PMC

HCC with Poor and Good Prognosis Differ in Their Copy Number Pattern. (A) Unsupervised hierarchical clustering of the weighted SCNAs profile 76 HCC cases revealed clusters C1, C2, C3 and C4. (B) Multidimensional scaling shows close positioning of clusters C1 and C3 as well as of clusters C2 and C4. (C) Kaplan-Meier survival analysis of these four clusters reveals that clusters C1 and C3 have good prognosis, whereas, clusters C2 and C4 have poor prognosis. The statistical p-value was generated by the Cox-Mantel log-rank test.

SH2D4A, SORBS3 and PROSC Inhibit Colony Formation and Cell Migration In Vitro. (A and B) Colony formation assay and (C and D) cell migration assay of Hep3B and HuH1 transfected with the vector control or chr 8p gene as indicated. Data represent averages ±SD. Colony formation and migration assays were performed in quintuplets for Hep3B and in triplicates for HuH1. ND: not determined.

The 10-’Driver’ Gene-Signature Can Predict Survival Outcome in HCC Tumor Specimens and in Breast Cancer. (A) Kaplan-Meier overall survival on the independent validation cohort2 from the LCI by predicted classification of G1 and G2 by SVM. (B) Kaplan-Meier overall survival based on the predicted classification of G1 and G2 by gene expression of the LEC validation cohort by SVM. (C) Forest plot of the hazard ratios for poor survival of six breast cancer studies with varying percentage of node-negative patients. HR (95%CI), hazard ratio (95% confidence interval); N−, node-negative; N+; node-positive.

SH2D4A and SORBS3 Inhibit Tumor Growth In Vivo. (A) Tumor incidence of Hep3B cells transfected with vector control, SH2D4A, SORBS3 or PROSC cDNA after subcutaneous injection into immune-compromised mice (n=10). Tumor incidence was observed twice per week. The log-rank p-value is indicated. (B) Growth curve of tumor xenografts of Hep3B cells transfected with vector control, SH2D4A, SORBS3 or PROSC (n=10). Data represent averages ± SEM. * p<0.05, ** p<0.005 by two-sided Student’s t-test. (C) Representative nude mice and (D) subcutaneous tumors 34 days after subcutaneous injection. The experiment had to be terminated after 34 days because of tumor burden. (E) Relative gene expression of Hep3B cells transfected with vector control, SH2D4A, SORBS3 or PROSC, respectively, and of subcutaneous tumors was determined by real-time qRT-PCR. The first two bars represent Hep3B cell lines prior to subcutaneous injections. Quantitative RT-PCR was performed in triplicates. Data represent averages ±SD.

Good (G1) and Poor (G2) Survival HCC Subgroups Differ in Distinct Genomic Areas (A and B) Frequencies of significant increases in SCNAs are plotted as a function of genome location of G1 and G2 HCC subgroups, respectively. Positive values indicate frequencies of samples showing copy number increases [log2 (copy number)>0.5] and negative values indicate frequencies of samples showing copy number decreases [log2 (copy number)<−0.5]. Chromosome boundaries and centromere position are indicated by vertical solid and dashed lines, respectively. Horizontal dashed blue lines indicate a frequency of 0.2. Grey and white boxes indicate tumor samples with or without chromosome 1q gain, respectively. (C and D) Differences (y axis) between frequencies of gain and loss across the genome for G1 versus G2 subtypes are shown. SCNA frequencies are plotted as a function of positioned location in the genome with positive values indicating higher frequencies in G1 versus G2. Horizontal dashed blue lines indicate frequency differences of 0.2. (E) A representative case with chromosome 8p deletion. Dots represent single probes, red dots represent amplified and green dots represent lost genomic regions.

Integration and Correlation of the Global SCNA and Gene Expression Profiles. (A) Schematic overview of the study design. (B) The density histogram shows the distribution of the Pearson correlation coefficients of gene expression and arrayCGH data from 60 tumor tissues. The blue line represents the density distribution of the 1000-fold random permutation of the data and the red line represents the density distribution of the Pearson correlation coefficients. (C) The density histogram shows the Pearson correlation coefficient of the gene expression of the non-tumor tissue and paired arrayCGH data of the cancerous tissue. (D) In the lower panel, frequencies of significant aberration in SCNAs are plotted as a function of genome location for 76 clinical specimens. Positive values indicate frequencies of samples showing copy number increases [log2(copy number)>0.5; shown in red] and negative values indicate frequencies of samples showing copy number decreases [log2(copy number)<−0.5; shown in green]. Chromosome boundaries and centromere position are indicated by vertical solid and dashed lines, respectively. Horizontal dashed blue lines indicate frequency of + and −0.5. The upper panel shows the position of correlating genes which are more than two-fold up (indicated in red) or down (green) regulated compared to normal liver.