Article Text

British Association for the Study of the Liver Meeting
  1. S. Clark,
  2. W. Bernal,
  3. J. A. Wendon
  1. Institute of Liver Studies, King's College Hospital, Denmark Hill, London SE5 9RS, UK
  1. H. M. Evans,
  2. P. J. Mckiernan,
  3. S. V. Beath,
  4. J. De Ville De Goyet,
  5. D. A. Mayer,
  6. J. A. C. Buckels,
  7. P. Davies3-1,
  8. D. A. Kelly
  1. The Liver Unit, Birmingham Children's Hospital, UK; 3-1The University of Birmingham, UK
  1. M. C. Wright,
  2. R. Issa,
  3. D. E. Smart,
  4. N. Trim,
  5. J. N. Primrose,
  6. M. J. P. Arthur,
  7. J. P. Iredale,
  8. D. A. Mann
  1. Liver Group, Cell and Molecular Medicine, Southampton General Hospital, Southampton SO16 6YD,UK
  1. D. M. Forton,
  2. S. D. Taylor-Robinson,
  3. P. Karayiannis,
  4. H. C. Thomas
  1. Hepatology Section, Dept of Medicine, ICSM, St Mary's Campus, Praed St, London, UK
  1. N. N. Mirza,
  2. P. F. Lalor,
  3. A. Williams,
  4. D. H. Adams
  1. Liver Research Laboratories, MRC Centre for Immune Regulation, University of Birmingham, UK
  1. T. T. Yee,
  2. A. Griffioen,
  3. C. A. Sabin12-1,
  4. G. Dusheiko12-2,
  5. C. A. Lee
  1. Haemophilia Centre and Haemostasis Unit, Royal Free Hospital and Depts of 12-1Population Sciences and Primary Care and 12-2Medicine, Royal Free and University College Medical School, University College, London, UK
  2. lee{at}rfhsm.ac.uk
  1. S. K. Jain,
  2. P. W. Pemberton,
  3. P. C. Burrows,
  4. A. Smith,
  5. R. F. T. McMahon,
  6. A. A. Aboutwerat,
  7. T. W. Warnes
  1. Liver Unit, Manchester Royal Infirmary, Manchester, UK
  1. “The European Liver Fibrosis Consortium” represented by W. M. Rosenberg,
  2. A. D. Burt18-1,
  3. M. Becka18-2,
  4. M. Voelker18-2,
  5. M. J. P. Arthur
  1. Cell & Molecular Medicine, University of Southampton, UK; 18-1Dept of Pathology, University of Newcastle, UK; 18-2Bayer AG, Wuppertal, D42096, Germany
  1. W. Bernal,
  2. P. Leckie,
  3. J. Wendon
  1. Institute of Liver Studies, King's College Hospital, Denmark Hill, London SE5 9RS, UK
  1. N. V. Naoumov20-1,
  2. D. Suri20-1,
  3. E. I. Rigopoulou20-1,
  4. T. Machell20-1,
  5. I. Mullerova20-1,
  6. S. Chokshi20-1,
  7. S. Rice20-2,
  8. R. S. Tedder20-2,
  9. R. Williams20-1
  1. 20-1Institute of Hepatology and 20-2Department of Virology, University College London, London WC1E, UK
  1. M. J. Lorite,
  2. D. J. Bevitt21-1,
  3. J. Benitez,
  4. N. McKie21-1,
  5. C. P. Day
  1. Centre for Liver Research and21-1Dept of Rheumatology, Medical School, Framlington Place, Newcastle upon Tyne, UK
  1. P. J. McKiernan,
  2. A. J. Baker25-1,
  3. D. A. Kelly
  1. Liver Unit, The Children's Hospital NHS Trust, Birmingham and 25-1Paediatric Liver Unit, King's College Hospital, London, UK. Pat.Mckiernan{at}bhamchildrens.wmids.nhs.uk
  1. K. Agarwal,
  2. A. K. Daly27-1,
  3. J. B. Leathart27-1,
  4. M. Hudson,
  5. C. P. Day
  1. Centre for Liver Research and27-1Dept of Pharmacological Sciences, Medical School, Framlington Place, Newcastle upon Tyne, UK
  1. C. A. Sabin28-1,
  2. V. Emery28-2,
  3. H. L. Devereux28-3,
  4. A. Griffioen,
  5. J. Bishop28-2,
  6. T.-T. Yee,
  7. E. Herrero28-2,
  8. C. A. Lee
  1. Haemophilia Centre and Haemostasis Unit, Departments of 28-1Primary Care and Population Sciences, 28-2Virology and 28-3Retrovirology, RF&UC Medical School, Rowland Hill Street, London, UK
  1. M. C. Wright,
  2. D. E. Smart,
  3. P. R. McCrudden,
  4. J. E. Trim,
  5. J. N. Primrose,
  6. M. J. P. Arthur,
  7. D. A. Mann
  1. Liver Group, Cell and Molecular Medicine, Southampton General Hospital, Southampton SO16 6YD,UK
  1. E. J. Williams,
  2. B. C. Cochrane,
  3. M. J. P. Arthur,
  4. R. C. Benyon
  1. Liver Group, Southampton General Hospital, Southampton SO16 6YD, UK
  1. C. M. Morland,
  2. D. H. Adams
  1. Liver Research Laboratories, University of Birmingham, Edgbaston, Birmingham, UK
  1. S. Clark34-1,
  2. S. Creighton34-2,
  3. M. Horner34-1,
  4. B. Portmann34-1,
  5. M. Cramp34-1,
  6. C. Taylor34-2
  1. 34-1Institute of Liver Studies and 34-2Department of Sexual Health, King's College Hospital, London SE5 9RS, UK. sarah.clark{at}kcl.ac.uk Abstract category: c) Viral Hepatitis
  1. R. Issa,
  2. N. Trim,
  3. T. Kendal,
  4. J. Riechen35-1,
  5. R. C. Benyon,
  6. J. Iredale
  1. Liver Group, Division of Cell and Molecular Medicine, Level D South Block 811, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK;35-1Department of Pharmacology, University of Berne, Switzerland
  1. A. J. Grant,
  2. C. Walker,
  3. S. Goddard,
  4. D. H. Adams
  1. MRC Centre for Immune Regulation, University of Birmingham, B15 2TH, UK
  1. D. M. Flynn,
  2. A. J. Strain,
  3. D. A. Kelly,
  4. H. A. Crosby
  1. Liver Laboratories, Queen Elizabeth Hospital, Edgbaston, Birmingham, UK; Birmingham Children's Hospital, Steelhouse Lane, Birmingham, UK
  1. N. Trim,
  2. J. E. Trim,
  3. D. A. Mann,
  4. J. P. Iredale
  1. Liver Group, Division of Cell and Molecular Medicine, Southampton General Hospital, Southampton, SO16 6YD, UK
  1. R. L. Jones,
  2. S. C. Afford,
  3. D. H. Adams
  1. Liver Research Laboratories, MRC Centre for Immune Regulation, University of Birmingham, Edgbaston, Birmingham, B15 2TH, UK.r.l.jones{at}bham.ac.uk
  1. J. Goldblatt44-1,
  2. P. J. S. Taylor44-2,
  3. T. Lipman44-2,
  4. O. F. W. James44-1,
  5. D. E. J. Jones44-1
  1. 44-1Centre for Liver Research and 44-2Dept of Primary Health Care, University of Newcastle
  1. S. J. Hodges,
  2. E. Rigney45-1,
  3. A. J. Lee,
  4. R. Eastell,
  5. D. Gleeson45-1
  1. Bone Metabolism Group, University of Sheffield, Sheffield, UK; 45-1Gastroenterology and Liver Unit, Royal Hallamshire Hospital, Sheffield, UK. dermot.gleeson{at}csuh.nhs.uk
  1. R. Rolla,
  2. S. F. Stewart46-1,
  3. E. Mottaran,
  4. D. Vay,
  5. M. Vidali,
  6. M. Sartori46-1,
  7. C. Rigamonti46-1,
  8. C. P. Day46-2,
  9. E. Albano
  1. Dept of Medical Sciences and46-1Medical Clinic, University of East Piedmont, Novara, Italy, 46-2Centre for Liver Research, Medical School, University of Newcastle, Newcastle upon Tyne, UK
  1. P. Vyas,
  2. R. Schilling,
  3. E. Rigopoulou,
  4. R. Williams,
  5. N. V. Naoumov
  1. Institute of Hepatology, University College London, London WC1E 6HX, UK
  1. A. Aboutwerat,
  2. P. Pemberton,
  3. P. Burrows,
  4. A. Smith,
  5. S. Jain,
  6. T. W. Warnes
  1. Liver Unit, Manchester Royal Infirmary, Manchester, UK
  1. D. Gleeson,
  2. M. P. Bradley,
  3. J. Sorrell52-1,
  4. R. J. Peck52-2,
  5. G. W. Duff
  1. Gastroenterology and Liver Unit,52-2Dept of Radiology, Royal Hallamshire Hospital; 52-1Division of Molecular & Genetic Medicine, University of Sheffield 52-150, Sheffield, UK. e-mail: dermot.gleeson{at}csuh.nhs.uk
  1. T. Zolfino,
  2. S. Norris,
  3. P. M. Harrison,
  4. I. G. McFarlane
  1. Institute of Liver Studies, King's College Hospital, Denmark Hill, London SE5 9RS, UK.
  1. R. McCrudden,
  2. D. Smart,
  3. M. Wright,
  4. J. T. Trim,
  5. C. Benyon,
  6. Y. De Klerk,
  7. D. Mann,
  8. J. P. Iredale
  1. Liver Group, Cell and Molecular Medicine, Southampton General Hospital, Southampton SO16 6YD
  1. M. Khalil,
  2. H. J. F. Hodgson,
  3. T. Ryder55-1,
  4. C. Selden
  1. Centre for Hepatology, Dept. Medicine, Royal Free Campus-UCL, Rowland Hill St., London, NW3 2PF, UK;55-1Dept. Histopathology, Queen Charlotte's Hospital-ICSM, London, W12 0NN, UK
  1. T. M. Rahman,
  2. H. J. F. Hodgson
  1. Centre for Hepatology, Royal Free &University College Medical School, Royal Free Campus, Rowland Hill Street, London NW3 2PF, UK.t.rahman{at}ic.ac.uk
  1. D. F. Wallace,
  2. A. P. Walker,
  3. A. Pietrangelo61-1,
  4. M. Clare61-2,
  5. A. B. Bomford61-2,
  6. J. S. Dooley
  1. Centre for Hepatology, Department of Medicine, Royal Free and University College Medical School, Rowland Hill Street, London, NW3 2PF, UK; 61-1Dipartimento di Medicina Interna, Università di Modena, Via del Pozzo 71, 41100 Modena, Italy; 61-2Institute of Liver Studies, Guy's, King's and St Thomas' School of Medicine, Bessemer Road, London, SE5 9PJ, UK
  1. T. M. Rahman,
  2. H. J. F. Hodgson
  1. Centre for Hepatology, Royal Free & University College Medical School, Royal Free Campus, Rowland Hill St. London NW3 2PF, UK.t.rahman{at}ic.ac.uk
  1. X. Zhou,
  2. R. Issa,
  3. R. McCrudden,
  4. C. J. Hovell,
  5. J. P. Iredale,
  6. R. C. Benyon
  1. Liver Group, Cell and Molecular Medicine, Southampton General Hospital, Southampton SO16 6YD, UK.
  1. S. Clark,
  2. D. Burt,
  3. J. Miell,
  4. J. A. Wendon,
  5. L. Gnudi
  1. Institute of Liver Studies, King's College Hospital, London SE5 9RS, UK; Department of Diabetes, Endocrinology and Metabolic Medicine, Guy's Hospital, London SE1, UK; Department of Diabetes, Endocrinology and Metabolic Medicine, King's College Hospital, London SE5 9RS, UK
  1. R. Jalan,
  2. S. Olde Damink68-1,
  3. P. C. Hayes,
  4. A. Lee
  1. Liver Unit and Scottish Liver Transplantation Unit, Royal Infirmary of Edinburgh, UK; 68-1Institute of Hepatology, RFUCMS, London, UK
  1. I. Gehrke,
  2. P. John70-1,
  3. D. A. Kelly,
  4. S. V. Beath,
  5. P. J. McKiernan,
  6. J. de Ville de Goyet
  1. Liver Unit and 70-1Department of Radiology, Birmingham Children's Hospital, Birmingham, UK
  1. A. Mahmood,
  2. S. Moodie,
  3. L. Ang,
  4. J. D. Maxwell,
  5. G. E. Levin,
  6. C. Finlayson
  1. St.George's Hospital and Medical School, London SW17 0RE, UK
  1. J. M.C. Stenner,
  2. R. Jazrawi,
  3. A. Verma,
  4. H. Ahmed,
  5. J. D. Maxwell,
  6. T. C. Northfield
  1. St George's Hospital Medical School, London SW17 0RE, UK.jmcstenner{at}hotmail.com
  1. J. M. C. Stenner,
  2. A. E. A. Joseph,
  3. C. F. Finlayson. J. D. Maxwell
  1. St Georges Hospital and Medical School, London, SW17 ORE, UK.jmcstenner{at}hotmail.com

Statistics from Altmetric.com

The following abstracts were presented at the British Association for the Study of the Liver Meeting, University of Southampton, UK, September 7–8, 2000.

1. MURINE HISTIDINEMIA: AMELIORATION USING RETROVIRAL GENE TRANSFER

Introduction: Histidinaemia is an autosomal recessive disorder in man affecting the hepatic enzyme histidase resulting in elevated plasma and urinary histidine. A mouse model with a similar phenotype lends itself to investigating methods of gene therapy.

Aim: To revert the biochemical phenotype in by in vivo retroviral gene transfer.

Methods: A cDNA encoding normal histidase was cloned into the LNCX retroviral construct under the control of the CMV promoter. Virus was concentrated and administered to histidinaemic mice primed for hepatocyte transfection by partial hepatectomy. HPLC flow scintillation was used to assess in vivo catalysis of14C-histidine. Hepatic cytosol was examined for histidase expression by Western analysis and enzyme activity by conversion of histidine to urocanic acid.

Results: Infected animals showed a 33%±15% enhancement of metabolism as shown by in vivo histidine breakdown at 7 days (p<0.05 c.f. controls). Phenotype amelioration continued until 14 days but returned to pre-operative levels at 21 days post infection. Histidase activity from liver tissue from transduced animals ranged from 46.1±13.3 to 81.8±22 pmols of urocanic acid formation/min/mg total protein compared with 8.8±3.7 in histidinaemic controls and 1315±47.4 observed in wild type controls.

Conclusion: Murine histidinaemia provides a model for testing gene therapy for hepatocyte enzyme defects. Therapeutic retroviral gene transfer was sufficient to ameliorate the enzyme deficiency in this model and restored enzyme function in excess of 14 days. Immune reaction to expressed normal protein in this model is unlikely from previous hepatocellular transplantation studies and down regulation of the therapeutic gene by CMV promoter silencing is a likely explanation of early loss of enzyme function.

2. THE PROGNOSTIC IMPLICATIONS OF HYPERAMMONEMIA IN ACUTE LIVER FAILURE

Background: Extensive animal research has suggested that ammonia (NH4) has a pivotal role in the pathogenesis of hepatic encephalopathy (HE) in acute liver failure (ALF). Clinically the role of NH4 in HE remains less clear although two recent studies have suggested significant correlation with HE and cerebral herniation.

Aim: To examine sequential NH4levels in ALF and correlate these with HE and outcome.

Methods: 66 patients were studied prospectively. Arterial NH4 levels were measured on alternate days (Blood Ammonia Checker II, Menarini, UK), non-heparinised blood drawn into EDTA and analysed immediately (normal range 10–47 μmol/l). Statistical analysis was by Anova, Spearman rank correlation, and multiple linear regression.

Results: Of the 66 patients studied, aetiology was paracetamol in 48 (73%) patients, acute hepatitis B infection in 4 (6%) and other causes for 14 (21%). 12 patients underwent emergency liver transplantation (OLT), 3 patients dying post OLT. Mortality was 39% in the 54 patients treated medically. 51 (77%) patients required ventilation for grade 3–4 HE and 45 (68%) patients developed renal failure. The NH4 on admission for non-survivors and survivors was 96 μmol/l (28–232) and 80 (18–166) respectively. NH4 correlated positively with bilirubin and INR (p<0.02) but not with creatinine or lactate. The median NH4 within the first 24 hours of admission for grade 0–1 HE was 80 μmol/l (range 18–147), 104 (59–232) for grade 2–3 HE and 101 (29–286) for grade 4HE, with no significant difference between groups. Eight (12%) patients had an episode of pupillary abnormality during their admission, 6 of whom subsequently died. Post mortems finding detected cerebral oedema in only 2 of these patients. In patients with pupillary abnormalities, the median NH4 was 73.5 μmol/l (29–210). The NH4 in patients with no pupillary abnormalities, was 70 μmol/l (13–286). 20 (30%) patients had ICP bolts inserted. NH4 levels for ICP measurements of <15mmH20, 15–25 and >25 were 68 μmol/l (13–262), 79 (11–232) and 61 (46–62) respectively. NH4 levels were 97 μmol/l (28–198), 59 μmol/l (18–232) and 71 μmol/l (14–215) in patients receiving no enteral nutrition, patients receiving 50% of their nutritional requirements, and patients receiving greater than 75% of their requirements respectively. No increase in NH4was detected related to enteral feeding, a trend towards a higher NH4 being detected in patients receiving no enteral nutrition 48 hours post admission. The NH4 was 125 μmol/l (50–175) in patients not fed, as compared to 70 (12–227) in those fed (p<0.05). Conclusion: Arterial NH4 is unable to predict outcome in ALF or the grade of HE, and is not associated with the presence of cerebral oedema, as measured by the ICP and post mortem findings. There is no evidence that feeding increases circulating NH4 levels and may, in fact, be beneficial. Early enteral feeding is therefore recommended.

3. HISTOLOGY OF LIVER ALLOGRAFTS FIVE YEARS FOLLOWING PAEDIATRIC LIVER TRANSPLANTATION

Introduction: Five year survival following paediatric liver transplantation (OLT) is >85% but the long term histological outcome of the transplanted allografts is unknown.

Aims: To evaluate histological findings in liver allografts five years post OLT and to correlate these with donor and recipient variables.

Methods: A retrospective review of case notes and histology from protocol biopsies at five years post OLT. Donor details were obtained from the National database. Logistic regression was used to identify the characteristics significantly associated with abnormal histology.

Results: Between 1983–1995, 123 children (61M; 62F) received liver transplants, of which, 16 were re-grafts. Primary immunosuppression was cyclosporin, azathioprine and corticosteroids for 3 months with cyclosporin monotherapy after 12 months. Histological data were available from 90 (44M; 46F) (73%) of the patients of which 74 (83%) were protocol biopsies from asymptomatic children with normal liver biochemistry.

There is a positive and statistically significant association between larger donor weight and hepatitis or fibrosis at 5 years (p=0.025) irrespective of donor age and whether the graft was whole or reduced. Male grafts transplanted into female recipients had a higher incidence of both chronic hepatitis and fibrosis (p=0.04; odds ratio 7:1 95% CI 1.1–45.4). Female into male and sex-matched grafts had no excess abnormal histology.

Abstract 3, Table 1

Conclusions: The incidence of abnormal histology at 5 years post OLT was higher than anticipated. Male–female donor recipient grafts had a higher incidence of abnormal histology at five years, which may represent low grade chronic rejection. There was a significant association between abnormal histology and higher donor weight which is unexplained and requires further study.

4. NON-CYTOLYTIC INHIBITION OF HEPATITIS B VIRUS REPLICATION IN HUMAN HEPATOCYTES

Background: Studies in HBV transgenic mouse revealed that interferon-γ (IFNγ) secreted from T lymphocytes can inhibit HBV replication without killing infected cells. We investigated the relevance of this non-cytolytic antiviral mechanism in human HBV infection.

Results: We developed anin vitro model where PBMC from 16 HBsAg(+) patients were co-cultured with a cell line supporting full HBV replication. The effector and target cells were separated by a membrane which allows transfer of soluble factors only and the secreted IFNγ was measured by ELISA. Stimulation of PBMC, resulting in IFNγ levels greater than 170 pg/ml, was associated with 70–90% reduction of cytoplasmic HBV DNA compared to controls. The levels of IFNγ correlated with the degree of HBV DNA reduction (r=0.8). Next, we analysed the effect of recombinant IFNγ (300 and 3000 pg/ml) on naturally infected human hepatocytes, isolated from 11 HBeAg+/HBV DNA+ patients. The number of HBV DNA copies per cell was quantitated by Amplicor Monitor (Roche) and viral transcription (HBV RNA) by RT-PCR. HBV DNA ranged between 4 and 600 copies per hepatocyte. Recombinant IFNg reduced HBV DNA between 14 to 76% and HBV RNA between 51 to 100% in comparison with controls. Cytotoxicity, measured by LDH release in the culture supernatants, was low (<9%) and similar to the spontaneous release from untreated hepatocytes.

Conclusions: These results demonstrate that IFNγ produced by stimulated lymphocytes from chronic HBV carriers is able to inhibit HBV replication. These data provide direct evidence of IFNγ-induced suppression of both HBV transcription and replication in human hepatocytes without cell lysis.

5. SUSEPTIBILITY TO ALCOHOLIC LIVER DISEASE IS DETERMINED BY GENETICALLY DETERMINED DIFFERENCES IN THE IMMUNE RESPONSE TO ACETALDEHYDE ADDUCTS

Twin studies suggest that genetic factors play a role in susceptibility to alcoholic liver disease (ALD). The immune response to protein adducts arising during ethanol metabolism may be involved in disease pathogenesis. We have recently reported associations between ALD and polymorphisms of the T cell molecule, CTLA-4, and the IL-10 promoter, both of which would be expected to increase the magnitude of immune responses. IL-4 favours a Th2-mediated response and a Q576R “increase-of-function” mutation in the IL-4 receptor gene has recently been described. We therefore determined the frequency of this mutation in 267 patients with biopsy-proven advanced fibrotic ALD, 125 heavy drinkers with normal liver/steatosis only and 179 healthy controls. 48% of cirrhosis/fibrosis patients possessed at least 1 copy of the mutant allele versus 33% of controls (OR:1.9[1.3–2.9], p=0.002) and 32% of the heavy drinkers (OR:2.1[1.3–3.2], p=0.002). 19% of ALD patients had all 3 “at-risk” alleles of the IL-4 receptor, IL-10 and CTLA-4 polymorphisms versus 2% of heavy drinkers (OR:12[3–49], p<0.0001).

We then determined whether these polymorphisms were associated with an increased Th2-mediated antibody response to ethanol-derived acetaldehyde adducts. Titres of anti-adduct antibodies were determined by ELISA in the serum of 59 genotyped patients with ALD and were significantly higher in patients with all three “at-risk” polymorphisms compared with those with none (p<0.005) and were also associated with the presence of each individual polymorphism (IL4R, p<0.01; IL-10, p<0.05; CTLA-4, p<0.005).

These results suggest that, in some heavy drinkers, susceptibility to fibrotic ALD is due to a genetically determined enhanced Th2 immune response to ethanol-derived adducts. This is most likely attributable to the pro-fibrotic cytokine mileau associated with a Th2 response.

6. GLIOTOXIN STIMULATES THE APOPTOSIS OF HEPATIC STELLATE CELLS: A MECHANISM FOR MODULATING LIVER FIBROSIS

It is broadly accepted that hepatic stellate cells (HSC) play a central role in the development and resolution of liver fibrosis. In response to liver damage, HSCs “activate” to a myofibroblast-like (α-smooth muscle actin +ve) phenotype which is widely regarded to be responsible for the majority of extracellular matrix protein deposition in liver fibrosis. We recently demonstrated that the spontaneous recovery from fibrosis observed in rats after treatment with the hepatotoxin carbon tetrachloride for 4 weeks is associated with a reduction in liver collagen and tissue inhibitor of metalloproteinase (TIMP) expression that correlates with an increase in activated HSC apoptosis. The fungal metabolite gliotoxin, which initiates the apoptosis of activated rat and human hepatic stellate cell in vitro, was administered to rats with liver fibrosis to determine if stellate cell apoptosis may be stimulated by gliotoxin in vivo, and whether an enhanced rate of apoptosis of activated stellate cells may modulate liver fibrosis. Seven weeks of CCl4 treatment resulted in marked increases in serum alanine aminotransferases activities; liver damage as judged by histological examination of H&E stained tissue sections; marked increases in liver collagen deposition as judged by sirius red staining of tissue sections and extensive α-smooth muscle actin positive staining of tissue sections. Gliotoxin treatment did not give rise to any apparent hepatoxic effects and did not affect either serum alanine aminotransferases activities or liver damage as judged by histological examination of H&E stained tissue sections in CCl4-treated animals. However, gliotoxin treatment resulted in reduced levels of collagen (sirius red) staining and an approximately 50% reduction in α-smooth muscle actin positive staining in liver sections from rats treated with CCl4. TUNEL staining demonstrated that gliotoxin stimulated a marked increase in TUNEL positive nuclei. These data indicate that gliotoxin stimulates the apoptosis of activated rat hepatic stellate cells in an in vivo rat model of liver fibrosis and that this does result in a modulation in the severity of fibrosis as judged by a reduction in liver collagen.

7. MODERATE HYPOTHERMIA FOR UNCONTROLLED INTRACRANIAL HYPERTENSION IN ACUTE LIVER FAILURE (ALF)

Background: Increased intracranial pressure (ICP) complicating acute liver failure (ALF) has a mortality of about 90% in patients who do not respond quickly to treatment with mannitol and ultrafiltration. This study evaluates the safety and efficacy of moderate hypothermia as a treatment for uncontrolled increase in ICP in patients with ALF.

Methods: Fourteen consecutive patients [age 24 (range 16–54), 9 females, 8 candidates for orthotopic liver transplantation (OLT)] with who fulfilled the Kings College criteria for poor prognosis and had uncontrolled increase in ICP [defined as persistently elevated ICP of >25 mmHg for 1 hour or more despite 2 separate treatments with mannitol (1g/Kg body weight over 20 min) and removal of 500 ml of fluid by continuous veno-venous hemofiltration] were studied. All patients required sedation and ventilation for Grade III-IV encephalopathy. Patients were monitored continuously, using a Swan-Ganz catheter for measuring cardiovascular haemodynamics, intracranial pressure was monitored using a subdural catheter (Camino) and cerebral blood flow was measured using the Kety-Schmidt method. A reverse jugular catheter allowed sampling of blood. Core temperature was reduced to 32–33°C using cooling blankets.

Results: Eight of the nine patients who were candidates for OLT were successfully bridged to transplantation with a mean of 32 hours (range 8–120) hours of hypothermia. The five patients who were unsuitable candidates for OLT died following rewarming. ICP prior to cooling was 54 (range 23–67) mmHg and this was reduced in all patients, to 19 (range 10–22) mmHg (p<0.01) and the cerebral blood flow decreased from 97 ml/100g/min (range 22–154) to 32 ml/100g/min (range 24–76) (p<0.01). Cerebral perfusion pressure increased (p<0.01) mmHg and cardiac index decreased significantly (p<0.01). During hypothermia there 4 relapses of increase in ICP in 3 patients which responded to treatment with mannitol, further fluid removal or thiopentone infusion. Arterial ammonia and cerebral uptake of ammonia were significantly reduced with cooling. No adverse effects of hypothermia were observed.

Conclusions: Moderate hypothermia is useful in the treatment of uncontrolled increase in ICP in patients with ALF and may serve as a bridge to OLT. The mechanism by which hypothermia reduces intracranial hypertension is through reduction of cerebral blood flow and ammonia delivery and extraction by the brain.

8. IDENTIFICATION OF BRAIN-SPECIFIC QUASISPECIES VARIANTS OF HEPATITIS C VIRUS (HCV) CONSISTENT WITH VIRAL REPLICATION IN THE CENTRAL NERVOUS SYSTEM

Chronic HCV infection causes an impaired quality of life, often without significant liver disease. We therefore considered whether HCV infects the central nervous system.Post-mortem brain, liver, and serum were analysed for HCV RNA from 4 HCV+ve patients. Extracted RNA was amplified by RT-PCR and the 5' untranslated region (5'UTR) and the hypervariable region of E2 (HVR1) targeted with appropriate primers. The HVR1 PCR products were cloned and between 8 and 18 colonies were sequenced. The mean genetic distance was calculated for all pairs of sequences in each tissue. Maximum likelihood phylogenetic trees were constructed. HVR1 amplicons were obtained from serum and liver in all 4 patients and from brain in 2 patients (genotypes 3a and 4). In patient 1, the nucleotide sequence of the predominant clone in each tissue was identical and accounted for 31% of the brain, 40% of the liver and 67% of the serum sequences. However, 8 clones from brain (69%) and 6 clones from liver (60%) were not seen in the serum. This is reflected in the increased genetic diversity of the brain and liver clones compared to serum (mean genetic distances: brain 0.091, liver 0.088, serum 0.034, p<0.001). The maximum likelihood tree showed clustering of brain and liver clones with bootstrap values of between 85 and 93. There were 3 amino acid (aa) sequences (31%) unique to brain tissue and 4 aa sequences unique to liver (40%). In patient 2, there were 3 distinct brain and 9 distinct liver clones. None of the brain clones was seen in liver. When aa sequences were compared, the dominant sequence in brain and liver (75% and 30% respectively) was identical. However, 2 aa sequences were unique to brain. This study demonstrates the presence of distinct HCV quasispecies in brain material for the first time. The presence of both nucleotide and aa sequences unique to the brain suggests that serum contamination of brain specimens is an unlikely explanation for these findings. Infection of the CNS by HCV may explain the excess of neuropsychological symptomatology in these patients. These results raise the possibility of the CNS as a sanctuary site allowing immune evasion and viral persistence.

9. ENDOTHELIAL CARBOHYDRATE LIGANDS SUPPORT BOTH L-SELECTIN DEPENDENT AND L-SELECTIN INDEPENDENT T CELL ADHESION TO REJECTING HUMAN LIVER TRANSPLANTS

Background: Primary adhesion, the initial step of the adhesion cascade which brings the circulating leucocyte into contact with the vessel wall, is crucial for subsequent lymphocyte extravasation into liver tissue during allograft rejection. Most primary adhesion is dependent on the recognition and binding of carbohydrate-based endothelial adhesion molecules by lymphocyte receptors. The glycoprotein core of these endothelial adhesion molecules is modified by the attachment of carbohydrate side chains that present fucose, sulphate, and sialic acid subunits which can act as recognition sites for lymphocyte adhesion receptors. Two such molecules which have been shown to mediate primary adhesion of lymphocytes via carbohydrate ligands to lymph node and tonsil endothelium are the lymphocyte surface receptor L-selectin and the endothelial adhesion molecule vascular adhesion protein–1 (VAP-1).

Aims: In this study we investigated the role of L-selectin and VAP-1 in promoting primary adhesion of lymphocytes to hepatic endothelium in rejecting human liver allografts.

Methods: Non-static adhesion assays were developed using frozen sections of human liver biopsies as the substrate for T cell binding to liver endothelium.

Results: On tissue from patients with rejection, lymphocytes bound to central hepatic veins (CV), sinusoids (SC) and portal tract vessels (PT) and this adhesion was VAP-1 and L-selectin dependent. Combinations of 1B2 mAb raised against VAP-1 and DREG-56 mAb raised against L-selectin reduced T cell adhesion to CV (60% inhibition), SC (64% inhibition) and PT (27% inhibition) of control levels. The use of neuraminidase to remove sialic acid residues from the tissue endothelium revealed that all VAP-1 mediated adhesion was sialic acid dependent, whereas L-selectin mediated both sialic acid dependent and independent binding.

Conclusions: The presence of functional L-selectin ligands and VAP-1 on endothelium in rejecting liver allografts suggest that carbohydrate dependent pathways direct trafficking of lymphocytes into rejecting liver allografts. This raises the possibility of glycan-based immunomodulation therapies to prevent graft rejection.

10. COMPARISON OF TRENDS IN MORTALITY RATES FOR PRIMARY LIVER TUMOURS WORLDWIDE, 1979–1997

Background: The age-standardised mortality (ASMR) rate per 100 000 population for primary liver tumours (International Classification of Disease-9 155) is increasing in several countries. We have previously reported that in England & Wales this is mainly due to an increase in ASMR from intrahepatic cholangiocarcinoma. We further analysed mortality statistics from North America, Japan, Australia, and Europe to compare death rates of histologic subcategories of liver tumours in these countries.

Methods: ASMR for subcategories of liver tumours (ICD-9 155), tumours of the gall bladder (ICD-9 156) and pancreas (ICD-9 157) from 1979–97 were obtained from the World Health Organisation (WHO) mortality database. The standard population used was the world standard population.

Results: The most striking and consistent finding was an increase in ASMR for intrahepatic cholangiocarcinoma across the countries studied, although the rate of rise was not uniform. The largest ASMR rises were seen in Australia (0.1 to 0.7), England & Wales (0.2 to 0.9), Spain (0.0 to 0.6), France (0.0 to 0.3), Japan (0.1 to 0.3, men only) and USA and Canada (0.2 to 0.6). There were also significant rises in ASMR for hepatocellular carcinoma in some countries particularly Japan (9.4 to 16), Italy (4.4 to 8.9) and France (2.4 to 7.6).

Conclusion: We present a hitherto unreported rise in mortality rates from intrahepatic cholangiocarcinoma across 4 continents. This cannot be readily explained by diagnostic transfer from other tumour types, or by improved diagnosis and hence increased reporting. It may be related to local environmental and/or dietary factors affecting people with a susceptible genotype.

11. FUNCTIONAL CHARACTERISATION OF FULL LENGTH HEPATITIS B VIRUS (HBV) GENOMES FROM LIVER AND SERUM OF PATIENTS WITH FIBROSING CHOLESTATIC HEPATITIS

Background: Fibrosing cholestatic hepatitis (FCH) is a unique syndrome with rapidly progressive liver failure, originally described in patients with HBV recurrence after liver transplantation. Typically, patients with this syndrome have very high accumulation of viral proteins in hepatocytes, but little inflammation. In some cases, enhanced signals for HBV DNA and RNA have been demonstrated in the liver using in situhybridisation, thus indicating that enhanced viral replication could be a major factor for the development of FCH. The aim of our study was to investigate the replication capacity and protein expression of HBV strains isolated from patients with FCH.

Methods: Twelve clones of full length HBV DNA were generated from the liver and serum of 4 HBsAg+ liver transplant recipients with clinical and histological features of FCH. The precore/core and pre-S regions of the HBV genome were sequenced and functional characterisation was carried out by transfection into HuH-7 cells, in comparison with HBV not associated with FCH (wild type clones). HBV replication in the cells was monitored by quantitation of HBV DNA (Southern blot) and HBV RNA (Northern blot) using a Bio-Rad imager. Intracellular and extracellular HBsAg expression was analysed by immunostaining and ELISA, respectively.

Results: HBV DNA sequencing demonstrated that some FCH clones have precore stop codon mutations and/or pre-S deletions, however no common mutation was identified. Quantitation of HBV replication in cells transfected with FCH clones showed lower levels of HBV DNA (between 24–43%) and HBV RNA (between 48–85%) of those found in cells transfected with the wildtype HBV. The intra- and extracellular levels of HBsAg from FCH clones were similar or lower to the wild type clones. Interestingly, the addition of prednisolone to the cell cultures markedly increased the expression of HBsAg from two liver derived FCH clones, which was 3–8 fold higher than the effect on the wildtype HBV.

Conclusion: Increased replication capacity is not a necessary feature of the HBV strains responsible for the development of FCH. However, these strains differ from the wild type virus in their response to corticosteroids, which is characterised by a marked enhancement of viral protein expression.

12. THE NATURAL HISTORY OF HCV IN A COHORT OF HAEMOPHILIC PATIENTS INFECTED BETWEEN 1961–1985

This study describes the long term follow up of haemophilic patients infected with HCV between 1961–85. Clinical and treatment records from 310 patients with inherited coagulation disorders treated with blood product before 1985 were reviewed. Standard survival analysis methods were used to model progression to liver failure and death. 298/305 (98%) of patients tested were anti-HCV positive. 27 (9%) individuals consistently HCV PCR negative were considered to have cleared the virus. By 1 September 1 1999, 223/310 (72%) were alive, 26 (8%) had died a liver related death, and 61 (20%) had died from other, predominantly HIV related causes. The Kaplan-Meier progression rates to all cause and liver disease related death 25 years from exposure to HCV were 47% (95% CI 34–60) and 19% (95% CI 10–27), respectively. After 13.3 years from 1985, by which time all patients had seroconverted to HIV, the progressions to all cause and liver related deaths respectively were 8% (95% CI 4–13) and 3% (95% CI 0.4–6) for those HIV negative; and 57% (95% CI 48–66) and 21% (95% CI 13–31) for those HIV positive (p=0.0001). Using Cox proportional hazard models, the adjusted relative hazard of death for individuals co-infected with HIV compared to those infected with HCV alone was 19.47 (95% CI 9.22–41.10); 0.99 (95% CI 0.39–2.53), 3.47 (95% CI 1.40–8.63), 9.74 (95% CI 3.91–24.26) for the age groups at infection 10–19, 20–29, >30 years respectively compared with age group <10 years and 2.7 (95% CI 1.36–5.15) for genotype 1 compared to other genotypes. Whilst 25 year follow up of 310 haemophilic patients has shown the potentially lethal combination of HIV and HCV co-infection, HCV singly infected individuals show slow progression of liver disease.

13. MEMORY CD8+ CELLS REVERT TO A MORE STABLE CD45RA+ PHENOTYPE AND RESIDE WITHIN THE LIVER; AN ALTERNATIVE FATE TO ACTIVATION INDUCED CELL DEATH

The intrahepatic T cell population remains poorly characterised both in normal and inflamed liver and little is known of the fate of liver infiltrating T cells.

Methods & results: We examined freshly isolated liver derived and autologous peripheral blood (PB) T cells from normal donors and patients with cirrhosis due to chronic hepatitis C or PBC. T cell phenotype was established using 4 colour flow cytometry and antibodies to T cell receptors (CD3,CD4,CD8), differentiation markers (CD45RA,-RB,-RO), adhesion molecules, chemokine receptors (CR) and intracellular cytokines. T cells from inflamed livers were more terminally differentiated (CD45RBdull) and poorer producers of IFN-γ than T cells isolated from normal liver. CD45RA+CD8+ cells were consistently found within both normal and inflamed liver. This subset was L-selectinlowCD69highCCR7lowCCR5highgiving it a distinct and previously undescribed phenotype that distinguishes it from CD45RA+ cells in PB. The integrin LFA-1, expressed at high levels on antigen primed T cells, was biphasic on PB CD45RA+CD8+ cells but detected at high levels on all liver derived CD8+CD45RA+ T cells. PB CD8+CD45RA+LFA-1high cells were indistinguishable from primed (CD45RO+) CD8 T cells in terms of a) CR usage (CCR7lowCCR5highCXCR3highCXCR4low), b) intracellular expression of IL-2 & IFN-γ_ c) telomere length. CD8+CD45RA+ cells were detectable at only low frequency among T cells isolated from normal donor lymph nodes taken from the porta hepatitis, suggesting preferential homing to liver rather than local draining lymph node.

Summary: 1) Defective secretion of IFN-γ by terminally differentiated T cells in HCV infected liver may favour viral persistence 2) We describe a novel subset of CD8+CD45RA+LFA-1high T cells in human liver. We propose that this subset arises from antigen-primed CD45RO+ cells that have reverted to a more stable, resting phenotype and which contain long term T cell memory.

Conclusion: Memory “reversion” may provide an alternative fate for antigen primed T cells, rather than clonal senescence and activation induced apoptosis within the liver.

14. TUMOUR NECROSIS FACTOR α GENE POLYMORPHISMS AND DECOMPENSATED ALCOHOLIC LIVER DISEASE (ALD)

Background: Candidate genes to explain the variability in predisposition to ALD include that for tumour necrosis factor (TNF)α Biallelic promoter-region polymorphisms of the TNFA gene exist at positions -238 and -308; the rarer allele of the -308 polymorphism has greater promoter activity than the common allele. Data on allelic frequencies of these polymorphisms in ALD are limited and, in the case of -308, conflicting.

Methods: We studied 2 groups of Caucasian heavy drinkers (>60U/wk(M) or 40U/wk(F) for >5yr) (1): Patients (n=151; 105M, age 48±SD10yr) had decompensated liver disease (Childs Grade B or C; Maddreys DF 30.2±30.6), a more severe phenotype than in other studies; other liver diseases were excluded by serum testing and, in 47 cases, by liver histology (2): Controls (n=94, 73M, age 48±9yr) had no clinical liver disease, normal serum bilirubin, albumin and prothrombin time, no abnormality or fatty infiltration on ultrasound, and in 78 cases, a PGA index of ⩽5 (90% exclusive of cirrhosis; Gastro 1991;

100:1397). Subjects were genotyped at theTNFA -308 and -238 loci using the Taqman system.

Results: see table; 1=common 2=rarer allele. Summary: The association between decompensated ALD and carriage of the high-production allele of TNFA-308, is consistent with a role for genetic variation in TNFα production in pathogenesis of ALD.

Abstract 14, Table 1

15. OXIDATIVE STRESS IS RELATED TO HEPATIC FIBROSIS RATHER THAN NECROINFLAMMATION IN CHRONIC HEPATITIS C (HCV)

Introduction: HCV causes hepatic inflammation, necrosis and fibrosis by cytopathic and immunologically mediated mechanisms. Free radicals, generated by inflammatory or non-inflammatory processes, may activate stellate cells leading to fibrosis. Inflammation and fibrosis can be assessed both histologically (necroinflammatory score and fibrosis stage) and using serum markers (ALT and pro-collagen peptide, PIIINP). We examined their correlation with oxidant stress. Patients and methods: Fasting blood and urine samples were obtained from 29 patients with HCV. PIIINP was determined by radio-immunoassay (Orion Diagnostica). Lipid peroxidation was assessed by serum malondialdehyde (MDA) and urinary 8-isoprostane (8-IP). Serum antioxidants and liver function tests were measured by standard methods. All values expressed as means: normal range (NR); p values by Spearman rank. Paraffin embedded liver biopsy sections were graded using the Ishak histological activity index.

Results: PIIINP (6.41:NR 1.9–4.2μg/l) was elevated, selenium (82.7:NR 86.7–110.3μg/l) and vitamin A (2.49:NR 2.8–4.7μmol/l) were decreased and vitamin E (9.88:NR 9.8–15.7 mg/l) within the normal range. Lipid peroxidation markers were significantly raised (MDA 2.73:NR 1.4–2.24 μmol/l; 8-IP 0.89:NR 0.04–0.32μg/g creatinine). Markers of oxidant stress did not correlate with histological necroinflammatory activity. Histological fibrosis correlated significantly with both PIIINP (p<0.005) and 8-IP (p<0.05); negative correlations were seen with the antioxidant vitamins A (p<0.001) and E (p<0.05). PIIINP correlated negatively with selenium (P<0.01) and vitamin A (P<0.05).

Conclusions: Our results show that, in HCV, oxidant stress is related to the stage and progression of fibrosis but not to necroinflammatory activity. Fibrosis may therefore be linked to the generation of free radicals by mechanisms other than necroinflammation such as steatosis and iron overload which are known to be features of hepatitis C.

16. CROSS-REACTIVITY TO LACTOBACILLUS DELBRUECKII/HUMAN PYRUVATE DEHYDROGENASE COMPLEX-E2 CHARACTERISES PRIMARY BILIARY CIRRHOSIS

We have previously suggested (Nature1992;358:377–8) that molecular mimicry could account for antimitochondrial cross-reactivity targeting the inner lipoyl domain of pyruvate dehydrogenase Complex E2 subunit (PDC-E2), the major autoantigen of primary biliary cirrhosis (PBC). Reactivity to biotinylated peptides (Mimotopes Ltd, Victoria, Australia) spanning homologous sequences of the inner lipoyl domain PDC-E2213-227 core sequence—KLSEGDLLAEIETDK—and microbial sequences from proteins unrelated to PDC (table 1[t3], similar aa underlined in bold) and an irrelevant control peptide was tested by ELISA in 50 PBC patients (median age 56, range 26–84, 43 female) all AMA positive (median titre 1/640, range 1/40–1/10240), and 120 pathological and 20 demographically matched healthy controls.

Double reactivity to the PDC-E2 peptide and the peptide from the lactase protein of Lactobacillus delbrueckii(subsp. bulgaricus) was found in 29/50 (58%) PBC patients, but only in 3/120 (2.5%) pathological and in none of the healthy controls (p<0.001, for all). Double reactivity toLactobacillus delbrueckii/PDC-E2 homologous peptides was significantly more frequent compared to double reactivity with other microbial antigens sharing sequences with human PDC-E2 (p<0.001 for all). Pre-incubation of double reactive sera with solid phase lactase Lactobacillus delbrueckiipeptide, PDC-E2 peptide, recombinant PDC-E2 protein (Pharmacia & Upjohn, UK), control peptide or control protein (cytochrome P450IID6, Pharmacia & Upjohn, UK) decreased reactivity to theLactobacillus delbrueckii peptide by up to 87%, 79%, 82%, 8%, and 11%, respectively. Our results show that cross-reactivity between lactase of Lactobacillus delbrueckii/human PDC-E2 is highly specific for primary biliary cirrhosis.

This work was supported by the Edgar Ernest Grant-PBC Foundation, UK

Abstract 16, Table 1

17. AUTOANTIBODIES TO SOLUBLE LIVER ANTIGEN/UGA SUPPRESSOR TRNA-ASSOCIATED ANTIGENIC PROTEIN ARE SPECIFIC TO AUTOIMMUNE LIVER DISEASE

Background: Antibodies to soluble liver antigen (SLA) have been suggested to define type 3 autoimmune hepatitis (AIH). SLA is a 50 kDa cytosolic protein comprising of 422 amino acid, the sequence of which has been recently deposited and is 99% identical to the sequences encoding UGA suppressor tRNA-associated antigenic protein (Accession No. AJ238617), an antigen reported to be target of autoimmunity in AIH type 1. Aims &

Methods: To investigate the potential diagnostic value of antibodies to conformational epitopes of SLA/UGA suppressor tRNA-associated antigenic protein, we established a sensitive radioligand assay using the cDNA sequence of this protein to express the target antigen eukaryotically. Sera from 61 patients with AIH (35 type 1 and 26 type 2), 37 with other autoimmune liver diseases [17 anti-nuclear and/or anti-smooth muscle (ANA/SMA) antibody positive sclerosing cholangitis and 20 primary biliary cirrhosis], 82 with non autoimmune liver diseases, 46 with non hepatic autoimmune disorders, and 40 healthy controls were tested.

Results: Reactivity to SLA/UGA suppressor tRNA-associated antigenic protein was present in 54% patients with type 1 AIH, 31% with type 2 AIH, 41% with ANA/SMA positive sclerosing cholangitis, but in none of those with primary biliary cirrhosis or virus induced liver disease. Only 2 patients with non-autoimmune liver disease and 2 with non-hepatic autoimmune disorders were positive. None of the healthy controls was positive.

Conclusions: Reactivity to SLA/UGA suppressor tRNA-associated antigenic protein is strongly linked with AIH and ANA/SMA positive sclerosing cholangitis, but does not identify a third type of AIH. This work was supported by the Children's Liver Disease Foundation, Birmingham, UK

18. SERUM MARKERS PREDICT HISTOLOGICAL LIVER FIBROSIS

The current method for diagnosing hepatic fibrosis is by the histological analysis of a liver biopsy, but this may be hazardous and is subject to sampling error. We have therefore developed a panel of serum markers of liver fibrosis.

Methods: We have investigated the relationship between serum levels of ten markers and histological liver fibrosis in 605 patients undergoing liver biopsy for the investigation of liver disease at 13 centres. All biopsies were analysed locally and by 3 central pathologists using the HAI and Scheuer scoring systems. Monoclonal antibodies were used to detect 10 markers of fibrosis in sandwich immunoassays performed in an automated analyser employing fluorescein labelled capture antibodies and alkaline phosphatase labelled detection antibodies. The correlation between panels of the serum marker concentrations and histological fibrosis score was investigated in 154 samples to determine the optimal combination of markers that predict fibrosis in discriminant analysis. The performance of this panel was then tested in a 451 further samples.

Results: We have developed two algorithms incorporating the immunoassays, one for the Scheuer scoring system and one for the modified HAI (Ishak) scoring system. Biopsy scores were grouped into two categories, “mild” (Scheuer 0–1) and “moderate/severe” (Scheuer 2–4) fibrosis to distinguish clinically important categories. Rates of agreement between serum markers and histological fibrosis were 77 % and 74 %, respectively for Scheuer and HAI, yielding kappa values of up to 0.41 and 0.40, respectively (p<0.0001). Further analysis was performed by dividing biopsy scores into paired categories over a range of scores and calculating the corresponding receiver operating characteristic curves. The area under curves obtained ranged between 0.74 and 0.91.

Conclusions: We have shown that an algorithm of serum markers of liver fibrosis can accurately predict the extent of liver fibrosis on liver biopsy. Importantly this algorithm can accurately detect the development of clinically significant liver fibrosis (Scheuer stage 2–4 compared with stage 0–1). We suggest that the algorithm of serum markers will be a useful tool in the assessment and management of chronic liver disease.

19. HIGH VOLUME HAEMOFILTRATION IN ACUTE LIVER FAILURE: A BRIDGE TO TRANSPLANTATION

Background: Established acute liver failure (ALF) is associated with severe acidosis, rapidly progressive haemodynamic instability and multiple organ systems failure. In many cases deterioration is so rapid that liver transplantation is impossible. We have examined the use of high volume veno-venous haemofiltration (HVHF) in critically ill patients with advanced ALF to achieve metabolic and haemodynamic stability and enable successful transplantation.

Patients and Methods: 18 patients fulfilling transplantation criteria with paracetamol induced ALF were studied. Median age was 31 yrs (range 19–59), INR 4.6 (1.7–15), AST 4662 (455–17080), pH 7.28 (7.01–7.49), lactate 9.6 mMol/l (3.1–23) and APACHE II 25 (18–37). 12 patients (67%) were receiving vasopressor support with noradrenaline at 0.1 mcg/kg/min (0.03–0.5) and all were in anuric renal failure. HVHF was commenced on the second day of admission after using buffer free dialysate at 4000ml/hr (3500–6000) with concurrent NaHCO3 infusion and AN69hf filter (Hospal). Comparison was made with 18 historical controls who received conventional haemodiafiltration on the second day of admission at a rate of 1000 ml/hr (800–2000) matched for age, sex, liver injury, and APACHE II score. Haemodynamic and biochemical variables were compared 6 hours before filtration and at 6 hour intervals after therapy had been commenced. Statistical comparison utilised non-parametric tests, and figures show mean (SEM).

Results: HVHF resulted in a rapid correction of pH (p<0.006) (fig 19-1) with significant reductions in both serum lactate and base deficit within 12 hours, occurring more rapidly than those seen in controls. Over the study period mean arterial pressure remained stable in both patients and controls. However the proportion of HVHF patients requiring vasopressors fell and that of controls rose; at 24 hours an increased proportion of controls required vasopressor support (40% v 88%), and their median vasopressor requirement (0.15 v 1.7 mcg/kg/min) was significantly higher (p<0.05, fig 19-2). A trend toward a reduction in the incidence of cerebral complications was also noted in HVHF patients. Nine of 18 (50%) HVHF patients survived to undergo transplantation as compared with 4 of 18 (22%) controls.

Figure 19-2

Vasopresser requirement

Conclusions: The use of HVHF is associated with increased metabolic and haemodynamic stability in the critically ill ALF patient awaiting transplantation. It uses readily available technology and appears practical and effective in supporting the patient to eventual transplantation.

20. COMBINATION TREATMENT WITH LAMIVUDINE PLUS INTERLEUKIN-12 SHOWS GREATER ANTIVIRAL ACTIVITY IN CHRONIC HEPATITIS B

Background/Aim: Lamivudine has potent antiviral activity against HBV by directly inhibiting viral DNA synthesis. Interleukin-12 plays a critical role for effective immune control of viral infections by stimulating cellular immune responses and interferon-γ (IFNγ) production. The aim of our study was to evaluate the antiviral and immunostimulatory activity and safety of combination treatment with lamivudine plus recombinant human interleukin-12 (IL-12).

Study design/Methods: 15 patients with chronic hepatitis B (all HBeAg+/HBV DNA+) were randomised into 3 groups for 24 weeks of treatment: Group 1: lamivudine (LAM) monotherapy 100 mg daily; Group 2: LAM plus IL-12 (200 ng/kg s.c. twice weekly) beginning with LAM alone between baseline and treatment week 4 (TW4), LAM plus IL-12 between TW4 and TW16 and IL-12 alone between TW16 and TW24;Group 3: LAM plus IL-12 (500 ng/kg) scheduled as in Group 2. Serum HBV DNA levels were monitored by bDNA assay (Chiron) and the number of viral copies/ml was quantitated by PCR (NGI assay). HBV-specific T cell reactivity was assessed prospectively by enumeration of IFNγ-producing helper (CD4+) and cytotoxic (CD8+) T lymphocytes (ELISPOT assays) and T cell proliferation. Phenotypic activation markers were monitored by flow cytometry.

Results: Patients receiving LAM plus IL-12 showed greater reduction in HBV DNA copies/ml, compared with those on LAM monotherapy—the mean log10 reduction between TW4-TW16 in Group 3 was 2.2 (range 1.0–2.9); in Group 2 was 1.4 (0.9–2.8) and in Group 1 was 0.6 (0.02–1.1). This enhanced antiviral effect was not associated with significant ALT rises, only 1/10 patients developed hepatitis flare (ALT>3x baseline) during IL-12 treatment. All patients receiving LAM plus IL-12 showed a marked increase in the number of HBcAg-specific, IFNγ-producing CD4+ T cells, which paralleled the increase in T cell proliferation. The frequency of IFNγ-producing CD8+ T cells (core epitope 18–27) also increased 5 to 8 fold in comparison with baseline. The expression of activation markers (CD25, HLA-DR) on CD4+ T cells showed an early and dose-dependent increase after starting IL-12.

Conclusions: Interleukin-12 potentiates the antiviral effect of lamivudine and stimulates HBV-specific T cell reactivity without increasing hepatocellular damage. This combination has greater antiviral activity against HBV and may be useful for treatment of patients with chronic hepatitis B.

21. THE ADAMS: A MODULAR PROTEIN FAMILY WITH METALLOPROTEINASE AND PROTEOGLYCAN DEGRADING ACTIVITIES EXPRESSED BY THE HEPATIC STELLATE CELL

Liver fibrosis is associated with a change in the composition of the extracellular matrix from basement membrane type IV to interstitial type I collagen. This change plays a key role in the activation of hepatic stellate cells (HSC), and it has been attributed to the secretion of matrix metalloproteases (MMPs) by Kupffer cells and HSC themselves. However, HSC also secrete tissue inhibitors of metalloproteases (TIMPs) to which all MMPs to date are sensitive. The recently described ADAM family of MMP-related enzymes are alternative candidates responsible for matrix remodelling in liver fibrosis since they possess type IV collagenase activity and are largely TIMP insensitive. They have other potentially fibrogenic effects including proteoglycan-degrading activity, which could release TGFβ from decorin in the ECM, and TNFα convertase (TACE) activity. Both TGFβ and TNFα activate HSC in vitro. We have therefore studied the expression of the ADAM family members, ADAM10 and ADAMTS1, in HSC.

HSCs were isolated from male Sprague Dawley rats. RNA analyses were carried out by RT-PCR using rat-(ADAM10) or mouse-specific primers (ADAMTS1). Proteins were analysed by Western blotting using specific antisera and ADAM10 activity was determined with fibrin zymography in the presence of TIMP-1. ADAM10 RNA, protein and enzyme activity were present in day 14 cells. ADAMTS1 RNA was present in day 2, 4, 8, and 14 cells, and the sequence of the PCR product was homologous to published sequences from other species.

The presence of ADAM10, which has type IV collagenase and TACE activity, and ADAMTS1, a putative proteoglycanase, in HSC suggests that these enzymes could play an important role in ECM remodelling and the release/activation of soluble factors known to play a role in the initiation and perpetuation of HSC activation.

22. THE KINETICS OF ACUTE HEPATITIS B VIRUS INFECTION IN HUMANS

We characterised the kinetics of acute hepatitis B virus (HBV) infection in humans, using data from seven patients acutely infected from a unique single source outbreak before seroconversion, incorporating both the incubation and clinical phases of the disease. From serial measurements of serum HBV DNA, we calculated viral doubling times, and estimated the time between infection and peak viral level. We have shown that HBV replicates rapidly with doubling times ranging between 2.2 and 5.8 days (mean 3.71.5 days). Shorter viral doubling times were associated with an increase in severity of acute hepatitis. After a peak viral load in serum of nearly 1010 HBV DNA copies/ml is attained, clearance of HBV DNA follows a two or three phase decay pattern with an initial rapid decline characterised by a mean half-life (t1/2) of 3.71.2 days, similar to the t1/2 observed in the non-cytolytic clearance of cccDNA for other hepadnaviruses. The final phase of virion clearance occurs at a variable rate (half-life of 4.8 to 284 days) and may relate to the rate of loss of infected hepatocytes. Free virus has a mean t1/2of at most 1.20.6 days. We estimate a peak HBV production rate of at least 1013 virions/day and a maximum production rate of an infected hepatocyte of 200 virions/day. At least 3.2 point mutations were deduced to be produced per day, vastly more than the possible number of different single base changes in the HBV genome. Hence, we would expect all possible single base changes to be produced/day at peak infection. This has important consequences for the development of mutations that may escape the immune response, and also for mutations that provide resistance to reverse transcriptase inhibitors.

23. INTERORGAN AMMONIA METABOLISM FOLLOWING ACUTE VARICEAL BLEEDING IN PATIENTS WITH CIRRHOSIS OF THE LIVER: HYPERAMMONEMIA IS CAUSED BY RENAL AMMONIA RELEASE

Background: Variceal bleeding in patients with cirrhosis of the liver stimulates ammoniagenesis and induces hepatic encephalopathy. Present therapy is primarily directed at colonic ammonia production. The present study was designed to determine the contribution of the kidneys and the splanchnic bed in whole body ammoniagenesis during an acute variceal bleeding.

Material and Methods: Eight patients with cirrhosis of the liver (2F/6M, mean age 48.7 (±3.5) yrs, mean Pugh 12.5 (±0.7), all ALD) that underwent an emergency transjugular intrahepatic portosystemic stent-shunt (TIPSS) insertion to control their variceal bleeding were studied 21.5 (±7.4) hours after the start of the bleed and had an estimated blood loss of 7.8 (±0.8) units. Blood was sampled before the shunt puncture/placement from a femoral artery, the right renal vein and a hepatic vein. Also, directly after TIPSS insertion, blood was sampled from the portal-drained-viscera; samples were taken from the portal vein, the left gastric vein, the splenic vein, the superior mesenteric vein (SMV), and the inferior mesenteric vein (IMV). Ammonia release (positive value) or uptake (negative value) was calculated as venous-arterial concentration differences. Renal V-A difference is expressed per two kidneys and hepatic V-A difference represents ammonia handling of the splanchnic area.

Results: Before TIPSS insertion, only the kidneys released ammonia, resulting in a significant difference between splanchnic area and renal ammonia handling (p=0.015). After TIPSS placement, the portal-drained-viscera significantly released ammonia, originating from the IMV and the SMV.

Conclusion: Hyperammonemia following acute variceal bleed in patients is mainly caused by renal ammonia release. This study confirms previous findings in a controlled post-simulated bleeding study. New therapeutic strategies that aim to diminish ammonia production after variceal bleeding should aim at renal ammonia production.

Abstract 23, Table 1 Venous-arterial ammonia differences (uM) and significance from zero

24. HEPATITIS C VIRUS INFECTED HEPATOCYTES AND APOPTOSIS

Aims: To investigate the relationship between the intrahepatic burden of HCV and chronic disease. Methods: We developed an intracellular staining technique based on isolation of hepatocytes from fresh tissue biopsies and staining using an antiserum directed against the HCV NS3 protein and FACs analysis of the infected cells.

Liver histology was analysed without knowledge of the intracellular staining results and a contemporary serum specimen was obtained to measure HCV RNA by PCR (Amplicor).

Results: Liver biopsies from 34 patients (24 with chronic HCV and 10 with other liver diseases) were studied. In HCV positive patients there was a wide variation in the proportion of cells expressing HCV antigens (2.5–42%). There was a positive correlation between the number of infected cells and viraemia but no correlation with hepatic histology. In the hepatitis C group, the number of apoptotic hepatocytes assessed by direct microscopy detection correlated inversely with histological damage (p<0.01). It is noteworthy that the number of apoptotic hepatocytes was higher in patients with HCV (mean: 6 cells/field) than in a HBV control group (mean: 2 cell/field)

Conclusions: In chronic hepatitis C the proportion of HCV infected hepatocytes is variable and the hepatic burden of HCV per se does not influence disease progression. In patients, the severity of chronic HCV infection was inversely related with apoptosis of infected cells, suggesting that HCV might inhibit apoptosis.

25. BILIARY ATRESIA IN THE BRITISH ISLES—OUTCOME OF A NATIONAL COHORT AFTER UP TO 7 YEARS

All cases of biliary atresia in the British Isles diagnosed between March 1993 and February 1995 have been followed prospectively. Analysis after a median follow up of 3 years showed that outcome was better, with higher overall survival and survival without liver transplantation, in surgical centres managing >5 cases yearly (McKiernan PJ, et al. Lancet2000;355:25–9). Aim: To describe the current outcome of a national cohort of children with biliary atresia.

Subjects: 93 children with biliary atresia diagnosed between March 1993 and February 1995. Median follow up period 5 years (range 0.25–7.3).

Results: 15 children (16%) have died. 10 died following unsuccessful Kasai portoenterostomy and 4 following liver transplantation. 39 (42%) have undergone liver transplantation at median age 1 year (0.5–6). 5 year actuarial post transplant survival is 90%.7 year actuarial survival without liver transplantation was 40% overall, and was more likely in children managed in centres treating >5 cases yearly, (51% vs 26%, P<0.002). Where the Kasai portoenterostomy was successful in clearing jaundice (n=50), 7 year actuarial survival without liver transplantation was 74%. Where the Kasai portoenterostomy was unsuccessful (n=41) only 1 child has survived without liver transplantation.

Conclusions:

  • Most children with biliary atresia will eventually require liver transplantation.

  • If the Kasai portoenterostomy is successful few children will need liver transplantation before 7 years old.

  • Children with biliary atresia should be managed in experienced centres.

26. TACROLIMUS/CERIVASTATIN INTERACTION STUDY IN LIVER TRANSPLANT RECIPIENTS

Introduction: Although there is evidence that statins may have favourable effects upon rejection of solid organ transplants, to date no studies have addressed the role of statins in ameliorating acute cellular rejection in liver transplantation. A randomised placebo controlled trial of the HMG-CoA reductase inhibitor cerivastatin commenced prior to transplantation and continuing for one year post transplant is underway.

Immunosuppression after liver transplantation comprises combination therapy with tacrolimus, azathioprine, and prednisolone. It is not known how these drugs and, in particular, tacrolimus affect the pharmacokinetics of cerivastatin and vice versa. The objective of this study, therefore, was to evaluate the single dose pharmacokinetics of cerivastatin in this patient population.

Methods: 8 patients (4f/4m; 20–60y) with steady-state tacrolimus levels after liver transplantation, received a single dose of 0.2mg cerivastatin sodium. Plasma concentrations of parent drug and active metabolites M-1 and M-23 were measured by HPLC with fluorescence detection. R

esults: Cerivastatin was well tolerated; no serious adverse events were observed. The following table summarises the key pharmacokinetic characteristics of cerivastatin (geom. means/gSD (range)) compared with data obtained from healthy young male subjects:

Abstract 26, Table 1

Mean cerivastatin and metabolite AUC and Cmax values were approximately 50% higher in liver transplant patients on tacrolimus compared with previous studies in healthy subjects. Cerivastatin and metabolite half lives remained unaffected, indicating that multiple once daily dosing will not lead to significant accumulation of the drug in these patients. Tacrolimus trough concentrations assessed before and after dosing of cerivastatin remained constant.

Conclusion: First pharmacokinetic interaction investigations in liver transplant recipients on tacrolimus steady state treatment receiving cerivastatin resulted in only moderately elevated cerivastatin systemic exposure, not being prohibitive for the use of cerivastatin in this patient population. The data are in contrast to pronounced interactions reported for most statins, including cerivastatin, when combined with cyclosporin.

27. POLYMORPHISMS OF IMMUNOREGULATORY GENES: A STEP TOWARDS INDIVIDUALLY TAILORED IMMUNOSUPPRESSION?

At present, immunosuppression regimes used in transplant medicine remain largely empirical since there is currently no way to predict individual dose requirement. Recently, functional polymorphisms have been described in several “immunoregulatory” genes, some associated with immune mediated liver diseases. The aim of this study was to determine if these polymorphisms could predict the risk of rejection and/or infection following orthotopic liver transplantation (OLT).

Methods: 195 patients transplanted consecutively on our unit with a follow-up >6 months were genotyped for: exon 1 CTLA-4 polymorphism, -592 IL-10, -590 IL-4,-238/-308 TNFα promotor polymorphisms; Q576R IL-4 receptor polymorphism. Patients were classified as “immune responders” if they had 1 or more episodes of acute rejection requiring treatment and no episodes of major sepsis. Their genotypes were compared with 2 groups of “immune non-responders” (those suffering major sepsis and those with no episodes of rejection) and 128 local healthy controls.

Results: Significant results were only seen for the CTLA-4 polymorphism. As expected, more patients than controls possessed at least 1 copy of the mutant G allele (62%v 46%, OR: 1.9[1.2–3.1], p=0.006). However this frequency was higher in the “immune responders” (73%, 59/81) than in either the “septic group” (53%, 31/58); OR: 2.3[1.5–4.8], p=0.021, or the non-rejectors (55%, 40/72); OR: 2.1[1.1–4.2], p=0.028. Smaller, non-significant differences were also seen for the IL4 polymorphisms and 2 TNFα polymorphisms.

Conclusions: The CTLA-4 polymorphism linked to susceptibility to various “immune” diseases is associated with an increased risk of rejection and a low risk of major sepsis following OLT. It seems likely that, in future, prior to OLT, an individual's immune responsiveness will be assessed via genotyping for several immunoregulatory genes and their immunosuppression “tailored” accordingly.

28. LONGITUDINAL ANALYSIS OF HCV RNA LEVELS IN 85 HIV-VE MEN WITH BLEEDING DISORDERS INFECTED WITH HCV

Despite the high prevalence of HCV infection worldwide, little is known about the natural history of hepatitis C virus (HCV) RNA levels over the entire course of infection. The aim of this study was to describe the natural history of HCV RNA levels in 85 HIV-ve men with bleeding disorders infected with HCV for up to 30 years. HCV RNA levels were measured in yearly serum samples using a branched DNA assay (Chiron, lower limit of detection 0.2 x 106 Eq/ml). Dates of first exposure to HCV were known for all 85 men and ranged from 1965–1984 (median 1977). The majority of men had haemophilia A/B, and were infected with HCV genotype 1 or 3. Genotypes could not be assessed in 13 (15.3%) patients. RNA levels increased over time: in the second year after exposure, 53% of individuals had undetectable levels and no patients had levels > 7 log10(Eq/ml). By 20 years, these proportions had changed to 23% and 32% respectively. The RNA level correlated strongly with both the ALT (correlations between 0.41 and 0.71, depending on stage of infection) and AST (0.20–0.51) levels, although moderate correlations existed with the other markers. Individuals with haemophilia A had significantly higher HCV RNA levels at 1–5 (p=0.09), 5–10 (p=0.001), 10–15 (p=0.001) and 16–20 (p=0.0002) years after first exposure to HCV than those with other disorders. Whilst there appeared to be some relationship between HCV genotype and HCV RNA levels, this became non-significant after excluding patients from the analysis who could not be genotyped. Relationships between the HCV RNA level and age at first exposure to HCV or bleeding severity were not significant. HCV RNA levels increase over the course of HCV infection and may be useful for monitoring HCV disease. However, studies with clinical outcomes are required to assess whether HCV RNA levels can provide useful prognostic information in addition to that provided by other markers, such as ALT.

29. EVIDENCE OF THE GENERATION OF HBV-SPECIFIC CD8+ CELLS IN SECONDARY LYMPHOID ORGANS DURING LAMIVUDINE

Background: Antiviral treatment with lamivudine in patients with chronic hepatitis B infection is associated with an increase of circulating HBV-specific T cells. The origin of these cells remains uncertain: efflux of HBV-specific T cells from the liver or reconstitution of new HBV-specific T cells from lymphoid organs are both possibilities. Methods: In a patient with chronic hepatitis B and hepatocellular carcinoma, the frequency, function, and phenotype of HBV-specific CD8+ lymphocytes from peripheral blood and liver before lamivudine, and at the time of transplantation from the blood, explanted liver, and a porta hepatis lymph node, 3 months after starting lamivudine, were studied using HLA-peptide tetrameric complex and flow cytometry analysis.

Results: The frequency of HBV-specific CD8+ cells was extremely low in the blood (0.004% of CD8+ cells), and undetectable in the liver, before lamivudine. On lamivudine, a similar increase in frequency of these cells was detectable in blood, liver, and lymph node (0.05–0.09% of CD8+ cells), with no sign of preferential distribution between the different anatomical compartments. Functionally, only HBV-specific CD8+ cells present in the lymph node were able to expand after antigen-specific stimulation. Phenotypic analysis of this HBV-specific CD8+ population showed a chemokine receptor pattern (CCR-3 and CCR5 -ve) different from cells derived from liver and blood (usually CCR-5 +ve).

Conclusions: Our data suggest that the reconstitution of the peripheral HBV-specific T cell repertoire during antiviral treatment does not derive from recirculation of cells sequestered in the liver but is likely to arise from a population of HBV-specific CD8 cells originating in secondary lymphoid organs. The potential of these cells to expand supports the rationale for immunotherapy based upon augmenting the HBV-specific immune response.

30. ACTIVATORS OF THE PXR PREVENT THE TRANS-DIFFERENTIATION OF HUMAN HEPATIC STELLATE CELLS: AN ANTI-FIBROTIC ROLE FOR THE HPXR?

Liver fibrosis is characterised by an accumulation of extracellular matrix protein that with increasing severity impairs the normal functioning of the liver. It is a common response to liver damage mediated by a variety of mechanisms including xenobiotic damage and viral infection. Hepatic stellate cells play a central role in the development and resolution of liver fibrosis. In response to liver damage, the stellate cells trans-differentiate to a myofibroblast-like (α-smooth muscle actin +ve) phenotype which is responsible for the majority of extracellular matrix protein deposition in liver fibrosis. Using an in vitro model of liver fibrosis, we have observed that several steroid hormones modulate the trans-differentiation of human hepatic stellate cells although the same effect was often stimulated by a steroid receptor agonists and antagonists. The PXR is a recently identified orphan nuclear receptor that regulates the inducible expression of cytochrome P450 3A sub-family genes in response to certain pregnane steroids and several drugs including rifampicin and metyrapone. We have examined the activation of human hepatic stellate cells in vitro in the presence of activators of the human PXR. RNA was prepared from activated human hepatic stellate cells and primers designed to amplify a 500bp region of the human PXR mRNA by RT-PCR gave rise to a single DNA band on agarose gels of the expected size. Fragments amplified from 2 human preparations were cloned and sequenced and were found to be identical to the published human PXR. Expression at the protein level was confirmed by Western blotting and identified a single of band at the predicted ∼50kDaltons in 3 separate activated human stellate cell extracts. Addition of PXR activators to 5 preparations of human hepatic stellate cell cultures—rifampicin, lovastatin, clotrimazole, and metyrapone—resulted in an inhibition in cell proliferation and an inhibition in the expression of α-smooth muscle actin after 15 days of culture. These data indicate that the PXR is expressed in human hepatic stellate cells and that it may anti-fibrogenic and a novel drug target for treatment of liver fibrosis as well as fibrotic diseases in other tissues.

31. ACTIVATION OF TGF-Β1 BY HSC IN VITRO AND CONSEQUENCES FOR CELL PROLIFERATION AND SURVIVAL

TGF-β1 is a potent profibrotic cytokine upregulated in liver fibrosis. It is secreted in a latent form with activation being critical in controlling its activity. Although HSC express latent TGF-β1 following liver injury the precise method of activation and cell types involved is unclear. Plasmin is required in a variety of cells. We have shown that HSC express urokinase plasminogen activator necessary for plasmin production. We hypothesise that HSC can autonomously activate TGF-β1 and thus regulate its profibrotic effects.

Conditioned media (CM) from culture activated primary and passaged rat HSC were assayed for active and latent TGF-β1 using an ELISA. HSC proliferation was assessed by 3H-thymidine incorporation and apoptosis by acridine orange staining and assessment of nuclear morphology. Collagen synthesis was determined by collagenase-sensitive3H-proline incorporation per μg of DNA. Components of the plasminogen activating system were added to the CM and the effects determined.

Primary HSC cultured for 3–14 days secreted and activated TGF-β1, with only 1.5–5% in the active form, demonstrating that HSC can activate TGF-β1 without accessory cells. Exogenous plasmin increased the amount of active TGF-β1 by 40–300%, this being inhibited by the serine protease inhibitor aprotonin. Exogenous plasminogen also increased active TGF-β1, with plasminogen and plasmin increasing total TGF-β1 in the CM. This may reflect direct stimulation of cell production or cleavage from extracellular matrix. In primary and passaged HSC, plasmin consistently increased DNA content, mean±SE 301±37% of control (p<0.0001), whereas plasminogen had a smaller effect, 140±14% (p<0.05). Activation of TGF-β1 by HSC may also be important in regulating HSC number, as active TGF-β1 (10ng/ml) inhibited HSC proliferation by 38%, decreased cellular DNA by 21% and increased apoptosis by 40.8% over 24 hours. TGF-β3 was also produced by primary HSC as determined by ELISA. TGF-β3 may be important in similar regulatory processes to TGF-β1, as addition of 10ng/ml to HSC inhibited proliferation by 37% and increased collagen synthesis to 173±31% of controls (p<0.05), compared with 216±38% for 10ng/ml of TGF-β1. We conclude that HSC activate TGF-β1 autonomously by a plasmin dependent mechanism. Although, TGF-β1 and -β3 enhance collagen synthesis, their effects on the cell cycle may constrain expansion of HSC number following liver injury.

32. PROTECTION AGAINST TAUROLITHOCHOLATE—INDUCED CHOLESTASIS BY MURICHOLATE, TAUROURSODEOXYCHOLATE AND DBCAMP: INVOLVEMENT OF INTRACELLULAR SIGNALING MECHANISMS

Background: Taurolithocholate (TLC)- induced cholestasis is partially protected by tauroursodeoxycholate (TUDCA). It has been suggested that the effect of TUDCA may involve activation of protein kinase C (PKC) but this has not yet been validated in TLC-induced cholestasis. The possible protective effect of muricholic acid (MA) and cAMP has not yet been studied in TLC-cholestasis.

Aim: to study: (i) the effect of MA on TLC-induced cholestasis (ii) the effect of DBcAMP on TLC-cholestasis (iii) the effect of agents affecting intracellular signaling on the hepatoprotection seen with TUDCA and MA.

Methods: Hepatocyte couplets were isolated as described previously (Hepatology 1991;

14:180), and their function assessed as canalicular vacuolar accumulation of the fluorescent bile acid, cholyl-lysyl-fluorescein (cVA of CLF) (Hepatology 1999;

29:471). The effects of the following agents on TLC-cholestasis and protection with TUDCA or MA were analysed: BAPTA/AM (Ca+2 chelator); Staurosporin (STA) and H7 (both PKC inhibitors); KT2750 (PKA inhibitor) and DBcAMP (PKA activator/Ca+2-elevating compound). All these agents had no independent effect on cVA of CLF in controls.

Results: TLC induced significant reduction of cVA of CLF, which was not affected by BAPTA/AM, STA, H7, or KT2750. TUDCA, MA, and DBcAMP all significantly protected against TLC-induced cholestasis. This effect was almost completely abolished by BAPTA/AM, STA, and H7 in couplets protected with TUDCA and significantly impaired in couplets protected with MA. BAPTA/AM significantly impaired the protection with DBcAMP. KT2750 had no significant effect on hepatoprotection with TUDCA, MA, or DBcAMP. The protective effect of TUDCA or MA was not affected by the presence of DBcAMP.

Conclusions: (i) hepatoprotection with TUDCA in TLC-cholestasis may involve activation of PKC (ii) MA had, similar to TUDCA, protective properties in TLC-cholestasis, and the mechanism may also involve, but to lesser extend, activation of PKC; (iii) PKA inhibition does not affect protection with TUDCA or MA; (iv) DBcAMP protects against TLC-cholestasis; (v) the effect of DBcAMP is most probably mediated by Ca+2 rather than by PKA activation.

33. ENA-78 SECRETION BY HUMAN LIVER BILIARY EPITHELIAL CELLS: WHAT ROLE DOES IT PLAY?

Epithelial derived neutrophil-activating factor-78 (ENA-78) is an ELR+ member of the CXC group of chemokines and a neutrophil chemotactic factor similar to IL-8. ENA-78 was originally thought to be exclusively expressed by epithelial cells, but it has since been shown to be produced by endothelial cells.

We have investigated whether ENA-78 is produced in the human liver and by which cells. Biliary epithelial cells (BEC), sinusoidal endothelial cells (SEC) and hepatocytes were isolated from human liver and cultured in vitro. All cell types cultured were found to secrete ENA-78 when unstimulated, with significantly increased levels after IL-1 or TNF stimulation. However, biliary epithelial cells secreted much higher levels of ENA-78 than sinusoidal endothelial cells and hepatocytes:

Abstract 33, Table 1

Also, biliary epithelial cells secreted higher levels of ENA-78 than IL-8, whereas for sinusoidal endothelial cells and hepatocytes levels of IL-8 were much higher than for ENA-78. This relatively high secretion of the neutrophil chemoattractants ENA-78 and IL-8 by biliary epithelial cells, suggests an important role in initiating an acute inflammatory response against ascending biliary infection, and provides a mechanism to explain the neutrophil infiltrate that is centred in intrahepatic bile ducts in allograft rejection. Both ENA-78 and IL-8 also have angiogenic properties which may be relevant to the appearance of neovessels in portal tracts during chronic inflammation.

34. REACTIVATION OF LATENT HEPATITIS B VIRUS INFECTION WITH HIV RELATED IMMUNOSUPPRESSION

Background: Cellular immune responses are important in the control of hepatitis B virus (HBV) infection and reactivation of HBV infection after apparent resolution has been described with immunosuppression. The effect of HIV related immunosuppression on HBV reactivation is not clear. Furthermore the effect of highly active anti-retroviral treatment (HAART) on HBV reactivation has not been described.

Method: Of 211 HIV infected patients seen with evidence of prior HBV infection (HBcAb positive) but without evidence of HBV carriage (HBsAg negative), we have identified 4 cases with HBV reactivation.

Results: All 4 HBcAb positive cases became HBsAg positive having been previously HBsAg negative. All 4 were male, 2 were Caucasian and 2 African. Three had AIDS and 1 had symptomatic HIV. In all 4 cases, HBV reactivation occurred in the context of advanced HIV related immunosuppression. HBV reactivation was recognised because of repeat serological testing to investigate abnormal liver function tests. One patient died of non-liver related problems 10 months after HBV reactivation. The other 3 patients are alive and 2 have gone on to clear HBsAg with the use of HAART that did not include lamivudine. In 2 cases HBV reactivation was recognised at the time of an acute hepatitic illness developing several months after commencing HAART when CD4 counts had increased and HIV viral load fallen. The hepatitic illness preceded re-clearance of HBsAg in both cases. In the remaining 2 cases, who were not receiving HAART at the time, HBV reactivation resulted in only minor liver enzyme abnormalities.

Conclusions: Reactivation of apparently resolved HBV infection can occur with HIV related immunosuppression and can be a cause of liver enzyme abnormalities. The development of an acute hepatitic illness and re-clearance of HBsAg whilst on HAART regimens not including HBV replication inhibitors suggests that HBV reactivation had gone unrecognised until the immune reconstitution with HAART once again permitted immune mediated viral control and clearance.

35. APOPTOSIS OF HEPATIC STELLATE CELLS (HSC) CRITICALLY CONTRIBUTES TO RECOVERY FROM BILIARY FIBROSIS

Introduction: Activated HSC are central to the pathogenesis of liver fibrosis, both as a source of the fibrillar collagens that characterise fibrosis and matrix degrading metalloproteinases and their inhibitors, the TIMPs. In this study we have investigated whether HSC apoptosis is involved in recovery from biliary fibrosis.

Methods: Bile duct ligation (BDL) was performed on 15 rats. After 21 days of BDL, 13 animals underwent a further laparotomy at which an R-Y choledochal-Jejunostomy was performed to effect biliary drainage. Livers were harvested at 21 days post BDL (peak fibrosis, n=2) and at 1 day (n=3), 2 days (n=4), 7 days (n=4) and 42 days (n=2) of recovery post biliary-jejunal anastamosis. Sham controls (n=4) were also harvested at 21 days.

Livers were stained with H&E, Sirius red, reticulin, and α-smooth muscle actin (αSMA). Liver sections were TUNEL stained as a marker of apoptosis. Nonparenchymal TUNEL positive figures were then counted (50 fields of X40 per slide) by a blinded observer. In addition, dual staining for α-SMA and TUNEL was undertaken in representative liver sections.

Results and conclusions: Following biliary reanastamosis, 21 days after ligation, a progressive resolution of biliary fibrosis was apparent. In association with this resolution there was a 5-fold decrease in activated HSC determined by αSMA staining. TUNEL staining indicated that the loss of activated HSC resulted from an increase in the rate of apoptosis. HSC apoptosis thus plays a critical role in the spontaneous recovery from biliary fibrosis.

36. A PROSPECTIVE STUDY OF LIVER DYSFUNCTION IN PREGNANCY IN SOUTH WEST WALES

Background and aims: Liver disease is uncommon in pregnancy but has serious consequences. Its frequency varies according to country and ethnic origin and is not documented in Wales. We have prospectively determined incidence, causes, and outcome of liver dysfunction in pregnancy in an obstetric unit in South Wales. Methods: A central laboratory identified all abnormal liver tests from patients in antenatal clinics and wards of an obstetric unit serving a population of 250 000 between April 1999 and March 2000. Abnormal liver tests were defined as bilirubin >25μmol/L, AST >40U/L or γGT >35U/L. Patients were studied prospectively on receipt of blood results and medical advice was provided to obstetricians. Clinical course of mother and fetus/infant was recorded.

Results: During the twelve-month study there were 3511 deliveries. 102 patients had abnormal liver tests: elevation of AST in 84 (median 69;range 41–4123), γGT in 39 (51;36–278) and bilirubin in 15 (29;26–155). Thrombocytopaenia was seen in 28 (118; 23–149 x 109/L), raised uric acid in 21 (0.45;0.41–0.70 μmol/L) and raised bile acids in 13 (51;17–179 μmol/L). Hepatobiliary ultrasound scan (USS) was abnormal in 6. Liver biopsy was not justifiable in any patient. Causes of liver dysfunction were determined by symptoms, clinical circumstances, laboratory tests, and USS. There were 147 diagnoses amongst 102 patients: pre-eclampsia 55, intrahepatic cholestasis of pregnancy (ICP) 15, HELLP syndrome 14, sepsis 12, post Caesarean section 20, hyperemesis gravidarum 8, bile duct stones 2, placental pathology 8, drugs 4, acute fatty liver 2, diabetes 5, hepatic haematoma 1, and chronic hepatitis C 1. Elevation of AST was the predominant abnormality in ICP. There were no maternal deaths but there was one intrauterine death at 30 weeks associated with pre-eclampsia. 34 patients required delivery by induction or Caesarean section to improve hepatic function. 21 babies required admission to special care unit.

Conclusions: Liver dysfunction was seen in 102 of 3511 pregnancies during a twelve month period giving an annual incidence of 1 in 34 pregnancies. Transient but sometimes serious effects on maternal and infant morbidity were observed but there were no maternal deaths and only one stillbirth.

37. NEO-ANGIOGENESIS AND LYMPHOID ORGANISATION WITHIN THE LIVER IN CHRONIC HEPATITIS C INFECTION

Lymphoid aggregate formation within portal tracts has long been recognised in association with chronic hepatitis C infection although the function of such structures remains unknown. We used immunohistochemistry to examine cryopreserved liver sections from normal liver donors and patients with HCV cirrhosis who had undergone liver transplantation. A panel of monoclonal antibodies were used to identify T cells, B cells, dendritic cells, and vascular endothelial structures within liver. The expression of chemokines (IL-8, IP-10, and MIG) known to influence angiogenesis were also examined. T cells were identified within HCV infected liver predominantly within expanded portal tracts. CD4+ cells were evenly distributed throughout portal areas whereas CD8+ cells were mainly located at the interface between portal tract and hepatocyte lobules. Lymphoid aggregates were present within portal tracts in the majority (19/20) of livers examined. Aggregates comprised a central core of B cells (CD45RA+CD40+CD22+CD20+) surrounded by CD4+ and CD8+ cells. Using confocal microscopy, an interdigitating network of dendritic cells (MAdCAM-1+) was identified within these lymphoid structures, as were a smaller population of activated (CD83+) dendritic cells. Endothelial markers (E-selectin, P-selectin, CD31, CD34, CD36 VAP-1) identified neovessels within the inflamed portal tracts, often in close proximity to lymphoid aggregates that were not present within normal donor liver. Portal vascular endothelium strongly expressed the angiogenic chemokine IL-8 but not the angiostatic chemokines IP-10 and MIG, which were predominantly expressed by sinusoidal endothelial cells within hepatocyte lobules. In summary, the distribution of angiogenic and angiostatic chemokines favours portal tract neo-angiogenesis. Lymphoid aggregate formation within inflamed portal tracts may result from local recruitment of T cells via such neovessels. The presence of a dendritic cell network within organised lymphoid structures would permit local antigen presentation and perpetuation of the chronic immune response in hepatitis C infection.

38. HEPATIC EXPRESSION OF SECONDARY LYMPHOID CHEMOKINE MAY PROMOTE THE RECRUITMENT OF α4β7+ LYMPHOCYTES INTO THE LIVER

Background: The interaction between mucosal adressin cell adhesion molecule (MAdCAM-1) on endothelium and the integrin α4β7 on lymphocytes activated in the gut, is crucial in establishing lymphocyte homing patterns. Secondary lymphoid chemokine (SLC) expressed predominantly on lymph node high endothelial venule (HEV) has been reported to stimulate α4β7 mediated lymphocyte adhesion to MAdCAM-1 under flow and blockade of the SLC receptor, CC chemokine receptor 7 (CCR7) selectively inhibits T-cell recruitment to Peyer's patch HEVs. Others and we have described the presence of functionally active MAdCAM-1 on endothelium in chronic inflammatory liver diseases but the induction of SLC in liver disease has not been reported.

Method & results: SLC was detected in human liver tissue by immunohistochemistry. SLC was detected in association with dendritic cells in portal tracts in normal liver and on hepatic portal vein endothelium in chronic inflammatory liver disease. We used flow cytometry of liver derived lymphocytes to determine the expression of the SLC receptor CCR7 on liver infiltrating T cells. A small population of CD3+CCR7+ T cells were detected in normal liver whereas in primary sclerosing cholangitis up to 20% of T cells were positive. Because SLC is a potent chemotactic factor for naïve T cells entering lymph node we also determined expression of the naïve cell marker CD45RA. Three colour FACS analysis demonstrated that 80% of the α4β7+ T cells in peripheral blood and 45% of the α4β7+ T cells in the liver were CD45RA+.

Summary: SLC is demonstrable on hepatic portal endothelium and in association with dendritic cells in chronic inflammatory liver disease. The expression of its receptor CCR7 on both peripheral blood and liver derived α4β7+/CD45RA+ lymphocytes suggests that SLC could be attracting this subset of T cells from the circulation and retaining them at sites of chronic inflammation within the liver.

39. SUBCELLULAR LOCALISATION OF THE X PROTEIN IN HBV INFECTED HEPATOCYTES

The X gene product (HBx) is the least well understood protein produced by the four open reading frames which constitute the HBV genome. Little is known about its functional mechanisms and although interactions with several nuclear and cytoplasmic proteins have been demonstrated in vitro, there is no consensus as to where HBx localises in infected hepatocytes. We have analysed the expression and intracellular distribution of HBx in human liver biopsies using an anti-HBx rabbit polyclonal antiserum. HBx was detected in a high proportion (69%) of samples from patients with chronic HBV infection. Detection of HBx correlated with the absence of cirrhosis and serum e-antigenaemia. HBx was predominantly detected in the cytoplasm; however, it was also found in the nuclei of up to 20% of positively stained hepatocytes, either exclusively nuclear or localised both in the nucleus and cytoplasm within the same cell.

Further, the intracellular distribution of HBx was analysed in transfected Huh-7 cells by confocal microscopy, using the monoclonal antibody 16F1. A substantial nuclear detection was confirmed in a significant proportion of HBx expressing cells. The cytoplasmic distribution was analysed in detail, suggesting that HBx is associated with mitochondria, but not with the endoplasmic reticulum, plasma membrane, or lysosomes. We did not observe apparent changes in transmembrane potentials.

These data provide novel insights into the compartmentalisation of HBx and may prove important for future evaluations of its functions. Clarification of the role of HBx in the viral life cycle and its mechanisms of action could lead to it becoming a novel target for antiviral therapy.

40. HEPATITIS C VIRUS (HCV) CELLULAR IMMUNE RESPONSES ARE INDUCED DURING COMBINATION THERAPY BUT ARE NOT SUSTAINED

Background: Treatment of chronic HCV infection with combination therapy (alpha-interferon and ribavirin) leads to sustained virological response in 30–65% of patients. Cross sectional studies suggest that enhanced T helper responses may be seen following interferon (IFN) therapy, suggesting that the mechanism of action of combination therapy includes augmentation of antiviral T cell responses

Aims: 1. To characterise prospectively, T helper (TH) and cytotoxic T lymphocyte (CTL) responses and γ-IFN production during therapy. 2. To identify important TH and CTL epitopes associated with viral control. 3. To identify mechanisms of viral persistence in patients who fail to control virus.

Method: T cell responses to HCV peptides and proteins were characterised prospectively in 16 patients, before, during, and after combination therapy (10 time points per patient). TH responses were analysed using proliferation and α-IFN ELIspot assays against HCV proteins (NS3, NS3/4, NS4, NS5) and pools of peptides 20 AA in length, spanning the whole of HCV core region. HCV specific CTL responses were analysed in HLA-A2 patients using γ-interferon ELIspot against a panel of 6 HLA-A2 restricted HCV epitopes.

Results: Of 13/16 patients who have completed therapy, 7 were HCV RNA negative at the end of treatment, 4 patients became RNA negative but relapsed during therapy, and 2 patients failed to become PCR negative at any time. Enhanced proliferative TH responses were seen in 12/13 patients. Maximum TH responses were observed 12 weeks after the onset of therapy and responses to HCV core appeared to dominate. TH γ-IFN production was demonstrated in 6 patients and was associated with virological response in 3 of these. Enhancement of TH responses were not associated with outcome: all of the patients who were RNA positive at the end of treatment mounted at least transient TH responses and 3 of these were associated with γ-IFN production. Of note, HCV specific CTL γ-IFN production could not be demonstrated in the HLA-A2 patients (0/5).

Conclusion: Combination therapy enhances HCV specific TH responses. However, these responses are usually transient and frequently not associated with γ-IFN secretion. Strategies that further augment and sustain T cell responses might have an important role to play in treatment success.

41. CO-LOCALISATION OF JAGGED 1 AND CYTOKERATIN 19 IN HUMAN FETAL AND PAEDIATRIC LIVER.

Mutations in Jagged 1 have recently been implicated in the pathogenesis of Alagille syndrome, resulting in paucity of bile ducts. However, altered expression of Jagged 1 may also be involved in pathogenesis of other cholestatic liver disorders, such as extrahepatic biliary atresia (EHBA), where there is abnormal bile ductular proliferation. To explore this further we have used double immunofluorescent staining with cytokeratin 19 (CK 19) and Jagged 1 in fetal, paediatric normal, EHBA and α1 antitrypsin deficiency (α1AT) liver. Frozen sections of fetal (10–16 weeks), normal (4–12 years), EHBA and α1AT livers were incubated with antibodies to Jagged 1 and CK 19, followed by secondary immunofluorescent antibodies for double staining. Co-localisation was visualised by confocal microscopy. Jagged 1 staining was seen in all tissues. In fetal liver, it stained portal vein and portal tract mesenchyme. At 10 weeks gestation co-localisation with the biliary epithelial cell (BEC) marker CK 19 was seen on a few cells within the ductal plate, which increased in number to include most ductal plate cells by 16 weeks gestation. In normal liver, Jagged 1 stained bile ducts, vascular endothelium, and hepatocyte membranes. By co-localisation with CK 19, Jagged 1 expression was restricted to the lumenal surface of BECs. Jagged 1 expression was more intense on all ductular proliferative cells and was seen over the entire cell. Expression of Jagged 1 mRNA by RT-PCR was confirmed in all tissues examined. The pattern of Jagged 1 expression in fetal liver highlights its importance in the normal development of bile ducts. The loss of lumenal localisation in ductal proliferative cells seen both in EHBA and α1AT indicates a loss of polarity in these cells. The upregulation of Jagged 1 and lack of polarity suggest that Jagged 1 may play a role in the abnormal ductular proliferation and failure of these cells to organise into lumenal ducts.

42. THE NEUROTROPHIN NGF DOWNREGULATES TRANSCRIPTIONAL ACTIVITY OF NFKB IN HEPATIC STELLATE CELLS

The resolution of experimental liver fibrosis is associated with the apoptosis of activated hepatic stellate cells (HSCs). This apoptosis may result from induction by specific ligands such as the neurotrophin nerve growth factor (NGF), which promotes apoptosis of HSCs by the activation of p75, a member of the death domain bearing TNF receptor superfamily. It has been suggested that cells may resist apoptosis via activation of the transcription factor NFκB. Since NGF potently stimulates apoptosis of HSCs we hypothesised that this may be associated with a downregulation of NFκB binding in the nuclei of HSCs.

In serum deprived, activated primary cultures, treatment with TNFα (30 mins, 10ng/ml) stimulated an increase in NFκB binding as determined by EMSA using a wild type NFκB probe. In contrast, NGF treatment (30 mins, 10ng/ml) was found to reduce basal NFκB binding activity and CAT reporter activity from an NFκB consensus sequence in parallel experiments (n=5). This effect was rapid and NFκB returned to pretreatment levels within 3 hrs. A combination of TNFα and NGF (10ng/ml for 30 mins) resulted in NFκB binding and reporter activity comparable to control (untreated with either NGF or TNF). Caspase 3 activity in HSC was significantly elevated by treatment with NGF and TNF for 24 hrs (5.0±1.5 and 5.7±1.2 fold control activity, respectively; n=4 p<0.05). Over the same period a combination of NGF and TNF resulted in a greater caspase 3 activity (10.4±3.1 fold control activity; n=4 p<0.05). Analysis of DNA from representative cultures confirmed that treatment with NGF and TNF induced oligonucleosomal fragmentation characteristic of apoptosis. Combined NGF and TNF treatment was associated with DNA fragmentation greater that that resulting from NGF or TNF alone.

These data demonstrate that NGF mediates downregulation of NFκB, an effect that is associated with apoptosis. Moreover, HSC apoptosis appears to be enhanced with both relative and absolute reduction in nuclear levels of activated NfκB.

43. CD40 AND CD40 LIGAND EXPRESSION IN HUMAN HEPATOCELLULAR CARCINOMA

Introduction: CD40 is a member of the TNF receptor superfamily, found on B lymphocytes, dendritic cells and tumour cells of epithelial origin. Its functions include an important role in apoptosis. We have previously shown that CD40 engagement on HepG2 cells, a primary liver tumour cell line, induces apoptosis. Hepatocellular carcinoma is heavily infiltrated by lymphocytes (TIL), most of which are T cells and there is tumour cell MHC class 1 expression, making it a potential target for the development of immunotherapeutic approaches to treatment.

Aim: To investigate CD40 and CD40ligand expression in human HCC.

Methods: Frozen tissue sections were stained using immunohistochemical techniques for CD40 expression. Fresh TIL were isolated from resected HCC and analysed by flow cytometry for CD40 and CD40ligand expression. TIL were expanded in high dose interleukin 2 in vitro and CD40 and CD40ligand expression analysed by flow cytometry in the expanded cells. PCR was performed to detect CD40ligand mRNA in expanded TIL.

Results: We have detected CD40 expression on HCC tumour cells and a proportion of infiltrating lymphocytes. In fresh TIL, a small proportion of the lymphocytes express CD40. These are non-T cell lymphocytes. CD40 expression is also present on a subset of cells which may be tumour dendritic cells. CD40ligand expression is absent in fresh TIL, but a small amount is detectable on expanded TIL. Expression of CD40ligand mRNA was confirmed in this population by PCR.

Conclusion: These results suggest that CD40 on HCC tumour cells is a potential target for immunotherapy. The expression of CD40ligand in IL-2 expanded TIL could have implications for adoptive immunotherapy strategies using autologous TIL.

44. PREVALENCE OF FATIGUE IN A GEOGRAPHICALLY BASED PRIMARY BILIARY CIRRHOSIS: COMPARISON WITH COMMMUNITY CONTROLS

Primary biliary cirrhosis (PBC) is an autoimmune chronic cholestatic liver disease. Previous patient surveys have suggested that fatigue is a symptom frequently experienced by PBC patients, and is an important quality of life issue. To date, however, studies of the true prevalence of fatigue, and its resulting impact on quality of life, have been hampered by difficulties in the quantification of fatigue and its impact on quality of life. Moreover, those studies which have been performed have largely been based on groups of patients attending tertiary referral clinics, an approach which may result in over representation of patients with advanced, excessively symptomatic or otherwise difficult to manage disease, with an attendant exaggeration of the true impact of fatigue. A failure to compare fatigue severity in patients with PBC with appropriate community recruited controls has also limited our ability to assess the extent to which fatigue as experienced by the PBC population is materially different to that experienced by the normal population. In this study we set out to apply a fatigue impact/severity assessment tool (the 40 item Fisk Fatigue Severity Score (FFSS)) to a geographically based PBC population and community recruited age/sex matched normal controls. The PBC study group consisted of the population of PBC patients resident in the NE1–25 postcode districts of Newcastle-Upon-Tyne in North East England (total n=202). Controls (2 per PBC patient) were chosen from the registers of 2 primary care practices in the same area and matched for age and sex with their index PBC patient. All subjects were asked to complete a postal questionnaire. For PBC patients this included the FFSS and the Rand MOS depression screener. Controls were asked to complete the FFSS. The FFSS has been previously validated for use in PBC patients (J Hepatol2000;32:368–73) and the Rand MOS has been validated in community based populations for self-completion. Relevant clinical data for all PBC subjects was obtained from their medical records. Median total FFSS was significantly higher for PBC patients (median 41 (max possible 160, range 0–150) than matched community controls (21 (0–119), p=0.001 (Mann-Whitney)). The FFSS contains three individual domains assessing cognitive (max 40), physical (max 40) and psychosocial (max 80) aspects of fatigue. Scores for all three domains were significantly higher for PBC patients than for controls (cognitive fatigue: 9 (0–39) v 4 (0–30), p<0.001; physical fatigue 16 (0–38) v 8 (0–40), p<0.0005; psychosocial fatigue 20 (0–73) v11(0–61), p=0.005). Depression scores for PBC patients showed only a weak positive correlation with FFSS scores (r2 = 0.33, p = <0.0001). Conclusion: Fatigue severity scores are significantly higher in a geographically based PBC population than in normal, community based matched controls. This study confirms the importance of fatigue as a symptom of PBC in general and not just the more selected populations seen in tertiary referral clinics. The apparent excess of physical fatigue in the PBC patient group is likely to be a consequence of the FFSS design rather than a true association. The findings of this study suggest only a limited contribution of depression to the fatigue reported by PBC patients.

45. UNDERCARBOXYLATION OF SERUM OSTEOCALCIN IN PRIMARY BILIARY CIRRHOSIS (PBC); REVERSAL BY VITAMIN K1

Background: Patients with PBC have lower bone mineral density than matched controls and 10–30% have osteoporosis. Vitamin K deficiency may occur in PBC and might contribute to osteoporosis, as carboxylation of osteocalcin (Oc), a major bone protein, implicated in mineralization, is vitamin K-dependent. We studied 12 patients (11F, age 59±9yr) with PBC. Fasting serum was taken at weekly intervals for 4 weeks, before (weeks 1,2) and during (weeks 3,4) vitamin K1 2 mg/day, given orally. Vitamin K1 levels were measured using electrochemical detection coupled to HPLC. Total and undercarboxylated (Uc) Oc levels were measured using commercial kits. Values were compared to those in healthy female postmenopausal controls (age 68±8yr).

Results: see table; mean+SD (number); all units μg/L apart from %UcOc. +: Mean of two baseline (week1,2) and post vitamin K (week 3,4) values.

Abstract 45, Table 1

Conclusions: We confirm that compared with healthy controls, patients with PBC have lower serum total Oc; they also have lower serum vitamin K1 and a higher serum undercarboxylated Oc fraction. Vitamin K1 2 mg/day orally is well absorbed in PBC and markely reduces serum undercarboxylated Oc. Vitamin K supplementation might thus help to prevent osteoporosis in PBC.

46. ALCOHOL-INDUCED OXIDATIVE STRESS CONTRIBUTES TO THE DEVELOPMENT OF IMMUNE REACTIONS ASSOCIATED WITH ALCOHOLIC LIVER DISEASE

Acetaldehyde and hydroxyethyl adducts with proteins have been shown to induce specific antibodies in alcoholic liver disease (ALD) patients. Proteins modified as a consequence of oxidative stress are also known to be immunogenic. As oxidative damage is associated with ALD, we investigated the presence of an immune response towards human serum albumin (HSA) adducts with several lipid peroxidation products in ALD patients. Antibody titres to HSA modified by reaction with malondialdehyde (MDA), 4-hydroxynonenal (HNE), arachidonic acid lipid peroxides (LOOH), oxidised cardiolipin and hydroxyethyl radicals (HER) were measured using ELISA in 50 pts with alcoholic and 40 with viral cirrhosis and 20 controls with low alcohol intake. A separate cohort of 51 biopsied heavy drinkers (28 with hepatitis +/- cirrhosis, 23 fatty liver only) was also studied. In the first cohort, IgG titres to all four lipid peroxidation-derived antigens were significantly higher (p<0.001) in alcoholic cirrhotics compared to non-alcoholic cirrhotics or controls, and were also associated with disease severity (Maddrey's DF index >90 v <90, p<0.002). ALD pts with high titres to one adduct, had high titres to all the others (inter-group correlation coefficients 0.54–0.78). In the second cohort titres against MDA, HNE, LOOH, and HER adducts were higher in pts with advanced ALD compared to those with fatty liver. The fatty liver pts and controls had similar antibody titres to all antigens except HER where titres were significantly higher in pts with fatty liver (p<0.05).

These results suggest that antigens resulting from the interaction of liver proteins with lipid peroxidation products contribute to the development of immune responses associated with the presence and severity of alcoholic liver disease. This immune response may contribute to disease pathogenesis.

47. COMPARING HLA DRB1 VALINE/GLYCINE-86 DIMORPHISM AND LYSINE-71 AS PRIMARY MHC-ENCODED SUSCEPTIBILITY DETERMINANTS IN PRIMARY SCLEROSING CHOLANGITIS AND TYPE 1 AUTOIMMUNE HEPATITIS: A STUDY OF 448 ADULT PATIENTS

Introduction: In type 1 autoimmune hepatitis (AIH) two different models of MHC-encoded disease susceptibility have been suggested both are functionally relevant. One of these, based on HLA DRB1 alleles encoding lysine at position 71, accounts for the diseases in adult patients and the second, more recent model, based on DRB1alleles encoding valine at position 86 may account for susceptibility in children, especially in South America. The two models suggest that the paediatric and adult forms of type 1 AIH though histologically similar may arise from the interaction of different environmental and genetic factors. A key difference between children and adults with type 1 AIH is the strong overlap with primary sclerosing cholangitis (PSC), indeed the HLA DRB1 alleles associated with paediatric AIH are those which are found in adult patients with PSC.

Aims: To test both lysine-71 and valine-86 as models of disease susceptibility in adult AIH and adult PSC.

Methods:HLA DRB1alleles were determined in a total of 298 adult patients with type 1 AIH and 150 adult patients with PSC using standard PCR-based techniques. The distribution of the DNA and amino acid sequences encoded by these alleles were compared with those of 236 healthy controls.

Results: The etiologic fraction (EF) which is a measure of the relative strength of association and describes the total load attributable to each gene, is used for comparing results. For type 1 AIH: the principal HLA DRB1associations were DRB1*0301 (EF = 0.35),DRB1*0401 inDRB1*0301 -ve patients (EF = 0.49). In PSC: they were DRB1*0301 (EF = 0.28) andDRB1*1301 (EF = 0.08). Although bothDRB1*0301 andDRB1*1301 encode valine 86, DRB1*0401 does not. Conversely DRB1*0301 andDRB1*0401 encode lysine 71 butDRB1*1301 does not. For allDRB1 sequences the lysine-71 model has the highest EF value in adult type 1 AIH (EF = 0.71, p<10-8) compared with valine-86 (EF = 0.396). In PSC however valine-86 has the highest EF value (0.62, p = 0.00004) whilst the EF value for lysine-71 was only 0.35 which is marginally higher than that forDRB1*0301 alone.

Conclusion: The lysine-71 model is the best model to account for type 1 AIH in adult patients accounting for 71% of the genetic risk. In contrast valine-86 which accounts for 62% of the genetic risk may be a good model for MHC-encoded disease susceptibility in PSC. This similarity in the immunogenetic predisposition to both adult PSC and paediatric-onset type 1 AIH may suggest a common pathogenesis or similar/related environmental triggers.

48. BLOCKING OF IL-10 RECEPTOR—A NOVEL APPROACH TO STIMULATE VIRUS-SPECIFIC T CELL REACTIVITY IN CHRONIC HEPATITIS C VIRUS (HCV) INFECTION

Background: The development of a strong and multispecific T cell response to HCV antigens is important for both spontaneous and treatment-induced resolution of HCV infection. Interleukin-10 (IL-10) is an immunosuppressive cytokine that may contribute to impaired virus-specific T cell responses present in patients with chronic hepatitis C.

Aim: To determine whether blocking IL-10 signalling would stimulate HCV-specific T cell reactivity in these patients and thus be of potential therapeutic benefit.

Methods: PBMC were isolated from 28 HCV RNA+ patients with chronic hepatitis C and incubated with HCV antigens (core, NS3, NS4, NS5) alone and in the presence of a monoclonal antibody to IL-10 receptor (MAb3C5.2) at doses of 0.1, 1.0, and 10.0 μg/ml. Antigen-specific T cell proliferation was measured by3H thymidine uptake. The number of HCV-specific, IFNγ producing CD4+ T cells was determined by an ELISPOT assay. IL-10 and INFγ levels in the supernatants were measured by ELISA.

Results: Significant T cell proliferation to at least one HCV antigen was found in 7/28 patients. With the addition of MAb3C5.2 the strength of this response was markedly increased in all 7 patients. Furthermore, MAb3C5.2 induced strong T cell reactivity, usually to more than one HCV antigen in another 12 patients, thus 19/28 patients showed significant HCV-specific T cell response with the IL-10R blockade (McNemar's test, p=0.0003). Control experiments with an isotype-matched MAb showed no effect on proliferation. At 10 μg/ml MAb3C5.2 increased the frequency of response to HCV core from 5/28 to 11/28 patients (p=0.02) and to NS3 from 1/24 to 6/24 patients (p=0.03). Increased proliferation in the presence of MAb3C5.2 was always associated with increased IFNγ levels in the supernatant, as well as an increased number of IFNγ producing T cells (ELISPOT assay). The addition of recombinant IL-10 to cultures with significant proliferation profoundly inhibited 3H-thymidine uptake. Importantly, this effect was reversed by the addition of neutralising antibody to IL-10, thus confirming the inhibitory activity of IL-10 on HCV-specific T cell proliferation.

Conclusions: These results provide the first evidence that selective blocking of IL-10 receptor results in potent activation of HCV-specific T cell reactivity. This approach may be useful to enhance the therapeutic response in patients with chronic hepatitis C.

49. CHARACTERISATION OF THE CELLULAR AND HUMORAL IMMUNE RESPONSES TO TT VIRUS (TTV) IN HEALTHY SUBJECTS AND PATIENTS WITH LIVER DISEASES

Background: TTV is a recently identified human DNA virus with a very high degree of genomic variability. The understanding of the role of TTV in human pathology has been complicated by the great heterogeneity of this new virus and detection of TTV infection varies considerably according to the primers used for amplification of TTV DNA. Furthermore, there is no information on the host immune response to TTV. The aim of this study was to investigate the presence and characteristics of T-cell and antibody responses to TTV in healthy subjects and patients with liver diseases.

Methods: Peripheral blood mononuclear cells and serum samples were obtained from 10 healthy subjects and 31 liver disease patients. We synthesised 40 overlapping peptides, corresponding to the amino acid (aa) sequences in open reading frame 1 of TTV genotype 1 (aa 451–770) and TTV genotype 2 (aa 451–520). These peptides were then used to test virus-specific T-cell proliferation (by3H-thymidine uptake) and antiviral antibodies (by ELISA). The presence of TT viraemia was sought by several sets of primers directed to coding and non-coding regions of the TTV genome. TTV DNA titre was determined by real time TaqManPCR.

Results: Significant TTV-specific T-cell proliferation (SI > 4) was found in 6/10 (60%) healthy subjects and 11/31 (35%) patients. Fine epitope mapping revealed an immunodominant region between aa 451–481, which was recognised by 14/17 (82%) cases with significant T-cell proliferative responses and showed cross-reactivity for TTV genotype 1 and 2. Importantly, this region is conserved amongst all other TTV genotypes. Antiviral antibodies to at least one TTV peptide were detected in 31/35 (88%) cases tested with the main B-cell epitope being between aa 571–590. The humoral response was usually broad, as reactivity to more than 10 peptides was found in 19/35 (55%) cases. Cases with high T-cell response showed significantly weaker antibody reactivity (p=0.004). Quantitation of serum TTV DNA by real time PCR using primers for a highly conserved non-coding region showed an inverse relationship between viral load and the breadth of antibody reactivity to TTV.

Conclusion: These results indicate for the first time considerable differences in the cellular and humoral immune responses to TTV and identify the corresponding immunodominant epitopes. A selection pressure from antiviral antibodies may have a role for generating the great diversity of TTV population.

50. DIRECT EVIDENCE FOR THE IMMUNOMODULATORY ACTION OF RIBAVIRIN ON T-CELL REACTIVITY TO HEPATITIS C VIRUS

Background: The addition of ribavirin to interferon-α treatment in patients with chronic hepatitis C has markedly increased the proportion of cases with a sustained therapeutic response. Experiments with mitogen-induced activation of peripheral blood mononuclear cells (PBMC) and indirectin vivo data suggest that ribavirin influences T cell proliferation and/or cytokine production.

Aim: To determine the effects of ribavirin on HCV-specific T cell reactivity and associated cytokine gene expression.

Methods: PBMC from 31 HCV RNA+ patients with chronic hepatitis C were incubated with recombinant HCV proteins (Core, NS3, NS4, and NS5) alone and in the presence of four concentrations of ribavirin (0.5, 1.25, 2.5, and 5.0 μg/ml). HCV-specific T cellproliferation was measured by3H-thymidine uptake. The effect of ribavirin on virus specific cytokine production was analysed by: (a) the level of IFNγ and IL-10 in the culture supernatants and (b) by direct enumeration of IFNγ and IL-10 producing CD4+ T cells using an ELISPOT assay. Cytokine mRNA (IFNγ, IL-10, IL-5, IL-12p35, IL-12p40, TNFα) levels in HCV-stimulated PBMC were quantitated usingTaqMan real time PCR.

Results: The low concentrations of ribavirin (0.5 and 1.25 μg/ml) induced de novo or enhanced existing T cell proliferative responses to HCV antigens in 9/31 patients (McNemar's test, p=0.01), while T cell proliferation in these cases was inhibited with the higher concentrations (2.5 and 5.0 μg/ml, p=0.01). The anti-proliferative effect of the higher concentrations was also observed with mitogen- and tetanus toxoid-stimulated PBMC and was not due to decreased cell viability. Quantitation of cytokine gene expression revealed that the presence of HCV-specific T cell proliferation correlates with high IL-12p35 mRNA (p=0.03) and high IFNγ mRNA (p=0.002), but low IL-10 mRNA (p=0.04). A ribavirin-induced increase in the proliferative response to HCV core was always associated with an increase in IFNγ mRNA (p=0.02) and a decrease in IL-10 mRNA (p=0.02). Levels of IFNγ protein in the supernatants correlated with levels of IFNγ mRNA (p=0.003) and paralleled the increase in T cell proliferation. The ELISPOT assay showed no difference in the number of IFNγ producing CD4+ T cells in the presence or absence of ribavirin.

Conclusions: Ribavirin augments HCV-specific T cell reactivity in a proportion of patients with chronic hepatitis C. This effect is associated with increased IFNγ and decreased IL-10 mRNA expression per cell.

51. INCREASED HEPATIC FIBROGENESIS (PIIINP) IS ASSOCIATED WITH DECREASED ANTIOXIDANT LEVELS IN PRIMARY BILIARY CIRRHOSIS (PBC)

Progressive hepatic fibrosis is a universal feature of PBC, but the mechanisms leading to increased collagen synthesis have yet to be completely elucidated. In a group of PBC patients, serum antioxidants, markers of oxidant stress and a marker of hepatic fibrogenesis were determined to examine the hypothesis that oxidant stress may be a factor in the development of hepatic fibrosis. The effect of ursodeoxycholic acid (URSO) treatment was also examined.

Patient & methods: Fasting blood and urine samples were obtained from 33 PBC patients, most of whom had normal bilirubin and synthetic liver function. All were non smokers, took no vitamin supplements and had not received URSO or colchicine for at least three months. A subset was also tested following 3 months of URSO treatment. Type III procollagen peptide (PIIINP) was determined by radioimmunoassay (Orion Diagnostica). Lipid perox-idation was assessed by serum malondialdehyde (MDA) and urinary 8-iso prostane (8-IP). Serum antioxidants and liver function markers were measured by standard methods. All values expressed as means with normal ranges (NR).

Results: PIIINP (7.23: NR 1.9–4.2μg/l). was significantly raised Serum antioxidants were decreased (Se 84.7: NR 86.7–110.3μg/l; Vit A 2.31: NR 2.8–4.7μmol/l.) and showed a strong negative correlation with PIIINP. Vit E (13.0: NR 9.8–15.7mg/l) was normal but also showed a strong negative correlation with PIIINP. Lipid peroxidation markers were significantly raised (MDA 3.23: NR 1.4–2.4μmol/l; 8-IP 1.24: NR 0.04–0.32 μg/g creatinine) but did not correlate with PIIINP. Bilirubin, AST, ALP, and albumin correlated with PIIINP, albumin negatively. Treatment with URSO decreased AST, ALT, ALP, and GGT and increased albumin, but had no effect on PIIINP, markers of lipid peroxidation or on antioxidant levels.

Discussion: We have previously shown that oxidant stress is an early feature of PBC. This study shows that, in a group of predominantly early stage patients, hepatic fibrogenesis. is significantly increased, inversely correlating with antioxidants levels. We have also shown that any beneficial effect of URSO is unlikely to be mediated by antifibrotic or antioxidant mechanisms. Antioxidant supplementation may be beneficial in early PBC.

52. INTERLEUKIN-1 (IL-1) CLUSTER GENE POLYMORPHISMS AND DECOMPENSATED ALCOHOLIC LIVER DISEASE (ALD)

Background: Candidate genes for the variability in predisposition to ALD include those for the pro-inflammatory cytokines IL-1 α and β and for the IL-1 receptor antagonist, IL-1ra. Biallelic polymorphisms in the1L-1A gene exist at positions +3954 and+4845, in theIL-1B gene at position-511, and in theIL-1-RA gene at position +2018. These polymorphisms are associated with altered IL-1 ratios, however, data on their allelic frequencies in ALD are limited.

Methods: We studied 2 groups of Caucasian heavy drinkers (>60U/wk(M) or 40U/wk(F) for >5yr) (1): Patients (n=151; 105M, age 48+10yr) had decompensated liver disease (Childs Grade B or C; Maddreys DF 30.2+30.6), other liver diseases were excluded by serum testing and, in 47 cases, by liver histology. (2): Controls (n=94; 73M, age 48+9yr) had no clinical liver disease, normal serum bilirubin, albumin and prothrombin time and no abnormality or fatty infiltration on ultrasound, and in 78 cases, a PGA index of ⩽5 (90% exclusive of cirrhosis (Gastro1991;100:1397)). Subjects were genotyped at the above loci using the Taqman system. 1=common and 2=rarer allele.

Results: for the IL-1B -511, see table.

Abstract 52, Table 1

None of the other 1L-1 gene polymorphisms was associated with ALD.

Summary: ALD showed a suggestive association with genotype 2:2 of the IL-1B -511 polymorphism; however, statistical significance was lost after correction for multiple testing.

53. FEATURES OF AUTOIMMUNE HEPATITIS IN PATIENTS WHO ARE NOT OF NORTHERN EUROPEAN CAUCASOID ETHNIC ORIGIN

Very little is known about the clinical expression of autoimmune hepatitis (AIH) in populations other than those of northern European caucasoid (NEC) or Japanese extraction, but it is our impression that patients from other ethnic backgrounds (non-NEC) have more severe disease. A retrospective analysis of all such patients attending our clinics since 1983 identified 12 (10 female: 6 African, 5 Asian, 1 Arabic; median age at presentation 30 y, range 12-58 y) for whom complete data were available. All were seronegative for antimitochondrial antibodies and for markers of hepatitis A, B, and C virus infection. Eleven satisfied international criteria for definite AIH and one for probable AIH (11 with ANA and/or SMA and one with anti-LKM1). Five of 6 HLA typed were DR3 or DR4 positive, one of whom had the extended haplotype A1-B8-DR4. Six presented acutely, and 9 (75%) had cholestatic biochemical features (serum ALP > 2x elevated). Liver biopsy revealed cirrhosis in 5 and 3 (25%) had mild biliary changes but no definitive features of PBC nor evidence of PSC on cholangiography. All received prednisolone (0.5 mg/kg/day) with azathioprine (1 mg/kg/day) as initial therapy according to a standard protocol. Four showed a complete initial biochemical response, 3 partial, and 5 no response. Four have required liver transplantation for intractable disease. Of 183 NEC patients with definite AIH (140 female) attending during the same period, median age at presentation was 45 y (range 6–79 y; p<0.05 v non-NEC patients). 35% had ALP > 2x elevated (p=0.014 vs. non-NEC), only one had biliary changes on liver biopsy (p<0.0005) and 95.0% showed a complete initial response to therapy (p<0.0005). There were no significant differences between the groups with respect to other parameters.

Conclusions: These findings suggest that non-NEC patients (other than Japanese) present at a younger age with a significantly higher frequency of cholestatic features and have a significantly poorer response to standard therapy than NEC patients. Whether these differences are due to cultural, environmental or genetic factors needs to be investigated.

54. TISSUE INHIBITOR OF METALLOPROTEINASE 2 IS UPREGULATED IN ACTIVATED RAT HEPATIC STELLATE CELLS

Introduction: Fibrotic injury in the liver is characterised by proliferation and activation of hepatic stellate cells (HSCs) which become the major source of the fibrillar matrix that characterises fibrosis. In addition activated HSCs but not quiescent cells produce tissue inhibitor of metalloproteinases type 1 (TIMP 1). A study was undertaken to characterise TIMP 2 production by rat HSCs.

Methods: Livers were harvested from rats; HSCs isolated and plated on uncoated tissue culture plastic. Lysates were prepared at 0, 3, 7, and 14 days after primary culture. Northern blots prepared from these timepoints were hybridised with a rat TIMP 2 cDNA probe. The expression of a human TIMP 2 promoter in HSCs was studied by subcloning the respective cDNA into a promoterless luciferase vector (pGL3-basic, Promega) and transfecting into HSCs. Either a 2.7 Kb TIMP 2 promoter construct (full promoter) or a 250bp promoter construct (minimal promoter) were used in the transfection of freshly isolated (day 0) or activated (day 7–10) HSCs. Cells were harvested and assayed for Luciferase activity. Results and

Conclusions: Expression of TIMP 2 was detected at day 0. Further, there was a clear up regulation in expression after day 3 (2x) and maximally at day 7 (4x; n=2). In experiments studying the TIMP 2 human promoter transfections in HSC results were normalised to the value obtained with the negative control. There was an 9 fold increase in promoter activity for the short TIMP 2 construct (minimal promoter) in comparison with transfection of this construct in quiescent HSCs (n=3).

Differences in expression of TIMP 1 and TIMP 2 promoter constructs may be attributable to the spatial distribution of the AP-1, PEA-3, and SP-1 motifs within each promoter.

55. ENHANCEMENT OF PROLIFERATION AND FUNCTION IN 3D SPHEROIDS OF HUMAN HEPATOCYTE CELL LINES BY HUMAN EXTRACELLULAR MATRIX, BUT NOT CHEMICAL DIFFERENTIATING AGENTS

Introduction: Proliferation and differentiation are normally inversely related. This impacts on the use of proliferating cells in a bioartificial liver. We have designed a novel approach to establishing 3-D cultures from human liver cell-lines, in which cells proliferate to form spheroids in uncoated alginate beads. Cells express many though not all differentiated functions at near in vivo levels.

Aim: To enhance performance further, without imperilling the clonal expansion that underlies the generation of spheroid colonies. We investigated both biological and chemical pro-differentiating agents: extracellular matrix (ECM) proteins, and Dimethyl sulphoxide (DMSO), butyrate and retinoids.

Methods: ECM was added to HepG2 cells in alginate beads on day 0 at 22, 44, or 88μg protein/ml. DMSO (1, 2%), sodium butyrate (1, 2mM) and retinoic acid (1, 10μM) were added at day 0 or 7. Proliferation, liver specific protein secretion and transmission electron microscopy (TEM) were analysed.

Results:Proliferation: ECM resulted in increased growth by day 8 with 22, 44, 88 μg ECM protein/ml i.e. 1.5 ± 0.1, 2 ± 0.1, and 2.5 ± 0.1 cf. control 0.88 ± 0.1 x106nuclei (p⩽0.005 cf. control, Students t-test). All chemical agents on d 0 inhibited proliferation and decreased cell viability. When added on day 7, no difference was observed in 3-D HepG2 cells on day 8, however by day 15, each agent was anti-proliferative, eg. day 15, 2% DMSO: 1.96 ± 0.2 cf. control 7.23 ± 0.6 x106 nuclei (p⩽0.005 cf. control). Maintenance of function: Higher levels of proteins were secreted by ECM cultures cf. controls by day 11, eg. 88 μg ECM /ml: albumin 22.58 ± 2.1 cf. ctrl 11.03 ± 1.6; fibrinogen 0.358 ± 0.03 cf. ctrl. 0.178 ± 0.02 μg/106 nuclei/24h (p⩽0.005). Hepatocytes had prominent microvilli, and desmosomes, and a loose finely filamentous matrix surrounded spheroids when cultured in ECM. None of the chemical differentiating agents increased protein secretion.

Conclusion: Human ECM increases both proliferation and performance of spheroid cultures of human hepatocyte cell-lines, which may prove useful for a bioartificial liver.

56. OXIDATIVE PHOSPHORYLATION DEFECTS IN WILSON'S DISEASE LIVER

Background: Wilson's disease (WD) is caused by mutations in a P-type ATPase and is associated with copper deposition in liver and/or brain. The WD protein is present in the trans-Golgi network and may also be imported into mitochondria. The WD protein functions as a P-type copper transporting ATPase in the Golgi but any action in mitochondria is at present unknown.

Methods: We have studied mitochondrial function and aconitase activity in WD liver tissue and compared the results with a series of normal and disease controls.

Results: There was evidence of severe mitochondrial dysfunction. Enzyme activities were decreased as follows: complex I by 62%, complex II+III by 52%, complex IV by 33%, and aconitase by 71%. These defects did not appear to be secondary to penicillamine use, cholestasis or poor hepatocellular synthetic function.

Conclusions: The results provide direct evidence for a defect of energy metabolism in WD. The pattern of enzyme defects suggests that free radical formation and oxidative damage, probably mediated via mitochondrial copper accumulation, are important in WD pathogenesis. These results provide a rationale for a study of the use of anti-oxidants in WD.

57. HEPATIC STELLATE CELLS IN ARCTIC ANIMALS

In physiological conditions in human and usual experimental animals, hepatic stellate cells store 80% of the total retinoid in the whole body as retinyl ester in lipid droplets in the cytoplasm, and regulate transport and storage of retinoid. More than 50 years ago, Rodahl reported that animals in the arctic area were able to store a large amount of retinoid in the liver. To investigate the cellular and molecular mechanisms in transport and storage of retinoid in these arctic animals, we performed this study in the Svalbard archipelago (situated at 80° N, 15° E). Arctic animals caught were 3 polar bears (Ursus martimus), 8 arctic foxes (Alopex lagopus), 6 bearded seals (Erignathus barbatus), 22 glaucous gulls (Larus hyperboreus), 5 fulmars (Fulmarus glacialis), 4 Brünnich's guillemots (Uria lomvia), 6 ringed seals (Phoca hispida), 5 hooded seals (Cystophora cristata), 6 puffins (Fratercula arctica), 5 Svalbard ptarmigan (Lagopus mutus hyperboreus) and 7 Svalbard reindeers (Rangifer tarandus platyrhynchus). Fresh organs, namely, the liver, kidney, spleen, lung, and jejunum were examined with gold chloride staining, fluorescence microscopy for autofluorescence of retinoid, hematoxylin and eosin staining, azan staining, Ishii-Ishii's silver impregnation, transmission electron microscopy, and high-performance liquid chromatography. Serum from each animal was analysed with high performance liquid chromatography. The arctic animals stored retinoid in hepatic stellate cells. Only a small amount of retinoid existed within other organs such as kidney, spleen, lung, and jejunum. Top predators among arctic animals stored 6–23 μmole retinyl ester per g liver which 20–100 times the levels normally found in other animals including humans. Plasma concentrations of retinol in these animals were 6.21 μM (glaucous gull), 3.77 μM (Svalbard ptarmigan), 2.97 μM (arctic fox), 0.67 μM (ringed seal), and 0.44 μM (bearded seal). These results indicate that the hepatic stellate cells in these animals have high ability for uptake and enough capacity for storage of retinoid.

58. ENHANCED EXPRESSION OF B7-1, B7-2, AND ICAM-1 IN HEPATIC SINUSOIDAL ENDOTHELIAL CELLS BY ISCHEMIA/REPERFUSION

Ischemia/reperfusion injury to the liver occurs in various clinical situations such as liver transplantation and surgery, trauma, hemorrhagic and endotoxic shocks, or thermal injury. Mononuclear and/or polymorphonuclear leukocytes mediate microvascular and parenchymal cell dysfunction after ischemia/reperfusion, and cell adhesion molecules are important for the development of the dysfunction mechanism. In this study, we examined the expression of cell adhesion molecules, B7-1 (CD80), B7-2 (CD86), and ICAM-1 after hepatic ischemia/reperfusion injury in the liver tissue.

To induce the hepatic ischemia in male Wistar rats, both portal vein and hepatic artery entering the left-lateral and median lobes were occluded with microvascular clamp for 30 min, and then reperfused. B7-1 protein was not detectable in the liver tissue from normal and sham-operated control rats by immunofluorescence staining. However, after ischemia/ reperfusion, B7-1 signals were detectable on cell membranes of sinusoidal endothelial cells. On the other hand, B7-2 and ICAM-1 proteins were constitutively expressed in sinusoidal endothelial cells in normal and sham-operated control rats, and they were increased 24 hours after ischemia/ reperfusion. The localisation of B7-1, B7-2, and ICAM-1 to endothelial cells was confirmed by comparison with the localisation of von Willebrand factor used as an endothelial marker. Dual fluorescence immunostaining of B7-1/ICAM-1 or B7-2/ ICAM-1 indicated that these cell adhesion molecules were colocalised on the endothelial cell surfaces. Results from RT-PCR and real-time RT-PCR analyses showed the upregulation of B7-1, B7-2, and ICAM-1 mRNA in the liver tissue after ischemia/reperfusion injury. These results suggest that hepatic sinusoidal endothelial cells involve as antigen-presenting cells in cell-mediated immune response in the liver, and that the upregulation of costimulatory molecule expression in endothelial cells has a pivotal role in liver dysfunction after ischemia/reperfusion injury.

59. LOW DOSE PENTOXIFYLLINE (PTX) REDUCES MORTALITY IN A THIOACETAMIDE (TAA) INDUCED MODEL OF ACUTE HEPATIC FAILURE (AHF)

Introduction: Pentoxifylline inhibits, release of TNF-α, platelet aggregation and adherence of leukocytes to endothelium. Previous studies report increased mortality following its use.

Aims: To investigate the dose dependant effect of pentoxifylline administration both pre- and post- hepatic injury in a characterised TAA model of AHF in the Wistar rat. Methods: We have previously reported a robust TAA induced model of AHF in the rat. AHF is induced by two intraperitoneal (i.p.) injections (500mg/kg 8 hours apart) of TAA. Four groups were studied (n=5).Group 1 received TAA only.Groups 2, 3, & 4 followed the protocol forGroup 1, however, Groups 2 and 3 were pre-treated for 5 days with once daily high dose PTX (300mg/kg i.p.) and low dose PTX (25mg/kg i.p.), respectively. Whereas, Group 4, commenced PTX (25mg/kg i.p.) post-TAA for 5 days. Mortality was determined at 96 hours in separate groups. (n=5)

Results: *p<0.05 Mean +/-S.D.

Abstract 59, Table 1

Conclusion: Pre-treatment with low dose PTX reduces hepatic injury and mortality. This contrasts with exacerbated hepatic injury and mortality following high dose pre-treatment and post-TAA treatment with low dose PTX, demonstrating dose dependent effects and differences in its temporal use.

60. MODULATION OF HEPATOCYTE FUNCTION IN A HEPATOCYTE CELL LINE FOLLOWING EXPOSURE TO LIVER FAILURE PLASMA

Introduction: There is great interest in the development of a device containing hepatocytes in an extracorporeal liver circuit for the treatment of patients with acute liver failure. This system requires maintenance of hepatocyte function in plasma. There is a paucity of data examining the effect of liver failure plasma (LFP) on function in human hepatocytes.

Aim: To characterise the functional capacity of a human hepatocyte cell line HHY41, when cultured in LFP.

Methods: HHY41 was cultured in either four normal plasma (NP) or four LFP (LFP 1-4) samples for 16 h. Controls were cultured in αMEM. HHY41 was assessed for toxicity, oxidative status, DNA and protein synthesis, MTT metabolism, CYP1A activity, caffeine metabolism and androstenedione metabolism.

Results: All LFPs induced toxicity in HHY41 (p<0.05-cf. control). NP had no toxic effect. Oxidative status was decreased by all 8 plasma samples (p<0.05). The greatest decrease was with two LFPs (10% of control values). DNA synthesis was increased by all NPs (p<0.05) and decreased by LFP 2 and 3 (p<0.005). Protein synthesis was maintained by NP but decreased by the same LFPs (p<0.0005). MTT metabolism was decreased by all 8 plasma samples with LFP 2 and 3 (p<0.0005) inhibiting MTT metabolism by 95%. CYP1A activity was increased two fold by NP (p<0.05). LFP 2 and 3 decreased CYP1A activity to 50% of controls (p<0.005). Caffeine biotransformation to 13X, 17X, and 37X metabolites was reduced by all LFPs (p<0.05) while NP only decreased 37X formation (p<0.05). Androstenedione metabolism was unaltered by NP while all LFP samples decreased the metabolism of androstenedione.

Conclusion: LFP had marked inhibitory effects on hepatocyte function and viability suggesting the presence of factors detrimental to hepatocytes in LFP. Implementation of procedures to remove toxins from LFP before exposure to hepatocytes in a BAL may be required to provide conditions of optimal hepatocyte function.

61. THE S65C MUTATION OF HFE MAY CONTRIBUTE TO IRON OVERLOAD IN C282Y HETEROZYGOTES

Background & Aims:HFE-related haemochromatosis is a common disorder of iron metabolism. Most affected individuals are homozygous for the C282Y mutation of HFE. Some are compound heterozygotes for C282Y and another missense substitution H63D. A small proportion of patients have neither of these genotypes. Recently the S65C substitution has been implicated in a mild form of haemochromatosis.

Methods: Genotype for the S65C mutation was determined in 55 individuals who were heterozygous for C282Y and negative for H63D and IVS3+1G→T, referred because of increased serum iron indices and/or a family history of haemochromatosis. A control group comprising 315 individuals was also studied. The phenotypes of individuals who were compound heterozygous for C282Y and S65C were assessed.

Results: Five individuals were found to be compound heterozygous for C282Y and S65C. Three referred for family screening had normal serum iron indices. Two were receiving venesection treatment for haemochromatosis, although a component of the disease may be non-HFE related in one case.

Conclusions: Our results suggest that some individuals who are compound heterozygous for C282Y and S65C may have elevated serum iron indices and evidence of iron overload. The penetrance of this genotype appears to be low and other genetic and environmental factors may influence the expression of iron loading in these individuals.

62. INHIBITION OF TERT-BUTYLHYDROPEROXIDE INDUCED OXIDATIVE STRESS IN THREE HUMAN HEPATOCYTE CELL LINES BY PRETREATMENT WITH 1,2:5,6-DIBENZYLANTHRACENE

Introduction: Oxidative stress is one mechanism by which hepatocellular injury is mediated.tert-butylhydroperoxide (t-BH) is used as a model of oxidative stress in rat hepatocytes, however little is known about its effect in human hepatocytes.

Aim: To establish a model oft-BH induced oxidative stress in three human hepatocyte cell lines. To investigate ift-BH induced damage can be reduced by treatment of cultures with Dibenzylanthracence (DBA).

Methods: HH25, HH29 and HHY41 cell lines were incubated in 0–1.5mM t-BH for 2 h. Cultures were assessed for toxicity, MTT metabolism, GSH content, and lipid peroxidation (malondialdehyde - MDA). Cultures were pretreated with 13μM DBA for 36 h.

Results:t-BH induced oxidative stress:t-BH caused dose dependent toxicity in HH25, HH29, and HHY41 (p<0.05), reaching 70% at 1.5mMt-BH. At or above 0.2mMt-BH caused a decrease in MTT metabolism (p<0.05) falling to 15% of controls at 1.5mM. GSH levels were decreased at or above 1mM t-BH and to 50% of controls at 1.5mM t-BH.t-BH caused a dose dependent increase in lipid peroxidation in HH25, HH29, and HHY41 and was significantly increased at or above 1 mM t-BH (p<0.05). Pretreatment with DBA: DBA pretreatment in HH25, HH29 and HHY41 decreased t-BH induced toxicity at or above 0.75 mM compared to non-DBA treated cultures (p<0.005) and partially abrogated the decrease in MTT metabolism, att-BH ⩾1mM (P<0.005) compared with non-DBA treated cell lines. DBA increased GSH levels two fold over basal levels for HH25, HH29, and HHY41. At all t-BH concentrations GSH levels were higher in the DBA treated as compared to non-DBA treated cultures (p<0.005). At allt-BH concentrations MDA levels were lower in DBA as compared with non-DBA treated cultures (p<0.05).

Conclusion: In this human hepaocyte system we have shown that DBA has a hepatoprotective role int-BH induced oxidative stress. This model may prove useful in elucidating the mechanism of hepatotoxicity by oxidative stress and for investigating methods of protection against this type of cell damage.

63. L-NAME AND AMINOGUANIDINE (AMG)—SELECTIVITY AND TEMPORAL DIFFERENCES IN iNOS INHIBITION IN A RAT MODEL OF ACUTE HEPATIC FAILURE (AHF)

Introduction: iNOS inhibition in AHF is controversial, fuelled by conflicting results from studies using varieties of pharmacological agents, animal models and temporal protocols.

Aims: We hypothesise that increased selectivity of iNOS inhibition (AMG>L-NAME) leads to reduced hepatic injury and this is subject to temporal constraints.

Methods: We have previously described a robust model of AHF in the Wistar rat. AHF was induced by two i.p. injections (500mg/kg, 8 hours apart) of thioacetamide (TAA). Six groups were studied (n=5). Group 1 received TAA alone. Groups 2, 3 &4 followed protocol forGroup 1, however, Groups 2 & 3 were pre-treated and Groups 4 & 5 post-TAA treated with iNOS inhibitors AMG (100mg/kg s.c.) and L-NAME (100mg/kg s.c.) for 5 days. Results: *?p<0.05, Mean+/- S.D.

Abstract 63; Table 1

Conclusion: AMG is a more selective iNOS inhibitor, significantly reducing parameters of AHF compared L-NAME. Beneficial effects are, however, negated if iNOS inhibitors are administered after TAA, exacerbating hepatic injury, increasing mortality, and thus demonstrating temporal constraints.

64. AGEING PROCESS AFFECTS GALACTOSAMINE ACTION ON ISOLATED HEPATOCYTES AND LIVER IN RAT

Galactosamine (GalN) is an aminosugar that by depleting the pool of uridine nucleotides inhibits transcription and translation, and leads to a reversible liver damage. Although GalN has been frequently used as a model hepatotoxin, the effect of ageing on its action has been only investigated by few authors. Therefore we assessed the effects of in vitro andin vivo GalN administration on the content of nucleotides and aminosugars in hepatocytes and liver of young (4 mo), adult (12 mo), and old (24 mo) male Wistar rats. Hepatocytes were isolated by the collagenase perfusion of the liver and incubated for 60 min with 10 mM GalN. Cell damage was estimated by decrease in the rate of protein synthesis measured as 3H-leucine incorporation into acid-precitable material, and by increase in LDH release into the medium. In an another set of experiments rats were injected intraperitoneally with GalN (0.2 g/kg bw) or saline; after two hours livers were freeze-clamped, immersed in liquid nitrogen, and stored at −80°C. Hepatocytes, and lyophilised blocks of liver, were extracted with perchloric acid, and assessed for the content of nucleotides and aminosugars by the ion-exchange HPLC. Alanine aminotransferase activity was measured in blood serum.

GalN increased LDH activity in the medium to a similar extent in hepatocytes of young and old rats. However, hepatocytes of old rats were more susceptible to the suppression of protein synthesis by GalN than cells of young ones, probably due to the decreased basal level of UTP that was by 60% lower (old 0.18±0.03, young 0.45±0.09, nmoles/mg protein, mean±SE, n=6, p<0.01). This decrease could be partially caused by the cell isolation procedure as the UTP content in the intact liver was lower by 28% in the group of old rats (young 2.62±0.14, old 1.90±0.21 nmoles/mg dry w., p<0.05, n=6). Thei.p. GalN administration decreased UTP content by 89%, 65%, and 55%, and increased the content of UDP-sugars by 305%, 175%, and 189%, in the liver of old, adult, and young rats, respectively. The activity of AlAT in blood serum and the hepatic content of ATP, ADP, AMP, NAD, and GTP did not differ significantly between the age groups of rats, and was not changed by GalN treatment. Lower hepatic UTP content may partially explain the increased sensitivity of hepatocytes of old rats to the toxic action of galactosamine, and possibly to other hepatotoxic compounds that decrease transcription in the liver. As the age related alterations of the hepatic UTP levels might result from decreased synthesis or/and increased degradation, the effect of ageing on uridine nucleotide metabolism in rat liver has yet to be elucidated.

65. EXPRESSION AND ACTIVATION OF GELATINASE A IN LIVER FIBROSIS

Liver fibrosis results from an imbalance of matrix synthesis versus degradation. Matrix metalloproteinases (MMPs) play a key role in the latter process. We and others have previously reported that hepatic stellate cells (HSC) produce MMPs during their activation to myofibroblasts following liver injury and during culture on plastic. We have recently shown that activated HSC produce gelatinase A (MMP-2 or 72 kD type IV collagenase) and activate this MMP by cleavage of its propeptide. This process correlates with HSC synthesis of membrane type MMP (MT1-MMP), considered an important biological activator of gelatinase A. We have reported that active gelatinase A is a mitogen for HSC and might contribute to HSC proliferation following liver injury. In this study, we have explored this possibility by examining expression of gelatinase A and MT1-MMP, and also gelatinase A activation in rat and human liver fibrosis.

Human liver explants from patients with primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), and biliary atresia (BAT) and normal liver were generously provided by the Birmingham Liver Unit. Livers were harvested from rats intoxicated with CC14 (0.1ml/100g twice weekly) as detailed. Expression of mRNA for gelatinase A in liver homogenates was examined by gelatin zymography.

Hepatic gelatinase A mRNA was increased significantly to 224–330% above normal liver expression in the three human liver diseases. Total hepatic gelatinase A protein showed mean increases of 170% (PBC), 230% (PSC), and 180% (BAT) above normal. This protein was localised to HSC in these livers by immunostaining. The active forms (66 and 62 kD) of gelatinase A were undetectable in normal liver homogenates, but represented a considerable proportion of the total gelatinase A in PBC (27% active), PSC (22% active), and BAT (23% active). There was also significantly enhanced expression of MT1-MMP mRNA in the fibrotic livers, this becoming expressed at 110% (PBC), 165% (PSC), and 103% (BAT) above normal liver expression. Rats treated with CC14 for 4 wk developed progressive liver fibrosis and hepatic expression of gelatinase A was increased 520% above normal after 4wk, although MT1-MMP mRNA was increased only modestly (66%). Pro and active gelatinase A, undetectable in normal liver, appeared after 1 wk of CC14 and at 4 wk was readily detectable with active gelatinase representing 12% of the total. In rats allowed to recover from the 4 wk of CC14 treatment, liver histology returned to virtual normality over the following 28 days. During this recovery, there was a progressive fall in the hepatic MT1-MMP and gelatinase A mRNA and gelatinase A protein such that they approached levels in normal liver. In a second model, rats were administered CC14 for 12 wk after which total hepatic gelatinase A protein was found to be increased by 22% over normal, with 36% of this being in the active form. However, in this model, cessation of CC14 dosing did not result in detectable reversal of fibrosis (ie recovery) over the following 28 days, and hepatic expression of both total and active gelatinase A did not diminish over this period.

In conclusion, increased synthesis and activation of gelatinase A is a feature of liver fibrosis in both humans and rats. Gelatinase A expression decreases in livers when there is a recovery from liver fibrosis, but not in livers when recovery is not apparent. Activated HSC are a source of gelatinase A. Our studies support the hypothesis that gelatinase A contributes to liver fibrogenesis and increased expression of its active form may sustain HSC proliferation in vivo.

66. GLUT 4 SKELETAL MUSCLE LEVELS IN ACUTE LIVER FAILURE

Background: Acute liver failure (ALF) is characterized by peripheral insulin insensitivity. The cellular molecular mechanisms responsible for impaired insulin action are currently unknown. Alterations in protein levels of the skeletal muscle glucose transporter GLUT 4 (the major insulin-regulated transporter) might represent a cellular mechanism of insulin resistance in ALF.

Aim: To study GLUT 4 protein expression in skeletal muscle and its time course in ALF.

Methods: An open muscle biopsy was taken from the quadriceps muscle of patients admitted with ALF (n=9) secondary to acetaminophen ingestion within 72 hours of the peak INR (range 3.5–10). The biopsies were repeated on day 7. Control muscle biopsies were obtained from subjects at the time of routine orthopaedic surgery (n=3). Biopsies were snap frozen in liquid nitrogen and stored at −70oC before analysis. Total protein lysates were then extracted from the muscle biopsies. GLUT 4 expression was determined by Western blotting using a monoclonal antibody to GLUT 4 (Biogenesis, UK), triplicate studies being performed. Immunoreactive bands were visualised by enhanced chemiluminescence (ECL) and band intensity determined by densitometric scanning. Values are expressed as the mean ± SEM. Comparisons were made using Mann Whitney U analysis.

Results: 6 of the 9 patients survived, 2 patients dying before the second biopsy being performed. GLUT 4 levels were increased in ALF; 3.5 fold (1111± 143) and 3.6 fold (1151 ± 214) on day 1 and 7 respectively, as compared with controls (320 ± 31) (p<0.05). No differences were detected between day 1 and 7, males and females, or survivors and non-survivors. Previous studies in ALF using hyperinsulinaemic euglycaemic clamping studies on day 1 and 7 showed a marked decrease in peripheral glucose utilisation.

Conclusion: In the presence of insulin resistance, a significant increase in GLUT 4 expression has been detected. GLUT 4 level changes appear to be poorly functional, as seen by impaired peripheral glucose utilisation, and could reflect impairment in recruitment and translocation of intracellular GLUT 4.

67. INTRA-RENAL ENDOTHELIN PRODUCTION MEDIATES DECREASE IN RENAL BLOOD FLOW FOLLOWING ACUTE INCREASE IN PORTAL PRESSURE GRADIENT

Background: Acute increase in hepatic intrasinusoidal pressure is known to decrease renal blood flow (RBF). The mechanism of this change is unlcear.

Aims: This study was performed to test the hypothesis that alterations in RBF following acute TIPSS shunt occlusion are mediated by intrarenal endothelin production.

Patients and Methods: Sixteen patients, haemodynamically and metabolically stable (mean age 54.7 years; 9M,7F) who underwent a TIPSS insertion to prevent recurrence of variceal hemorrhage at a mean of 2.65 months before this study were enrolled to undergo portography to assess the shunt function. The patients were divided into 2 groups: Group I (n=8) who underwent measurement of RBF (using PAH) and Group II(n=8), who underwent measurements of cardiac output (CO) and systemic vascular resistance (SVR). All 16 patients underwent measurements of mean arterial pressure (MAP), hepatic blood flow (HBF using ICG) and renal/ hepatic venous ET-1 and BigET-1 measurements (using radioimmunoassay. The assays and measurements were performed at baseline (T=0) and repeated 12 minutes after occlusion of the TIPSS shunt (T=12) with an angioplasty balloon. The arterial concentration of ET-1 and BigET-1 were assayed from a catheter placed in the femoral artery. The regional V-A differences for ET-1 and BigET-1 were calculated.

Results: All 16 patients had functional shunts with a PPG of 8.3(2.1) and 6.8 (2.9) mmHg in the 2 groups respectively. Following shunt occlusion, PPG, SVR, and HBF increased (p<0.01), and HR, CO, and RBF decreased significantly (p<0.03). Arterial ET-1 and Big ET-1 increased significantly (p<0.05), renal ET-1, and Big ET-1 decreased significantly (p<0.01) and hepatic ET-1 increased (p<0.01) but hepatic Big ET-1 was unchanged.

Conclusions:1. Following acute occlusion of TIPSS, there is an enhanced renal output of endothelin. Significant increases in renal BigET-1 V-A concentrations point to intrarenal production of endothelin in this setting. 2. The increase in SVR could possibly be due to an increase in arterial endothelin level. 3 While the RBF decreases, the HBF increases significantly, alluding to a possibly different ET receptor subtype density/stimulation in the two vascular territories.

68. MODERATE HYPOTHERMIA FOR UNCONTROLLED INTRACRANIAL HYPERTENSION IN ACUTE LIVER FAILURE (ALF)

Background: Increased intracranial pressure (ICP) complicating acute liver failure (ALF) has a mortality of about 90% in patients who do not respond quickly to treatment with mannitol and ultrafiltration. This study evaluates the safety and efficacy of moderate hypothermia as a treatment for uncontrolled increase in ICP in patients with ALF.

Methods: Fourteen consecutive patients (age 24 (range 16–54), 9 females, 8 candidates for orthotopic liver transplantation (OLT)) with who fulfilled the Kings College criteria for poor prognosis and had uncontrolled increase in ICP (defined as persistently elevated ICP of >25 mmHg for 1 hour or more despite 2 separate treatments with mannitol (1g/Kg body weight over 20 min) and removal of 500 ml of fluid by continuous veno-venous hemofiltration) were studied. All patients required sedation and ventilation for Grade III–IV encephalopathy. Patients were monitored continuously, using a Swan-Ganz catheter for measuring cardiovascular haemodynamics, intracranial pressure was monitored using a subdural catheter (Camino) and cerebral blood flow was measured using the Kety-Schmidt method. A reverse jugular catheter allowed sampling of blood. Core temperature was reduced to 32–33°C using cooling blankets.

Results: Eight of the nine patients who were candidates for OLT were successfully bridged to transplantation with a mean of 32 hours (range 8–120) hours of hypothermia. The five patients who were unsuitable candidates for OLT died following rewarming. ICP before cooling was 54 (range 23–67) mmHg and this was reduced in all patients, to 19 (range 10–22) mmHg (p<0.01) and the cerebral blood flow decreased from 97 ml/100g/min (range 22–154) to 32 ml/100g/min (range 24–76) (p<0.01). Cerebral perfusion pressure increased (p<0.01) mmHg and cardiac index decreased significantly (p<0.01). During hypothermia, there 4 relapses of increase in ICP in 3 patients which responded to treatment with mannitol, further fluid removal or thiopentone infusion. Arterial ammonia and cerebral uptake of ammonia were significantly reduced with cooling. No adverse effects of hypothermia were observed.

Conclusions: Moderate hypothermia is useful in the treatment of uncontrolled increase in ICP in patients with ALF and may serve as a bridge to OLT. The mechanism by which hypothermia reduces intracranial hypertension is through reduction of cerebral blood flow and ammonia delivery and extraction by the brain.

69. RESTORATION OF CEREBRAL BLOOD FLOW AUTOREGULATION AND REACTIVITY TO CARBON DIOXIDE IN ACUTE LIVER FAILURE BY MODERATE HYPOTHERMIA

Background: In patients with ALF and uncontrolled intracranial hypertension, moderate hypothermia (32°C) reduces ICP and CBF (Jalan et al. Lancet1999;354:1164–8). The mechanism by which hypothermia reduces intracranial pressure (ICP) is not clear. The purpose of this study was to test the hypothesis that moderate hypothermia reduced intracranial pressure by restoring CBF autoregulation. Patients and Methods: Nine patients (median age 32 (range 22–46); 2M, 7F) with uncontrolled intracranial hypertension (defined as persistently elevated ICP of >25 mmHg for 1 hour or more despite 2 separate treatments with mannitol (1g/Kg body weight over 20 min) and removal of 500 ml of fluid by continuous veno-venous hemofiltration) and ALF who fulfilled criteria for poor prognosis were studied. If the patients had uncontrolled intracranial hypertension, moderate hypothermia was instituted using a cooling blanket. CBF autoregulation and reactivity to carbon dioxide was evaluated before and 4 hours after cooling. CBF was measured using the Kety-Schmidt technique prior to and after elevating the resting mean arterial pressure by 20–30 mmHg by infusion of noradrenaline. An elevation in CBF by 10% or more with the change in mean arterial pressure was taken as the evidence for lost autoregulation. Reactivity of CBF to carbondioxide was measured with the PaCO2 set between 3.5 and 4 KPa and then after it was increased to between 5.5 and 6 Kpa.

Results: Significant reductions were observed in the ICP (median 46 (range 27–54) mmHg to 19 (15–22) mmHg, p<0.01); and CBF [median 111 (69–134) to 56 (38–67) ml/100g/min, p<0.05). The defective CBF autoregulation was restored with cooling. There was no significant change in CBF with the increase in pCO2in all patients prior to cooling but significant increase in CBF was noted after cooling (pre cooling: median CBF, 69 (range 74–78) mmHg at a median PaCO2of 3.9 (range 3.8–4.1), and 74 (range 69–80) mmHg at median PaCO2of 6.3 (range 5.9–6.4), p=0.4; post cooling: median CBF, 46 (range 38–57) mmHg at a median PaCO2of 4 (range 3.9–4.2), and 56 (range 57–63) mmHg at a median PaCO2of 6.1 (range 5.8–6.2), p<0.05).

Conclusions: The results of our study suggest that the improvement in ICP observed with hypothermia may be due to its effects upon CBF autoregulation and provides a tool to explore the mechanisms associated with the deranged autoregulation that is important in the pathogenesis of intracranial hypertension in ALF.

70. MR CHOLANGIOGRAM IN CHILDREN WITH HEPATO-BILIARY AND PANCREATIC DISEASES: WHEN IS IT USEFUL?

The assessment of the biliary tree and the pancreatic duct (PD) usually requires invasive techniques such as endoscopic (ERCP), percutaneous (PTC) or intra-operative cholangiogram (ICG). MR Cholangiogram (MRCP) is a non invasive, non ionising radiological investigation, which has been developed more recently. Data available about MRCP in adults with hepato-biliary and pancreatic problems have confirmed its diagnostic value. In children, only little data about MRCP are available and the conclusions are controversial. Unlike ERCP, PTC, and ICG, MRCP in children does not require general anasthesia and it is not associated with the risks of invasive imaging.

Aim: to determine the diagnostic value of MRCP in children with hepato-biliary and pancreatic diseases.

Method: indications and images of all children, who had MRCP between 9/1998 and 1/2000 were reviewed. Diagnostic findings and limitations of the MRCP, ultrasound scans (US), and ERCP, PCG, ICG were recorded and compared qualitatively.

Results: MRCP was performed on 22 children, 11 M/11 F, age range 6 weeks to 15.6y, median 10.5 y. Indications were: assessment of asymptomatic BD dilation found incidentally on US (n=3), upper abdominal pain/suspected pancreatic disease (n=6), cholestatic diseases (n=13). MRCP was not diagnostic in neonatal cholestatic disease in absence of BD dilation (4/13).The table below summarises the results of the diagnostic MRCP compared to US and conventional cholangiograms.

Conclusion: MRCP is a non invasive, non ionising investigation, which provides excellent images of the biliary tree and the PD. In our experience, MRCP in children has been a valuable and reliable diagnotis tool for BD dilation, obstruction or irregularities. It is particularly useful for the assessment of the PD. However MRCP is not yet helpful for the assessment of neonatal cholestatic disease in absence of BD dilation. In the hands of a skilled radiologist and with increasing experience MRCP should replace in most cases an invasive procedures for the diagnosis of hepato-biliary and pancreatic diseases.

Abstract 70, Table 1 Diagnostic MRCP compared with US and ChG

71. THYROID AUTOIMMUNITY IN HEPATITIS C VIRUS INFECTION IS GENOTYPE DEPENDENT

The degree of cross-reactive antibody binding has been shown to correlate with the number of identical aminoacids between viral and self epitopes (J Clin Invest1997;100:1114). The immunodominant epitope of Thyroid Peroxidase (TPO115-124) shares 6 identical amino acids (aa) with HCV2442-2451 of genotype 1a and 1b, 5 aa with HCVJA2441-2450 of genotype 2a containing a single Alanine (A) to Proline (P) aa substitution and 4 identical aa with HCVJ82464-2473 from genotype 2b containing an additional Leucine (L) to Methionine (M) aa substitution (fig 1, identical aa in bold, conservative substitutions underlined). To ascertain variation in antibody binding in relation to normally occuring single aa substitution, five highly cross-reactive TPO115-124/HCV2442-2451 sera from patients infected with genotypes 1a, 1b, and 2a and two additional sera from patients with genotype 2b were tested in ELISA against biotinylated peptides containing the TPO and HCV sequences. Of the five cross-reactive sera, serum 3 (genotype 1b) bound less efficiently to the peptide reproducing the HCVJA2441-2450 sequence of genotype 2a containing a single A to P aa substitution while all sera were unreactive against a peptide reproducing sequence HCVJ82464-2473 of genotype 2b and containing an additional L to M aa substitution. The two sera from patients with genotype 2b did not react with either TPO or HCV synthetic sequences.

Our results suggest that genotype-encoded aa differences normally occurring in equivalent viral sequences of distinct hepatitis C genotypes account for differences in anti-HCV immune response and cross-reactive autoimmunity. This may explain why virus/self cross-reactivity and thyroid autoimmunity are present in some HCV infected patients and not in others.

Abstract 71, Table 1 Patients code number and viral genotype

72. BOTH ALCOHOL ABUSE AND HETEROZYGOSITY FOR THE C282Y MUTATION INCREASE RISK OF HEPATIC SIDEROSIS: THE EFFECT IS NOT ADDITIVE

C282Y gene mutation and alcohol are associated with hepatic siderosis, when combined they may have an additive effect on siderosis score.

Aims: To determine HFE status and alcohol consumption of patients attending liver clinic and correlate these variables with the degree of siderosis.

Methods: 386 patients(141F and 245M) attending liver clinic were studied. Alcohol abuse was determined by a single observer using a standard questionnaire.Genomic DNA was extracted from EDTA blood, DNA fragments amplified by polymerase chain reaction and mutations determined by restriction fragment analysis using restriction endonucleases Rsa1 and Bcl1.A semiquantitative grading system from 0–4 was used for histological assessment of liver iron. Statistics were analysed using Mann-Whitney U test.

Results: When compared with wild type non-alcoholic group, both heterozygosity for C282Y gene mutation and alcohol were individually associated with greater mean siderosis scores although this difference reached statistical significance only with the alcohol group. However in the presence of both C282Y gene mutation and alcohol, no additional increase in siderosis score was seen. Across all genotypes alcohol was associated with significantly higher siderosis score (means 0.73 v 0.32, p=0.0005). As expected homozygosity for C282Y mutation was also associated with significant iron overload.

Conclusion: Both a history of alcohol abuse and heterozygosity for the C282Y mutation are associated with increased hepatic siderosis. However on subgroup analysis, the effect of these two risk factors did not appear to be additive although the patient numbers were relatively small.

Abstract 72, Table 1 Effect of C282Y gene mutation and alcohol on siderosis score

73. BILE ACID THERAPY IN ALCOHOLIC LIVER DISEASE

The hydrophilic bile acid UDCA is now the treatment of choice in primary biliary cirrhosis and has been shown to improve liver function tests (LFTs) in other cholestatic liver diseases including alcoholic liver disease. However there is no information on the effect of UDCA on fibrosis in alcoholic liver disease. Serum markers of fibrosis have been shown to correlate with hepatic fibrosis and can be performed serially to reflect changes in fibrosis.

Aims: To assess the effect of UDCA on serum markers of fibrosis, as well as routine LFTs in alcoholic liver disease.

Methods: 50 patients with biopsy confirmed alcoholic liver disease were randomised to six months of UDCA 10-15mg/kg/day or to placebo. Patients with decompensated liver disease, or an entry procollagen III peptide (PIIIP) below 0.5 μg/mL were excluded. Serum was collected at monthly checkups for analysis at the end of the trial. PIIIP was measured using a radioimmunoassay and hyaluronic acid by an enzyme linked binding protein assay

Results: 47 patients completed 3 months treatment (23 UDCA, 24 placebo) and 39 patients completed the six-month study period (20 UDCA, 19 placebo). The entry demographic characteristics of the two groups are well matched.

Conclusions: UDCA treatment improves conventional liver function tests and reduces PIIIP in alcoholic liver disease. This suggests that UDCA could have a beneficial anti fibrotic effect in this condition. A longer term study of UDCA therapy with survival and histological progression as end points is warranted.

Abstract 73, Table 1 Data expressed as percentage change from entry value

74. ULTRASOUND AND FIBROSIS MARKERS IN THE ASSESSMENT OF HEPATIC FIBROSIS

Background: Liver biopsy is commonly performed in patients with liver disease. The procedure is increasingly not required for diagnosis and can have serious morbidity and even mortality. Hence, a major indication for a biopsy is the assessment of the degree of fibrosis, an important factor in determining management and prognosis. Serum markers of fibrosis have been shown to correlate with hepatic fibrosis in a number of liver conditions but are not in clinical usage because of their lack of specificity. Ultrasonography has been used to assess fibrosis but has lacked sensitivity in detecting early fibrosis.

Aims: To determine if the combination of procollagen III N terminal peptide (PIIIP) and ultrasound assessment of fibrosis could be used to identify patients with significant hepatic fibrosis.

Methods: Patients attending for Day Case liver biopsy had a pre-procedure single operator (AEAJ) ultrasound that was retrospectively graded for fibrosis. At biopsy, serum was collected for later batched analysis of PIIIP using a commercial radioimmunoassay. PIIIP <0.5μg/mL was taken as no fibrosis. A single operator (CFF) staged the liver biopsy for fibrosis based on Ishak. Patients were designated “No Fibrosis” if Ishak stage 0 or 1 and for all other stages “Fibrosis”.

Results: Data were available on 66 patients. 25 alcoholic liver disease, 24 viral hepatitis, and 15 “Other” (eg. PBC, PSC, abnormal LFT's etc). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) was calculated for each modality and a combined modality of PIIIP and US against histologically defined fibrosis.

Conclusions: Significant fibrosis can be excluded with 100% confidence, in a range of liver diseases, in patients with a PIIIP <0.5μg/mL and a negative US for fibrosis. This means liver biopsy could be safely avoided in these patients.

Abstract 74, Table 1

View Abstract

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.