Volume 133, Issue 5, Pages 1475-1486 (November 2007) Genetically Distinct and Clinically Relevant Classification of Hepatocellular Carcinoma: Putative Therapeutic Targets Hiroto Katoh, Hidenori Ojima, Akiko Kokubu, Shigeru Saito, Tadashi Kondo, Tomoo Kosuge, Fumie Hosoda, Issei Imoto, Johji Inazawa, Setsuo Hirohashi, Tatsuhiro Shibata Gastroenterology Volume 133, Issue 5, Pages 1475-1486 (November 2007) DOI: 10.1053/j.gastro.2007.08.038 Copyright © 2007 AGA Institute Terms and Conditions
Figure 1 Overall chromosomal alteration profiles of the 87 HCCs revealed by the whole-genomic array CGH method. Chromosomal copy number alterations of each chromosome are presented as a clustered matrix in which the rows represent the individual chromosomal loci and the columns represent each HCC. Each cell in the matrix represents the chromosomal copy number of the relevant locus. The green, red, and yellow colors in the cells reflect chromosomal loss, gain, and amplification, respectively, as indicated in the color scale bar. In hierarchical clustering analysis, cosine correlation was used for calculating the similarity coefficient of each HCC case. Blue and red stars represent homozygous deletions and chromosomal amplifications, respectively. Gastroenterology 2007 133, 1475-1486DOI: (10.1053/j.gastro.2007.08.038) Copyright © 2007 AGA Institute Terms and Conditions
Figure 2 Validation experiments for newly identified chromosomal alterations. (A) To confirm a novel homozygous deletion on 4q35 (185.5–185.8 Mb) (Figure 1 and supplementary Table 2; see supplementary material online at www.gastrojournal.org), qualitative genomic PCR analysis for the genes located within this locus were performed. Novel homozygous deletions of the Caspase3 and IRF2 genes were confirmed clearly in this HCC case. t, DNA from HCC tissue; n, DNA from noncancerous tissue. (B) Similarly, a novel homozygous deletion on 14q32.1 (87.8–89.1 Mb) that contains the CHES1 gene was confirmed by the genomic PCR analysis. (C) To confirm a novel chromosomal amplification on 11q23.3–24, fluorescent in situ hybridization analysis was performed with an HCC in which our CGH analysis had indicated a chromosomal amplification at the relevant locus. Novel chromosomal amplification of the 11q23.3-24 locus was confirmed clearly in this HCC (red signals, left figure), and, interestingly, this HCC also indicated a chromosomal amplification at the CyclinD1 gene locus (red signals, right figure). Gastroenterology 2007 133, 1475-1486DOI: (10.1053/j.gastro.2007.08.038) Copyright © 2007 AGA Institute Terms and Conditions
Figure 3 (A) Hierarchical clustering of the 87 HCC tumors. Chromosomal loci with alterations in more than 20.0% of the cases were selected for hierarchical analysis (999 signal features). Cosine correlation was used for calculating the similarity coefficient of each HCC case. The data are presented in matrix format in which the rows represent the individual chromosomal loci and the columns represent each HCC. Each cell in the matrix represents the chromosomal copy number of the relevant locus. The green, red, and yellow colors in the cells reflect chromosomal loss, gain, and amplification, respectively, as indicated in the color scale bar. Mutational status of the p53 and β-catenin genes and viral infection status of each HCC case are also indicated at the top of the clustering matrix. Black cells indicate HCC cases with gene mutation. Blue and red colors indicate positivity for HCV and HBV, respectively. (B) Kaplan–Meier survival curves of the patients in clusters A and B. Differences in the survival probability of the clusters were tested by the log-rank test. (C) Chromosomal alterations that were altered differentially in clusters A and B. A > B and A < B indicate chromosomal loci that were detected significantly more frequently in either of the clusters (P < .05: (1) 2-sided Student t test of the signal intensities for each BAC clone between the clusters, (2) unifying these univariate P values among each chromosomal band by Fisher’s C method28, and (3) adjustments of the unified P values by the Simes procedure27 were performed as described in previous articles,27,28 and supporting data are indicated in supplementary Table 3 (see supplementary material online at www.gastrojournal.org). Gastroenterology 2007 133, 1475-1486DOI: (10.1053/j.gastro.2007.08.038) Copyright © 2007 AGA Institute Terms and Conditions
Figure 4 PCA plots of the chromosomal alterations in each subcluster. The PCA plot displays the position of each chromosomal alteration reflecting their approximate degree of correlation. Each box-shaped symbol represents a BAC clone mounted on our CGH array. Colors of the box-shaped symbols display the chromosomal locations of the BAC clones, as indicated at the right side of the figure. Gastroenterology 2007 133, 1475-1486DOI: (10.1053/j.gastro.2007.08.038) Copyright © 2007 AGA Institute Terms and Conditions
Figure 5 (A) Overview of the combinations of chromosomal alterations in each subcluster. -, chromosomal loss; +, chromosomal gain; ++, chromosomal amplification. The most pathognomonic chromosomal alterations in each subcluster are indicated as red-colored signages. (B) Cellular proliferation assay of human HCC cells with or without Rapamycin treatment. The y-axis indicates the ratio of cell counts of Rapamycin-treated cells and Rapamycin-untreated counterpart cells. **P < .01, calculated by 2-sided Student t test. The genome types of each line of HCC cells were determined by their chromosomal alteration profiles revealed by the array CGH method (for detail, see supplementary Figure 3; see supplementary material online at www.gastrojournal.org). Gastroenterology 2007 133, 1475-1486DOI: (10.1053/j.gastro.2007.08.038) Copyright © 2007 AGA Institute Terms and Conditions