Differential gene expression pattern in early gastric cancer by an integrative systematic approach

Abstract

To elucidate the molecular basis of early gastric cancer (EGC), the genome-wide expression pattern of cancer and normal tissues from 27 patients were analyzed by a microarray-based method. Using an integrative systematic bioinformatics approach, we classified the differentially expressed genes in EGC. Interestingly, the more highly expressed genes in EGC exhibited the most significant correlation with cell migration and metastasis. This implies that, even at the early stage of gastric cancer, the molecular properties usually observed in late-stage cancer are already present. Furthermore, we have found a novel association between the expression pattern and molecular pathways of EGC and estrogen receptor α (ERα)-negative breast cancer through cross-experimental analysis. These results provide new insights into the biological properties of EGC, as well as yielding useful basic data for the study of molecular mechanisms of EGC carcinogenesis.

Figure 1.

The GO analysis by BiNGO. (A) The upregulated molecular functions in EGC tissues. (B) The upregulated biological process in EGC tissues. (C) The downregulated molecular functions in EGC tissues. (D) The downregulated biological process in EGC tissues. The Information is presented as a percentage of x/n (x, the number of genes in a cluster annotated for a certain GO term; n, the number of genes in the reference set annotated for a certain GO term).

Figure 2.

The expression of MMPs in EGC tissues. (A) Expression levels of MMPs genes in microarray data. The vertical axis represents fold-change of the cancer tissues over normal tissues. (B) mRNA expression of MMPs using RT-PCR. Three pairs of non-cancerous and tumor tissues from the microarray analysis were used. (N, adjacent non-cancerous gastric tissue; T, EGC tissue).

Figure 3.

Hierarchical clustering. Genes up- or downregulated in EGC as well as 5 independent cancer types were used in the hierarchical clustering analysis. Each cancer type is presented with the following column side-bars: EGC (brown), ERα-negative breast cancers (orange), bladder cancer (grey), Ewing sarcoma (black), small cell lung cancers (yellow) and LNCaP prostate cancer cell lines (blue). Seven MMP genes are presented with row side-bars.

Figure 4.

The expressions (Z-scores) of 7 MMP genes. Z-score zero corresponds to the mean of all expressions in each cancer type. (EGC, the EGC tissues; ERα (-), ERα-negative breast cancer; Lung, lung small cell cancer; LNCaP, LNCaP prostate cancer cell lines; Bladder, bladder cancer; and Ewing, Ewing sarcoma).

Figure 5.

The upregulated genes belonging to the KEGG TLR signaling pathway (A) and cell cycle pathway (B) in our EGC tissues.