Laboratory for Clinical & Molecular Virology, University of Edinburgh, Summerhall, Edinburgh, EH9 1QH; 3MRC Human Genetics Unit, Western General Hospital, Crewe Road, Edinburgh EH4 2XU; 4John Hughes Bennett Laboratories, Western General Hospital, Crewe Road, Edinburgh EH4 2XU; 5Department of Pathology, University Medical School, Teviot Place, Edinburgh EH8 9AG, UK

Introduction: Stem cell plasticity has been revised in the light of studies in rodents, which suggests that these cells can differentiate down a multitude of different lineages. Adult human stem cells have been shown in studies utilising archival biopsy samples to share some of these pluripotent properties but the lack of suitable in vivo model systems has hindered detailed study of human stem cell plasticity. Recent reports have speculated that this presumed stem cell differentiation may simply be the product of cellular fusion between stem cells and pre-existing differentiated cells.

Aims: Our first aim was to see whether human stem cells could differentiate down the hepatocytic lineage when infused into an in vivo (NOD-SCID mice) model. Our second aim was to see whether these differentiated human hepatocytes in the mouse liver were due to cellular fusion of human stem cells and mouse hepatocytes.

Methods: Sub-lethally irradiated (250 Rad) NOD-SCID mice received a tail-vein infusion of 50×106 human cord blood derived mononuclear cells. Livers were removed, fixed in formaldehyde and embedded in paraffin blocks. Sections were then processed to detect the presence of human or murine chromosomes using FISH. Indirect immunofluorescence was also performed using either a 1:75 dilution of a mouse monoclonal antibody against cytokeratin 19, or a 1:25 dilution of mouse monoclonal antibody against human hepatocytes.

Results: Despite no evidence of hepatocyte damage after irradiation, FISH analysis revealed the presence of multiple mature hepatocytes of human origin in the murine liver. The frequency of human hepatocytic transformation seen in our experiments was low, with such cells constituting approximately 0.011% of the cells on each section. There was, however, no evidence of human biliary differentiation in any of the sections. We were unable to identify any cells, either hepatocytic or non-hepatocytic, that displayed fluorescent positivity for both human and mouse DNA, suggesting that cellular fusion of human and mouse cells had not occurred.

Conclusion: We have showed that infusion of human cord blood cells into NOD-SCID mice leads to their engraftment and differentiation down the hepatocytic lineage with no evidence of cellular fusion. We have therefore established an in vivo model with which the mechanisms of human stem cell plasticity can be fully characterised.


S.D. Ryder, G. Vautier, P. James.

Departments of Gastroenterology and Pathology, Queens Medical Centre, Nottingham NG7 2UH, UK

Cholangiocarcinoma (CC) is a common and usually fatal complication of sclerosing cholangitis (PSC). There is no predictive test currently available to identify patients at high risk of cancer. Tumour markers (CA 19.9 and CEA) have been widely used and have sensitivity and specificity of more than 80% for the identification of cancer in PSC. We have previously shown that k-ras and p53 mutations are common in PSC associated cholangiocarcinomas and established that these mutations can be detected in biliary cytology from PSC patients with cancer. The aims of this study were to determine if mutations in these genes predict cancer development in PSC and compare their utility with standard tumour markers. 56 patients with PSC were followed for a median of 5.5 years. Biliary cytology was taken at the time of cholangiographic diagnosis of PSC and mutations in p53 and k-ras were detected using SSCP analysis. Tumour markers were obtained 6 monthly during follow up. Patients developing CC within 6 months of initial presentation were excluded. Over 5.5 years 7 patients (12.5%) developed cholangiocarcinoma (all histologically proven). Cancers were diagnosed at 11, 14, 22, 27, 44, and 51 months after index ERCP.

Elevations in CEA and CA19.9 were seen in 14/56 patients during follow up. 9/56 have no evidence of CC and 5 who developed CC had elevated levels but none had abnormal values more than 6 months prior to the diagnosis being made. We conclude that detection of mutations in p53 or k-ras in biliary cytology specimens is a predictor of CC development and may be useful in selecting patients at high risk for liver transplantation. Tumour markers are useful diagnostically but have little predictive value.

Abstract 16


X. Zhou, J.P. Iredale, R.C. Benyon.

The Liver Research Group, IIR Division, Southampton General Hospital, Tremona Road, Southampton, SO16 6YD, UK

Apoptosis is a fundamental biological process crucial for organ development and for tissue homeostasis, and the resolution of pathological fibrosis. Because the recovery of liver fibrosis is associated with apoptosis of α-SMA positive hepatic stellate cells (HSC), the factor regulating survival of these HSC are of interest as potential therapeutic targets. Our previous studies have shown that activated HSC express integrin ανβ3, which regulates HSC activation and proliferation. In this study we investigated the role of ανβ3 integrin in apoptosis of activated HSC. Incubation of HSC with the ανβ3 antagonist echistatin increased the ratio of Bax/Bcl-2 expression 4.6-fold. Moreover, caspase-3 activity was increased 188% by echistatin and 170% by ανβ3 blocking antibody and cell viability was significantly reduced by acridine orange staining compared with control HSC. Both echistatin and ανβ3 blocking antibody induced DNA fragmentation in HSC detected by electrophoresis. Apototic gene array analysing HSC mRNA provided more evidence showing the network activation of apoptotic signals after inhibition of this integrin. Furthermore, echistatin induced apoptosis was associated with reduced adhesion rate of activated HSC to vitronectin, fibronectin or collagen I. Western blotting analysis demonstrated echistatin reduced FAKpp125 and MAPK (p42/44) expression, suggesting that the dissociation of integrin ανβ3 from its ligand binding regulates HSC apoptosis, which might be mediated via inhibition of intracellular FAK-MAPK signalling pathway. We suggest that during recovery from fibrosis, matrix degradation resulting in the loss of ανβ3 ligands, such as collagen I, fibronectin, and vitronectin may prime HSC for apoptosis.


R. Jalan1, S.W.M. Olde Damink2, D.N. Redhead3, A. Lee 3, P.C. Hayes 3, N.E.P. Deutz2.

1 Institute of Hepatology, University College London Medical School, London, UK; 2 Academic Hospital, Maastricht, The Netherlands; 3 Royal Infirmary, Edinburgh

Hypothesis: TIPSS insertion in cirrhotic patients may produce acute mental deterioration and lead to cerebral oedema and increased intracranial pressure. The cause of this is unknown but is thought to result from increased cerebral blood flow (CBF). This study tests the hypothesis that the increase in CBF observed after TIPSS insertion is the result of increased whole body NO production.

Methods: Nine patients (3F, 6M; mean age: 55.7 (± 3.8) years; mean Pugh score: 10.2 (± 1.0); time after bleed: 34.2 (± 5.8) hours) undergoing emergency TIPSS placement for control of variceal bleeding were studied. CBF was measured using the Kety-Smith method and cardiovascular haemodynamics using a Swan-Ganz catheter. Whole body NO production was measured using infusion of stable isotopes into a peripheral vein: a primed, continuous infusion of L-[guanidine-15N2] arginine (1 mg/kg bw/hour) and L-[ureido-13C;5,5–2H2] citrulline (0.1 mg/kg bw/min). Blood was sampled prior to and 60 min after TIPSS insertion and analysed using an automated LC-MS system. The whole body rate of production of arginine, citrulline, and NO were calculated.

Results: TIPSS insertion reduced portal pressure (p < 0.05) and resulted in a significant increase in CBF: 53.9 (± 3.4) to 68.3 (± 3.5) ml/100 g/min (p < 0.05) and cardiac output and a reduction in systemic vascular resistance (SVR) (p < 0.05 for both). The arterial arginine and citrulline concentrations did not change significantly. Whole body NO production increased significantly: 40.6 (± 11.2) to 71.2 (± 15.7) nmol/kg bw/min) (p < 0.05). The change in NO production correlated with the change in MAP (r = 0.89), SVR (r = 0.8), and CBF (r = 0.69).

Conclusion: This is first study that quantitates the rate of NO production in patients with cirrhosis and shows that TIPSS insertion results in peripheral and cerebral vasodilation, and increased whole body production of NO. The cause of this increase in NO production is unknown but is likely to be responsible for the haemodynamic disturbances.


M. Wright1, R. Goldin1, S. Hellier2, S. Knapp1, A. Frodsham2, B. Hennig2, A. Hill2, R. Apple3, S. Cheng3, H. Thomas1, M. Thursz1.

1 Hepatology Section, Division of Medicine A, Imperial College School of Medicine at St Mary’s Hospital, Praed Street, London W2 1NY, UK; 2 The Wellcome Trust Centre for Human Genetics, University of Oxford. UK; 3 Department of Human Genetics, Roche Molecular Systems Inc, Alameda, CA 94501, USA

Background: The rate of progression to cirrhosis varies among individuals chronically infected with the hepatitis C virus (HCV). Rapid progression may be considered as a specific phenotype and genetically as a complex trait, determined by host genetic, viral, and environmental factors. Clotting pathway activation in models of hepatic fibrosis suggests variation in coagulation may influence the rate of fibrosis. We hypothesised that polymorphisms of the coagulation factors II, V, and VII affect the rate of progression of HCV infected people to cirrhosis.

Methods: We studied the relationship between rate of fibrosis (calculated by dividing the fibrosis stage (Ishak modification of the Knodell scoring system) by duration of infection) and the genotypes of specific coagulation pathway genes in 352 white Caucasian patients infected with HCV. Genotyping was performed using reverse line blot hybridisation.

Results: Rate of fibrosis is faster in factor V Leiden heterozygotes (Arg560Gln) than the homozygous wild type (ANOVA p = 0.004). Disease association studies were performed comparing genotype frequency between individuals with slow fibrosis rates (not predicted to reach cirrhosis for 30 years or more) and those with fast rates (expected to reach cirrhosis in less than 20 years). A significant association was seen (Fishers exact test p = 0.02, odds ratio = 4 for fast progression to cirrhosis if heterozygous for factor V Leiden). No associations were seen between factor II or factor VII genotypes and fibrosis rate.

Conclusions: Possession of the factor V Leiden polymorphism significantly increases the risk of rapid disease progression in HCV, suggesting a role for the coagulation system in the pathogenesis of fibrotic liver disease. This finding suggests that liver inflammation resulting from HCV infection leads to activation of the coagulation system. In those patients with the factor V Leiden mutation the degree of activation is enhanced leading to increased thrombin activity and fibrin production. Thrombin is a stellate cell mitogen and therefore activation of the coagulation cascade may stimulate stellate cell activation and fibrosis.


A.E. Marshall1, S.M. Rushbrook1, L.S. Morris2, N. Coleman2, G. Alexander1.

1 University of Cambridge Department of Medicine; 2 Cancer Cell Unit, Hutchison/MRC Research Centre, Hills Road, Cambridge, CB2 2QQ, UK

Background: HCV-related cirrhosis is a leading indication for orthotopic liver transplantation in the UK. One year survival rate is comparable to other indications but graft infection is virtually universal. Progression of graft fibrosis is accelerated, as the median time to reach cirrhosis is 5–7 years compared with 20–30 years pre transplant. Some transplant patients develop cirrhosis within 2 years yet others remain fibrosis free and there are no reliable markers available to predict fibrosis before it occurs. Previous work has shown a correlation between hepatocyte expression of minichromosome maintenance protein 2 (Mcm2) and fibrosis stage in the non-transplant population. Mcm proteins are essential for DNA replication and are sensitive and specific markers for entry into the cell division cycle.

Methods: Standard immunohistochemical techniques were used to detect hepatocyte Mcm2 expression in early liver biopsies from 21 patients with recurrent HCV following OLT. Patients were divided into groups according to subsequent rate of fibrosis progression. All samples had no histological diagnosis other than recurrent HCV. Normal liver served as a negative control.

Results: Hepatocyte Mcm-2 expression was significantly higher in patients with subsequent rapid progression of fibrosis (RP) than in patients with intermediate (IP) or no progression (NP) as shown below.

Conclusion: In patients with recurrent HCV following OLT, increased hepatocyte expression of Mcm-2 in early liver biopsies is associated with subsequent graft fibrosis. Use of this method to predict fibrosis progression could allow optimisation of immunosuppressive therapy and targeting of antiviral treatment.

Abstract 20


P. Macdonald, D.E.J. Jones, J.A. Kirby.

Centre for Liver Research, University of Newcastle, UK

Primary biliary cirrhosis (PBC) is an autoimmune chronic cholestatic liver disease characterised by immunological destruction of intrahepatic biliary epithelial cells (BEC). The principal autoantigen in PBC, the mitochondrial enzyme pyruvate dehydrogenase complex (PDC), is normally sequestered from the immune system on the inner surface of the inner mitochondrial membrane. Aberrant expression of PDC is, however, seen on the surface of PBC patients’ BEC. The mechanism of this aberrant expression, and its contribution to the breakdown of tolerance to PDC seen in PBC, are at present unclear. Several proteins are released into the cytoplasm at an early stage during mitochondrial apoptosis, and at least one of these (Apo2.7) is subsequently exposed on the extracellular surface of the plasma membrane. In this study we investigated the release of immunoreactive PDC into the cytoplasm following mitochondrial apoptosis as a potential mechanism for aberrant PDC expression in PBC. Flow cytometry demonstrated minimal labelling of the cell surface of normal human cells by patient-derived affinity purified anti-PDC. Permeabilisation of the plasma membrane with digitonin did not enhance anti-PDC staining, confirming mitochondrial sequestration of this antigen. Apoptosis was then induced by treatment with staurosporine, verified after 2–3 hours by demonstration of phosphatidyl serine translocation and activation of caspases 3, 8, and 9, and confirmed to be mitochondrial by demonstration of cytochrome c in the cytoplasm by western blotting within 3 hours. Positive labelling of permeabilised cells with anti-PDC was seen within 6 hours of the induction of apoptosis, suggesting leakage of PDC into the cytoplasm. Significantly, labelling was seen of almost 50% of non-permeabilised cells, indicating additional cell-surface expression of the autoantigen. The surface location of the antigen was subsequently confirmed by immunofluorescence confocal microscopy.

Conclusion: PDC is released from the inner mitochondrial space during apoptosis, resulting in cytoplasmic and ultimately cell surface expression. These observations may help to explain the aberrant expression of PDC seen on BEC in PBC, and provide a novel potential route for immunological exposure to this antigen, which may contribute to tolerance breakdown.


S. Sen, N.A. Davies, R.P. Mookerjee, L. Cheshire, R. Williams, R. Jalan.

Institute of Hepatology, University College London, 69–75 Chenies Mews, London WC1E 6HX, UK

Background: The Molecular Adsorbents Recirculating System (MARS) is an extracorporeal liver support system using albumin dialysis. It has been shown to improve hepatic encephalopathy (HE) and systemic haemodynamics in liver failure. This study tests the hypothesis that these effects are due to altered ammonia and nitric oxide metabolism.

Methods: 14 alcoholics with acute-on-chronic liver failure were randomised into: group1, standard medical therapy (SMT) (n = 6), or group 2, MARS (n = 8). Blood and albumin dialysate (from MARS circuit) samples collected over 7 days were analysed for NOx (nitrate/nitrite) and ammonia (NH3), and compared with corresponding clinical changes.

Results: 7 day analysis: serum bilirubin reduced significantly (460±48 to 240±38 μmol/L, p < 0.001) in the MARS group, but worsened in the SMT group (239±55 to 327±69 μmol/L, p = 0.09). Grade of HE improved significantly in the 7 encephalopathic MARS patients (2.7±0.3 to 1.3±0.5, p = 0.008) but not in controls (1.8±0.3 to 1.5±0.6). 3 of the 5 MARS patients with renal dysfunction (none in SMT group) improved. There was associated reduction of plasma NOx (110.9±26.7 to 61.1±10.0 μmol/L, p = 0.03) and NH3 (77.4±11.4 to 62±8.5 μmol/L, p = 0.06) in the MARS group over 7 days, but not in controls. Over 14 individual MARS sessions: there was a significant improvement of HE grade (p = 0.001) and diastolic blood pressure (p = 0.03) but not of mean arterial pressure. This was associated with a significant reduction of plasma NH3 (6.2±13.4%, p = 0.05) and NOx (21±10%, p = 0.009), which correlated well with their concentrations in the dialysate (p = 0.04 and 0.02, respectively).

Conclusions: MARS causes a significant reduction of plasma NH3, mainly due to removal by the system, which is associated with an improvement of HE. The observed immediate improvement of diastolic blood pressure was associated with a significant reduction of plasma NOx (mainly due to a removal by the system). This may explain the long term haemodynamic improvement observed after several treatment sessions. Serum bilirubin and renal function also significantly improved. None of the positive biochemical or clinical changes were observed in the control group.


F. Oakley, M.M. Meso, D.A. Mann.

Department of Inflammation Infection and Repair, Southampton General Hospital, Tremona Road, Southampton, SO16 6YD, UK

Liver injury causes HSC to undergo a process known as “activation”. During the activation process the HSC phenotype changes from a quiescent retinoid storing cell to a proliferative myofibroblast. HSC activation is associated with an increase in the basal level of the active form of the transcription factor NFκB (Elsharkawy et al. Hepatology, 1999;30:761–9). As NFκB has previously been described to act as a anti-apoptotic survival factor, inhibition of NFκB may induce stellate cell apoptosis.

In its inactive form NFκB is sequestered in the cytoplasm by the inhibitory protein IκBα. Following stimulation by pro-inflammatory cytokines, viral proteins and UV radiation IκBα is rapidly phosphorylated at serine 32 and 36 by inhibitory kappa B kinases (IKK). Phosphorylated IκBα is then targeted for ubiquination and degraded by the proteasome, releasing active NFκB. The anti-inflammatory, immunosuppressive drug sulfasalazine, a known IKK inhibitor (Weber et al. Gastroentrology 2000;119:1209–18) was used to determine the role of NFκB in regulating stellate cell apoptosis.

Electromobility shift assay analysis revealed that treatment on day 7 HSC with 0.5, 1, and 2 mM sulfasalazine for 24 hours inhibited NFκB DNA binding activity, but not that of the transcription factors CBF1 and upstream TIMP1 binding element (UTE1), compared with control cells. Sulfasalazine treatment for 24 hours repressed the activity of two NFκB responsive gene promoters (IL6 and IκBα, but not a control 7xAP1 promoter, which is not NFκB responsive) compared to untreated cells. Treatment of activated HSC with increasing concentrations of sulfasalazine (0–2 mM) for 24 hours induced a dose-dependent increase in apoptosis visualised by both acradine orange staining and caspase 3 activation.

Sulfasalazine specifically inhibits activation of NFκB, down regulates NFκB gene expression, and promotes apoptosis in activated hepatic stellate cells.


E. Barnes1, N. Mathou1, I. Qattan1, D. Brown1, S. Bloor3, B. Clarke3, A. Dhillon2, G. Dusheiko1.

1 Centre for Hepatology, Royal Free and UniversityCollege Medical School, Royal Free Campus, London NW3 2PF, UK; 3 Visible Genetics Inc, Cambridge, UK; 2 Department of Histopathogy, Royal Free and University College Medical School, Royal Free Campus, London NW3 2PF, UK

Introduction: Amino acid mutations in the HBV polymerase (rt) gene that are associated with lamivudine resistance have been well characterised. The overlapping nature of the open reading frames (ORF) for the surface (S) and rt genes of HBV mean that mutations in the S gene may impact on the rt gene and vice versa. Recent reports have suggested that P120T S gene mutations induced by HBIG, in the presence of lamivudine induced rt gene mutations, lead to enhanced viral replication in vitro and a poor clinical outcome.1

Aim: To investigate the incidence and clinical significance of S gene mutations in patients treated with lamivudine.

Methods: 63 patients with chronic HBV infection were treated with lamivudine for a variable period (6 months to 8 years). Sequence analysis of the S and rt genes was performed on all patients, many at multiple time points while on lamivudine. Sequence determination employed the TRUGENETM HBV Genotyping Kit (Research Use Only v1.0) with the GeneLibrarianTM HBV software module for simplified analysis (Visible Genetics Inc., Toronto, Canada). Sequence analysis covered HBsAg (amino acids 100 to 227) and the rt gene (amino acids 110 to 278).

Results: 16 of the 63 patients studied developed characteristic lamivudine resistant mutations; 9/16 were M204V/L180M and 7 were M204I mutations. Nine patients had P120T surface gene mutations (corresponding to T128N rt), all of which were associated with the M204V/L180M or M204I mutations. Three patients had P120S mutations (corresponding to T128I rt), which were not associated with the lamivudine resistant mutations. Interestingly, 5 of these patients did not receive HBIG. The clinical course of the patients with lamivudine resistant/P120T mutations was variable.

Conclusions: P120T surface gene mutations are not uncommon, are often associated with the characteristic lamivudine-resistant mutations, may arise in the absence of HBIG, and are associated with a variable clinical outcome.



P.D. Richardson1,2, L. Augustin2, B.T. Kren2, L. Collins1, C.J. Steer2, J.W. Fabre1.

1 Dept of Clinical Sciences Institute of Liver Studies, King’s College Hospital, London, UK; 2 University of Minnesota Medical School MN, USA

Introduction: Non-viral gene therapy has many distinct advantages over viral based approaches, including ease of production, lower cost, reduced safety concerns, and less risk of adverse immunological events particularly with repeat administration. Despite these advantages long term gene expression remains a formidable hurdle to effective non-viral gene therapy strategies. Recently the development of the “sleeping beauty” transposon system has gone some way to addressing this problem. This plasmid based system codes for an enzyme (transposase “SB”) that binds to specific DNA inverse repeat /direct repeat (IR/DRs) sequences that flank an expression cassette (transposon) on the same plasmid. The transposase binds to the IR/DRs and “cuts” out the transposon from the plasmid and “pastes” it into the host cell’s genome at TA dinucleotide sites. This genomic integration ensures long term gene expression.

Methods: Two classes of plasmids expressing the fluorescent reporter proteins GFP and dsRed were studied. Group1, pTSB-GFP, and pTSB-dsRed plasmids containing either a GFP/dsRed expression transposon and a source of transposase within the same plasmid and Group 2, pT-GFP, and pT-dsRed plasmids containing the same expression transposons as group 1 but no source of transposase (ie non-integrative). The HuH-7 human hepatoma cell line was transfected by a non-viral system consisting of a synthetic peptide containing an integrin binding moiety and a 16 lysine chain for electrostatic binding of DNA. Gene transfer was optimised by the addition of a fusogenic peptide and lipofectamine to enhance endosomal escape. Cell toxicity was assessed using the MTT assay. Reporter gene expression was assessed by fluorescent microscopy and flow cytometry.

Results: 48 hours post transfection all plasmids demonstrated equivalent reporter gene expression. There was a gradual reduction over time in reporter gene expression for all the plasmids. No expression was noted for pT-GFP/pT-dsRed at 2 weeks, in contrast at two months 1.4% of the pTSB plasmids group were still expressing their reporter gene. There was no increase in toxicity in the pTSB plasmid groups compared to pT plasmids.

Conclusions: The combination of our non-viral delivery system and the integrative potential of the “sleeping beauty” transposon system offers potential to overcome a major limitation in non-viral gene therapy and may have a role in future gene augmentation approaches.


R.P. Mookerjee, N.A. Davies, S. Sen, S.J. Hodges, R. Williams, R. Jalan.

Institute of Hepatology, University College London, 69–75 Chenies Mews, London WC1E6HX, UK

Background: Tumour necrosis factor α (TNFα) is a key mediator in alcoholic hepatitis (AH) promoting inflammation and haemodynamic changes, possibly by modulating the nitric oxide (NO) synthases. This is thought to result in an increase in intra-hepatic resistance and an altered liver microcirculation.

Hypothesis: Administration of the chimeric anti-TNFα antibody, infliximab, to patients with AH would result in improvements in portal and systemic haemodynamics by altering NO metabolism.

Methods: 10 consecutive patients with biopsy proven AH (age 53±3 and discriminant function 58.6±14.6) received a single infusion of infliximab (5 mg/Kg). Cardiovascular haemodynamics (Swan-Ganz), hepatic venous pressure gradient (HVPG), and hepatic and renal blood flow (ICG and PAH, respectively, Fick principle) were measured before, 24 hours, and 28 days after infliximab. Clinical and biochemical profiles and NO production (modified Greiss Test (NOx)) were monitored.

Results: Significant improvements in bilirubin (p = 0.02), CRP (p = 0.003), and white cell count (p = 0.01) and all patients were alive at 28 days. HVPG decreased significantly in all patients by 24 hours (28.38±2.83 to 14.3±1.89 mm Hg, p = 0.001) with a sustained reduction at 28 days (12.8±1.99 mm Hg, p = 0.003). Mean arterial pressure increased significantly (71.6±1.72 to 81.1±3.15 mm Hg, p < 0.05), while systemic vascular resistance increased and cardiac index decreased non-significantly. Hepatic and renal blood flow increased significantly (506.2±42.9 to 646.3±49.2 ml/min, p = 0.0005 and 424.3±65.12 to 506.3±85.7 ml/min, p = 0.003, respectively) by 28 days. NO production demonstrated no correlation with the haemodynamic changes observed.

Conclusion: Anti-TNFα treatment produces a highly significant, early, and sustained reduction in HVPG, without a significant change in cardiac index. The observed reduction in HVPG is likely to be secondary to reduced intra-hepatic resistance. This suggests that TNF induced local vascular mediators may play a major role in the portal hypertension in these patients. Some of the observed clinical effects may also relate to the improvements in liver and renal circulations following anti-TNF treatment.


F. Murphy1, J. Waung1, J. Collins2, M.J.P. Arthur2, H. Nagase3, D. Mann3, R.C. Benyon3, J.P. Iredale3.

1 Liver Group, Southampton University, Southampton, UK; 2 Mucosal Immunology, Southampton University, Southampton, UK; 3 The Kennedy Institute, Imperial College, London, UK

Introduction: The activated hepatic stellate cell (HSC) plays a key role in liver fibrosis. During spontaneous recovery from experimental liver fibrosis loss of activated HSCs is through apoptosis. This has highlighted control of HSC apoptosis for further investigation. Inhibition of HSC apoptosis by TIMP-1 is mediated via MMP inhibition. N-cadherin, which mediates cell to cell contact, is expressed in HSC. Disruption of this interaction has been shown to promote apoptosis in melanoma cells. We studied the fate of N-cadherin in HSC during apoptosis.

Methods: Extracted rat HSC were cultured on plastic in the presence of serum until activated. Rat HSC were induced into apoptosis by gliotoxin or cycloheximide in the presence of TIMP-1 or a mutant TIMP-1 (T2G N-TIMP-1 with no MMP inhibitory activity). Protein extracts were studied by western blotting for N-cadherin. Apoptosis was quantified by examination of nuclear morphology after staining with acridine orange.

Results: TIMP-1 but not the T2G mutant TIMP-1 demonstrated a dose dependent inhibition of HSC apoptosis. Western blotting clearly demonstrated an intact 135 kDa form of N-cadherin but also 60 and 38 kD truncated forms of N-cadherin in protein extracts from rat activated HSC cultured in serum. Cleavage of N-cadherin was demonstrated with the appearance of a 60 kDa fragment after 3–4 hours exposure to cycloheximide or gliotoxin. This fragmentation was reduced by co-incubation with 5 nM recombinant TIMP-1 but not by the non-functional T2G mutant TIMP-1.

Conclusions: Activated rat HSC express intact 135 kDa N-cadherin that is degraded into smaller fragments during apoptosis induced by gliotoxin or cycloheximide. N-cadherin fragmentation is reduced by exposure to TIMP-1. Protein database analysis indicate N-cadherin contains a number of potential MMP cleavage sites. MMP cleavage of N-cadherin may be an important mechanism promoting HSC apoptosis in recovery from liver fibrosis.


G.J.M. Webster1,2, S. Reignat1, M. Lascar1, D. Brown2, G.S. Ogg3, J. Gotto2, G. Dusheiko2, R. Williams1, A. Bertoletti1.

1 Institute of Hepatology; 2 Centre for Hepatology, Royal Free and University College Medical School, London, UK; 3 Institute of Molecular Medicine , Oxford, UK

Background/Methods: During anti-viral immunity, CD8 cells specific for epitopes in different viral proteins are present, with a hierarchy of dominant and sub-dominant specificities. The hierarchy of HBV-specific CD8 response was analysed in 29 patients infected with HBV (11 resolved/18 chronic) using HLA-tetramer analysis and intracellular cytokine staining (directly and after in vitro expansion). 11 different HLA-A2 restricted epitopes were monitored. Twelve of the 29 patients were followed longitudinally for one year along with HBV-DNA and ALT values.

Results: Differences in CD8 epitope specificity were found in relation to HBV-DNA level. Core 18–27 was the dominant response in 9/12 (75%) patients with resolved infection, but in only 1/12 (8%) chronic patients with detectable HBV-specific CD8 response. However, envelope specificities were dominant in 8/12 (66%) chronic patients compared with only 1/12 (8%) patients with resolved infection. Epitopes that were subdominant in resolved infection (env 183–91, pol 816–24) were not only more likely to persist in chronic patients, but their responsiveness was not affected by HBV-DNA values. In contrast, the expansion of core 18–27 specific CD8 cells in vitro was inversely proportional to the quantity of HBV-DNA, being completely absent in patients with HBV-DNA > 107 copies/ml. Remarkably, the pattern of CD8 epitope hierarchy was not influenced by preferential intrahepatic localisation of CD8 cells or by viral mutations causing CD8 epitope inactivation.

Conclusions: The hierarchy of HBV-specific CD8 response in chronic hepatitis B differs from resolved infection in that CD8 cells specific for sub-dominant epitopes are more likely to persist. These data suggest that therapeutic vaccines designed to restore specific T cell response in chronic patients should be tailored towards sub-dominant specificities.


J. Ahmed-Choudhury, C. Wildman, C.L. Russell, D.H. Adams, S.C. Afford.

Liver Research Laboratories, MRC Centre for Immune Regulation, University of Birmingham, Institute of Clinical Science, Queen Elizabeth Hospital, Edgbaston, Birmingham, B15 2TH, UK

Background: CD40, a membrane receptor of the TNF-receptor superfamily, has been shown to be critically involved in Fas-mediated cholangiocyte apoptosis during chronic inflammatory biliary diseases such as primary biliary cirrhosis. However, the underlying molecular and signalling mechanisms remain unclear. Our recent work implicates the transcription factor AP-1 in CD40-induced cholangiocyte apoptosis, but also suggests involvement of other signalling pathways. Because STAT3 has been implicated in apoptosis in other systems we have studied the AP-1 and STAT3 signalling pathways in primary human cholangiocytes before and after CD40 ligation.

Methodology: Cholangiocytes were isolated from human liver tissue and cultured in the presence and absence of crosslinking mAb to CD40 (αCD40). Stimulation with TNF-α was used as a positive control. Functional activation of STAT3 and AP-1 transcription factors was assessed by electrophoretic mobility shift assays of cellular nuclear extracts. Western blot analysis was used to determine protein levels of the transcription factors. Apoptosis was assessed by in situ DNA end labelling and morphology. Pretreatment with selective inhibitors to JNK, ERK, and JAK2 (100 μM DMAP, 50 μM PD98059, and 50μM AG490, respectively), were used to investigate whether inhibition of AP-1 and STAT3 transcription factors affected apoptosis.

Results: Stimulation of cholangiocytes with αCD40 for 2 hours resulted in an increase in the levels of functional AP-1 (fourfold) and STAT3 threefold). This increase of both transcription factors was maintained over the entire culture period of 24 hours. A similar profile was found with protein levels of c-jun, c-fos (components of AP-1) and STAT3. Induction of apoptosis in these αCD40-stimulated cells was observed at 24 hours (% cells apoptotic 43 ± 6; n = 9), however, pretreatment of the cholangiocytes with individual selective inhibitors partially inhibited apoptosis (% cells apoptotic with: DMAP pretreatment 22 ± 4; PD98059 pretreatment 19 ± 5; AG490 pretreatment 21 ± 5; n = 7). Combining the selective inhibitors reduced apoptosis to basal levels (% cells apoptotic 8 ± 3; n = 7).

Conclusion: The activation of AP-1 and STAT3 signalling pathways are required to induce CD40-mediated apoptosis in primary human cholangiocytes and blockade of both pathways completely inhibits CD40 induced apoptosis.

Study Funded by BBSRC (6/C11113).


W. Abbott, H. Cooksley, R. Williams, N. V. Naoumov.

Institute of Hepatology, University College London, UK

Dendritic cells are antigen presenting cells specialised to regulate T-cell immunity and to produce interleukin-12 (IL-12). Biologically active IL-12 is a heterodimer (p70) composed of two subunits (p35 and p40) encoded by different genes. We investigated whether genetic polymorphisms and/or functional dysregulation in IL-12 gene expression might be associated with different outcomes of exposure to hepatitis B virus (HBV).

Methods: We recruited and studied simultaneously 14 trios consisting of a patient with chronic hepatitis B (serum HBV DNA+); a subject spontaneously recovered from HBV exposure (anti-HBc/anti-HBs+) and a healthy subject never exposed to HBV. The regions containing regulatory elements in the IL-12p40 promoter (785 base pairs) and IL-12p35 promoter (1050 base pairs) were sequenced in both directions. Monocyte-derived dendritic cells (DC) were generated from all 42 subjects and analysed for expression of IL-12p40 and p35 mRNA (by TaqMan™ PCR), of IL-12p40 and p70 protein (by ELISA) and cell surface molecules CD40, CD80, CD83, and CD86 (by FACS)—at baseline and after stimulation with CD40 ligand (CD40L) or IFN-γ plus LPS.

Results: The IL-12p40 promoter was identical to the published sequence in all subjects and no polymorphic sites were identified. Three novel nucleotide polymorphisms were identified in the IL-12p35 promoter of 2 patients with chronic hepatitis B. However, the IL-12p35 mRNA levels in DC from these patients were no different from the other subjects. Comparisons within the 14 trios showed no differences in baseline or stimulated IL-12 mRNA or protein levels and no abnormalities in CD40L-stimulated cell surface molecules. The IFNγ/LPS stimulated expression of CD80, CD83, and CD86 was markedly reduced in DC from patients with chronic hepatitis B compared to controls.

Conclusions: Polymorphisms or functional dysregulation of IL-12 genes do not contribute to differences in the outcome of HBV infection. The abnormal expression of co-stimulatory molecules on dendritic cells may account for impaired T-cell activation and identifies candidate genes for susceptibility to chronic HBV infection.


R.G. Ruddell, I. Yeung, M.J. Arthur, D.A. Mann.

Department of Inflammation Infection and Repair, Southampton General Hospital, Tremona Road, Southampton, SO16 6YD, UK

The 5-hydroxytyptamine2 (5-HT2) family of G-protein coupled receptors are important mediators of many physiological functions including smooth muscle contraction, platelet aggregation, and neurotransmission. 5-HT2 receptors have also been linked to fibrotic and proliferative pathways via the activation of extracellular regulated kinase (ERK).1 Here we report the expression patterns and function of 5-HT2A, 5-HT2B, and 5-HT2C receptors on the surface of the rat hepatic stellate cell. Whole cell protein and cDNAs were prepared from freshly isolated and activated rat HSCs ready for immunoblotting and PCR. cDNAs were primed with oligonucleotides specific to rat 5-HT2A, 5-HT2B, and 5-HT2C receptors and amplified for up to 40 cycles prior to resolution. Activated rat HSCs were also treated with a range of 5-HT2 specific antagonists before caspase 3 activity was determined. 5-HT2A receptors were found to be expressed on both freshly isolated and activated rat HSCs. 5-HT2B receptors were only expressed on activated rat HSCs whereas 5-HT2C receptors were absent from both cell preparations. Pretreatment of activated rat HSCs with certain 5-HT2 antagonists resulted in a highly significant increase in caspase 3 activity, which was significantly reduced when cells were also incubated in the presence of 5-HT. Since HSCs are a key mediator in the progression of liver fibrosis and cirrhosis, the 5-HT2A and 5-HT2B receptors expressed on rat HSCs represent a possible target for therapeutic agents. Both receptors bind specific ligands that can either promote or inhibit the signalling cascades linked to the receptors. 5-HT2 receptors also possess the ability to synergize with other signalling pathways amplifying their response to exogenous ligands.2 This study also shows that 5-HT2 receptor antagonists mediate activated stellate cell apoptosis.




MI Prince1, AC Chetwynd2, WL Craig1, JV Metcalfe1, OFW James1.

1 Centre for Liver Research, Newcastle University, NE2 4LL, UK; 2 Department of Mathematics, University of Lancaster LA1 4YF, UK

Background: Most studies of the prognosis of PBC have been reported from tertiary centres and may reflect an atypical spectrum of disease (referral bias). We have previously reported survival data from a large population based cohort of PBC patients, unadjusted for the effect of the presence of prevalent cases (Hepatology 2000;32:171A). We now report updated survival using improved multivariate analysis that accounts for this potential bias.

Methods: A geographically and temporally defined cohort of 770 patients with PBC alive from 1/1/1987 to 31/12/1994 identified through multiple methods were followed up by interview, case note, and death certificate review until death, transplant, or censor on 1/1/2000. Survival was analysed by Cox hazards regression adjusted for prevalent cases (Kieding, NATO advanced statistics workshop 1991).

Results: 5613 patient years of follow up (417 deaths, 39 transplants) were analysed. Median survival from date of diagnosis, adjusted for prevalent cases, was 9.3 years. Survival was worse than in local age-sex matched populations (standardised mortality ratio (SMR) = 2.9 (95% CI 2.6 to 3.2)). SMR was 1.7 (1.5 to 2.0) excluding liver deaths. Survival did not differ significantly between patients presenting with symptoms (n = 464) and initially asymptomatic (n = 301) patients (8.0 (6.6 to 9.0) and 9.6 years (8.6 to 10.5)). Symptom development in initially asymptomatic patients was not associated with worsening prognosis. In multivariate analysis a model including age at diagnosis, alkaline phosphatase, albumin, and bilirubin best predicted survival. This model explained 37% of observed variation in survival (R2M = 0.37). In comparison, the Mayo model also accurately predicted survival adjusted for the effect of prevalent cases (R2M = 0.18). Prothrombin time at diagnosis and initial histology did not independently affect survival. Use of UDCA was not associated with a statistically significant change in survival.

Conclusion: PBC is associated with increased mortality from both liver and non-liver related causes. Symptom status did not affect outcome in this cohort. The Mayo score remains acceptably accurate for use even in “non-tertiary” cohorts.


J. Mann, F. Oakley, D.A. Mann.

Department of Inflammation Infection and Repair, Southampton General Hospital, Tremona Road, Southampton, SO16 6YD, UK

The transcription factor NFκB is a key regulator of proliferation, differentiation, and cell fate in mammalian cells. In its inactive form NFκB is found in the cytoplasm bound to the inhibitory protein IκBα. NFκB is expressed at a low basal level in all cells and the inducible signalling pathway has been well characterised. However, during differentiation and phenotypic changes associated with chronic inflammation, fibrosis, and cancer, NFκB expression is re-programmed to a constitutively high level. We have identified a novel pathway that leads to the re-programming of NFκB in mammalian cells, regulated by a persistent down-regulation of IκBα protein. During the activation of hepatic stellate cells (HSC) we observed an increase in NFκB activity and a diminution of IκBα protein levels, which is unexpected as expression of IκBα is NFκB dependent. Previous reports show that some NFκB dependent genes contain an overlapping CBF1 site within the NFκB binding site. CBF1 is a transcriptional repressor and its up-regulation during HSC activation mirrors the decrease in IκBα protein. Sequence analysis of IκBα promoter revealed the presence of a CBF1 consensus sequence in the κB2 site, which is one of five NFκB binding sites in the promoter. Over-expression of CBF1 in HSC repressed IκBα promoter activity and decreased IκBκ protein levels in COS cells. CBF1 is reported to repress gene expression by recruiting histone deacetylases (HDACs). HSC transfected with the IκBα promoter and treated with trichostatin A (HDAC inhibitor) promoted an increase in IκBα promoter activity.

Binding of Notch1 cell surface receptor, to its ligand Jagged 1, promotes cleavage of Notch 1 intracellular region (IC), which then translocates to the nucleus converting CBF1 from a repressor to an activator of gene transcription. We confirmed expression of Notch1 in HSC by FACS analysis and RT-PCR. Over-expression of Notch1 IC in HSC resulted in an increase in IκBα promoter activity and an up-regulation of IκBα protein in Notch1 IC transfected COS cells. Co-culture of rHSC with L-cells expressing Jagged 1, but not parental L-cells increased the activity of a wild type, but not mutant CBF1 luciferase reporter. Physiological Notch 1 signalling occurs in rHSC and can potentially alter CBF1 regulated gene expression. The transcriptional activity of Notch1 IC can be further potentiated by recruiting p300 histone acetylase. We have shown in HSC that over-expression of p300 in the presence of Notch1 IC increases IκBα promoter activity, however over-expression of a competitor for p300, p53, abolishes this effect.

We have therefore identified a novel and potentially physiologically important pathway for altering basal levels of NFκB in a wide variety of cell types including HSC.


D. E. Smart, D.A. Mann.

Liver Group, Division of Infection, Inflammation and Repair, University of Southampton, Southampton General Hospital SO16 6YD, UK

Liver fibrosis is defined not only by a net increase in the overall content of fibrillary collagen, but also by a change in its distribution. This change in matrix metabolism is believed to result from a decrease in matrix degradation rather than an increase in deposition. Previous work has demonstrated that the increased expression of the tissue inhibitor of metalloproteinase-1 (TIMP-1) plays a fundamental role in the reduction in matrix degradation, as does the hepatic stellate cell (HSC) as its main source. The increase in TIMP-1 expression by the HSC is closely linked to their phenotypic change, occurring in response to liver injury. It is this link that allows TIMP-1 transcription to be used as a molecular tool for investigating the subcellular mechanisms associated with liver fibrosis. In previous work we have demonstrated that in HSC the classical c-Jun:c-Fos regulation of TIMP-1 transcription is replaced by interaction of Jun D homodimers with an activator protein-1 (AP-1) binding site in the minimal promoter. In the present study we have investigated the role of phosphorylation of JunD in the regulation of TIMP-1 gene transcription. A mutant JunD lacking an amino acid residue (Ser 100) previously identified as a substrate for Jun-N-terminal kinase (JNK) was unable to stimulate TIMP-1 promoter activity. However, transfection of dominant negative JNK or over-expression of JIP-1, a protein that sequesters JNK away from its target substrates, had no effect on TIMP-1 promoter function. In addition, we found that unstimulated activated HSC lack detectable background JNK activity yet express high levels of TIMP-1. Further investigations have indicated that TIMP-1 promoter activity is regulated by the phosphorylation of Jun D via ERK1/2 kinases, in contrast to the JNK regulation observed in NIH3T3 fibroblasts. This lack of JNK activity and subsequent lack of c-Jun phosphorylation may account for non-classical regulation of TIMP-1 transcription in the HSC.


C.J. Marek, M.C. Wright.

Department of Molecular and Cell Biology, University of Aberdeen, Aberdeen, UK

Although the liver is able to regenerate and grow in vivo, hepatocytes rapidly de-differentiate and fail to proliferate to any significant extent in vitro. This common response of hepatocytes to isolation and culture has limited their utility in for example bio-artificial liver devices. An unusual rat pancreatic cell line (AR42J-B13) has therefore been examined for its ability to trans-differentiate into hepatocytes. Pancreatic AR42J-B13 cells proliferated, were amenable to sub-culture and failed to express liver-specific genes under normal culture conditions. However, after treating cells with glucocorticoid, proliferation was inhibited. Over a period of 9 days of continuous glucocorticoid treatment, the cells changed their morphology from a relatively small round phenotype in which cells grew in clusters to an hepatocyte-like morphology (hepatic AR42J-B13 cells). This change in morphology correlated with an induction of cytochrome P450 2C11 (the major liver-specific cytochrome P450 isoform in rat) such that the levels of expression at 9 days were similar to those expressed in freshly isolated rat hepatocytes. Other cytochromes P450 were also expressed in hepatic AR42J-B13 cells, including CYP2E1 and CYP3A1/3A23. To determine if the expression of cytochrome P450 gene induction in AR42J-B13 cells was functional, cells were treated with the hepatotoxin paracetamol. This drug requires metabolism by cytochromes P450 for its toxicity. Paracetamol had little effect on pancreatic AR42J-B13 cells but resulted in significant loss of viability in hepatic AR42J-B13 cells, suggesting that the cytochromes P450 are functional within these cells.

These data indicate that this cell line has an unusual capacity to alter its gene expression profile in response to glucocorticoid. The mechanism(s) underlying this response are unknown. However, this work suggests that it may be possible to generate an unlimited supply of hepatocytes in vitro that can be stimulated to differentiate into hepatocytes when required. A similar human cell line could be essential to the realistic development of a bio-artificial liver for liver bridging support.


W.J.H. Griffiths, T.M. Cox.

Department of Medicine, University of Cambridge, Addenbrooke’s Hospital, Cambridge, UK

Mutations in the HFE gene predispose to haemochromatosis; HFE mediates transferrin-iron entry into cells in vitro, and specific expression in small intestinal crypts suggests HFE may be involved in determining iron absorption by mature enterocytes. Transferrin receptor 2 (TfR2) is a recently identified homologue of the ubiquitous transferrin receptor (TfR1) that mediates low affinity transferrin iron uptake. TfR2 gene mutations are associated with cases of non-HFE related haemochromatosis suggesting an additional key role for this receptor sub-class in the control of intestinal iron absorption. Expression of TfR2 and the precise interactions of HFE in the small intestine are unknown.

To investigate the functions of HFE and TfR2, the in situ localisation of these proteins was examined by confocal microscopy in intestinal tissue and cells from humans and mice. A panel of rabbit and avian polyclonal antisera was generated to specific peptide sequences of HFE and TfR2; antibodies were characterised by Western immunoblotting. In mouse and human duodenal sections, expression of HFE and TfR2 was restricted to crypt cells where they co-localised. Furthermore, the co-localisation signal was absent in humans and mice with HFE deficiency. TfR1 expression occurred throughout the villus epithelium and only partly co-localised with HFE in the crypts. In human Caco-2 cells, which have a small-intestinal phenotype, abundant vesicular staining of TfR2 was observed. TfR2 and TfR1 did not co-localise but HFE colocalised with TfR2 in an endosomal compartment following addition of iron-saturated transferrin to the culture medium.

HFE and TfR2 appear to interact in an endosomal transport pathway for transferrin iron uptake within duodenal crypt cells. These results support a novel molecular mechanism by which iron absorption is regulated according to body iron status and which is disrupted in hereditary haemochromatosis.


A. Semper, N. Libri, E. Sanders, W. Rosenberg.

Liver Group, University of Southampton, UK

Clearance of acute hepatitis C virus (HCV) infection is associated with vigorous Th1/Tc1 responses and in chronic hepatitis C (CHC) T cells are either unresponsive to HCV or exhibit a Th2/Tc2 phenotype. The impairment of HCV specific T lymphocyte responses in CHC remains unexplained. As potent antigen presenting cells, dendritic cells (DC) play a pivitol role in the development of antigen specific T cell responses. Cytokines produced by DC direct the polarisation of type 1 or type 2 T lymphocytes. IL-12 is the predominant DC derived cytokine that directs Th1 differentiation. Factors, including IL-10, that reduce IL-12 production by DCs, in turn cause DCs to promote Th2 differentiation. We have shown that peripheral blood DCs from CHC donors have reduced allostimulatory capacity that is associated with reduced cell surface expression of co-stimulatory molecules. This altered DC function may contribute to T cell unresponsiveness in CHC. HCV infection of DCs has been reported but it is not known whether this accounts for the altered DC phenotype in CHC. HCV infects other cell types through the cell surface tetraspanin CD81 and we have recently shown that peripheral blood DCs and monocyte derived DCs (mDC) express CD81. To investigate the consequences of CD81 engagement on DCs we have exposed mDCs to plate-bound anti-CD81 monoclonal antibody for 48 hours, prior to CD40 ligation using CHO cells transfected with human CD40L. After 24 hours of CD40 ligation, IL-10 and IL-12p70 production were measured by enzyme linked immunosorbent assays and DC phenotype was examined.

mDCs pre-treated with CD81 over a range of doses from 2.5–20 μg/ml showed enhanced IL-10 production in response to CD40L. IL-12p70 production was slightly reduced.

This alteration in the balance of cytokine production was accompanied by an alteration in DC phenotype. The observed changes in cytokine production by mDC induced by CD81 cross-linking are consistent with a phenotype that would drive T cells to promote type 2 responses, or to be unresponsive. Engagement of CD81 on DC by HCV during natural infection may initiate the cascade that culminates in T cell unresponsiveness observed in CHC.


Y. Ma1, D.P. Bogdanos1, M.J. Hussain1, U. Sharma1, G. Mieli-Vergani2, D. Vergani1.

1 Institute of Hepatology, University College London Medical School, Chenies Mews, London, UK; 2 Institute of Liver Studies, King’s College Hospital, Denmark Hill, London, UK

Background: Using a proliferative assay, six antigenic regions have been previously defined (J Hepatol 2001;34:Suppl I:212 on cytochrome P4502D6 (CYP2D6) the target of the humoral autoimmune response in autoimmune hepatitis type 2 (AIH-2).

Aims: To identify the cytokine profile of the CYP2D6 specific T cell response in patients with AIH-2.

Methods: T cell reactivity was investigated in 13 pediatric patients with AIH-2 using 61 overlapping icosameric peptides, spanning the full length of CYP2D6 and constructed using fmoc solid phase chemistry (Mimotopes UK Ltd, UK). Peripheral blood mononuclear cells (100 000/well) were cultured in the presence of 10 μMol of each individual peptide for 8 days in a 96-well plate and 36 T cell lines were established from 3 patients by weekly re-stimulation with the peptides. Levels of IFN-γ, IL-4, and IL-10 were measured by ELISA in the culture supernatant of day 7 and 28.

Results: At day 7, nine of 13 patients responded to the CYP2D6305–364 and CYP361–436 peptides producing elevated levels of IFN-γ (median 52 pg/ml, range 0–2136), IL-4 (median 28 pg/ml, range 0–440) and IL-10 (median 20 pg/ml, range 0–466), while four patients produced more IL-4 (median 54 pg/ml, range 0–1050) and IL-10 (median 37 pg/ml, range 0–2318) than IFN-γ (median 0 pg/ml, range 0–9), indicating a mixed Th1/Th2 profile. Cell lines cultured for 28 days with two cycles of re-stimulation responded to CYP2D6305–364 and CYP361–436 with a high IFN-γ (median 3276 pg/ml, range 661–7142) and a low IL-4 (median 0 pg/ml, range 0–154) and IL-10 (median 0 pg/ml, range 0–412) release.

Summary: The early (7 days) CYP2D6 specific response has a Th0 profile, but assumes a Th1 profile after specific re-stimulation over 28 days. The CYP2D6 specific T cell response, assessed through cytokine release, is more restricted than that observed in proliferation assays, being focused on two of the six antigenic regions. These two regions may be immunodominant and could be the focus of therapeutic immune manipulations.


S.M. Rushbrook, V. Sharma, M. Allison, C. Watson, G.J. Alexander.

University of Cambridge School of Clinical Medicine, Level 5, Box 157, Addenbrooke’s Hospital, Hills Road, Cambridge, CB2 2QQ, UK

Background: Hepatic artery complications post transplantation are a cause of increased morbidity and mortality. Risk factors for adult hepatic artery thrombosis identified previously include multiple transplantations, recipient CMV negative/donor CMV status, biopsy proven acute rejection within 1 week of transplantation, donor/recipient weight > 1.25, arterial anastomosis to an old conduit and transplantation for primary sclerosing cholangitis.

Method: 450 patients transplanted at one centre were analysed for the development of hepatic arterial complications, with the purpose of identifying risk factors.

Results: 16 patients of 450 analysed developed hepatic arterial complications (3.5%). 13/16 developed hepatic artery thrombosis, 1/16 had a hepatic artery stricture and 2 had evidence of non-thrombotic graft infarction: 13/16 (81.5%) occurred in males; 75% of hepatic arterial complications occurred within 1 month of transplantation. 32/450 patients were transplanted for primary sclerosing cholangitis. 7/32 of this group (21.8%) developed hepatic arterial complications. PSC counted for 7/16 (43.75%) of all patients identified in this analysis. This is in contrast to the 9/418 non PSC who developed arterial complications (p = < 0.05). CMV mismatch was not found to be a risk factor for arterial complications. 50% of patients had acute rejection. 18.75% of patients had chronic rejection and 12.5% of patients had greater than two transplants at the time of development of arterial complications.

Conclusions: Transplantation for PSC is significant risk factor identified for arterial complications. CMV mismatch was not a major risk factor as previously identified. It is of interest that PSC and chronic rejection may both have an arterial component to their pathogenesis, and this may be why they have an increased risk of arterial complications. This analysis would also suggest that new strategies (eg anticoagulation/ anti platelet therapy) might need to be considered in transplanting patients for PSC and chronic rejection. It may also suggest that PSC has a vasculitic component to its pathogenesis.


S. Saksena, O. James, W. Craig, C.P. Day.

Centre for Liver Research, University of Newcastle upon Tyne, UK

Background: Non-alcoholic fatty liver disease (NAFLD), either fatty liver (FL) or non-alcoholic steatohepatitis (NASH) is the likeliest diagnosis in patients with persistently abnormal liver blood tests unrelated to viral hepatitis or alcohol excess. Although patients with simple FL seldom progress to advanced liver disease, up to 25% of patients with NASH progress to cirrhosis within 10 years. In this study we sought to determine whether any clinical or laboratory features could distinguish between FL and NASH in an unselected group of patients with NAFLD, thus potentially avoiding the need for liver biopsy.

Methods: All patients with a diagnosis of NAFLD were prospectively included over 2 years. NAFLD was diagnosed on: (a) compatible histology, (b) weekly alcohol consumption less than “sensible limits”, and (c) exclusion of other causes of liver disease by appropriate blood tests (HCV/HBV serology, liver autoantibody screen, iron indices, and HFE genotyping). NASH was diagnosed when hepatocyte ballooning with or without an inflammatory cell infiltrate or fibrosis was present in addition to steatosis (cirrhosis excluded). All patients underwent full clinical and laboratory evaluation including measurement of BMI, waist-hip ratio (WHR), and insulin resistance by the homeostasis model assessment (HOMA) that utilises fasting glucose and insulin concentrations.

Results: Complete data were available in 112 patients: 63 FL and 49 NASH. The only features that differed between the 2 groups was presence of T2DM, which was more common in NASH (36.7% v 16%, OR 2.8 (1.2 to 6.7), p = 0.028) and a higher percentage of NASH patients had ALT greater than twice the upper limit of normal (41% v 16%, OR 3.7 (1.5 to 8.8) p < 0.005). A combination of T2DM and ALT greater than twice normal was present in 18% patients with NASH and none with FL (OR 29.8 (1.7 to 526) p < 0.001).There was no difference between the two groups in age at biopsy, sex distribution, mode of presentation, WHR, BMI, lipid profile, or indices of insulin resistance.

Conclusion: Presence of T2DM or ALT greater than twice normal increases the risk of NASH by four fold. The combined presence of both increases the risk of NASH by 30 fold. These two simple parameters may help in deciding the need for liver biopsy in unselected patients with NAFLD.


D. Kapoor, R. Williams, R. Jalan.

Institute of Hepatology, University College London Medical School, UK.

Aims: Ammonia is central in the pathogenesis hepatic encephalopathy (HE). We have recently shown that the kidneys are an important target of ammonia removal and that angiotensin II may be an important mediator. This study tests the hypothesis that in patients with HE precipitated by dehydration, volume repletion would improve HE and reduce plasma ammonia by increasing urinary ammonia excretion (UNH3V) by suppressing angiotensin II.

Methods: 8 consecutive patients with cirrhosis (age 52 (4.3), all Child C) were studied after a diagnosis of diuretic precipitated HE was made. The intravascular volume was restored using 4.5% albumin intravenously. Systemic hemodynamics, plasma renin activity (PRA) and angiotensin II (AII), renal plasma flow (RPF, using para-amino hippuric acid), glomerular filtration rate (GFR, using Inulin), HE grade, NCT-A, DSST and plasma, and UNH3V were evaluated at baseline and then at 24 and at 72 hours afterwards.

Results: 2.5 (0.2) L of albumin over 72 hours was required to raise the CVP from 1.1 (0.7) to 9.7 (0.8) mm Hg. At baseline, 6 patients had Gd II, HE and 2, Gd III, HE. At 72 hours, all of them reverted to Gd I, HE. Both NCT-A and DSST improved significantly (p < 0.01 each). After volume repletion, the RPF (416 (21) to 622 (42) ml/min) and GFR (32 (4) to 79 (4) ml/min) increased (p < 0.01). The plasma ammonia levels fell from 98 (7) to 52 (4) micmol/L (p < 0.001). The UNH3V and the fractional excretion of ammonia increased (p < 0.05). There was a significant reduction in the plasma levels of PRA (30.1 (3.5) to 10.1 (0.9) ng/ml/hr, p < 0.05) and AII (325 (21) to 143 (20) pg/ml, p < 0.05). The reduction in plasma ammonia correlated with the reduction in AII (r = 0.65), increase in UNH3V (r = 0.89), and urinary sodium excretion (r = 0.65). The reduction in AII correlated with increased UNH3V (r = 0.67) and urinary sodium excretion (r = 0.66).

Conclusions: In patients with cirrhosis and diuretic induced HE, volume repletion reduces plasma ammonia and mental status by increasing renal ammonia excretion, which is likely to be mediated by a reduction in angiotensin II.


D.M. Flynn1, H.A. Crosby2, J. de Ville de Goyet1, S.G. Hubscher2, A.J. Strain2, D.A. Kellyv.

1 Liver Unit, Birmingham Childrens Hospital; 2 Liver Research Laboratories, Queen Elizabeth Hospital, Birmingham, UK

Introduction: Mutations in the Jagged1 gene are causally associated with Alagille syndrome. Recently, mice doubly heterozygous for mutations in Jagged1 and the receptor Notch2 have recently been described as exhibiting a range of developmental abnormalities characteristic of Alagille syndrome. The expression of Notch receptors in paediatric liver disease is not known.

Aims: The aim of this study was to determine patterns of Notch (1–4) receptor expression in the livers of children with Alagille syndrome and to compare this with other cholestatic paediatric liver diseases.

Methods: Notch receptor expression was determined by RT-PCR on whole liver tissue in Alagille syndrome (n = 5; age 0.75–6 years), extrahepatic biliary atresia (EHBA n = 4; age 0.5–5 years), alpha one anti-trypsin deficiency (α1AT n = 4; age 0.75–13 years), normal (n = 4; age range 2–12 years). Notch receptors were localised on frozen liver sections by double immunofluorescence with either cytokeratin-19 (biliary epithelial cell marker) or CD31 (endothelial cell marker).

Results: All four Notch receptors were detected by RT-PCR in Alagille syndrome, while no consistent pattern was seen in EHBA and α1AT. Immunofluorescence studies showed that all Notch receptors colocalised with vascular endothelium in normal liver and neo-vessels in diseased tissue. Intact bile ducts, which were absent in EHBA and Alagilles livers, were positive for Notch1, 2, and 4 in normal and α1-AT tissue. In three out of 5 patients with Alagille syndrome, ductular proliferative cells were observed and notably were positive for Notch2 and 3. This was in contrast to α1-AT and EHBA, where ductular proliferative cells did not appear to express Notch 2 or –3, an observation consistent with the down-regulation of these Notches in adult cholestatic liver disease such as PBC, which we have previously reported. Notch2 and 3, however, were also present on stromal cells closely adjacent to the ductular reactive cells in all diseases.

Conclusion: These studies suggest that during liver disease pathogenesis, alterations in the expression pattern of Notch receptors, and therefore signalling, may lead to aberrant cellular morphogenetic responses in epithelium and endothelium.


S.Goddard, J. Youster, E. Morgan, D.H. Adams.

Liver Unit Laboratories, Queen Elizabeth Hospital, Edgbaston, Birmingham, UK

The route of antigen delivery has an important impact on the type of response generated. For instance, portal delivery is associated with a less aggressive immune response. The rate of rejection of allografts varies, with skin grafts being one of the most difficult to engraft and liver one of the more easy tissues to engraft. The reason for these differences is not clear, but may relate in part to differences in the type of immune response generated in the tissue. Dendritic cells (DC) are able to take up antigen in the tissue and migrate to draining lymph nodes, where they become very effective antigen presenting cells, able to initiate primary responses.

We have developed a method of isolating human liver DCs by overnight migration presented at BASL previously. We have compared the function of pure liver and skin DCs isolated in the same way, by measuring cytokine production, T cell proliferation, and T cell cytokine stimulation. We find that liver DCs produce IL-10 and MCP-1, and skin DCs do not. Neither cell type produces the functional form of IL-12. Both cell types stimulate T cell proliferation, but liver DC stimulated T cells produce less IFNγ and IL-4.

IL-10 maintains DC immaturity reducing their ability to efficiently stimulate immune responses; it is also shown to play a role in generating regulatory T cells. Using knock out models MCP-1 has been shown to play a role in T cell deviation to Th2 immunity. Our data therefore show that interstitial type DCs from the liver produce more down-regulatory cytokines than those from the skin, which also contains Langerhans type DCs. The function of these cytokines is supported by the data from cocultures showing that only skin DC stimulated T cells produce Th1 cytokines.


P.D. Richardson, X. Zhang, L. Collins, R.R. Mitry, A. Dhawan, J.W. Fabre.

Dept of Clinical Sciences, Institute of Liver Studies, Kings College Hospital, Denmark Hill, London, UK

Introduction: Transfection of double-stranded RNA (dsRNA) into a variety of cells can suppress expression of endogenous genes by a process termed RNA interference (RNAi). Large dsRNAs are processed by an RNase III endonuclease into smaller units that target homologous mRNA for degradation. This powerful tool has been utilised for targeted gene suppression. However, the presence of large dsRNA molecules in mammalian cells induces apoptosis, thus limiting its applications. Recently it has been demonstrated that small dsRNAs can be expressed in vitro from plasmids under the control of the human U6 pol III promoter and suppress gene expression without inducing apoptosis. In these studies we explore gene suppression in hepatoma cell lines and primary human hepatocytes with this technology.

Methods: RNAi was assessed in primary human hepatocytes and the HuH-7 and HepG2 human hepatoma cell lines. Cells were co-transfected with a constant amount of the luciferase reporter plasmid pGL3 and increasing amounts of an RNAi plasmid (SFF) targeting luciferase mRNA, and a control non-coding plasmid to maintain a consistent DNA concentration for all studies. Cells were transfected using a non-viral system consisting of a synthetic peptide containing an integrin binding moiety and a 16 lysine chain for electrostatic binding of DNA. Gene transfer was optimised by the addition of a fusogenic peptide and lipofectamine to enhance endosomal escape. Cell toxicity was assessed by the MTT assay. Luciferase activity was measured on a luminometer.

Results: Luciferase activity was suppressed in primary human hepatocytes by 64–85%, in HuH-7 cells by 75–83%, and in HepG2 cells by 84–88%. No toxicity was noted with the RNAi construct.

Conclusion: RNAi can markedly suppress gene expression in human hepatocytes. This powerful tool may have important implications for basic scientific studies and potentially offer new therapeutic options for conditions such as chronic viral hepatitis.


S. McKiernan1, S. Norris1, R. Vaughan2, A.J. Czaja3, P.T. Donaldson4.

1 Liver Unit, King’s College Hospital, London, UK; 2 Tissue Typing Laboratory, Guy’s Hospital, London, UK; 3 Mayo Clinic, Rochester, Minnesota, USA; 4 Centre for Liver Research, University of Newcastle, Newcastle-upon-Tyne, UK

MICA molecules, expressed exclusively on gastric and thymic epithelium, act as ligands for the NKG2D activity receptor on γδ T and NK (CD56+) cells. The normal liver has a large resident population of γδT, NK, and natural (N) T cells. Recent studies have identified two MICA alleles: MICA*008 and MICA*002, encoded within the HLA region on chromosome 6p21.3, as major determinants of disease susceptibility and resistance (respectively) in primary sclerosing cholangitis. The present study investigates the relationship between MICA polymorphism and type 1 autoimmune hepatitis. MICA genotyping was performed on genomic DNA from 55 adult patients of northern European ancestry all with type 1 AIH and 78 racially and geographically matched controls. Standard ARMS-PCR technique was used throughout. All patients were fully HLA genotyped for HLA A, B, DRB1, DQA1 and DQB1, as well as TNF -238, -308 polymorphisms. Overall there was no difference in the distribution of any of the MICA alleles at the phenotypic level. Patients were more likely to be homozygous carriers of the MICA*008 allele than controls (38.1% v 21.7%: p = 0.039, odds ratio = 2.22). This difference was entirely dependent upon linkage between MICA*008 and HLA DRB1*0301. These data are strong evidence that MHC encoded genetic susceptibility to type 1 AIH maps within the HLA class II region. This contrasts with our current findings in PSC where MICA polymorphism may play a major role. These findings have important implications for our understanding of the genetic basis of these two “autoimmune liver diseases” and may provide valuable insight into the pathogenesis of these idiopathic diseases. For example, the data above are suggestive evidence that class II restricted (CD4+) T cell responses predominate in type 1 AIH, whereas innate immunity or CTL (CD8+) T cell responses are more important in PSC.


A.E.R. Hicks, A.D. Burt, M.G. Thompson, C.P. Day.

Centre for Liver Research, Medical School, University of Newcastle-upon-Tyne, Framlington Place, Newcastle-upon-Tyne NE2 4HH, UK

Cholestatic liver diseases, characterised by obstruction of bile flow from the liver (eg biliary atresia and primary biliary cirrhosis), demonstrate progressive portal fibrosis (and ultimately cirrhosis) and are major indications for liver transplantation. Although indirect evidence has suggested that accumulation of toxic bile acids may be important in liver injury, the mechanisms whereby cholestasis leads to liver fibrosis are currently unknown. Because fibrosis is characterised by increased deposition of extracellular matrix proteins, particularly type I collagen, we hypothesised that bile acids may directly influence collagen synthesis by hepatic stellate cells (HSC), the principal fibrogenic cells of the liver.

This study, to determine the effects of bile acids on collagen biosynthesis, used HSC isolated from wild-type male Wistar rats, which were grown to confluence and passaged into multiwell plates. At 48 h HSC were serum-starved and incubated with 1 μCi 3H-proline. At 72 h, cells were treated with serum-free media alone, or media containing TGF-β1 (10 ng/ml) or bile acids (5–50 μM) for an additional 24 h. Media from each well was then assayed for 3H-proline incorporation into total protein and protein insensitive to collagenase treatment. DNA levels (μg/well) were then ascertained. TGF-β1 (10 ng/ml) and bile acids significantly increased total protein production by HSC when compared with control treated cells. Experiments to specifically identify effects on collagen synthesis demonstrated that while low dose (5–25 μM) cholic acid had no effect, 50 μM cholic acid significantly increased collagen production by 65% (compared with controls). Interestingly, lithocholic acid also significantly increased collagen production, but at the lower dose of 10 μM (collagen production was increased by 41% compared to controls). This study demonstrated that bile acids increase collagen biosynthesis by HSC and are therefore likely to play a key role in fibrosis in cholestatic disorders.


H.M. Evans, P.J. McKiernan, P. Davies, D.A. Kelly.

The Liver Unit, Birmingham Children’s Hospital, UK

Introduction: The calcineurin inhibitors cyclosporin (CyA) and tacrolimus (Tac) provide effective immunosuppression following liver transplantation (OLT) in children but may cause renal dysfunction. Mycophenolate mofetil (MMF) does not inhibit calcineurin and does not affect renal function. MMF has not yet been fully evaluated in children.

Aim: To evaluate the safety and efficacy of MMF in children with renal dysfunction following OLT.

Method: A retrospective review was performed of all children who commenced MMF between 1997 and 2001 for renal dysfunction post OLT. Glomerular filtration rate (cGFR) was estimated using the Schwartz formula ((40 × height (cm))/creatinine (μmol/l) and renal dysfunction was defined as cGFR < 65 ml/min/1.73 m2. From 1999, GFR was also measured using chromium EDTA slope clearance (CrEDTA). cGFR and liver function were measured at OLT, baseline (pre MMF transfer), 1, 2, 3, 6, and then 6 monthly following MMF. Results were analysed using the Wilcoxon signed rank test and general linear model analysis of variance to assess multi-factor effects.

Results: 46 children received MMF for renal dysfunction (21M; 25F). Median age at starting MMF was 10.7 years (0.9–18.1 years) and median time from OLT to starting MMF was 3.9 years (0–12.4 years). Median follow up was 1.7 years (0.2–5.4 years). 32 received MMF monotherapy, 4 MMF with prednisolone, 8 MMF with Tac (levels 2–4 ng/ml), 1 MMF with sirolimus and 1 MMF with CyA (levels 40–60 μg/l). There was a statistically significant increase in cGFR within 1 month of starting MMF (median increase 11.5%; p < 0.001; 95% CI 6.8 to 16.5) and between 1 and 2 months of treatment (median increase 4.6%; p < 0.01; 95% CI 1.2 to 7.9). The most significant change was seen in younger children (p = 0.022) and those who were closer to the time of OLT when transferred to MMF (p = 0.047). Four patients did not respond to MMF (median baseline cGFR 37.1, range 31.2 to 48.7) and required renal replacement therapy. Transient abnormalities in liver function were observed in 2, acute rejection associated with non-compliance in 1 and chronic rejection in 2, 1 of whom required re-transplantation. All patients who developed liver dysfunction had received MMF monotherapy. Side effects included significant bone marrow suppression in 3, headache in 1, and nausea in 1 but all responded to a decrease in MMF dose. Gastrointestinal bleeding requiring discontinuation of treatment occurred in 1.

Summary/Conclusions: Mycophenolate mofetil significantly improved renal function if started sooner after transplant and before irreversible renal dysfunction. In order to prevent potential rejection episodes, supplemental steroid therapy is required during the transfer from calcineurin inhibitors to MMF.

049 ASSOCIATION BETWEEN GENETIC POLYMORPHISM IN MATRIX METALLOPROTEINASE 2 (MMP-2, gelatinase A) and susceptibility to primary sclerosing cholangitis

S.N. Cullen, A. Armuzzi, T. Ahmad, J.H. Li, K. Ling, X. Yang, D.P. Jewell, R.W. Chapman.

Gastroenterology Unit, University of Oxford, UK

Background: Primary sclerosing cholangitis (PSC) is a disease of the intrahepatic and/or extrahepatic bile ducts that is characterised by concentric obliterative fibrosis and bile duct strictures eventually leading to biliary cirrhosis. The matrix metalloproteinase family of zinc-containing proteolytic enzymes is involved in mediating extracellular matrix degradation. Associations between polymorphisms of MMP-3 (stromelysin) and MMP-9 (gelatinase B) and susceptibility to PSC have recently been described. MMP-2 has an essential role in promoting cell invasiveness during tumour angiogenesis, arthritis and atherogenesis. This study assessed carriage of MMP-2 polymorphisms in relation to susceptibility to PSC.

Method: DNA was extracted from 80 patients with well-documented PSC and 130 healthy controls. Primers were designed to examine 3 polymorphisms in the MMP-2 gene, using a SSP/PCR method. PCR products were run on 1% agarose gels and read under UV light. PSC patients were compared with controls using 2×2 contingency tables and a χ2 test (with Yates correction). A Bonferroni correction for multiple comparisons was made using a factor of 3 (the number of polymorphisms tested).

Results: The F602F polymorphism was significantly associated with susceptibility to PSC compared with controls. The frequency of the mutant allele was 51% in the PSC patients compared with 31% in controls (pc < 0.001). No significant differences were found if patients with small duct PSC (n = 8) were analysed separately. Small duct PSC was defined as typical liver histology but a normal biliary tree on ERCP. There was no association between carriage of the mutant gene and progression of the disease to cirrhosis. No associations were seen with either of the other polymorphisms tested.

Conclusion: There is increased carriage of the F602F polymorphism in PSC patients compared with healthy controls. As this is a non-amino acid changing polymorphism, it is most likely that this is a marker for another functional polymorphism with which it is in linkage disequilibrium.


A. Marshall1, S. Rushbrook1, L.S. Morris2, I.S. Scott2, N. Coleman2, G. Alexander1.

1 University of Cambridge Department of Medicine, UK; 2 Cancer Cell Unit, Hutchison/MRC Research Centre, Hills Road, Cambridge, CB2 2QQ, UK

Background: Induction of entry into, followed by block of, cell cycle progression, in virus infected cells is a mechanism used to promote replication of a number of viruses including CMV, HIV, EBV, and HSV. Hepatitis B and C viruses might also use this mechanism because transfection of HCV core, NS3, NS4, or NS5A proteins or HBV X protein affects cell growth in vitro. Minchromosome maintenance proteins (MCM) are essential for DNA replication and are useful markers for cell cycle state in cell populations. Mcm-2 expression is increased in liver biopsies from patients with chronic HBV and HCV, indicating increased cell cycle entry. We have determined the proportion of hepatocytes within each cell cycle phase in chronic viral hepatitis and compared our findings with those in regenerating liver.

Methods: Liver biopsies from 69 patients with chronic HCV and 38 patients with chronic HBV were studied. Standard immunohistochemical techniques were used to detect Mcm-2, cyclin D1 (G1 phase), cyclin A (S), cyclin B1 (G2), phosphorylated histone 3 (M), and p21 (cyclin dependant kinase inhibitor). Controls were liver biopsies showing regeneration following drug-induced injury.

Results: The table shows the percentage of hepatocytes positive for Mcm-2 and each cell cycle phase marker. p21 was increased: HCV 17.7%, HBV 15.2%, and regeneration 23.5%.

Abstract 50

Conclusion: The data show that in chronic HBV and HCV significantly fewer cycling hepatocytes reach S phase in comparison with regenerating liver, consistent with G1 arrest. p21 is increased relative to Mcm-2 in chronic viral hepatitis versus regeneration and may be induced by HBV and HCV to mediate cell cycle arrest, which could facilitate viral replication but adversely affect cell function and hepatic regenerative capacity.


M.Warren, C. P. Day, C. Godfrey, O.F.W. James.

Centre for Liver Research, Medical School, University of Newcastle-upon-Tyne, Framlington Place, Newcastle-upon-Tyne NE2 4HH, UK

Aims: Little research has been performed into the quality of life (QoL) effects of chronic liver disease, especially alcoholic liver disease (ALD). Although QoL might be expected to decline with worsening ALD, the effects of ongoing drinking are not known.

Methods: Recently Younoussi et al described a tool for measuring QoL in ALD covering 6 health domains, and resulting in a final score of 1–7.1 This study aimed to use this tool to assess the QoL in patients with ALD, considering the effects of disease stage and alcohol consumption.

Results: Ninety (72%) completed questionnaires were returned from 66 (73%) males and 24 (27%) females, mean age 51years. Proportions reportedly abstaining, drinking “sensibly” (< 21 units and < 14 units/week in males and females, respectively) and to excess (> 21 and > 14 units/week) were 50%, 22%, and 28%. Six (7%) patients had fatty liver, 75 (83%) had biopsy proven or clinically evident cirrhosis. The stage of disease was unknown in 9 (10%). Increasing CP score correlated negatively with QoL score (r = −0.216, p < 0.05). QoL was also reduced significantly in excessive drinkers compared with abstainers (difference 1.16, p < 0.05) and sensible drinkers (difference 0.95, p < 0.05). There was no difference in QoL between abstainers and sensible drinkers. The effects of drinking behaviour on QoL were independent of CP score or stage of disease.

Conclusions: Although it is known that continuing heavy alcohol consumption adversely affects mortality, we have now shown that it also adversely affects QoL, however, sensible alcohol consumption does not impair QoL. Additionally, we have previously shown that sensible ongoing drinking does not adversely affect mortality.2 Although we must continue to strongly recommend abstinence from alcohol for our ALD patients, if confirmed, this study supports the notion that low (“sensible”) ongoing alcohol consumption neither affects QoL nor mortality in ALD. Perhaps treatment objectives and decisions may need to be modified in this light.




S.M. Curbishley, T. Lalor, D.H. Adams.

Liver Research Laboratories, University of Birmingham, Birmingham, UK

Chronic inflammatory liver disease is associated with increased expression of the IFNγ inducible chemokines IP-10, HuMig, and ITAC, and an associated infiltration of the liver by T cells expressing the receptor for these chemokines, CXCR3. We have investigated the role of liver-derived CXCR3 chemokines in recruiting and retaining lymphocytes by comparing the ability of human sinusoidal endothelial cells (HSEC) and biliary epithelial cells (BEC) to express CXCR3 ligands and adhesion molecules involved in lymphocyte recruitment. The ability of these chemokines and adhesion molecules to trigger and support adhesion under fluid flow has also been investigated.

Methods: Human BEC and HSEC were isolated from explanted livers at transplantation. Cells were grown to confluence before being stimulated with combinations of cytokines and the expression of cell surface CAM and chemokine secretion measured. For functional studies confluent monolayers of HSEC were cultured in glass capillary tubes before being perfused with lymphocytes at physiologically relevant shear rates and the pattern of adhesive events recorded.

Results: Stimulation with TNF-α or IL-1β induced expression of ICAM-1, E-selectin, and VCAM-1 on HSEC and ICAM-1 and VCAM-1 on BEC. TNF-β or oncostatin M alone failed to upregulate CAM on HSEC but were as effective as TNF-α in induction of ICAM-1 on BEC. Both HSEC and BEC secreted IP-10, ITAC, and HuMig in response to IFN-γ and secretion increased further if TNF-β, TNF-α ,or IL-1β were added. In contrast, oncostatin M suppressed the IFN-γ response. BEC consistently secreted higher levels of IP-10 and HuMig compared with HSEC. Functional studies demonstrated a reduction in total lymphocyte adhesion following incubation with a CXCR3 blocking mAb or inhibition of GPI receptors with pertussus toxin. Incubation with ICAM-1 and VCAM-1 blocking mAb further reduced adhesion. Incubation with recombinant human IP-10, MIG, and ITAC increased total adhesion when compared with untreated controls.

Conclusions: On activation with cytokines both HSEC and BEC express CXCR3 ligands and CAM that promote lymphocyte recruitment. However, cell-specific and cytokine specific-responses suggest that the local microenvironment will be critical in determining the nature and distribution of the lymphocytic infiltrate in chronic inflammation. We have shown for the first time that CXCR3 ligands co-operate with ICAM-1 and VCAM-1 to promote lymphocyte adhesion to endothelial cells under flow.


C.R. Brooks, W.M. Rosenberg, S.I. Khakoo.

Liver Unit, Southampton General Hospital, Tremona Rd, Southampton, UK

NK cells comprise approximately 45% of the intrahepatic lymphocyte compartment. Their functions are controlled by activating and inhibitory cell surface receptors, of which the KIR form a major component. The extracellular domains of specific pairs of activating and inhibitory KIR are too closely related to be distinguished by monoclonal antibodies. Therefore, to accurately determine KIR receptor expression on intrahepatic NK cells, we have developed a quantitative PCR assay using Taqman technology. Taqman assays were developed for the eight 2Ig domain KIR genes. The assays were validated on a panel of plasmids, each containing a cDNA for one KIR. Cross-reactivity was observed in only 3 out of 64 possible combinations tested and only when the cross-reacting KIR was present in greater than 1000-fold excess. cDNAs from 11 intra-hepatic NK cell clones were then tested. Expression of each KIR was standardised to that of KIR2DL4, which is ubiquitously expressed by all NK cells. Individual NK cell clones expressed KIR in a clonal fashion. However, a hierarchy of KIR expression level was observed in those clones positive for specific KIR: KIR2DL1 (up to 54 copies per 2DL4 copy) > 2DL3 > 2DS4 > 2DS2 > 2DL2 > 2DS3 (less than 0.08 copies per 2DL4 copy). A similar hierarchy was observed in four NK cell clones derived from the peripheral blood of a second individual.

Despite the KIR family working as a functional unit maintaining the balance between self-tolerance and immune responsiveness, our data suggest that the transcription of individual KIR genes may be independently regulated. In tandem with flow cytometry this technique will increase our understanding of the role of NK cells and their receptors in liver immunity.


D.M. Forton, P. Karayiannis, N. Mahmud, S.D. Taylor-Robinson, H.C. Thomas.

Hepatology Section, Imperial College, St Mary’s Campus, Praed St, London, UK

Specific HCV variants have been found in postmortem brain matter, consistent with viral replication. Mutations within the IRES result in changes in IRES activity depending on the cell type used for experimental transfection, suggesting that the IRES may be important for cellular tropism. This study examines the translational activity of brain derived HCV IRES RNA sequences. Postmortem brain, liver, and serum were analysed for HCV RNA from 2 HCV+ve patients (PI and PII). Extracted RNA was amplified by RT-PCR, using a proofreading enzyme. The IRES region, comprising the whole 5’ untranslated region and the coding sequence for the first 24 amino acids of the core protein was targeted with appropriate primers. The PCR products were cloned and 25 to 30 clones from each compartment were sequenced. The normalised genetic complexity (nCx) of the IRES sequences in each compartment was calculated. A bicistronic construct containing two reporter genes (Renilla and Firefly luciferases) was used to test IRES translation efficiency. Translation of the Renilla luciferase, upstream from the IRES of interest, occurs in a cap-dependent manner, whereas the downstream Firefly luciferase is under the control of the HCV IRES. In vitro IRES translational efficiencies of the chosen IRES sequences were tested, using the TNT coupled reticulocyte lysate system and the dual luciferase assay kit (Promega), by comparing the measured luciferase ratios. In PI, the nCx of the brain and liver IRES sequences (both 0.24) was greater than that of serum (0.16), whereas in PII, the nCx of serum was equivalent to brain (0.35 v 0.30). In both patients, the master sequence was identical in all compartments. However, 24% and 35% of the brain-derived clones in P1 and PII respectively, were not seen in the other compartments. The mutations in the IRES in the brain variants resulted in changes in translational activity of between +11% and –86% compared with the master IRES sequence, suggesting that the IRES may be important in facilitating viral replication in cells other than hepatocytes, such as cells of the CNS.


N. Holden, P.F. Lalor, D.H. Adams.

Liver Research Laboratories, Department of Medical Science, University of Birmingham, Birmingham, UK

One of the major contributing factors to hepatic graft failure is activation of the endothelial cells within the transplanted organ and subsequent recruitment of immune cells. Animal models of reperfusion injury show that platelets are sequestered in the liver following transplantation1,2,3 and that the degree of platelet adhesion to grafted liver tissue correlates to survival.2 We propose that adhesion of platelets to exposed sub-endothelial matrix or damaged endothelial cells in grafted tissue results in adhesion of leukocytes to the sequestered platelet layer. This hypothesis is supported by evidence that shows that adherent platelets and neutrophils are commonly observed in close proximity following hepatic damage due to LPS treatment,4 endotoxaemia5 or reperfusion injury.1 Thus, we used static and blood flow based adhesion assays in conjunction with tissue sections and cultured endothelial cells to assess whether resting platelets could bind specifically to liver endothelia. Our data show that platelets bind to both liver endothelial cell in situ and in culture under conditions of flow and that the adhesion is mediated by the integrins Gp1b, avb3, and allbb3, and that deposition of fibrinogen may contribute to platelet adhesion to endothelial cells in liver tissue but not cultured endothelia. Adhesion occurred under both static conditions and physiological blood flow and was highest on liver endothelial cells when compared with cells from other tissues. Therefore, in conclusion we believe that the deposition of adherent platelets in grafted organ microvasculature may provide an adhesive substrate for the recuitment of inflammatory cells.







A. Mancuso1, K. Fung1, M. Mela1, J. Tibballs2, A. Wathinson2, A.K. Burroughs1, D. Patch1.

1 Liver Transplantation Unit and Hepatobiliary Medicine, Royal Free Hospital, Pond Street, London; 2 Department of Radiology, Royal Free Hospital, Pond Street, London

Background: TIPS provides decompression of the congested liver in Budd-Chiari syndrome (BCS) irrespective of inferior vena cava obstruction—its great advantage over side to side porto-caval shunt, and also obviates the surgical mortality particularly in acute BCS. Thus, TIPS is considered the treatment of choice in many centres.

Aim: to evaluate the long term outcome of TIPS in BCS patients, particularly those presenting with acute liver failure.

Patients: Fifteen patients (7 males) with BCS underwent TIPS; average age was 39.5 (range 20–73). Aetiology was myeloproliferative disease (5), paroxysmal nocturnal haemoglobinuria (3), antiphospholipid syndrome (1), protein C and S deficiency (1), rectal adenocarcinoma hepatic metastasis (1), and cryptogenic (4). The indication was liver failure in 8 and chronic BCS in 7 (refractory ascites 3, variceal bleeding 3, progressive fibrosis 1). 12 of 15 were in Child-Pugh C class.

Results: All 3 hepatic veins were occluded in 5 patients. Only one patient had unsuccessful TIPS placement (percutaneous hepatic vein dilatation, alive). In the acute liver failure group, 4 of 8 died within few days after TIPS placement: ruptured liver (on fibrinolysis because of immediate TIPS thrombosis) (1); ruptured portal vein (1); progressive liver failure (1); and acute pulmonary oedema (1). There were two significant technical complications: one portal vein rupture (died) and one intrahepatic haemorrhage (resolved). The 10 who survived with TIPS had an average follow up of 18 months (range 3–39). All were anticoagulated post procedure. In all, ascites resolved and liver function improved. One died after 16 months because of the original metastatic disease. Four patients had TIPS dysfunction: balloon dilatation resolved this in 3, one had a repeat TIPS. No patient has required liver transplantation.

Conclusion: TIPS can be performed in BCS even if all hepatic veins are thrombosed, obviating surgical shunting and transplantation, and long term results are good. However, selection of patients in acute BCS is clearly important as mortality is high (50%), 3 of 4 depths as a complication of the primary disease or TIPS; in these cases liver transplantation may result in a better outcome.


K.L. Nash, A.M.L. Lever, G.J.M. Alexander.

Department of Medicine, University of Cambridge, Addenbrooke’s Hospital Hills Road, Cambridge, UK

Introduction: Efficient gene therapy is likely to require long term, organ specific therapeutic gene expression. Lentiviral vectors based on HIV-1 are promising gene delivery vehicles due to their ability to integrate transgenes into non-dividing cells. Many experimental vectors express transgenes under the control of the cytomegalovirus (CMV) immediate-early gene promoter. Although this promoter directs strong gene expression in vitro, it may be rapidly shut off in vivo. This study explores the potential of HIV-1 based vectors to transduce hepatocytes and compares gene expression from different promoters.

Methods and Results: Replication deficient HIV-1 based vectors were produced by transient co-transfection of 293T cells with vector and helper constructs. Cultured hepatocytes were transduced with vectors expressing green fluorescent protein (GFP) under the control of the CMV promoter. Transduction efficiencies exceeding 50% were obtained with gene expression persisting for over 3 months.

The CMV promoter in the vector construct was substituted by the alpha-1 antitrypsin gene promoter or promoters derived from the hepatitis B virus (HBV) genome. Most of these promoters gave high level GFP expression when transfected into 293T cells or hepatocytes. Upon hepatocyte transduction, however, only the CMV, alpha-1 antitrypsin, and HBV core promoters were able to produce GFP expression. Addition of the HBV enhancer 2 element improved the transducing ability of the HBV surface promoter and increased specificity for hepatocytes.

Conclusions: Lentiviral vectors can be used to direct long term transgene expression in liver cells. Promoters derived from the alpha-1 antitrypsin gene or HBV are alternatives to the CMV promoter for liver-specific gene expression.


S.W.M. Olde Damink1,4, C.H.C. Dejong2,4, D.N. Redhead3, P.B. Soeters4, P.C. Hayes1, N.E.P. Deutz4, R. Jalan1,5.

1 Liver Unit, Royal Infirmary, Edinburgh, UK; 2 Dept of Surgery, Royal Infirmary, Edinburgh, UK; 3 Dept of Radiology, Royal Infirmary, Edinburgh, UK; 4 Dept of Surgery, Univ. Maastricht, The Netherlands; 5 Institute of Hepatology, RFUCLMS, London, UK

Patients with liver cirrhosis have diminished urea synthesis capacity, depriving the body of its main route of ammonia detoxification. In experimental animals we showed that the kidneys actively diminish renal venous ammonia release after creation of a portacaval shunt. This study was designed to study the consequences of TIPSS placement on arterial ammonia and renal ammonia metabolism.

Material and Methods: Six patients with cirrhosis of the liver (3F/3M, mean age 60.5 (±4.6) years, mean Pugh 7.5 (±0.7), aetiology 5 ALD, 1 PSC) that underwent a TIPSS insertion were studied. Blood was sampled prior to the shunt puncture/placement and one hour hereafter from a femoral artery and the right renal vein. Renal plasma flow was measured using a primed, continuous infusion of para-aminohippuric acid (PAH). Ammonia, urea and PAH were determined spectrophotometrically, amino acids by HPLC. Substrate fluxes were calculated as venous-arterial concentration differences. Positive flux represents production. Additionally, no significant differences could be observed between pre-post TIPSS fluxes of urea, glutamine, and alanine across the kidney or arterial pH.

Conclusion: TIPSS placement induces hyperammonemia, although the kidneys release less ammonia into the systemic circulation. The cause of this post-TIPSS hyperammonemia is unclear but could be due to increased portal-systemic shunting of gut produced ammonia. This study confirms our previous data in animal models that the kidney plays an important role in ammonia homeostasis.

Abstract 58


C.J. Marek1, T. Monaghan2, D.E. Smart2, J.N. Primrose2, M. Koruth3, M. J.P. Arthur2, D.A. Mann2, G.I. Murray1, M.C. Wright1.

1 Department of Molecular and Cell Biology, University of Aberdeen, Aberdeen, UK; 2 Liver Group, Southampton General Hospital, Southampton, UK; 3 Aberdeen Royal Infirmary, Aberdeen, UK

The PXR is an orphan nuclear receptor known to regulate a variety of genes associated with xenobiotic metabolism in hepatocytes. RT-PCR and Western blotting have indicated that human hepatic stellate cells also express high levels of PXR mRNA and protein. Hepatic stellate cells are pivotal in liver fibrogenesis in that they trans-differentiate from a quiescent (vitamin A storing) to a myofibroblastic (α-smooth muscle actin positive) phenotype in response to chronic liver damage. The myofibroblast-like phenotype of stellate cells primarily contributes to fibrotic accumulation in the liver. Using an in vitro model of liver fibrosis, in which hepatic stellate cells undergo a phenotypically identical process of trans-differentiation to that in vivo, human hepatic stellate cells were treated with reported activators of the PXR. PXR activators inhibited the trans-differentiation of human hepatic stellate cells as determined by the levels of expression of α-smooth muscle actin versus glial fibrillary acid protein (a constitutive marker of stellate cells). The effect of the potent rodent PXR activator pregnenolone 16α-carbonitrile (PCN) in an in vivo model of liver fibrosis was therefore examined. Rats were administered 1 ml CCl4/kg body weight twice weekly by i.p. injection (mixed 1:1 with olive oil) to cause liver fibrosis with or without administration of 25 mg PCN/kg body weight once a week. Dosing continued for 6 weeks before animals were sacrificed and serum and liver samples prepared. Serum samples were analysed for the liver enzyme AST and liver sections were stained with H&E; anti-α-smooth muscle actin (to detect trans-differentiated pro-fibrogenic stellate cells) or sirius red to detect fibrotic scarring. There was no significant difference in serum AST levels between CCl4-treated and CCl4+PCN-treated rats. Blinded histological scoring of H&E stained liver sections did not detect a difference in apparent necrosis between CCl4-treated and CCl4+PCN-treated rats. These observations suggest that PCN did not modulate the hepatotoxicity of CCl4. However, PCN treatment significantly reduced the extent of α-smooth muscle immunostaining and sirius red staining in liver sections. These data indicate that PCN is anti-fibrogenic and suggest that the PXR is a potential target for anti-fibrogenic therapy.


M.I. Prince1, H.C. Mitchison2, D. Ashleigh3, D. Burke3, M. Edwards4, M.G. Bramble4, O.F.W. James1, D.E.J. Jones1.

1 Centre for Liver Research, Newcastle University, NE2 4LL, UK; 2 Sunderland Royal Hospital, SR4 7TP, UK; 3 Cumberland Infirmary, Carlisle CA2 7HY, UK; 4 Middlesborough General Hospital Green Lane, TS5 5AZ, UK

Background: Although fatigue is the commonest symptom in PBC, no controlled therapeutic trial has specifically addressed this. We previously reported that oral antioxidants reduced fatigue in PBC patients in an uncontrolled trial (J Gastro Hep 1999;14:1034–40). We aimed to assess this in a controlled trial.

Methods: 61 patients with PBC diagnosed on standard criteria were enrolled from 4 centres. Exclusion criteria were depression, hypothyroidism, anaemia, and renal failure. Patients received 12 week courses of oral antioxidants (L-selenomethionine (75 μg), vitamins A (3 mg), E (50 mg), and C (150 mg), L-methionine (375 mg), and ubiquinone (25 mg)) and identical placebo in random order separated by a 4 week wash out period. Primary outcome was fatigue measured by the Fisk score (FIS), a fatigue scale previously validated in PBC. Secondary outcomes were; biochemical parameters, depression assessed with the Beck’s Depression Index (BDI), symptoms of itch, bone pain, hypochondrial pain, and dry mouth or eyes measured on VAS and Likert scales. Analysis used the Wilcoxon rank sum test.

Results: There were no changes in FIS during active treatment (median change in total score = 0, p = 0.61, physical domain = 0, p = 0.74, cognitive domain = 0, p = 0.96, social domain = 1, p = 0.47). Small improvements in FIS were noted during placebo treatment (median change total score = 3, p = 0.03, physical = 1.5, p = 0.01, cognitive = 1, p = 0.03, social = 2, p = 0.07). Similar results were obtained when analysis was performed excluding patients with above median (= 12) initial BDI scores, excluding confounding by depression. Neither medication was associated with changes to other symptoms or biochemical parameters.

Conclusions: There was no evidence that oral antioxidant supplementation reduced PBC associated fatigue. However, importantly this study shows that trials of symptomatic therapy for fatigue are feasible and should be conducted.


M. Warren, C.P. Day.

Centre for Liver Research, Medical School, University of Newcastle-upon-Tyne, Framlington Place, Newcastle-upon-Tyne, UK

Alcohol excess is the commonest cause of chronic pancreatitis (CP) in the developed world, with postmortem studies showing that up to a third of excessive drinkers with liver damage will have evidence of CP. Despite this, our clinical impression is that CP is a treatable and often overlooked diagnosis in patients with alcoholic liver disease (ALD), with diarrhoea and weight loss attributed to the intestinal effects of alcohol and the development of diabetes considered an expected complication of cirrhosis. The aim of this study was to determine the prevalence of pancreatic dysfunction in an unselected group of patients with cirrhotic ALD and no known pancreatic disease.

Patients with clinical or histological evidence of alcoholic cirrhosis were included in the study. Patients with a history of pancreatitis or previous investigations for pancreatic dysfunction were excluded. Clinical and laboratory data were collected, including body mass index (BMI) and current drinking status. Pancreatic function was assessed by faecal elastase measurement, a recently described test for CP, with a reported sensitivity of 77–100%, and specificity of 83–93% for the diagnosis of moderate/severe CP.

Forty-seven stool samples were collected, 34 (74%) from males. Mean age was 54 years, 52% were smokers. The mean Child’s Pugh score was 7 (range 5–12) and the mean BMI was 27.8 (±5.3) kg/m2. An abnormal faecal elastase result was present in 19.2% of the patients (95% CI 10 to 32.5%), 6 of which (12.8%) showed severe pancreatic exocrine insufficiency. Neither age, sex, stage of liver disease, BMI, drinking status, nor the presence of diabetes were associated with an abnormal elastase result. These data suggest that unsuspected, but significant, pancreatic insufficiency is present in around a fifth of unselected patients with cirrhotic. In view of the fact that it is a treatable condition and that no clinical or biochemical markers accurately predict its presence, we suggest that consideration should be given to screening ALD patients for CP with faecal elastase.


P.F. Lalor1, S. Cheung1, F. Berditchevski2, D.H. Adams1.

1 Liver Research Laboratories, Department of Medical Science, University of Birmingham, Birmingham, UK; 2 Cancer Studies, Department of Medical Science, University of Birmingham, Birmingham, UK

Primary solid tumours in humans exhibit characteristic patterns of metastatic spread that cannot soley be attributed to patterns of blood flow. Therefore, we hypothesised that expression of adhesion receptors on metastatic tumour cells may influence their routes of dissemination in a manner analogous to the tissue-specific recirculation of leukocytes. Between 50 and 71% of patients with primary carcinoma of the breast subsequently develop metastases in the liver and so we investigated the adhesion of breast cancer cells to liver endothelium and the role of the CD82 molecule in this process. CD82, a member of the tetraspannin family, acts as a regulator of other adhesion receptors, and has been implicated in the metastatic process. We performed adhesion assays using tissue sections and cultured liver endothelial cells to investigate whether such tumour cells bound specifically to the liver, and if so what molecules were responsible for the interaction. We have shown that MCF-7 breast cancer cells bind preferentially to the liver and specifically localise to sinusoidal endothelial cells. The adhesion event was mediated in part by α3 and β1 integrin adhesion receptors and was dependent on the expression of the tetraspannin protein CD82. When we performed adhesion assays under conditions that model the blood flow in the liver we demonstrated that CD82 may act by strengthening integrin-mediated adhesion of tumour cells to endothelial cells. Thus tumour cells expressing CD82 were more strongly attached to the endothelial cells and less likely to be detached by blood flow. This may, therefore, increase the likelihood of these cells penetrating the liver parenchyma and forming a metastasis.


S. Edwards, P.F. Lalor, G.E. Rainger, D.H. Adams.

Liver Laboratories, Queen Elizabeth Hospital, Birmingham, UK

Background: The fenestrae in sinusoidal endothelial cells permit the exchange of solutes from the sinusoid, via the space of Disse, to the protruding microvilli of hepatocytes. This relationship is reciprocated by hepatocyte secretion of the endothelial growth factor VEGF, which maintains sinusoidal endothelial cell growth and differentiation. We hypothesised that the close relationship between sinusoidal endothelial cells and hepatocytes in vivo will influence lymphocyte adhesion by modulating adhesion receptors expressed on the endothelial surface. In order to address this hypothesis, we developed a coculture system in which hepatocytes and endothelial cells were cultured on either side of a porous membrane and the ability of lymphocytes to adhere to the endothelium studied under fluid flow.

Methods: The hepatoblastoma line HepG2 and human umbilical vein endothelium (HUVEC) or primary human hepatic sinusoidal endothelial cells (HSEC) were cocultured on tissue culture inserts plumbed into a flow assay by means of a parallel plate. Lymphocytes were perfused over the endothelial surface, in the presence and absence of endothelial blocking antibodies under resting conditions or after stimulation with TNF-β (10 ng/ml 24 hours). Lymphocyte adhesion was calculated per mm2/106 cells perfused.

Results: The level of lymphocyte adhesion to cocultured endothelium was more than double that to endothelium alone and was dependent on endothelial P-selectin, ICAM-1, E selectin, and VCAM-1. The adhesion seen in unstimulated cocultures was comparable to that seen with TNF-β stimulated endothelium alone. Stimulation of cocultures with TNF-β further increased adhesion that was mediated by ICAM-1, E selectin, and VCAM-1. In cocultures with HSEC but not HUVEC adhesion was reduced using small molecule inhibitors of the endothelial adhesion molecule VAP-1, which is preferentially expressed on hepatic endothelium.

Summary: Coculture with an hepatocyte cell line increases the level of lymphocyte adhesion compared with endothelial cells cultured alone. When human sinusoidal endothelial cells are used this adhesion is dependent on ICAM-1, VCAM-1, VAP-1, and E-selectin.


A. Alsowmely, A. Grasso, H. Hodgson, A. Burroughs, B. Davidson, A. Quaglia, A. Dhillon, K. Rolles.

Royal Free Hospital and Royal Free and University College Medical School, Rowland Hill Street, London, NW3 2PF, UK

Background: Orthotopic liver transplantation (OLT) is a potentially curative treatment option for primary HCC. In cirrhosis replacement of the cirrhotic organ also reduces the likelihood of further de novo HCC. During the period 1990–2000 at this centre 84 patients underwent liver transplantation for HCC, using accepted selection criteria (single tumours < 5 cm in diameter, or limited numbers of smaller tumours).

Aim: To identify factors that determine graft and patient survival in this group of patients undergoing OLT for HCC.

Method: Retrospective analysis of database on 84 consecutive patients. All HCC were biopsy proven (pre or postoperatively). Results were tested for significance by univariate analysis of Kaplan-Meier Survival curves. 2-sided p value of < 0.05 was considered statistically significant.

Results: 16 HCCs were identified at explant. At analysis, 54 patients were alive and 30 had died. For the whole group mean follow up was 26 months, median 17 months. By univariate analysis, improved survival was associated with (a) incidental finding of HCC, eg actuarial 3 years survival of 93% v 58%; p < 0.02 for difference in K-M curve; (b) female sex, 3 years survival of 91% v 58%; p < 0.02 for difference in K-M curve; and (c) short length (< 3 days) ITU stay with 3 years survival of 69% v 25%; p < 0.02 for difference in K-M curve. No difference was found in survival between different age groups of patients, different tumour size (< 30 mm v >30 mm), multifocality, viral and non-viral causes, CMV status, immunosuppressive agents used, and interval between tumour diagnosis and transplantation (> 90days cf < 90days). One or more episodes of acute rejection was associated with a trend towards longer survival but non-significantly (p = 0.067).

Conclusion: Incidental HCC has excellent survival, but pretransplant diagnosis of HCC results in worse survival not necessarily related to tumour characteristics or waiting time.


G.P. Aithal1, J.B.S. Leathart2,3, T. Singh Dang3, C.P. Day2, A.K. Daly2,3.

1 Queen’s Medical Centre University Hospital, Nottingham; 2 Centre for Liver Research; 3 Department of Pharmacological Sciences, University of Newcastle Medical School, Framlington Place, Newcastle-upon-Tyne

Diclofenac is a non-steroidal anti-inflammatory drug that causes rare, but serious, hepatotoxicity. Both metabolic idiosyncrasy and immunologic mechanisms are believed to contribute to liver injury. Our studies indicate that polymorphism in the main diclofenac metabolizing cytochrome P450, CYP2C9, does not account for susceptibility to hepatotoxicity. However, diclofenac also undergoes glucuronidation by UGT2B7. A polymorphism in the coding region of UGT2B7 resulting in a His268(H) to Tyr(T) substitution is tightly linked to an upstream T-161C, which is associated with variation in morphine 6-glucuronidation activity. Using PCR-RFLP, we genotyped 26 patients (21 females; age 24–70 years) with diclofenac-hepatotoxicity and 98 healthy controls for codon 268 polymorphism. UGT2B7 codon 268 genotypes in patients and controls are summarised in Table 1. The odds ratio for development of hepatotoxicity in subjects possessing the YY genotype was 2.69 (95%CI 1.11 to 6.50; p = 0.04). The altered glucuronidation associated with YY genotype may effect the metabolism of diclofenac and its oxidative metabolites from the hepatocyte, increasing the formation of protein adducts and therefore hepatotoxicity.

Abstract 65


D.P. Bogdanos1, Y. Ma1, S. Bansal2, P. Cheeseman2, G. Mieli-Vergani2, D. Vergani1.

1 Institute of Hepatology, University College London Medical School, Chenies Mews, London; 2 Dept of Child Health and Institute of Liver Studies, King’s College Hospital, Denmark Hill, London, UK

Background: Liver cytosol antibody type 1 (LC1) recognises a cytosolic liver protein of ∼60 kDa and was originally reported todefine a subset of autoimmune hepatitis (AIH) either negative for other autoantibodies or positive for anti-liver kidney microsomal type 1 (LKM1) antibody, the serological hallmark of type 2 AIH (AIH-2). Antibodies to a 60 kDa cytosolic liver protein were later described in ANA and/or SMA positive type 1 AIH (AIH-1). The molecular target of LC1 has been recently described as the liver enzyme formiminotransferase cyclodeaminase (FTCD).

Aim: To determine the prevalence of anti-FTCD antibodies in autoimmune liver disease using a molecularly based assay.

Methods: Antibodies against recombinant FTCD, expressed with a Baculovirus vector in insect cells (Euroimmun UK Ltd, Cardiff, UK) were tested by ELISA; ELISA positivity was verified in Western blot using human liver cytosol fraction.

Patients: Sera from 129 paediatric patients (87 female, median age 8.4 years, range 0.2 months–17.9 years) with autoimmune liver disease (57 AIH-1, 40 autoimmune sclerosing cholangitis (ASC) and 32 AIH-2 were collected at the time of diagnosis and stored as frozen aliquots at −70°C. 245 (209 children and 36 adults) with liver disease: 34 α1-antitrypsin deficiency, 33 Wilson disease, 38 Alagille’s syndrome, 40 biliary atresia, 15 HBV, and 49 HCV infection, 20 primary biliary cirrhosis, 16 SLE, and 74 normal subjects (60 children) were studied as controls.

Results: Five of the 57 (9%) patients with AIH-1 and 9 of 32 (29%) with AIH-2 were positive for anti-FTCD antibodies and immunofixed a 60 kDa liver cytosolic protein in Western blot. The patients with ASC and both pathological and normal controls were negative. The frequency of anti-FTCD antibodies in AIH-2 was significantly higher than in AIH-1 (p = 0.02).

Summary: Akin to LC1, anti-FTCD antibodies are specific for autoimmune hepatitis, being, however, significantly more frequent in type 2 AIH. No specific clinical manifestation was associated to the presence of LC1.


D.E.J. Jones, R. Walter, M.I. Prince, M. Hudson.

Centre for Liver Research, University of Newcastle, UK

It has previously been thought that the prevalence of portal hypertension (PHT) in patients with PBC is low, and that variceal haemorrhage is a rare and well-tolerated event. We have recently demonstrated, however, in a large patient cohort, a point prevalence for PHT in PBC of 25%. Moreover, the one year survival in PBC patients following the identification of oesophageal varices was only 63% regardless of bleeding history. We believe, based on these data, that the development of PHT and its sequelae in PBC is therefore a significantly less benign event than has previously been thought to be the case. Furthermore, we believe that the adverse outcome associated with PHT development in PBC should influence timing of liver transplant in this patient group. The Mayo prognostic model is widely used to predict outcome in PBC and guide timing of transplantation. In the current study we have addressed the extent to which the Mayo model predicts the adverse outcome associated with the development of portal hypertension and varices in our patient population. Retrospective data were obtained from clinical case records and database records for our cohort of 294 PBC patients developing PHT. Mayo risk scores and corresponding estimated survivals were calculated using steady state values obtained immediately prior to the development of the complication under study. Comparison between predicted and subsequent actual survival was made using Kaplan Meier analysis. Predicted and actual survival were identical in patients not developing PHT (survival from disease diagnosis) and in patients developing PHT but not varices (from PHT diagnosis). Actual survival was, in contrast, significantly worse (p < 0.001) than predicted survival in patients developing varices even in the absence of variceal haemorrhage.

Conclusion: The Mayo prognostic score fails to predict the adverse outcome associated with the development of oesophageal varices in PBC even in the absence of variceal haemorrhage. The Mayo risk score should not be used to guide timing of liver transplantation in PBC patients developing varices.


P. Haigh1,2, K. Tang1,2, B. Chiva1,2, N. Tatman1,2, W. Lees2, M. Novelli2, D. Maxwell3, C. Tibbs3, I. Murray-Lyon4, A. McNair5, M. Glynn6, R. Williams1,2, N. Naoumov 1,2.

1 Institute of Hepatology; 2 UCL Hospitals; 3 St. George’s Hospital; 4 Imperial College; 5 Greenwich Hospital; 6 Royal London Hospital

The presence of cirrhosis is a negative predictor of treatment response in chronic hepatitis C and such patients are frequently considered unsuitable for antiviral therapy. Accumulating evidence, however, shows that combination therapy involving pegylated interferons is more effective at achieving HCV clearance and has beneficial effects on liver histology.

Aims: To prospectively evaluate the efficacy of Viraferon-Peg plus Ribavirin on viral clearance and liver architecture, assessed by histology and contrast ultrasonography.

Methods: 22 patients (M/F:14/8) with histologically proven cirrhosis (all HCV RNA+) were recruited from 8 centres across London. Mean age was 49 years (range 29–59); estimated duration of HCV infection was 21.5±7.5 years; 11 patients (50%) were infected with genotype 1, 10 (45.5%) genotype 3, and 1 (4.5%) genotype 4. Patients were treated with 1.5 μg/kg Viraferon-Peg qw plus Ribavirin 800 mg daily for 48 weeks. Virologic response was assessed serially by Roche Amplicor (qualitative), as well as TaqMan PCR (quantitative). A histological assessment was performed before therapy and at TW48 by a single pathologist. Contrast ultrasonography was performed at baseline, TW24 and TW48 by a single radiologist, blinded to the treatment response.

Results: 12/22 (54.5%) patients achieved virologic end-of-treatment response (3 with genotype HCV-1; 8 with HCV-3; 1 with HCV-4). In 3/12 cases ALT remained elevated despite undetectable HCV RNA. Biochemical response was seen in 11/22 (50%) of whom 73% were HCV RNA negative. Mean ALT levels during treatment were significantly lower that at baseline (p = 0.008) for virologic responders. Contrast ultrasonography showed no progression of liver changes in 19/20 cases. Liver histology at TW48 showed no reduction of fibrosis irrespective of treatment outcome. 3 patients (13.6%) dropped out due to IFN-related side effects. 12/19 (63%) cases required dose adjustment due to leukopenia or thrombocytopenia.

Conclusions: In patients with compensated HCV cirrhosis combination treatment with Viraferon-Peg plus ribavirin is effective in achieving viral clearance. To maximise treatment efficacy in this group frequent monitoring and dose adjustments are required.


S.F. Stewart1, A.K. Daly2, Y. Chen2, J.B. Leathart2, B. Haugk1, D.E.J. Jones1, A.D. Burt1, E. Albano3, C.P. Day1.

1 Centre for Liver Research, Medical School, Framlington Place, Newcastle upon Tyne; 2 Dept of Pharmacological Sciences, Medical School, Framlington Place, Newcastle upon Tyne; 3 Dept of Medical Sciences and Medical Clinic University of East Piedmont, Novara, Italy

Background: As yet largely circumstantial evidence suggests that the immune response plays a role in the pathogenesis of alcoholic liver disease (ALD). Support for this hypothesis comes from our work showing that a functional polymorphism of the T-cell surface molecule, CTLA-4, associated with susceptibility to several autoimmune diseases, is also associated with susceptibility to ALD. In vitro studies show that this A-G polymorphism affects the inhibitory function of CTLA-4, and results in augmented immune responses. In this study we sought to determine whether the G allele influenced disease phenotype by scoring the liver biopsies from genotyped patients for the degree of lymphocyte, neutrophil and macrophage infiltrate.

Methods: 20 patients homozygote for the mutant allele (GG) were age and sex matched with 40 patients with AG genotype and 40 with AA genotype. Their biopsies were then scored semi-quantitavely in a totally blinded fashion for lymphocyte, macrophage, and neutrophil infiltration from 0/3 (no infiltration) to 3/3 (extensive infiltration).

Findings: 16 patients had a lymphocyte score of 3/3, 33 had a score of 2/3 and 51 had a score of 1/3. 15/16 (94%) with a score of 3/3 possessed at least one copy of the G allele, versus 26/51 (51%) patients with a lymphocyte score of 1/3 (OR 14.4 (1.75–314), p = 0.006). Neither neutrophil nor macrophage scores associated with the G allele.

Interpretation: These data further support a role for immune mechanisms in the pathogenesis of ALD, and suggest a mechanism whereby possession of the CTLA-4 G allele may predispose to liver injury. The identification of different genetically predetermined phenotypes opens up the possibility of individualised therapy with an improved risk-benefit ratio for this devastating disease.


T. Shonde1, J. Collier2, A. Lorton2, I. Murray Lyon3, A. Freedman4, M. Wiselka5, K. Jack5, C. Hayward6, G.R. Foster1.

1 The Liver Unit, Imperial College Faculty of Medicine at St Mary’s; 2 Dept of Gastroenterology, John Radcliffe Hospital, Oxford; 3 Chelsea and Westminster Hospital, London; 4 University Hospital of Wales, Cardiff; 5 Department of Infection and Tropical Medicie, Leicester; 6 Roche Products Ltd, Welwyn Garden City

Combination therapy with interferon and ribavirin cures only 40% of treated patients and has significant side effects. Therapy with interferon and the antiviral agent amantadine is well tolerated and in a recent meta analysis was shown to be superior to therapy with interferon alone (Mangia et al, unpublished data). To determine whether triple therapy with interferon/ribavirin and amantadine was more effective than dual therapy we conducted a UK wide multi-centre study in which patients with moderate/severe chronic hepatitis C were randomised to receive either interferon plus ribavirin or interferon/ribavirin and amantadine hydrochloride. Untreated patients were randomised to receive dual or triple therapy and virological response at the end of therapy was assessed by HCV RNA determination by PCR.

To date a total of 103 patients have been enrolled and 70 have completed 48 weeks of therapy. End of treatment virological response rates were increased in patients receiving triple therapy (24/39 – 62%) compared with those treated with interferon and ribavirin (13/38 – 48%), although this was not statistically significant.

Hence preliminary data indicate that triple therapy with interferon/ribavirin and amantadine may be superior to therapy with standard combination therapy and the sustained response rates are awaited with interest.


M.A. Morsy1, P.J. Norman2, M. Rela1, N.D. Heaton1, R.W. Vaughan2.

1 Liver Transplant Surgical Service, King’s College Hospital, London; 2 South Thames Tissue Typing Unit, Guy’s Hospital, London

Introduction: Although it is generally accepted that the liver is heavily populated with intrahepatic lymphocytes (IHL) in HCV infection, the differential role of IHL subsets in HCV pathogenesis is unknown. We aim to study the IHL subsets in HCV infection compared with normal liver.

Method: IHL were isolated from liver biopsies collected from 15 donor livers and 14 explanted HCV cirrhotic recipient livers at the time of transplantation, using enzymatic dispersal method optimised in our laboratory. IHL were purified for direct 3-colour FACS analysis, using anti-CD2 and anti-CD56 magnetic microbeads for T and NK/NKT cells respectively. IHL were stained with a panel of monoclonal antibodies for T, NKT, and NK cell phenotyping.

Results: The total numbers of intrahepatic T, NKT, and NK cells were higher in HCV patients than donors. CD4+ and predominantly CD8+ T cell subsets constituted significant high proportions in HCV group compared with donors, 40% versus 27% (p = 0.01) and 56% versus 45% (p = 0.02), respectively. However, there was no significant difference in CD4:CD8 ratio. γδTCR+ cells were significantly higher in HCV patients with 44% compared with 23% in donors (p = 0.02). The majority of T cells expressed CD45RO in both groups. CD45RO was expressed on 92% of CD4+ cells in HCV versus 70% in donor livers (p = 0.001), While CD45RA was expressed on 9% versus 14%, respectively (p = 0.03). NKT cells (CD3+CD56+) constituted 57% of T cells in HCV group compared with 24% in donors (p = 0.02). NK (CD3-CD56+) cells constituted 70% of CD56+ cells in HCV patients versus 44% in donors (p = 0.03). CD158a and CD158b were less expressed on NK cells in HCV patients than donors. CD161 was expressed on 68% versus 55% of NK cells in HCV patients and donors respectively. The cytotoxic effector (P38) was expressed on 89% of NK cells in HCV patients compared with 76% in donors (p = 0.04). CD69+ NK cells constituted 52% in HCV group in comparison with 17% in donors (p = 0.04).

Conclusion: CD4+, CD8+, γδ+, NKT, and NK cells are significantly increased in HCV related end stage liver disease compared with normal liver. These cell subsets may play a role in HCV pathogenesis.


M. A. Morsy1, P. J. Norman2, M. Rela1, N. D. Heaton1, R. W. Vaughan2.

1 Liver Transplant Surgical Service, King’s College Hospital, London; 2 South Thames Tissue Typing Unit, Guy’s Hospital, London

Introduction: The innate and adaptive immune cells are essential host intermediate in a sequence of mechanisms involved in the development of alcoholic liver disease. However, the role of intrahepatic lymphocytes (IHL) in the development of alcohol related chronic liver disease (ALD) is a subject of considerable debate. We aim to study the IHL subsets in human liver with ALD compared with normal liver.

Method: IHL were isolated from liver biopsies collected from 15 donor livers and 12 explanted recipient livers with ALD at the time of transplantation, using enzymatic dispersal method optimised in our laboratory. IHL were purified for direct 3-colour FACS analysis, using anti-CD2 and anti-CD56 magnetic microbeads for T and NK/NKT cells respectively. IHL were stained with a panel of mAb for T, NKT, and NK cell phenotyping.

Results: The numbers of intrahepatic T cells were lower in ALD patients than donors with 46% and 72% of the gated population, respectively. The CD4+ T cells were significantly higher in ALD group compared with donors, 41% versus 27%, respectively ( p= 0.01), while there was no significant difference in CD8+ T cells with 43% versus 45%, respectively. The CD4:CD8 ratio was 1:1.2 for ALD and 1:1.72 for donors (p = 0.03). γδTCR was expressed on 14% of T cells in ALD group compared with 23% in donors (p = 0.03). The majority of T cells were of the “memory” subset (CD45RO) in both groups. However, 86% of CD4+ and 68% CD8+ cells expressed CD45RO in ALD compared with 69% (p = 0.01) and 36% (p = 0.002) in donors respectively. NKT cells (CD3+CD56+) constituted 29% of T cells in ALD group compared with 24% in donors. NK (CD3-CD56+) cells constituted 73% of CD56+ cells in ALD patients versus 46% in donors (p = 0.01). There was no significant difference in the numbers of NK or NKT cells that expressed CD94, NKG2A, CD158a, CD158b, KIRp70, NKB1, CD161, p38, CD16, CD11b, and CD11c. The CD69 activation receptor was expressed on 58% of NK cells in ALD group in comparison with 17% in donors (p = 0.03).

Conclusion: NK and CD4+ T cells are significantly increased in ALD, while the total T cell population and γδ+ T cells are decreased in ALD compared to donor livers. These cell subsets appear to play a role in ALD pathogenesis.


D.A.J. Neal1, M.J. Brown2, I.B. Wilkinson2, G.J.M. Alexander1.

1 Department of Medicine and 2 Clinical Pharmacology, Addenbrooke’s Hospital, Cambridge, UK

Background: Arterial stiffness is an important determinant of cardiovascular risk. It can be measured using the technique of pulse wave analysis, which allows accurate recording of radial artery pressure waveforms and generation of the corresponding central aortic waveform. From this the central aortic pressure and augmentation index (AI) are derived. AI is expressed as a percentage of the central pulse pressure and represents arterial stiffness. Hypertension develops in over 50% of patients after liver transplant. Determining arterial stiffness in hypertensive patients may provide useful information on vascular compliance and also on treatment of hypertension.

Methods: 21 hypertensive liver transplant recipients were commenced on the dihydropyridine calcium-channel blocker amlodipine. 8 patients who were intolerant of or unresponsive to amlodipine were treated with the beta-adrenoceptor antagonist bisoprolol and the angiotensin-converting enzyme inhibitor lisinopril in a cross-over study. Systolic blood pressure (BP) and AI were measured before and after a median of 10 weeks on each drug. No patient received any additional antihypertensive therapy including diuretics.

Results: AI fell from 23.6 ± 2.4% to 16.4 ± 2.3% (p < 0.001) with amlodipine while BP fell from 156 ± 1.8 to 132.6 ± 2 mm Hg (p < 0.001). AI increased from 16.7 ± 4.9% to 25.6 ± 3.4% (p = 0.015) with bisoprolol while BP fell from 158.4 ± 4.9 to 141.9 ± 3.2 mm Hg (p = 0.019). In contrast lisinopril reduced AI from 19.7 ± 1.7% to 12.9 ± 3.1% (p = 0.021) and BP fell from 155.7 ± 2.3 to 133 ± 4.6 mm Hg (p = 0.003). 40% experienced lower limb oedema with amlodipine. Bisoprolol and lisinopril were both well tolerated. Creatinine clearance was unchanged after treatment with lisinopril.

Discussion: Amlodipine significantly reduces AI but is unsuitable as first line treatment of hypertension in that 40% of patients do not tolerate it. Lisinopril, being associated with a reduction in AI and improved vascular compliance, is preferred to bisoprolol for treatment of hypertension after liver transplant.


A. Semper, N. Libri, E. Sanders, W. Rosenberg.

Liver Group, University of Southampton

The ability to clear hepatitis C virus (HCV) infection and the course of chronic hepatitis C (CHC) are both thought to be determined by the nature of the immune response to the virus. Clearance of acute HCV infection is associated with Th1/Tc1 responses while in CHC, T cells are either unresponsive to HCV or exhibit a Th2/Tc2 phenotype. The patterns of antigen specific immune responses are, in part, controlled by dendritic cells (DCs), the body’s professional antigen presenting cells. DCs prime and educate T lymphocytes, determining their functional phenotype, resulting in the induction of effector T cells of Th1/Tc1, Th2/Tc2 phenotypes or non-responsive T cells. In the peripheral blood two main DC populations circulate, namely myeloid and plasmacytoid DCs. These two lineages of DCs exert different influences on the maturation of T cells. Plasmacytoid DCs are thought to be important in anti-viral responses as they are potent producers of interferon-alpha.

Infection of DCs by HCV has been reported, but the means by which infection occurs has not been explained. One known cell-surface receptor for HCV is CD81. We have investigated the expression of CD81 on the cell-surface of circulating peripheral blood dendritic cells (PBDCs). Four colour flow cytometry was used to identify CD81 on plasmacytoid (lineage-, HLA-DR+, CD123+) and myeloid (lineage-, HLA-DR+, CD11c+) PBDC. CD81 expression was consistently present on the majority of myeloid DC in both non-infected (93.4±2.2% and CHC donors (97.0±0.6%). In contrast the frequency of plasmacytoid DC expressing CD81 was markedly lower; and lower on non-infected donors compared to those with CHC (34.3±2.9% v 43.7±6.8%). Differential expression of CD81 on these lineages of PBDC may account for variability in the immune responses in acute and chronic HCV infection and provide opportunities for the modulation of the immune response in CHC.


S. Choudhury1, Mike Hubank2, Humphrey Hodgson1, Clare Selden1.

1 Department of Medicine, Royal Free Campus, Royal Free; University College Medical School, London, UK; 2 Molecular Haematology, Institute of Child Health, London

Background: Stress response proteins are involved in adaptation to adverse conditions, eg hypoxia, but may also be induced as an adaptation to increased cell performance. We have developed an alginate-encapsulation system in which proliferating HepG2 cells develop in 3-dimensional cell spheroids and upregulate expression of hepatocyte-specific functions. Compared with monolayer culture (M) per cell function rises progressively over 8 days, but then diminishes, despite maintained viability and proliferation, by 15 days. This allows analysis of stress response proteins as function improves and then decreases.

Hypothesis: Expression of stress response proteins occurs both as an adaptive response to improved performance and as a result of induced stress, eg hypoxia.

Aim: To investigate expression of stress proteins in monolayer (M), and spheroidal cultures with optimal (day 8), and diminished function (day 15).

Methods: RNA from M and d8 and d15 spheroids was subjected to microarray analysis (U95Av2 Affymetrix gene chip; n = 3 for each condition). Candidate stress proteins were analysed by Western blot and functionally.

Results: Comparing 8d spheroids with M, significant (> 2.5 fold) changes were seen in over 300 genes. Heme oxygenase-1 protein expression at d8 was > 2 × cf. M or d15. MAPKinase proteins activated by hypoxia, MEKK3, MAPKp49, and MKK3b were least abundant in d8 spheroids. A ∼2.5 fold increase of hypoxia-inducible factor protein (HIF 1α) was observed in d15 and M cf. d8. Genes with HIF-1α response elements showed an increase in d15 cf. d8, eg vascular endothelial cell growth factor (3.22 fold) and enolase (3.40 fold). Catalase specific activity (units/min/mg protein) was increased (0.017±0.002 d8 cf. 0.0075±0.001 M). GST activity decreased (0.0041±0.001 at d8 cf. 0.0016±0.0003 at d15).

Conclusion: Stress protein expression with enhanced differentiated function may be an adaptive response to metabolic activity. Changes between d15 and d8 are associated with hypoxia related stress, and contribute to the diminished levels of cell function observed at later time points of 3-D culture.


K.A. Perrow1, J.J. Taylor1, O.F.W. James2, M.F. Bassendine2, D.E.J. Jones2, P.T. Donaldson2.

1 Department of Oral Biology, School of Dental Science; 2 Centre for Liver Research, School of Clinical Medical Sciences, Faculty of Medical Science, University of Newcastle-upon-Tyne NE2 4HH, UK

Genetic studies have not only identified several loci on chromosome 6p but also a number of genes on chromosome 2q as important regions encoding both disease susceptibility and progression in PBC. The caspase 8 gene (CASP8) maps to 2q33-q34. Recent studies indicate that apoptosis is an important process in immune regulation in PBC. Caspase 8 is a key enzyme in the extrinsic apoptotic pathway and the recognition sequence shares significant sequence homology with the critical epitope (IEDTK) of PDC-E2, the major autoantigen in PBC. Making CASP8 an important positional and functional candidate for investigation. The present study investigates the relationship between recently described CASP8 polymorphism1 and PBC. Two single nucleotide polymorphisms (SNPs), tag 1, a C/T SNP in intron 2, and tag 2, a C/G SNP in exon 9 (His>Asp substitution) have been determined in 181 well characterised PBC patients and 195 healthy, geographically and racially matched controls.

There was no significant difference in the phenotype, genotype, or allele distribution comparing patients and controls for the tag 1 SNP (see tables). However, patients were significantly more likely to have the tag 2 G allele compared to controls (OR = 1.93; p = 0.044). These data suggest that either tag 2 contributes to overall genetic susceptibility in PBC or this SNP is in close linkage with an as yet unidentified susceptibility allele on chromosome 2q, candidates include other CASP8 or CASP10 alleles.

Abstract 76, Table 1

Abstract 76, Table 2



G .Pietrosi1, A.E. Marshall1, S.M. Rushbrook1, L.S. Morris2, S. Davies3, N. Coleman2, G. Alexander1.

1 University of Cambridge Department of Medicine; 2 Cancer Cell Unit, Hutchison/MRC Research Centre, Hills Road, Cambridge CB2 2QQ; 3 Histopathology Department

Background: Hepatocytes are highly differentiated cells with a remarkable ability to proliferate in response to acute or chronic injury. Our aim was to evaluate the cell cycle-regulating proteins involved in G1/S phase transition in non viral chronic liver disease.

Methods: We studied the expression of the minichromosome maintenance protein 2 (Mcm-2, a sensitive and specific marker for cell cycle entry), p21 (a cyclin dependent kinase inhibitor), and the cyclins D1 (early G1-phase marker), and A (S-phase marker) using a standard immuno-histochemical technique in liver biopsies from 20 patients with chronic alcoholic-related liver disease (ALD; 13 with steatohepatitis and 7 without), 10 patients with haemochromatosis (HFE) and normal controls.

Results: The normal controls were negative for all markers. Expression of Mcm-2 and p21 was increased in ALD and HFE and the mean percentage of positive hepatocytes was high in all stages of fibrosis. Increased D1 expression was found according to the degree of fibrosis (see table).

Conclusions: These results indicate that in ALD and HFE a significant proportion of hepatocytes have entered cell cycle. Cell cycle progression was observed in the presence of active ALD with steatohepatitis and in HFE with extensive iron deposition. These findings support a direct role for alcohol and iron in cirrhosis and HCC through effects on cell cycle.

Abstract 77


I. Gee, C.J.E. Watson, G.J. Alexander.

Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ

Introduction: Following liver transplantation, long term immunosuppression with tacrolimus and cyclosporin is associated with renal impairment in up to 33% of patients, with renal replacement therapy required in up to 12% of patients. Withdrawal of calcineurin inhibitor therapy often results in improvement in creatinine, but may also predispose to acute rejection. Sirolimus, a new immunosuppressant that does not act by calcineurin-inhibition, is equipotent with cyclosporine and tacrolimus but not nephrotoxic.

Methods: Twenty one liver transplant recipients had their calcineurin inhibitor substituted with sirolimus monotherapy at 6–83 months because of calcineurin inhibitor toxicity, principally nephrotoxicity. 13 (62%) patients continued on sirolimus monotherapy long term but 8 (38%) patients discontinued sirolimus after initiation because of side effects (mouth ulcers, pneumonitis, lethargy, rash, dizziness, blurred vision, pruritus, mucositis). 12/13 (92%) of those continuing sirolimus were converted because of established renal impairment (creatinine range 123–244 μmol/L). Renal function was recorded prior to changing immunosuppression and 12 months after changing.

Results: 8/12 (67%) patients converted for renal impairment who continued sirolimus had a lower serum creatinine 12 months after changing from a calcineurin inhibitor to sirolimus. The mean fall of creatinine was 45 μmol/L (range 27–84).

Conclusion: 67% of patients with calcineurin inhibitor toxicity demonstrated an improvement in serum creatinine when switching to sirolimus. This represents a marked improvement in the renal function of these patients in whom treatment options are limited. The creatinine continues to improve for at least 12 months after withdrawal of the calcineurin inhibitor. There is clearly a concern over the number of patients discontinuing therapy because of side effects although most side effects were reversible on discontinuation.


C. McRae, M.I. Prince, J. Goldblatt, W.L. Craig, D.E.J. Jones.

Centre for Liver Research, University of Newcastle, UK

Fatigue is a disabling symptom affecting a significant proportion of patients with the autoimmune liver disease primary biliary cirrhosis (PBC).1 The severity of fatigue experienced by PBC patients is not related to conventional parameters of disease severity. We have previously demonstrated that the majority of individuals presenting with anti-mitochondrial antibody (AMA) in the context of normal liver function tests (LFT) have histological features of early PBC2 and, over prolonged follow up, typically go on to develop classical cholestatic disease.3 Given the previously demonstrated equal prevalence of fatigue in early and late stage PBC we set out to study the extent to which this novel group of “pre-PBC” patients experience fatigue as a clinical problem. A total of 31 individuals were identified from our comprehensive patient database and from clinical records who were AMA+ve at a titre of ≥1:40 but who had had normal LFTs from the point of original AMA positivity in the absence of any immunomodulatory therapy (including UDCA). Fatigue severity was assessed using the Fatigue Impact Scale (FIS) previously validated for self-completion in PBC. Subjects were matched with age and sex-matched normal community and PBC patient (defined by the presence of AMA in the context of abnormal LFTs and supportive histology) case controls. 25/31 (81%) of the study group returned completed fatigue assessment tools. Mean FIS scores (max possible 160) were significantly higher in the AMA +ve normal LFT patients than in the age and sex matched normal controls (60 ± 39 v 33 ± 30; p =0.01). Mean FIS was actually higher in the study group than in the matched classical PBC controls (48 ± 38), although this did not reach statistical significance.

Conclusions: The association PBC and fatigue extends to patients who are AMA positive but who have normal LFTs. This observation supports the view that these patients have an early form of classical PBC. The possibility of mild PBC-spectrum disease should be born in mind in individuals presenting with fatigue in the wider population in whom liver disease has, in the past, been excluded as an aetiology by the presence of normal LFTs.





C. McRae, M. Prince, O.F.W. James, C.P. Day, M. Hudson, D.E.J. Jones.

Centre for Liver Research, University of Newcastle, Newcastle-upon-Tyne, UK

Chronic cholestasis is frequently complicated by pruritus and related symptoms. The observation that central opioidergic tone is increased in cholestasis has led, in light of the ability of opiates, to induce pruritus in normal individuals, to the rational use of opiate antagonists for the treatment of pruritus and related symptoms in cholestasis. The efficacy of opiate antagonists has been demonstrated in randomised controlled trials, and investigators in the field have strongly advocated that they be used early in the hierarchy of treatments for cholestatic symptoms. Limitation to their use has already been demonstrated in the form of an opiate withdrawal reaction. In this study we describe a further side effect of oral opiate antagonist use in cholestasis, namely pain, and outline a single centre’s experience of the efficacy and tolerability of these agents in practical clinical use. All 14 patients who had been prescribed the opiate antagonist naltrexone for the control of symptoms of cholestasis in our unit were identified from hospital pharmacy records and case notes were reviewed. 4/14 (29%) patients experienced opiate withdrawal reactions. 13/14 (93%) patients reported significant improvement in their cholestatic symptoms. Of the 13 patients continuing the treatment because of favourable response 1 discontinued treatment because of liver transplantation. 7/13 (54%) clinical responders discontinued opiate antagonist therapy because of side effects. In 3/13 cases the symptom precipitating treatment withdrawal was pain, in each case resulting from exascerbation of a previously well-controlled painful condition (post-herpetic neuralgia, gout, and liver capsular pain). The other side effects necessitating treatment withdrawal were opiate withdrawal reaction (n = 2), GI disturbance (n = 1), and neuro-psychiatric effects (n = 1). All withdrawals occurred within the first 26 weeks of therapy.

Conclusions: Only a minority of cholestatic patients were able to tolerate long term opiate antagonist therapy for symptom control. Pain, presumably resulting from blockade of natural opiate algesic effect, was the single commonest reason for discontinuation of treatment. Although undoubtedly efficacious in cholestasis symptom control, the usefulness of opiate antagonist therapy will, in practice, be limited by their side effects.


C.L. Russell, D.H. Adams, S.C. Afford.

Liver Research Laboratories, University of Birmingham, Birmingham UK

Background: Human intrahepatic endothelial cells (HIEC) and cholangiocytes (cholangios) increase expression of adhesion molecules that are involved in the recruitment and retention of leukocytes at inflammatory sites within the liver. Some evidence exists to suggest that the response to proinflammatory cytokines (PIC) varies depending on the cell type. However, this has not been addressed in depth with primary cells. Insights into such cell specific responses to PIC is critical for our understanding of how the inflammatory response is orchestrated in the liver. In the present study we have investigated expression of E-selectin (HIEC) and ICAM-1 and VCAM-1 (HIEC and cholangios) in response to a panel of PIC secreted by activated T cells and macrophages.

Methodology: Primary cultures of HIEC or cholangios were incubated with optimal concentrations of TNFα, CD40L, IL–17, or IFNγ either alone or in various combinations for up to 6 hours. Cells were then analysed for expression of E-selectin, ICAM-1, and VCAM-1 using cell based ELISA or flow cytometry.

Results: The table indicates which PIC or combination induced statistically significant increases in adhesion molecule expression. The effects of cytokine combinations was additive rather than synergistic.

Abstract 81

Conclusions: All adhesion molecules were induced in both cell types by TNFα, although HIEC were more sensitive to TNFα than cholangios. There were marked differences between cell types in their responses to individual PIC or specific PIC combinations. The responses to PIC appeared to be cell rather than adhesion molecule specific. These data have implications for regulation of liver inflammatory responses.


G.A. Clarke.

University of Leicester Hospitals NHS Trust

Background: We have previously demonstrated that inhibition of oxidosqualene-lanosterol cyclase by Ro 48–8071 up-regulates the secretion of all major biliary lipids in C57L/J mice that possess mutant Lith alleles. It has also been shown that inhibition of 3 hydroxymethylglutarate-Co A reductase (HMGR) upregulates phosphatidyl choline (PC) and cholesterol (Ch) secretion, without significant changes in bile salt (BS) secretion. As secretion of PC and Ch is not wholly independent of bile salts (BS) and ABCB11 (the bile salt export pump) is a candidate Lith gene, a possibility exists that this phenomenon might be artefact of mutant Lith alleles. AKR/J mice lack mutant Lith alleles and have been previously employed as controls in studies involving C57L/J mice.

Aims: To determine biliary lipid secretion in AKR/J mice treated with Ro 48–8071, simvastatin ( a HMGR inhibitor), and controls.

Methods: Male, 8 week old AKR mice were acclimatised for 2 weeks. Group Ro was administered. Ro 48–8071 30 mg/kg-1/day-1 (n = 5). Group Simva was administered simvastatin 10 mg/kg-1/day-1 (n = 3) and controls (n = 6) fed diet alone. At 2 weeks, animals were anaesthetised and an acute biliary fistula was fashioned. Bile was collected for 1 hour. Biliary lipids composition was analysed using standard techniques and biliary BS, PC, and Ch secretion rates were calculated.

Results: See table.

Conclusions: Both simvastatin and Ro 48–8071 up-regulate PC secretion rates in a statiscally significant manner. Ch secretion rates are also increased. In contrast to C57L mice possesing mutant Lith alleles, simvastatin increases but Ro 48–8071 decreases BS secretion, although these results do not achieve statistical significance. We conclude that inhibitor mediated up-regulation of PC secretion is not influenced by mutant Lith alleles.

Abstract 82


G.J.M. Webster1,2, S. Reignat1, M. Lascar1, D.Brown2, G.S. Ogg3, G. Dusheiko2, R. Williams1, A. Bertoletti1.

1 Institute of Hepatology, 2 Centre for Hepatology, Royal Free and University College Medical School, London; 3 Institute of Molecular Medicine, Oxford

Background: The pathogenesis of liver damage in HBV infection has been thought to be directly mediated by virus-specific CD8 cells. This belief has also led to the assumption that acute flares in chronic hepatitis B patients are caused by virus-specific T cell response.

Methods: We longitudinally analysed T cell response, HBV-DNA, and ALT values in 3 patients with chronic hepatitis B during 5 separate episodes of ‘hepatic flare’. Core, envelope, and polymerase CD8 cells were studied directly ex vivo and after in vitro expansion using HLA-tetramers and intracellular cytokines. T cell proliferation was tested against nucleocapsid and envelope proteins.

Results: Peaks of ALT (> 1000 U/L ) were associated with either sharp decline or increase of HBV-DNA. No significant variations in direct ex vivo frequency of circulating HBV-specific CD8 cells were found before, during, or after the ALT peak, either with tetramers or intracellular cytokines. Nevertheless, HBV-specific CD8 cells were present in these patients, as core 18–27 specific CD8 cells were found in the liver of one patient, and envelope and polymerase-specific CD8 cells in the circulation of all 3. These HBV-specific CD8 cells appeared completely indifferent to the fluctuations of HBV-DNA and ALT values, because no changes in direct frequency or expansion capacity were linked to the episodes of flare. Similar results were obtained with T cell proliferation analysis. No signs of T cell proliferative response were associated with acute flares. A significant core-specific T cell response was found in one patient only after HBeAg seroconversion and associated with HBV-DNA values < 106 copies/ml.

Conclusions: We were unable to find evidence of “HBV-specific T cell recovery” associated with episodes of acute flares during chronic hepatitis. The apparent indifference of the adaptive immune response during these episodes suggests that other factors, including activation of innate components of the immune system (NK and NK-T cells) require careful investigation.


M. Wright1, R. Goldin, A. Fabre, C. Trepo, H. Thomas, M. Thursz, on behalf of the HENCORE Collaboration.

Hepatology Section, Faculty of Medicine, Imperial College, London W2 1NY, UK (HENCORE is the Hepatitis C European Network for Cooperative Research)

Introduction: The rate of development of liver fibrosis in HCV infection varies between individuals, accounting for the variation in duration of progression to cirrhosis. The aims of this study were: (a) to determine whether fibrosis progresses linearly through the grading scales; and (b) to identify factors that influence the rate of fibrosis.

Methods: HCV infected patients who had had a liver biopsy were identified. Biopsies were scored using the modified HAI (Ishak) and METAVIR systems, which were compared. Patients were treatment naive at first biopsy. Fibrosis rate was defined as: fibrosis stage/infection duration. Demographic features were examined for relationship to fibrosis rate using univariate and multivariate analysis. A subgroup of patients with 2 biopsies was examined to test the assumption that fibrosis progresses in a linear fashion.

Results: 917 patients were included. Male gender (p = 0.0001), older age at infection (p = < 0.0001), and viral genotype 1 (p = 0.005) were all associated with a rapid rate of fibrosis. On multiple linear regression they accounted for 29.5% of the variability in fibrosis rate (r2 = 0.295). METAVIR and Ishak scores were highly correlated (r = 0.935, p < 0.0001) with similar reproducibility (κ = 0.79 Ishak, 0.93 METAVIR). In 137 patients who had 2 biopsies the predicted probability for an increase of 1 on the fibrosis score was too low to assess linearity.

Conclusions: Demographic features account for a minority of fibrosis rate variability. The Ishak and METAVIR scoring systems are equivalent. Linearity of fibrosis progression cannot be assessed in biopsies only a few years apart.


R.R. Mitry, R.D. Hughes, C. Terry, S. Lehec, G. Mieli-Vergani, P. Muiesan, M. Rela, N.D. Heaton, A. Dhawan.

Institute of Liver Studies, King’s College Hospital, London, UK

Hepatocyte transplantation (HTx) may be an alternative mode of treatment to orthotopic liver transplantation for patients with acute liver failure (ALF) and liver-based metabolic diseases. Isolation of high quality human hepatocytes is essential for clinical application of HTx. Toxic substances accumulating in the circulation of patients with ALF may be detrimental to hepatocyte function after cell transplantation.

Methods: Twenty hepatocyte isolation procedures were performed with donor tissue not used for orthotopic transplantation, using a collagenase perfusion technique (Strom et al). Isolated cells were either plated, or cryopreserved in UW solution/10% DMSO. Collagen-coated culture plates were seeded with viable fresh or thawed cells, and incubated for 24 h. Serum was obtained from 12 patients with ALF due to paracetamol overdose and 6 normal controls (NS). The culture medium was replaced with fresh medium containing pooled or individual sera at concentrations ranging from 0–80%. Plates were incubated for 24 h, and the cell activity/function was analysed using MTT, [14C]leucine incorporation, and cytochrome P-450 (CYP1A1/2) using EROD assays.

Results: Mean cell yield was 3.6×109±SD5.3×109 cells, with a cell viability of 56.9±24.7%. With both fresh and cryopreserved hepatocytes and serum concentrations of up to 80%, there was no difference in the effects on overall metabolic activity (MTT), [14C]leucine incorporation, or CYP1A1/2 activity with pooled ALF serum as compared to pooled NS. With the individual samples, there was no change in the overall metabolic activity (MTT) of hepatocytes with ALF serum (0.14±0.03; n = 12) compared with NS (0.14±0.02; n = 6). There was a significant 22% increase in CYP1A1/2 activity (14.6±0.40 pmol resorufin/min to 17.8±0.02 pmol resorufin/min; p = 0.014) and 25% increase in [14C]leucine incorporation (774±159 cpm to 969±100 cpm; p = 0.009), with ALF serum as compared with NS.

Conclusions: The findings of this study suggests that the metabolic function of both fresh and cryopreserved hepatocytes will not be impaired on transplantation into patients with ALF due to paracetamol overdose.


S.I. Khakoo, C.R. Brooks, M. Howell1, G.J. Alexander2, W.M. Rosenberg.

Liver Unit and 1 Dept of Immunology, Southampton General Hospital, Tremona Rd, Southampton, UK; 2 Dept of Medicine, Addenbrooke’s NHS Trust, Cambridge, UK

Introduction: NK cells are important for the innate immune response to viruses. Their functions are controlled by inhibitory and activating cell surface receptors. One family of these receptors are the killer cell Ig-like receptors (KIR). Different individuals have different numbers of KIR. Most of this diversity derives from the activating receptors, which initiate cytotoxicity. Individuals can have between 1 and 6 activating KIR, and up to 7 inhibitory KIR. The aim of this project was to determine the role of this diversity in the outcome of hepatitis C virus infection.

Methods: The KIR receptor genotype was determined by PCR-SSP in 58 individuals with resolved HCV infection (anti-HCV positive, PCR negative), 131 individuals with chronic HCV infection (anti-HCV positive, PCR positive) and 91 healthy unexposed controls.

Results: Overall similar numbers of KIR were found in all three groups. The mean number of activating KIR was 2.4+1.3, 2.6+1.6, 2.5+1.4 in the resolvers, chronics, and unexposed groups, respectively. The frequencies of KIR were similar in all populations, with the exception of KIR2DS3, which was found in resolvers at less than half the frequency of the chronically infected (X2 = 5.4, p = 0.12 (with Bonferoni correction), OR = 2.92 (95% CI 1.08 to 8.32)).

Conclusion: Our study does not suggest a role for most of the KIR loci in determining the outcome of HCV infection. Further studies of KIR2DS3 and KIR allelic polymorphism may be informative.

Abstract 86


S.S. Nijjar, S. Kumar, H. Schrewe, A.J. Strain.

School of Biosciences, University of Birmingham, Birmingham, UK

Background: Activins are members of the transforming growth factor-β (TGF-β) superfamily, which act as growth and differentiation factors in many cells and tissues. Biologically active forms of the activins consist of homo- and heterodimers of two closely related inhibin β subunits (βA and/or βB). Other members of this family include the recently identified inhibin βC and βE subunits. The activins mediate their biological function through transmembrane receptors that consist of heteromeric complexes of type I (A and B) and type II (A and B) receptors. Recent reports have shown that activin A acts as an autocrine inhibitor of liver DNA synthesis and can induce apoptosis in rat hepatocytes in vitro as well as in vivo.

Aim: The aim of this study was to investigate possible autocrine or paracrine signaling mechanisms mediated by these growth factors.

Methods: Using RT-PCR we examined in detail the cellular localisation of activin subunit and receptor mRNA in purified liver cell isolates. Semiquantitative RT-PCR using total RNA isolated from whole tissue, was used to determine whether activin subunit mRNA expression was perturbed in disease. The localisation of inhibin βC protein in liver tissue sections was also investigated by immunohistochemistry.

Results: Inhibin βA transcripts were detectable in all biliary epithelial (BEC), liver endothelial (LEC), and hepatocyte cell isolates. Inhibin βB expression was seen in occasional LEC and hepatocyte isolates, but never in BEC. However, inhibin βC and βE mRNA displayed a much more restricted pattern of expression, only being detectable in hepatocytes. The expression of activin type IA, IB and IIB receptors was found to be ubiquitous, whereas transcripts for the activin type IIA receptor were only seen in some BEC. Inhibin βA and βC expression did not display any discernable differences in the levels of expression between normal and diseased liver. The expression of inhibin βB was detectable in one of five normal livers, however expression appeared to be elevated in all liver tissue with primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). High levels of inhibin βE mRNA were detectable in normal liver tissue, however this expression was dramatically reduced in liver tissue with PBC and PSC. In normal liver tissue, weak inhibin βC subunit immunoreactivity was localised to bile ducts, the endothelium of hepatic arteries and portal veins, hepatocytes, and the sinusoids. However, in PBC strong immunoreactivity was present on bile ductules, small neovessels in portal regions, and the sinusoidal endothelium. A similar pattern of expression was seen in liver tissue with PSC or alcoholic liver disease, although the staining was less intense.

Conclusion: Taken together these data suggest that activin mediated signaling plays a significant role in the regulation of cell function in normal adult liver as well as during disease.


M. Metzner1, P. Maulayah2, C. Benshemesh3, M.E. Cramp1,2.

1 Plymouth Postgraduate Medical School, Derriford Hospital, Plymouth, UK; 2 Institute of Liver Studies, King’s College Hospital, London, UK; 3 Gateway Clinic, Lambeth NHS Healthcare Trust, Landor Rd, London SW9, UK

Background: Conventional, allopathic treatment for hepatitis C virus (HCV) infection is associated with significant side effects and is contra-indicated, ineffective, or not needed in a large number of cases. Many patients turn to alternative treatments for their disease and anecdotally claim significant benefit from their use, although to date there is little clinical evidence of efficacy.

Methods: Patients with documented chronic hepatitis C were randomly allocated to receive either traditional Chinese medicine (in the form of a herbal infusion) taken daily or no treatment for 6 months. Patients were assessed at baseline and after 3 and 6 months at King’s College Hospital in London. Treatment response was monitored by measurement of liver function tests, urea, creatinine, full blood count, and viral load at 0, 3, and 6 months. Symptomatic response was assessed by means of standard questionnaires measuring quality of life, fatigue anxiety, and depression completed at similar time points.

Results: 28 patients were entered into the study, 15 received treatment and 13 did not. The groups were matched for age, sex, duration of illness, previous interferon therapy (interferon naive), and alcohol intake. 10 patients in each group completed all follow up visits. No significant reduction in ALT and AST levels or viral load was seen with the use of traditional Chinese medicine in patients with HCV infection over the 6 month study period. No adverse reaction to the herbal preparations used was noted. There was a trend to improvement in symptoms with 5 of the 10 treated cases showing substantial falls in fatigue and anxiety scores.

Conclusion: Traditional Chinese medicinal herbs were well tolerated and improved symptoms in some patients with chronic HCV infection. This symptomatic benefit does not appear to be associated with improvements in either liver biochemistry or viral load. This pilot study confirms the need for a large randomised trial to determine the benefit of traditional chinese medicine in the symptomatic treatment of chronic HCV infection.


M. Hussain1, B. Portmann2, G. Mieli-Vergani2, D. Vergani1.

1 Institute of Hepatology, University College London; 2 Institute of Liver Studies, King’s College Hospital, London, UK

Background: The type of cells comprising interface hepatitis in autoimmune hepatitis (AIH) has been investigated on frozen tissue using a limited number of monoclonal antibodies.

Aim: (1) To establish the conditions for the detection of the following cell markers on deparaffinized liver tissue: CD4 (helper/inducer), CD8 (suppressor/cytotoxic), perforin, and granzyme B (markers of cytotoxicity present on CD8 and NK cells), CD79a (B cell receptor), and CD138 (plasma cells). (2) To analyse the cell composition of interface hepatitis.

Patients: Fifteen consecutive biopsies from children with AIH (9 AIH-1, median age 11.2 years and 6 AIH-2, median age 7.2 years) were investigated.

Method: Antigen retrieval was obtained through the combined use of high temperature (600W microwave oven for 25 minutes at full power) and appropriate buffers, Tris/EDTA (pH 9.0) being used for CD8, granzyme B, perforin, CD79a, and CD138 cells and 1mM EDTA (pH 8.0) for CD4 cells. The primary antibody was applied for one hour at room temperature and visualised with a polymeric conjugate (EnVision Dako).

Results: Using this novel approach, we found that within interface hepatitis CD8 T cells are expressed in equal or higher numbers than CD4 T cells; CD8 T cells are numerous also in the lobules where they express the cytolytic marker perforin; the percentage of perforin positive cells, far outnumbering granzyme B positive cells, correlates significantly with the levels of transaminases; cells of the B linage expressing the CD79 markers are present within the periportal infiltrate at high density and co-localise at the periphery of the infiltrate with the less frequent CD138 positive plasma cells. In AIH type 1, the number of plasma cells and B cells are significantly correlated with the levels of anti-nuclear and anti-smooth muscle antibodies.

Conclusion: The correlation between the biochemical extent of liver damage and the number of cytotoxic perforin positive cells suggests a pathogenic role for these cells.


P. Collins, S. Choudhury, M. Hubank1, C. Selden, H. Hodgson.

Centre for Hepatology, Department of Medicine, Royal Free Campus, Royal Free and University College Medical School, Rowland Hill Street, Hampstead, London, UK; 1 Molecular Haematology, Institute of Child Health, London, UK

Background: Mitochondrial uncoupling proteins (UCPs) are a family of proteins whose properties include the ability to collapse the mitochondrial proton gradient and uncouple respiration from ATP synthesis. Five UCPs have so far been discovered, although the exact cellular function of most of them remains obscure. Recent data have suggested that UCP gene expression may be up-regulated in response to increased reactive oxygen species (ROS) generated in mitochondria. Increases in mouse liver UCP2 mRNA are seen in chronic obesity, and after chronic alcohol ingestion—both conditions associated with increases in oxidative stress. It has been suggested that UCPs may function in these situations as endogenous antioxidants within the liver and protecting against cellular damage and preventing apoptosis.

Aim: To investigate changes in UCP mRNA levels and ROS generation in a hepatocyte cell line grown in monolayer and encapsulated within a 3-D alginate bead culture system.

Methods: RNA microarray and semi-quantitative RT-PCR was used to detect changes in UCP1–5 mRNA levels in HepG2 cells grown in monolayer (M) and in alginate beads at a time of maximal cell synthetic performance (day 8) and at a time when synthesis and ATP levels decline (day 15). ROS generation was estimated by incubation with 2,7-dichlorofluorescin diacetate followed by visualisation with confocal laser microscopy.

Results: UCP 1, 3 and 4 mRNA were not detected in any HepG2 cells studied. UCP-2 and -5 mRNA levels were both increased over twofold in cells cultured in alginate beads. UCP-5 expression continued to increase with time (d15 compared with d8). ROS associated fluorescence (measured on an analogue scale per cell) was visibly increased in 3-D cultures compared to monolayer.

Conclusions: UCP-2 and -5 mRNA levels are increased in cells that are producing increased amounts of ROS. UCP up-regulation may be a part of the cellular response to stress and be an attempt to limit ROS mediated damage. Long term UCP over-activity could, however, contribute to the decrease in ATP and synthetic function observed in our culture system after 15 days.


S.M. Rushbrook, T. Woodall, G.J. Alexander.

University of Cambridge, School of Clinical Medicine, Level 5, Box 157, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK

Background: Disorders of Th1 immune responses are reported in chronic hepatitis C infection (HCV), an impairment of dendritic cell responses has also been characterised. 1–25 (OH) 2 vitamin D has been shown to impair dendritic cell and Th1 immune responses in vitro. The purpose of this study was to measure vitamin D levels in patients with hepatitis C with active viral replication at the various stages of liver fibrosis.

Methods: Patients with hepatitis C virus infection were selected if they were hepatitis C mRNA positive. 111 patients who fullfilled these criteria were subdivided on the basis of their liver biopsy fibrosis stage: stage 0 (n = 20), stage 1 (n = 5), stage 2 (n = 16), stage 3 (n = 31), stage 4 (n = 11), stage 5 (n = 28). A 25-hydroxylated vitamin D level was then measured on a fasting blood sample. All samples were collected between Jan and April 2002. These values were compared with a standard laboratory reference range, for both the same time of year and geographical area (n = 657).

Results: For the time period stated above and the geographical area, the mean vitamin D level for a control population was 21 ug/L with a range of 9–47 ug/L. See table.

Abstract 91

Conclusions: Throughout the spectrum of fibrosis stage in hepatitis C induced liver disease there is a significant reduction in the level of 25OH vitamin D. Although malnutrition in this susceptible population is one possible explanation for the results shown, we speculate that vitamin D might be utilised intrahepatically, to suppress effective immune responses in the liver and potentate viral replication.


L. Cheshire, M.E. Orme1, A. van Angen, R Williams2, R. Jalan2.

1 The Lewin Group (Quintiles (UK) Ltd), Station House, Bracknell, UK; 2 Institute of Hepatology, Royal Free and University College Medical School, London, UK.

Background and Aims: Variceal haemorrhage occurs in about 40% of patients with cirrhosis within 2 years of diagnosis. Despite substantial improvements in management, mortality from an episode of bleeding remains high and treatment of the patient with variceal bleeding is resource intensive. At present, there are no comprehensive data about the healthcare cost of managing a cirrhotic patient who presents with a bleed in the UK. The aim of this pharmacoeconomic study was to develop a decision–analytical model that could be used to evaluate the economic consequences of current practice in the UK from the perspective of the NHS, which would allow assessment of the average treatment costs and cost effectiveness, and to identify the main cost drivers. The model was developed to cover the first 42 days.

Methods: The model was designed to assess healthcare utilisation of a variceal bleed; the treatments used were based on the UK guidelines (Gut 2000;46:Suppl 3:III1–III15); definition of events was based upon the Baveno II criteria; information on possible clinical outcomes were based upon literature review and trial data, and physician interviews (assuming 50% Child class patients, it was estimated that 36% would be free of bleeding, 60% will have initially controlled bleeding and 59% will survive the initial period). Costs were collected from a bottom-up approach—individual items identified and appropriate costs applied (BNF; ‘Unit costs of health and social care’ published by the University of Kent (PSSRU, 2000)).

Results: For this period of time, the model estimates that the total cost of treatment is 15 595. 60% of costs are due to hospital admissions, the average length of stay being 14 days. 15% of costs are the cost of disposables, such as blood products. The average cost of a patient with no further bleeding was 42 666; initially controlled bleeding was 25 600; and the average cost of a patient surviving 42 days was 28 468.

Conclusions: The key cost driver is the length of hospital stay and any intervention that leads to earlier discharge and reduces blood product usage would have an impact upon the costs.


R.Harry, D. Grieve1, M. Bowles, N. Heaton, P. Mueisan, M. Rela, J. Wendon, A. Shah1.

Institute of Liver Studies and 1 Department of Cardiology, King’s College Hospital, London, SE13 5DD; UK

Introduction: Peripheral vasodilatation and a hyperdynamic circulation accompany the development of human cirrhosis (Sherlock, 1958). This is associated with hyporesponsiveness to α-vasoconstrictors in vivo (Trewby, 1977) and in the hepatic artery studied in vitro (Smith, 1998). The same cardiovascular changes occur in acute liver failure (ALF) (Wendon, 1992) but the response to vasoconstrictors has not been studied in vitro.

Methods: Endothelial-denuded rings of hepatic artery from patients undergoing orthotopic liver transplant for ALF (n = 6) and from organ donors and patients undergoing hepatic resection (controls: n = 6) were studied in an organ bath. The response to 80 mmol/L potassium chloride was assessed. After washing, rings were incubated with either 0.1 mM L-NMMA (nitric oxide synthase inhibitor), 0.1 mM ZnPP (heme oxygenase inhibitor), or vehicle control for 30 min. Cumulative dose response curves to phenylephrine (PHE) were constructed.

Results: There was no difference in the maximal response to phenylephrine between the two patient groups (ALF median 1.34 g/mg tissue (IQR 1.08–1.6)); controls 1.05 g/mg tissue (1–1.4)). Incubation with L-NMMA and ZnPP did not affect the maximal response to PHE in either patients or controls. There was no correlation between bilirubin, AST, albumin, lactate, cardiac index, or SVRI and the response to phenylephrine, but there was a significant correlation between INR and maximal tension in response to PHE (Spearman rank rho 0.89; p < 0.05).

Conclusion: Vascular reactivity to phenylephrine in the endothelial denuded hepatic artery is unchanged in human ALF and unaffected by nitric oxide and carbon monoxide. Hepatic artery hyporesponsiveness to vasoconstrictors may not underlie the cardiovascular collapse seen in human ALF.


S. Nielsen, M.F. Bassendine, A. Burt, G.L. Toms.

The Centre for Liver Research, The University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, UK

Serum and liver macerates from a patient with common variable immunodeficiency and chronic HCV infection six weeks post-transplant were analysed by isopycnic centrifugation on isotonic iodixanol (optiprep) density gradients. All of the HCV RNA fractionated at the top of the gradient, in fractions of density of < 1.13 g/ml and was precipitable with manganese chloride and heparin (Mn/Hep) indicating that it is all associated with host beta-lipoprotein. Treatment with desoxycholate released putative hepatitis C virions with a density of 1.13 g/ml and treatment with 0.18% NP40 released putative virus cores with a density of 1.21 g/ml and a sedimentation coefficient of 150S. Northern blotting of Mn/Hep precipitates revealed a single band of HCV RNA of 9.4 kb corresponding in size to the full HCV genome. SDS-polyacrylamide gel electrophoresis and western blotting with monoclonal antibodies revealed a single 20 kDa band of core protein and a single 31 kDa E1 band reduced to 20 kDa after deglycosylation with endoglycosidase F. Both core and E1 bands co-migrated with corresponding bands derived from a recombinant vaccinia virus system expressing the HCV structural protein genes. Anti-E2 Mabs blotted a single 62 kDa band from Mn/Hep precipitates, approximately 7 kDa smaller than the anti-E2 staining band in the recombinant vaccinia virus system. Following deglycosylation, E2 glycoprotein ran with an apparent molecular weight of 36 kDa, co-migrating with E2, which forms a minor band in the deglycosylated vaccinia virus recombinant. No band equivalent to the major 40 kDa E2/P7 band in the deglycosylated vaccinia virus system was observed in the Mn/Hep precipitates, indicating that E2/P7 is not a structural protein. Under reducing conditions both E1 and E2 ran as monomers, suggesting that the two glycoproteins are not disulphide linked in the virion.


I. Kyrlagkitsis1, M. Metzner2, H. Smith1, E. Depla3, E. Sablon3, M.E. Cramp1,4.

1 Institute of Liver studies, King’s College Hospital London, UK; 2 Plymouth Postgraduate Medical School, Derriford Hospital, Plymouth UK; 3 Hepatitis Program, Innogenetics NV Gent, Belgium

Background: HCV RNA negativity in blood in anti-HCV positive patients may indicate viral clearance despite the presence of viral RNA in the liver of some cases. The time course of the humoral response to HCV in those who apparently clear viraemia is not known. We have measured specific antibodies titers against HCV in sequential blood samples from anti-HCV positive individuals with persistently negative HCV RNA.

Patients/Methods: We included all outpatients seen over a period of 6 years who tested positive for HCV antibodies but persistently negative for HCV RNA and who had been followed up for more than 2 years. All patients had spontaneously cleared detectable viraemia. Anti-HCV was tested by third generation ELISA and confirmed by recombinant immunoblot assay with HCV RNA tested by polymerase chain reaction (Roche Amplicor). Titres for specific HCV antibodies (core, E1, E2, NS3, NS4, NS5) were measured using a research line immunoassay (INNO-LIA IV, prototype version, Innogenetics) in sequential sera stored from the patients’ annual follow up.

Results: We identified 46 patients with a mean follow up period of 3.2 years (range 2–7). A statistically significant drop in core and E1 titers was noticed between the beginning and the end of follow up (p = 0.001 and p = 0.024, respectively) but not in the rest of the antibodies (p = ns). The longer the follow up the more substantial the drop in core and E1 antibodies. Fifteen patients demonstrated a decrease in antibodies to all HCV regions while in the rest titres either increased (4 patients) or were erratic (27 patients).

Conclusions: Overall, HCV antibody positive individuals without detectable HCV RNA demonstrate a slow but significant drop in core and E1 antibody titres. Most cases had preserved antibody titres over a period of several years, raising the possibility that they have continuing priming of their humoral immune response by undetectable residual infection. In contrast, a sub-group had a falling titre of antibodies to all HCV regions suggesting complete clearance of the infection with gradual loss of humoral response over time.


M. Hewett, M.I. Prince, W.L. Craig, B. Davidson, M.L. Schmid, M.F. Bassendine.

Liver Unit, Freeman Hospital, High Heaton, Newcastle NE7 7DN, UK

Background: Weight loss is frequently recorded in trials of α-IFN for chronic HCV. However, little is known about the impact of this in practice.

Patients and Methods: We performed a retrospective study of 100 randomly selected patients receiving α-IFN therapy for a minimum duration of 12 weeks for chronic HCV. Data were extracted by review of hospital notes and analysed by Spearman’s correlation coefficient and Mann-Whitney U test.

Results: 60 patients were male. Median age was 30 years. 53 patients received α-IFN monotherapy (doses 3MU tiw (n = 28), 4.5MU tiw (n = 15), higher dose (n = 8), pegylated IFN (n = 2)). 47 received combination therapy (α-IFN 4.5MU tiw and amantadine 100 mg bd (n = 9), α-IFN 3MU tiw and 400–600 mg ribavirin bd (n = 38)). 53 patients had been IV drug users (IVDU). 16 were taking methadone. Full clinical records and weights were available for 98 patients. The median (IQR) initial weight and body mass index (BMI) were 72.5 kg (65.3–82.2) and 25.3 (22.8–28.0). 87 patients (89%) lost weight during therapy. Median weight loss during therapy was 3.2 kg (1.7–5.2). This was 4.4% (2.5–6.8) of initial weight. Maximum weight loss was 12 kg. Percentage weight loss was not related to initial BMI (r = −0.02, p = 0.87). Median percentage weight loss was not related to sex (M 4.0%, F 4.9%, p = 0.50), social status (live alone 4.4%, live with partner 4.4%, p = 0.67), dose of α-IFN (3 MU 4.4%, > 3 MU 4.4%, p = 0.78), ribavirin (+R 4.4%, −R 4.4%, p = 0.86) or amantadine use (+A 8.9%, −A 4.1%, p = 0.10). Weight loss was not higher overall in IVDU (IDVU 5.0%, other 3.7%, p = 0.21) but was higher in those on methadone (+M 8.3%, −M 4.0%, p = 0.001). Patients taking methadone initially weighed less (BMI +M 23.5, −M 25.8, p = 0.04). 26 patients received less than 6 months, 26 received less than 12 months and 46 completed 12 months of therapy. No patient stopped therapy due to weight loss. 15 patients received formal dietary evaluation and advice.

Conclusions: Weight loss is very common with α-IFN therapy in HCV. It may be exacerbated by the anorexic effects of methadone. Further studies are needed to evaluate therapeutic interventions for this.


J. Gotto1,2, G.J.M. Webster1,2, S. Reignat2, E. Barnes1, J. Jenkins3, D. Mesogiti3, D. Brown1, A. Bertoletti2, G.M. Dusheiko1.

1 Royal Free and 2 University College Medical School, London; 3 GlaxoSmithKline, Middlesex

Background: Pre-treatment ALT level strongly predicts response to lamivudine (LAM), with 64% of patients HBeAg seroconverting when ALT > 5 × ULN, but 5% when ALT < 2 × ULN. A transaminase flare thought to be due to immune rebound frequently follows withdrawal of prednisolone (PRED). This may be of therapeutic benefit for those with normal or minimally elevated ALT when followed by LAM.

Methods: 15 adult non-cirrhotic patients with highly replicating chronic Hep B (HBeAg +ve) and serum ALT < 2 × ULN were given sequentially, each for 2/52; PRED 30 mg/day, PRED 15 mg/day, a withdrawal period then LAM 100 mg/day for 52/52. The HBV DNA, HBV serology and ALT were measured longitudinally. The HBV-specific CD8 response was analysed in the 7 HLA-A2 patients concurrently using HLA-tetramers and intracellular cytokine staining directly ex vivo and after in vitro stimulation. Core, envelope, and polymerase specificities were tested.

Results: HBV DNA was not affected by PRED but decreased significantly with LAM. Only 1/15 patients had an ALT flare > 200 IU/ml (249 IU/ml) and only 2/15 patients lost HBeAg. Despite the lack of flare induction 5/7 HLA-A2 +ve patients demonstrated an HBV-specific CD8 response following therapy. Recovery of the antigen specific CD8 T-cell response was mostly present after 16–20 weeks of therapy. Different peptides were recognised at different time-points and none was recognised constantly in any patient. There was no difference in peptide response between patients with a baseline ALT < 40 IU/ml compared with ALT 40–80 IU/ml. Interestingly, the majority, and the strongest, responses were to envelope antigens, particularly the subdominant envelope 183–191 epitope (seen in 4 patients).

Conclusions: Prednisolone withdrawal did not induce ALT flares in patients with normal or minimally raised baseline ALT, and is unlikely to enhance HBeAg seroconversion rates on lamivudine therapy in these patients. Despite the lack of ALT flares, an HBV-specific CD8 response, particularly to envelope antigens, was seen in 5/7 HLA-A2+ve patients.


K.T.T. Fung1, F.T.W. Li1, M.L. Raimondo1, D. Maugil2, J. Tibballs2, A.F. Watkinson2, A.K. Burroughs1.

1 Liver Transplantation and Hepatobiliary Unit, 2 Department of Radiology, The Royal Free Hospital, Pond Street, London, UK

Aims: To review systematically the evidence currently available for determining the best radiological imaging for characterising hepatocellular carcinoma (HCC) when it is suspected to have developed in a cirrhotic patient by routine ultrasonography (US).

Methods: To include prospective cohorts with either prospective or retrospective review of films by radiologists, and retrospective cohorts. The studies were classified according to whether cirrhotic patients alone constituted the whole study population, whether hepatocellular carcinoma per se or various kinds of focal hepatic lesion were the focus of interest, and the reference standard used for the diagnosis of HCC, which was either clinical follow up for tumour growth, confirmatory imaging, needle biopsy, surgical biopsy, or explanted liver. Literature search was performed from 1995 to 2001.

Results: 138 studies involving 6509 patients fulfilled the selection criteria. In 112 studies involving 5797 patients, they either included patients without cirrhosis or reported a subgroup of HCC results from a study of focal hepatic lesion. This made it difficult to extract data of HCC for cirrhotic patients alone. 26 studies provided relevant results for this review. Group 1: 10 studies used explanted liver as the reference standard of diagnosis of 520 HCC nodules in 159 cirrhotic patients. All studies except one either found no statistically significant difference or had no direct comparison of lesion-by-lesion sensitivity between different modalities of imaging, including US, conventional, or spiral computerised tomography (CT), lipiodol CT, dynamic gadolinium-enhanced magnetic resonance imaging (MRI), and angiography. Across these studies, sensitivity varied greatly even for the same modality of imaging. However, nodule-by-nodule specificity for US, conventional CT, spiral CT, and iodised-oil CT were more than 92% in 4 studies in which the data were available. Group 2: 16 studies included 553 patients limited to HCC among cirrhotic patients with surgical biopsy, needle biopsy, or imaging as reference standard for diagnosis. Grey-scale and microbubble Doppler US, spiral CT, lipiodol CT, CT hepatic arteriography, CT during arterial portography, gadolinium, or superparamagnetic iron oxide-enhanced MRI were studied. However, no definite conclusion could be drawn to the optimal imaging modality as no one technique was superior.

Conclusion: Despite publication of 138 articles involving over 6500 patients in the past 7 years, the evidence for choosing the best modality of imaging of HCC in cirrhotic patients is still inadequate. This is due to the paucity of good prospective (whether randomised or not) studies with sufficient number of patients. Large multi-centre trials for comparing modalities of imaging, especially non-invasive ones, among cirrhotic patients, preferably with explanted liver as the reference standard for diagnosis, are urgently needed.


S.F. Stewart, E. Albano1, C.P. Day, D.E.J. Jones.

Centre for Liver Research, Medical School, Framlington Place, Newcastle-upon-Tyne; 1 Dept of Medical Sciences and Medical Clinic University of East Piedmont, Novara, Italy

Background: Serum antibodies reactive with neo-antigens generated during ethanol metabolism have been identified in patients with alcoholic liver disease (ALD), although the role played by these antibodies (if any) in target cell damage is unclear. In contrast, the presence of liver infiltrating lymphocytes in ALD, and disease associations with polymorphisms in T-cell regulatory genes, suggest a role for T-cell mediated responses, although little is currently known about their antigen specificity. In this study we sought to characterise peripheral blood mononuclear cell (PBMC) T-cell proliferative responses to two such neo-antigens (acetaldehyde (AcA) and malondialdehyde (MDA) adducted to albumen) in patients with advanced ALD and control heavy drinkers with no liver disease (NALD).

Methods: PBMC were isolated and challenged with optimised concentrations of antigen and the control pan-T-cell stimulant OKT3 in vitro. Peak proliferative response was assessed by tritiated thymidine incorporation and results expressed as stimulation indices (SI).

Findings: Of 20 patients with advanced ALD, 8 (40%) had significant responses (SI of greater than 1.5) to at least one antigen (1 to AcA only, 4 to MDA only and 3 to both) compared to 0/10 (0%) of the NALD group (p = 0.029). Furthermore, the mean proliferation to MDA was significantly higher in the ALD group than in the NALD group (1.38 ± 0.7 v 0.94 ± 0.16; p = 0.04). Maximal proliferation to OKT3 was similar between the groups.

Interpretation: These data provide the first evidence for T-cell proliferative responses to antigens derived during ethanol metabolism in patients with ALD. They further suggest that the development of these responses is a susceptibility factor for the development of advanced disease.


D.A. Price, K. Agawal, M. Hewett, W.L. Craig, M. Schmid, A. Turner, A.D. Burt, M.F. Bassendine.

Liver Unit, Freeman Hospital, Newcastle-upon-Tyne, UK

HCV is now recognised as one of the leading causes of liver disease and the burden of illness in the UK is expected to increase as a result of injecting drug use. In this study we have looked at the demographics of our current clinical case load.

Methods: Retrospective analysis of the past 400 patients seen at our centre in whom chronic HCV infection has been found to be the main aetiological factor in their biopsy proven chronic liver disease. 256 (64%) were male, mean age at initial biopsy was 38.1 ± 12.1 years compared with 38.2 ± 11 years in females. The initial histology was divided on the basis of fibrosis stage (Ischak score) into mild (stages 0 and 1), moderate (stages 2, 3 and 4), and advanced (stages 5 and 6). 191 patients (61.2% male, mean age 35.1 ± 10.3 years) had mild fibrosis. 134 patients (69.4% male, mean age 36.2 ± 8.3 years) had moderate fibrosis and 75 patients (75% male, mean age 49.2 ± 13.8 years) had advanced fibrosis. Cirrhotic patients were significantly older (p < 0.0001) than patients with mild or moderate fibrosis. In this cohort HCV genotyping has been performed on 222. Genotype 1 was found in 112 (50.4%), genotype 3 in 91 (41%), genotype 2 in 9 (4.1%). There was no association between fibrosis stage and HCV genotype. Follow up liver biopsies have been performed in 116 of these patients. 35 (30.2%) have shown fibrosis progression on routine scoring, with a median rate of fibrosis stage progression in these ‘fast fibrosers’ of 0.36 per year. Hepatocellular cancer was proven on biopsy in 12 (83% male) of these 400 patients, at a mean age of 60.2 ± 11.4 years.

Summary: This retrospective analysis of HCV related chronic liver disease seen in clinical practice confirms previous natural history studies in that 30.2% would be expected to progress to cirrhosis in 25.2 years (if progression linear), cirrhosis presents at 49.2 years and HCC at 60.2 years. It emphasises the importance of anti-viral and/or anti-fibrotic therapy in controlling the future burden of disease.


I. Gee1, A. K. Trull1, S. C. Charman2, G. J. Alexander1.

1 Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK; 2 MRC Biostatistics Unit, Cambridge, UK

Introduction: Sirolimus is a new immunosuppressive drug that is being used increasingly in transplant recipients. It has been observed that some patients develop bacterial sepsis during treatment.

Methods: We have developed a physiological in vitro model to investigate the effects of therapeutic concentrations of sirolimus on the neutrophil oxidative burst (NOB)—a mechanism by which immunity to bacterial and fungal infection may be impaired. Whole blood from 24 healthy subjects was equilibrated with 0, 1, 5, 10, and 50 μg/L sirolimus or 60 mg/L propofol (a known inhibitor of neutrophil function) for 2 hours at 37°C. The cells were washed and the neutrophils stimulated with phorbol myristate acetate. NOB was measured by flow cytometry using the fluorescent marker dichlorofluorescein.

Results: A significant mean inhibition of NOB (95 % CI of mean % inhibition did not overlap zero) was found with 50 μg/L sirolimus (mean 6.3; CI 1.5 to 11.1 %) and 60 mg/L propofol (mean 5.1; CI 0.4, 9.8 %). 10 μg/L sirolimus also inhibited NOB (mean 4.6; CI –1.3, 10.6; NS) but inhibition was < 1.5% at lower concentrations. Repeated measures ANOVA confirmed a linear relationship between sirolimus concentrations and inhibition of NOB (p = 0.01). The 2 hour incubation period had no effect on red blood cell integrity or white cell viability in vitro.

Conclusion: Sirolimus had a dose-dependent inhibitory effect on NOB but this was not significant at low therapeutic concentrations. This may partly explain the predisposition to sepsis in patients receiving sirolimus and emphasises the importance of monitoring blood sirolimus concentrations as a guide to dosage adjustment. The 2 hour exposure of blood cells to sirolimus in these studies probably underestimates the in vivo effects of sirolimus and we are currently investigating the effect of chronic oral dosing with sirolimus on NOB.


C.P. O’Brien1,2, M. O’Brien2,3, M. Curry1,2, C. O’Farrelly2,3, J. Hegarty1,3, P.T. Donaldson4.

1 National Liver Transplant Unit; 2 Education and Research Centre, St Vincents University Hospital, Dublin; 3 Conway Institute, U.C.D.; 4 Centre for Liver Research, University of Newcastle, Newcastle-upon-Tyne, UK

The outcome following infection with HCV is highly variable. The determinants of disease progression and outcome following HCV infection are poorly defined. Stromelysin is a member of the matrix metalloproteinase (MMP) family of enzymes which regulate extracellular matrix degradation and fibrosis. Recent studies suggest that a functional polymorphism in the promoter region of the stromelysin gene (MMP3) is associated with both increased susceptibility (risk) and progression of the fibrotic liver disease, primary sclerosing cholangitis.1 The aim of the present study was to determine whether this polymorphism is indicative of prognosis and outcome following HCV infection. The MMP3 -1171 5A/6A polymorphism was determined in a homogenous cohort of 111 Irish women infected with HCV genotype 1b after the administration of contaminated anti-D immunoglobulin in 1977/8. A standard ARMS-PCR technique was used throughout. The overall genotype distribution was similar to that of racially matched controls. There was no significant difference in genotype distribution between those with persistent viraemia and those with spontaneous (untreated) viral clearance. Of the 72 HCV RNA + patients, 31 (43%) had established fibrosis on liver biopsy. In these patients the 6A/6A genotype was less common in those with established fibrosis compared with patients without (OR = 0.32). These data suggest that the 6A/6A genotype exerts a protective effect against the development of fibrosis in HCV patients. Although this effect did not reach statistical significance at the 5% CI (p = 0.06) in our small series, HCV infection is common and therefore this association may have considerable clinical impact, if it can be extended to a larger, and more heterogenous HCV population.