Integrative genomics analysis of chromosome 5p gain in cervical cancer reveals target over-expressed genes, including Drosha

Abstract

Background: Copy number gains and amplifications are characteristic feature of cervical cancer (CC) genomes for which the underlying mechanisms are unclear. These changes may possess oncogenic properties by deregulating tumor-related genes. Gain of short arm of chromosome 5 (5p) is the most frequent karyotypic change in CC.

Methods: To examine the role of 5p gain, we performed a combination of single nucleotide polymorphism (SNP) array, fluorescence in situ hybridization (FISH), and gene expression analyses on invasive cancer and in various stages of CC progression.

Results: The SNP and FISH analyses revealed copy number increase (CNI) of 5p in 63% of invasive CC, which arises at later stages of precancerous lesions in CC development. We integrated chromosome 5 genomic copy number and gene expression data to identify key target over expressed genes as a consequence of 5p gain. One of the candidates identified was Drosha (RNASEN), a gene that is required in the first step of microRNA (miRNA) processing in the nucleus. Other 5p genes identified as targets of CNI play a role in DNA repair and cell cycle regulation (BASP1, TARS, PAIP1, BRD9, RAD1, SKP2, and POLS), signal transduction (OSMR), and mitochondrial oxidative phosphorylation (NNT, SDHA, and NDUFS6), suggesting that disruption of pathways involving these genes may contribute to CC progression.

Conclusion: Taken together, we demonstrate the power of integrating genomics data with expression data in deciphering tumor-related targets of CNI. Identification of 5p gene targets in CC denotes an important step towards biomarker development and forms a framework for testing as molecular therapeutic targets.

Figure 1

Identification of chromosome 5p genomic alterations in cervical cancer.A-B. Copy number alterations in log2 ratio of chromosome 5 identified by 250 K NspI SNP array. Each vertical column represents a sample with genomic regions representing from pter (top) to qter (bottom). Prefix "T" indicates primary tumor; "CL" indicates cell line. The blue-red scale bar (-1 to +1) at the bottom represents the copy number changes relative to mean across the samples. The intensities of blue and red indicate relative decrease and increase in copy numbers, respectively. G-banded ideogram of chromosome 5 is shown on the extreme right in each panel. A. Identification of 5p gain as the most frequent genomic alteration in CC. All tumors that exhibited 5p gain are shown. Inferred copy number view of HeLa cell line showing 5p gain from normal (2N) (red line) is shown on right. B. 5q deletions in CC. All tumors that exhibited 5q deletions arranged from largest to smallest deletion are shown. Inferred copy number view of T-1875 showing 5q deletion from normal (2N) (red line) is shown on right. C-F. Fluorescence in situ hybridization (FISH) identification of 5p gains in invasive cancer (C-E) and high-grade cervical intra-epithelial lesion (HSIL) (F). Green signals represent 5p15.2 probe and Red signals represent probe mapping to 5q31 region used as control. Panels C-D represent invasive CC cases with increased copies of 5p (green) compared to 5q (red). Note the concordance of SNP copy number changes (panel A) and increased copies of 5p by FISH (panels C-E) for tumors T-207, T-218, and T-1981, respectively. Panel F showing 4 copies of chromosome 5p and 5q (tetrasomy 5) on a pap smear from HSIL.

Figure 2

Supervised analysis of over expressed genes identified as a consequence of gain of chromosome 5q in cervical cancer. Significantly differentially expressed genes were identified by filtering all the over expressed genes on chromosome 5p between tumor that showed gain of 5p and tumors with out 5p gain. In the matrix, each row represents the gene expression relative to group mean and each column represents a sample (shown on Top). T, represents primary tumor; CL, represents cell line. The dendrogram on left shows unsupervised clustering of genes differentially expressed between tumors with and without gain. The names of genes are shown on right. The scale bar (-2 to +2) on the bottom represents the level of expression with intensities of blue represents decrease and red for increase in expression. The groups within tumors shown at top represent no gain of chromosome 5p (I) and 5p gain (II).

Figure 3

Relative expression of differentially expressed genes as a consequence of 5p gain in relation to GAPDH in normal and tumors with and without gain of 5p gain. Genes are shown on top left-side corner of each panel.