CIRRHOSIS

s 34

Fridu 11, I 3 April

in all the studied normal liver tissue samples. However, compared with these samples (n = 5), the expression of SFRPs I and 5 decreased I2 and 16 folds, respectively, in cancer cell lines (p=0.003 for both). SFRP3 mRNA expression in cancer cell lines was not significantly different from normal liver tissue expression. Interestingly, the expression of SFRP 1 and 5 were 3 to 7 folds lower in HCC (n=7) than in matching non tumor livers (n= 7) (p = 0.03 both, Mann-Whitney’s U). In addition, SFRPl and 5 mRNA expression were strongly correlated (Spearman’s R=+0.90; n=21; p i0.000001), whereas SFRPl and 3 (Spearman’s R=+0.48; n = 2 1 ; p = 0.029) and sFRP3 and 5 (Spearman’s R=+0.47; n = 21; p = 0.03) were moderately correlated. Moreover, in the HepG2 cell line carrying a constitutively activated form of mutated b-catenin, SFRPl and 5 suppressed Wntib-catenin signalling and inhibited in vitro tumor growth through induction of cell death. Furthermore, SFRPl and 5 suppressed Wntibcatenin activation induced by Wnt-1, Wnt-3a, by a constitutively activated form of b-catenin (dn 29-48) and by LiCI 20 mM in Huh-7 and HEK293T cells. These findings indicate that SFRPs may inhibit the Wntib-catenin signalling pathway in the presence of activating b-catenin mutations or in response to enhanced Wnt activity in the tumor microenvironment. Furthermore, SFRP silencing is a newly identified molecular mechanism involved in tumor progression.

MAPPING A SEX-HORMONE SENSITIVE GENE DETERMINING FEMALE RESISTANCE TO LIVER CARCINOGENESIS IN A CONGENIC F344.BN-HCS4 RAT M.R. De Miglio’, P. Virdis’, D.F. Calvisi’, M. Frau’, M.R. Muroni’, M.M. Simile’, L. Daino’, G.M. Careddu’, E. Sanna-Passino’, R.M. Pascale’, F. Feo’ . ‘Division of Exprrimrntul Puthology und Oncology, Department of Biomedical Sciences, University of Sussari, Sussuri; 2Dioision of Surgeq Institute of Eterinaiy Pathology and of Sussuri, Sussuri, ltuly Obstetrics, Uniuer~sih~ E-mail: [email protected] Background and Aims: Hepatocellular carcinoma (HCC) is prevalently male associated in humans and rodents, but the genes responsible for female resistance to hepatocarcinogenesis remain unknown. Hepatocarcinogenesis is controlled by various genes in susceptible F344 and resistant BN rats. Seven hepatocarcinogenesis susceptibility (Hcs) loci, determining rat susceptibility to HCC, have been identified. Among them, allelic variant of Hcs4 locus in the BN rat strain (B alleles), on centromeric RNO16, controls neoplastic liver nodule volume. To investigate whether Hcs4 locus is implicated in the determination of rat genetic traits, we constructed the F344.BN-Hcs4 recombinant congenic strain (RCS). Methods: RCS was generated by introgressing a 4.41 cM portion of Hcs4 from BN strain in an isogenic F344 background. Male and female congenics were gonadectomized, with some gonadectomized animals being treated with h-estradiol 17-acetate and testosterone. Rats were initiated by diethylnitrosamine and treated according to the “resistant hepatocyte” model. Livers were excised and processed for H&E staining and proliferating cell nuclear antigen (PCNA) immunohistochemistry. Genotyping was done using 128 polymorphic microsatellite markers. Gene expression was assessed by using reverse transcription-PCR. Results: Lesion volume and positivity for PCNA were much higher in lesions of F344 than BN rats, of both sexes. These parameters were lower in females than males. Lesion volume and PCNA values of male RCS were similar to those of F344 rats, but corresponded in females to those of BN females. Carcinomatous nodules and HCC developed, at 32 and 60 weeks, respectively, in male F344 and congenics and, rarely, in F344 females. BN and congenic females developed only eosinophiliciclear cells nodules. Gonadectomy of congenic males, followed by (i-estradiol administration, caused development of preneoplastic lesions comparable to those from BN females. Administration of testosterone to gonadectomized females led to development of preneoplastic lesions as in F344 males. Conclusions: The data indicate a role of homozygous B alleles at Hcs4 in determination of phenotypic patterns of female RCS, and presence at

Hcs4 locus of a high penetrance gene(s), activated by estrogens and inhibitedunaffected by testosterone, conferring resistance to females, whose B alleles provide higher resistance.

Parallel Session 9: CIRRHOSIS AND ITS COMPLICATIONS

PROSPECTIVE ASSESSMENT OF LIVER STIFFNESS FOR THE NON-INVASIVE PREDICTION OF PORTAL HYPERTENSION C. Bureau, S . Metivier, J.M. Peron, M.A. Robic, 0. Rouquet, E. Dupuis, J.P. Vinel. Service d’He~)uto-Gastro-EnterologieFederation Digestive Purpan, Tonlouse, France E-mail: [email protected] Introduction: Prognosis of patients with cirrhosis strongly depends on the presence of portal hypertension (PHT). Several studies have shown that liver stiffness (LS) measured by Fibroscan’e is correlated with liver fibrosis. LS was also found to be associated with PHT in patients after liver transplantation. The aim of our work was to prospectively assess the accuracy of LS for the prediction of PHT in patients with chronic liver diseases. Methods: Between I li2005-10i2006, LS measurement was performed in all patients who had a transjugular liver biopsy. PHT was defined as a porto-systemic pressure gradient (PPC) >5 mmHg and clinically relevant PHT (CRPHT) as a PPG >lOmmHg. Endoscopy was planned in all patients with cirrhosis. LS was evaluated for the prediction of PHT, CRPHT and for the presence of varices. Results: 150 patients (90 males); mean age 54 years, were included. Alcohol and C virus were the main causes of liver disease. 99 (66%) had a PPG > 5 mmHg, 76 (51%) a PPG > 10 mmHg, 88 (59%) had cirrhosis, 70 patients (47%) had OV, large in 47 of them (31%). LS value was correlated with PPG (r=0.858; p 10.01). In multivariate analysis, prothrombin index and LS were the 2 parameters associated with a PPG > S m m H g and a PPG > 10mmHg. AUROC for the prediction of a PPG > 5 and 10mmHg were for LS: 0.948 [0.912-0.9841 and 0.945 [0.904-0.9871 respectively. Considering only patients with cirrhosis, only prothrombin index and LS were also associated with a PPG > 1 OmmHg. AUROC for predicting a GPH > IOmmHg were 0.904 [0.810-0.9971 and 0.865 [0.768-0.9611 for LS and prothrombin index respectively. Finally, choosing a cutoff value of 23 kPa, 92% of the patients were correctly classified as having CRPHT. LS and prothrombin index were less accurate for the prediction of OV Conclusion: Liver stiffness is useful for the prediction of PHT in patients with chronic liver disease. Further prospective studies are needed to assess the accuracy of LS for predicting complications related to PHT.

COMPARISON BETWEEN HEPATIC VENOUS PRESSURE GRADIENT (HVPG) AND LIVER STIFFNESS MEASUREMENT (LSM) IN PATIENTS WITH HCV-RELATED FlBROSlSlClRRHOSlS F. Vizzutti’, U. Arena’, R.G. Romanelli’, L. Rega’, A. Petrarca’, S. Moscarella’ , F. Marra’, G. Laffi’ , M. Pinzani’ . ’Diprtimrnto di Medicinu Internu, Uniurr~sitd di Firenze, Firrnze; Unit2 di Radiodiagnosticu, Dipartimento di Fisioputologiu Clinica, linioersitd di Firrnzr, Firenzr, ltuly E-mail: [email protected] Background: Measurement of HVPG is a standard method for the assessment of portal pressure and its complications. It is well established that

Parallel Session 9: CIRRHOSIS AND ITS COMPLICATIONS the threshold value of the HVPG for the formation ofvarices is 10 mmHg, while 12 mmHg is the threshold for the appearance of other complications, such as variceal bleeding and ascites. LSM has been recently proposed as a non-invasive technique for cirrhosis complications prediction. Aim: Aim of the present study was to evaluate the diagnostic accuracy of LSM compared to that of HVPG in a group of patients with HCV related CLD with different degrees of fibrotic evolution. Patients and Methods: We studied 61 consecutive patients: 47 cirrhotic, and 14 non-cirrhotic (10 F2 and 4 F3, according to METAVIR). On the same day, all patients underwent HVPG determination and LSM performed by FibroScan. Exclusion criteria included: i) BMI 2 3 5 , ii) ascites or other complications of cirrhosis, including cardiopulmonary and renal involvement, iii) transaminases > I OULN, iv) past/ongoing antiviral therapy and/or beta-blockers, v) active alcohol intake, vi) co-infections. Results: A strong relationship between LSM and HVPG measurements was found (r=0.81, p <0.00001). However, while the correlation was excellent for HVPG values less than 10 and 12 mmHg (r= 0.8 I , p i 0.0003 and r = 0.91, p < 0.00001, respectively), linear regression analysis became less optimal for HVPG values 310mmHg (3 =0.35, p <0.0001) and hardly reached statistical significance for values over 12 mmHg (3 = 0.17, 0.02). For the prediction of HVPG 2 10 and 2 12 mmHg, the AUROC pi were 0.99 and 0.92, respectively, and the cut-off values of LSM were 13.6 kPa (sensibility 97%) and 16.3 kPa (sensibility 94%), respectively. In cirrhotic patients LSM positively correlates with the presence of 0.002). The AUROC for the prediction of esophageal varices (EV) (p i EV was 0.76 and a LSM cut-off value was of 17.6 lcPa (sensibility 90%). However, no correlation between LSM and EV size was observed. Conclusions: In conclusion, LSM represents a non-invasive tool for the identification of CLD patients with an HVPG over 10-12mmHg and could be employed for screening patients to be subjected to standard investigations including endoscopy and hemodynamic studies.

ACUTE HEMODYNAMIC RESPONSE TO BETA-BLOCKERS FOR THE PREDICTION OF LONG-TERM OUTCOME IN PRIMARY PROPHYLAXIS OF VARICEAL BLEEDING. FINAL RESULTS OF A PROSPECTIVE STUDY C. Aracil, A. Colomo, I. Ordas, C. Guarner, X. Torras, A. Gallego, C. Villanueva, J. Balanzo. Hospital de la Sunta Cren i Sunt Pun, Burcelonu, SIiuin E-mail: [email protected]; [email protected] A decrease of the hepatic venous pressure gradient (HVPG) to 3 12 mmHg or by ~ 2 0 % from baseline defines hemodynamic response to pharmacological treatment of portal hypertension. It is well established that responders have a lower risk of variceal bleeding and better survival. Accordingly, non-responders may benefit from rescue therapy. However, to determine response, a second hemodynamic study 1-3 months after the baseline is required. Assessing response in a single study may prevent a period of risk (advancing rescue therapy in non-responders) and may avoid a second invasive procedure. The aim of this study was to investigate whether acuteresponse to beta-blockers may provide sufficient prognostic information on long-term outcome in a single hemodynamic study. Methods: A total of 88 cirrhotic patients with large varices and without previous bleeding were included. A baseline hemodynamic study was performed in all, propranolol was administered then (0.15 mg/kg/i.v) and measurements were repeated 20min. later. Nadolol was subsequently administered chronically. A second hemodynamic study was performed 1-3 months later. Response was defined as a decrease in HVPG to > I2 mmHg or 220%. Results: In a mean follow-up of 3 6 f 2 7 months, 14 patients (16%) had variceal bleeding and 14 died. Thirty (34%) were responders in the acute study. Acute responders had a significantly lower probability of bleeding (23% vs 7% at 30 months, P=0.01) and better survival (31% vs 7% at 30m, P i 0.01) than acute non-responders. The second hemodynamic study was performed in 60 patients; 18 of them (30%) were responders. There

s35

was a significant correlation between the acute and chronic changes of HVPG (r=0.7, P <0.001). Although 7/20 (35%) patients were acute but not chronic responders, 5 of them (7 I %) had a chronic decrease of HVPG > 10%. Using ROC curve a decrease of HVPG 2 10% from baseline was the best cut-off to predict bleeding, both for acute and chronic studies. Only 4/40 (10%) patients had chronic but not acute response. Conclusions: Our results suggest that assessing acute hemodynamic response to beta-blockers in a single study may provide an accurate prognostic information on long-term risk of variceal bleeding. This may avoid a second invasive procedure and safe risk-time for rescue therapy of non-responders.

H T H E IMPACT OF EPISODES OF HEPATIC ENCEPHALOPATHY (HE) ON NEUROCOGNITIVE FUNCTION AFTER LIVER TRANSPLANTATION E.U. Sotil’, J. Gottsteinl, C. Randolph’, A.T. Blei’ . ‘Diuision of Heputology, Northwestern Uniusersity, Northwextern Memorial Hospitul, Chicugo, IL; ’Depurtnzent of Nenrology, Loyola liniversity, Loyob linioersiiy Hospital, Chicugo, IL, USA E-mail: [email protected]

Background and Aims: Abnormalities in neurocognitive function have been reported after liver transplantation. However, the influence of the pre-transplant neurological status on post-liver transplant (LT) outcomes has not been examined. If the influence of HE pre-transplant can be still seen after the procedure, the basic paradigm of the pathogenesis of HE (a reversible condition) should be questioned. We examined this problem in a carefully selected group of patients. Methods: Of 24 patients who underwent LT, repeated episodes of H E had been documented pre-LT in group 1 (n= 12) while none in group 2 (n= 12). Age-matched controls were also examined (n=20). Patients from the 2 LT groups were matched by age; they were similar with regards to educational level, etiology of original liver disease, time from listing to LT (average 587 days), time from LT to testing (average 541 days), immunosuppressive medications post-LT as well as the use of medications for diabetes and hypertension. English was the first language in ~ 9 0 % of subjects. Tools to investigate patients included administration of the critical flicker frequency, a neurophysiological assessment (Hepatology 35:357, 2002); 5 tests included in the Psychometric Hepatic Encephalopathy Score (J Hep 34:768, 2001) and the Repeatable Battery for Assessment of Neuropsychological Status (RBANS, J Clin Exp Neuro-psycho1 1998; 20:3 10-9). Quality of life was assessed with the SF-36 questionnaire focusing on physical and mental domains. (Mean*SD)

Post-LT with pre-LT ITI:

Post-LT and no Controls pre-LT ITI:

Age (years) CFF (frequency) PHES Digit symbol (# correct) Number Connect-A (sec) Number Connect-B (sec) Serial Dotting (sec) Line connection (sec) RBANS (total scale) SF-36 Physical Mental

54.0*7.0 40.8*2.3*

54.0*9.0 40.9*2.3

52.0*9.0 43.6*3.0

41.2*8.9* 34.0*8.3 *’ 94.4*32.6* 53.8*11.3 81.1*23.7 84.6* 10.5*

48.7*8.9 22.3*5.5 77.3*37.5 58.7*20.6 76.5*19.1 95.2M.7

54.6M.4 19.6*3.9 54.5*17.0 54.8*18.1 70.7*26.2 50th percentile= 100

45.5*12.4 54.7+11.9

49.3*14.3 52.7+11.1

50th percentile= 50 50th percentile= 50

* p i 0 05 vs controls

vb

‘0 05

group 2

Conclusion: Patients with repeated bouts of HE before LT exhibited functional abnormalities post-LT when tested by a sensitive neurophysiological test (CFF) and two different validated neuropsychological batteries