ABSTRACT: Purpose: Hepatoblastoma (HB) is a malignant pediatric hepatic progenitor cell tumor with uncertain etiology. To discover genomic biomarkers associated with this malignancy, we conduct the study. Experimental Design: A panel of ten cases and two pairs of identical twins for HB were screened using high-resolution array-based comparative genomic hybridization (array CGH) technology. Genomic DNA from the peripheral blood was subject to our laboratory investigation. Results: In total of the twelve HB cases tested, we identified four high recurrent copy-number variations (CNVs) including gain on 1p13.3 (25%) and losses on 5p15.33 (33.3%), 16q12.2 (33.3%), and 19q13.42 (25%). Additionally, the twin study indicated that microdeletions of 3p12.1 and 12p13.2 correlated with hepatoblastoma occurrence. Among these candidate CNVs, 5p15.33 and 16q12.2 not only the most prevalent genomic deletion in this survey, but also encompassed hepatic expressed gene, Zinc finger, DHHC-type containing 11B (ZDHHC11B) and Zinc finger, DHHC-type containing 11 (ZDHHC11) map to 5p15.33; carboxylesterase 4-like (CES4) map to 16q12.2. Furthermore, the microdeletion of 5p15.33 was associated with high mortality rate (3/3; P = 0.018) in patients with native liver. Currently, the only surviving patient with 5p15.33 microdeletion is who has receipt liver transplantation. Conclusions: These results suggested that losses on 3p12.1, 5p15.33, 12p13.2 and 16q12.2 may be pathogenic for HB. In addition, HB patient who affected by 5p15.33 microdeletion seems to carry high risk of mortality unless undergo liver transplantation. Thus, 5p15.33 microdeletion was suggested as potential prognostic biomarker for poor survival in hepatoblastoma. A panel of ten cases and two pairs of identical twins for HB were screened using high-resolution array-based comparative genomic hybridization (array CGH). Genomic DNA from the peripheral blood was subject to our laboratory investigation. For array CGH experiment, we used commercial human male and female genomic DNA as control samples.In order to build up list of HB candidate CNVs, a strategy organized into two phases was designed. In phase one, the copy-number polymorphisms (CNPs) found in healthy control were excluded. Thus, two array hybridization experiments were performed with gender-matched DNA pools from healthy Taiwanese and Caucasian (Promega, Madison, WI, USA). Then the CNVs observed in these two arrays including microdeletions on 3q26.1, 4q13.2, 7q34, 9q33.2, 19q13.41, and Xp11.4, and amplification on 1p36.12, 1p22.1, 2p24.1, 7p22.3, 8q24.3, 9q34.11, 10q26.3, 11p15.4, 14q32.34, 17p11.2, 17q25.3, 19p13.3, 20q11.22, 21q22.3, 22q11.21, and 22q13.33 which were excluded from subsequent data analysis. We also excluded the CNVs observed in the healthy siblings of the two identical twins. In phase two, CNVs with altered patterns (i.e. gain or loss) that were inconsistent among patients were extracted from candidate list.
