C.-J. Chen

Synergism between p53 and Mcl-1 in protecting from hepatic injury, fibrosis and cancer.

BACKGROUND & AIMS Mcl-1-deficient hepatocytes are prone to undergo apoptosis. The tumor suppressor protein p53 plays an important role in apoptosis control as well as other cellular responses. This study was initially aimed to examine whether p53 was involved in Mcl-1 deficiency-induced apoptosis of hepatocytes. METHODS Hepatocyte-specific Mcl-1 knockout (Alb-Mcl-1(-/-)) mice and Alb-Mcl-1(-/-) mice in wild-type or p53-deficient background were generated and characterized. RESULTS Alb-Mcl-1(-/-) mice were viable, but their liver cells were prone to undergo apoptosis and manifested a slightly elevated level of p53. To examine the role of p53 in Alb-Mcl-1(-/-) livers, Alb-Mcl-1(-/-) mice without p53 (DKO mice) were characterized. Unexpectedly, although p53-deficient mice appeared to be developmentally normal, DKO mice were highly susceptible to neonatal death (∼60%). Further analysis revealed that such an early lethality was likely due to hepatic failure caused by a marked reduction of fully-differentiated hepatocytes at the perinatal/neonatal stage. Moreover, those DKO mice that did survive to adulthood manifested more severe liver damage than Alb-Mcl-1(-/-) mice, suggesting that p53 was activated in Alb-Mcl-1(-/-) livers to promote cell survival. Microarray followed by quantitative PCR analysis suggested that p21(Waf1/Cip1), one p53 target gene with apoptosis-inhibitory function, is likely involved in the protective role of p53 in Alb-Mcl-1(-/-) livers. Moreover, we demonstrated that loss of p53 promoted liver fibrosis and tumor development in Alb-Mcl-1(-/-) mice. CONCLUSIONS This study revealed an unexpected synergism between Mcl-1 and p53 in protecting from hepatic injury, fibrosis, and cancer.

Thymic epithelial β-catenin is required for adult thymic homeostasis and function

The role of β‐catenin in thymocyte development has been extensively studied, however, the function of β‐catenin in thymic epithelial cells (TECs) remains largely unclear. Here, we demonstrate a requirement for β‐catenin in keratin 5 (K5)‐expressing TECs, which comprise the majority of medullary TECs (mTECs) and a progenitor subset for cortical TECs (cTECs) in the young adult thymus. We found that conditionally ablated β‐catenin in K5+‐TECs and their progeny cells resulted in thymic atrophy. The composition of TECs was also aberrantly affected. Percentages of K5hiK8+‐TECs, K5+K8–‐TECs and UEA1+‐mTECs were significantly decreased and the percentage of K5loK8+‐TECs and Ly51+‐cTECs were increased in β‐catenin‐deficient thymi compared with that in the control thymi. We also observed that β‐catenin‐deficient TEC lineage could give rise to K8+‐cTECs more efficiently than wild‐type TECs using lineage‐tracing approach. Importantly, the expression levels of several transcription factors (p63, FoxN1 and Aire), which are essential for TEC differentiation, were altered in β‐catenin‐deficient thymi. Under the aberrant differentiation of TECs, development of all thymocytes in β‐catenin‐deficient thymi was impaired. Interleukin‐7 (IL‐7) and chemokines (Ccl19, Ccl25 and Cxcl12) levels were also downregulated in the thymic stromal cells in the mutants. Finally, introducing a BCL2 transgene in lymphoid lineages, which has been shown to rescue IL‐7‐deficient thymopoiesis, partially rescued the thymic atrophy and thymocyte development defects caused by induced ablation of β‐catenin in K5+‐TECs. Collectively, these findings suggest that β‐catenin is required for the differentiation of TECs, thereby contributing to thymocyte development in the postnatal thymus.

722 INCIDENCE AND DETERMINANTS OF SPONTANEOUS HEPATITIS B SURFACE ANTIGEN SEROCLEARANCE: A COMMUNITY-BASED LONG-TERM FOLLOW-UP STUDY

722 INCIDENCE AND DETERMINANTS OF SPONTANEOUS HEPATITIS B SURFACE ANTIGEN SEROCLEARANCE: A COMMUNITY-BASED LONG-TERM FOLLOW-UP STUDY J. Liu, H.-I. Yang, M.-H. Lee, S.-N. Lu, C.-L. Jen, L.-Y. Wang, S.-L. You, U.H. Iloeje, C.-J. Chen, R.E.V.E.A.L.-HBV. Genomics Research Center, Academia Sinica, Taipei, Taiwan R.O.C.; Emory University Rollins School of Public Health, Atlanta, GA, USA; Graduate Institute of Epidemiology, National Taiwan University, Taipei, Department of Gastroenterology, Kaohsiung Chang-Gung Memorial Hospital, Kaohsiung, MacKay Medical College, Taipei, Taiwan R.O.C.; Research and Development, Bristol-Myers Squibb, Wallingford, CT, USA E-mail: jessica.b.liu@gmail.com

O016 : Hepatitis B surface antigen and DNA levels can identify inactive carriers and predict lower risk for hepatocellular carcinoma and cirrhosis among genotype B and C chronic hepatitis B carriers

O015 QUANTITATIVE MEASUREMENTS OF MUTATIONS IN BASAL CORE PROMOTOR REGION SUGGEST A LOSS OF IMMUNE TOLERANCE IN HBeAg POSITIVE CHRONIC HEPATITIS B INFECTIONS G. Rosenberg, J. Bayliss, L. Yuen, A. Gaggar, K. Kitrinos, M. Subramanian, E. Gane, H.L. Chan, R. Hammond, S. Bowden, P. Revill, S. Locarnini, A. Thompson. VIDRL, Melbourne, Australia; Gilead Sciences, Foster City, CA, United States; New Zealand Transplant Unit, Auckland, New Zealand; The Chinese University of Hong Kong, Hong Kong SAR, Hong Kong, China; St Vincent’s Hospital, Melbourne, Australia E-mail: gillian.rosenberg@mh.org.au

P675 RISK PREDICTORS FOR LIVER CANCER AND CIRRHOSIS AMONG CHRONIC HEPATITIS B PATIENTS WITH UNDETECTABLE HEPATITIS B VIRAL LOADS

term trajectories of serum HBVDNA levels in a treatment-naive cohort of chronic hepatitis B patients. Methods: A total of 3140 participants from the REVEAL-HBV cohort were included in this analysis. They were anti-HCV-seronegative and free of liver cirrhosis at study entry. Long-term trajectories of serum HBVDNA levels were determined using group-based trajectory methods and Cox proportional hazards models. Results: Serum HBsAg levels at study entry was significantly associated with long-term HBVDNA trajectories (P 10 copies/mL, the higher the baseline serum HBsAg levels were the less likely follow-up serum HBVDNA levels decreased to lower levels. The adjusted odds ratio (95% confidence interval) of having serum HBsAg levels at study entry >1000 IU/mL for participants with HBVDNA trajectories of persisting at 300– 10 copies/mL was 2.43 (2.01–2.95), 2.70 (2.01–3.63), 3.50 (2.67– 4.58), 7.96 (5.97–10.62), and 22.69 (14.78–34.82), respectively, in comparison with participants with undetectable serum HBVDNA at study entry or during follow-up. Conclusions: Baseline serum HBsAg levels may predict long-term viral load trajectories in chronic hepatitis B patients.

687 SERUM IP-10 LEVEL ASSOCIATED WITH HEPATOCELLULAR CARCINOMA DEVELOPMENT IN PATIENTS WITH CHRONIC HEPATITIS B

correlation remained significant in patients with early-stage HCC or with normal serum AFP level. Multivariate analysis indicated that LARP1 expression might be an independent prognostic indicator of the survival of patients with HCC. Furthermore, we found that ectopic expression of LARP1 induced, while silencing LARP1 inhibited, cell invasion and anchorage-independent growth ability. Moreover, we demonstrated that the upregulation of LARP1 could reduce the expression of e-cadherin and induce the expression of fibronectin. Conclusions: Our findings suggest that the LARP1 protein is a valuable marker of HCC progression and may represent a novel prognostic biomarker and therapeutic target for the disease.

1397 SINGLE NUCLEOTIDE POLYMORPHISMS ASSOCIATED WITH SEROCLEARANCE OF HEPATITIS B SURFACE ANTIGEN IN CHRONIC HEPATITIS B PATIENTS WITH SEROCLEARANCE OF HBV DNA

as fibrolamellar hepatocellular carcinoma. No guidelines exist regarding the role of liver biopsy in the diagnosis of FNH. We sought to determine the diagnostic accuracy of imaging for FNH in children without comorbidities, as compared to the gold standard liver biopsy. Methods: 238 consecutive patients (age <18 years) referred to SickKids with a liver mass that underwent biopsy between 1998– 2011 were retrospectively ascertained. Individuals with a history of malignancy, liver disease, liver transplant, or syndromes with an increased risk of malignancy were excluded. Diagnostic imaging and biopsy reports were reviewed and compared. Results: After excluding 170 cases using the criteria listed above, 68 children with a liver mass, diagnostic imaging, and biopsy data were reviewed. Based on histology, the most common diagnosis was hepatoblastoma (33/68). 19 cases of FNH were found in children without comorbidities (28%) after excluding 6 cases with history of malignancy (2), liver disease (1), liver transplant (1) or underlying syndromes (2). Of the 19 children, their mean age at diagnosis was 11.4±5.3 years, they had female preponderance (74%) and a mean follow-up of 2.5±2.2 years. Of these, 17/19 met standard criteria for FNH on both imaging and histopathology. In 2/19 cases of biopsy-proven FNH, imaging did not suggest FNH. The sensitivity of imaging modalities for FNH diagnosis was 75% for ultrasonography (16/17 cases), 85.7% for MRI (14/17 cases), and 83.3% for CT (12/17 cases). All children whose imaging primarily suspected FNH had this diagnosis confirmed on histology, except for 1 child who had hepatocellular adenoma. The overall performance of imaging for the diagnosis of FNH was sensitivity 89.5%, specificity 97.9%, positive predictive value 94.4%, and negative predictive value 95.9%. Conclusions: In this cohort of children with liver mass and no co-morbidities, a diagnosis of FNH by imaging was highly specific; no malignancies would have been missed by foregoing biopsy and diagnosing FNH based on imaging alone.

475 DIFFERENCES IN THE ROLE OF QUANTITATIVE HEPATITIS B SURFACE ANTIGEN BETWEEN HBEAG SEROPOSITIVE AND SERONEGATIVE INDIVIDUALS WITH CHRONIC HEPATITIS B

2004. They were tested for HBV-related seromarkers at study entry and regularly followed using high-resolution abdominal ultrasonography and serological examinations. The incidence of LC was detected by ultrasonography, computerized linkage with national health insurance profiles, and medical chart reviews. We allocated two-thirds participants (2,228) for model development and the other one-third (1,114) for model validation. The cumulative risk for LC by predictors were estimated by Kaplan-Meier methods. There were three models developed: model 1 included age, gender, ALT, HBeAg, and HBV DNA level; model 2 included quantitative HBsAg in addition; and model 3 further included HBV genotypes. The Cox’s proportional hazards models were utilized to estimate the regression coefficients of each predictor for LC and each coefficient was converted to an integer as a simple risk score. The area under receiver operating curve (AUROC) was used to evaluate the performance of each model. Results: The cumulative risk for LC were 9.8% for genotype B or B+C and 21.0% for genotype C (p < 0.001). Participants with elevated serum levels for HBsAg had an increased cumulative risk of 4.8% for <100 IU/mL, 8.8% for 100–999 IU/mL, and 16.2% for ≥1000 IU/mL (p < 0.001 for trend). Participants with higher total risk scores had an increasing predicted risk for LC. By applying the risk score system, participants in validation set could be categorized as low-, medium-, and high-risk for LC accurately (all p-values <0.001). The AUROCs ranged 0.75–0.80 for the three models. Conclusions: HBV genotypes and serum levels of HBV DNA and quantitative HBsAg had excellent long-term predictability of LC risk. They may be used to derive valid risk prediction models.

Abstract B17: Synergism between p53 and Mcl-1 in hepatocyte maturation and survival

Mcl-1 plays an apical role in many cell survival programs. Mice without Mcl-1 die at early embryonic stage. A conditional knockout approach has thus been employed to study Mcl-1 functions in the liver. Mice without Mcl-1 in hepatocytes (Alb-Mcl-1 −/−) were viable but manifested some defects in the mature liver, including enlarged cell size, enhanced apoptosis and proliferation of liver cells. An increased level of p53 was also observed in adult Alb-Mcl-1 −/− mouse livers. We thus generated Alb-Mcl-1 −/− mice in p53-deficient background to examine whether p53 was involved in Mcl-1 deficiency-induced hepatocyte apoptosis. Loss of p53 in Alb-Mcl-1 −/− mice (DKO mice) resulted in a very high frequency of neonatal death. Further analysis revealed that such early lethality was likely due to hepatic failure caused by a marked reduction of fully-differentiated hepatocytes at the perinatal/neonatal stage. On the other hand, those DKO mice that did survive to adulthood manifested much more severe liver cell damage than Alb-Mcl-1 −/−mice of the same age, suggesting that p53 is activated in Alb-Mcl-1 −/− livers to help resolve Mcl-1 deficiency-induced hepatocyte damage. Last, we demonstrated that p53 enhanced-hepatocyte survival is a cell-autonomous effect, and that such effect is mediated in part through transcriptional up-regulation of p21 Waf1/Cip1 in Alb-Mcl-1 −/− livers. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the Second AACR International Conference on Frontiers in Basic Cancer Research; 2011 Sep 14-18; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2011;71(18 Suppl):Abstract nr B17.

117 SEROMARKERS FOR LONG-TERM PREDICTION OF HEPATOCELLULAR CARCINOMA RISK IN PATIENTS WITH CHRONIC HEPATITIS C

(p=0.05) score: ≤3 = 0, or ≥4 = 2. Two prognostic groups were identified in the TC according to score values ≤2 or ≥3, with 5-yr risk of recurrence of 16.7±2.6% and 47.5±4.7%, respectively (p < 0.0001). The model was validated in the 2 validation cohorts with, in particular, a 5-yr risk of recurrence of 9.0±1.7% (score ≤2) vs 47.3±9.5% (score ≥3), p < 0.001, in the ABM VC. In both cohorts, a score ≤2 was associated with better 5-yr overall survival: RVC: 80.6±3.5% vs 66.7±10.3% (p =0.06) and ABM VC: 67.7±3.4% vs 49.1±7.4% (p =0.002). In patients with score ≤2, recurrence and survival were similar whether patients met or not Milan criteria. Conclusions: 1. AFP level is highly predictive of HCC recurrence after LT. 2. A prognostic score including AFP significantly improves the identification of patients with expanded criteria at low risk of recurrence. 3. A score ≤2 could be proposed for selection of these patients.