Transcriptome profiling of the cancer, adjacent non-tumor and distant normal tissues from a colorectal cancer patient by deep sequencing

Abstract

Colorectal cancer (CRC) is one of the most commonly diagnosed cancers in the world. A genome-wide screening of transcriptome dysregulation between cancer and normal tissue would provide insight into the molecular basis of CRC initiation and progression. Compared with microarray technology, which is commonly used to identify transcriptional changes, the recently developed RNA-seq technique has the ability to detect other abnormal regulations in the cancer transcriptome, such as alternative splicing, novel transcripts or gene fusion. In this study, we performed high-throughput transcriptome sequencing at ~50× coverage on CRC, adjacent non-tumor and distant normal tissue. The results revealed cancer-specific, differentially expressed genes and differential alternative splicing, suggesting that the extracellular matrix and metabolic pathways are activated and the genes related to cell homeostasis are suppressed in CRC. In addition, one tumor-restricted gene fusion, PRTEN-NOTCH2, was also detected and experimentally confirmed. This study reveals some common features in tumor invasion and provides a comprehensive survey of the CRC transcriptome, which provides better insight into the complexity of regulatory changes during tumorigenesis.

Figure 1. Differential expression analysis of cancer, adjacent non-tumor and distant normal tissue.

A: The scatter plot for global expression between samples; the Pearson correlation coefficient is shown; B: Hierarchical clustering of differentially expressed genes (DEGs) among samples; C: Venn diagram to illustrate the overlapped DEGs between samples; D: Volcano plots for all the genes in each comparison. The red and blue dots indicate that up- and down-regulated DEGs were significant at q values less than 0.01.

Figure 2. The functional enrichment of cancer-specific dysregulated genes identified in the Gene Ontology analysis.

We only chose GO categories that enriched cancer-related dysregulated genes but did not enrich dysregulated genes that were identified when comparing normal tissue with adjacent non-tumor tissue. The cancer-specific dysregulated genes were categorized as significantly up- or down-regulated genes in cancer tissue. The level of significance is indicated by different colors. The “s1”, “s2” and “s3” indicators denote “cancer”, “adjacent non-tumor” and “distant normal” tissues, respectively.

Figure 3. The differentially expressed genes detected by RNA-seq are confirmed by qRT-PCR.

qRT-PCR was performed for five genes that are identified as differential expressed genes between CRC and other two tissues. The expression level of each gene was normalized to the level in normal tissue. The “s1”, “s2” and “s3” indicators denote “cancer”, “adjacent non-tumor” and “distant normal” tissues, respectively.

Figure 4. Analysis of differential exon skipping

(DES) events among samples. A: Venn diagram of the number of DES events; B: The overlap between differentially expressed genes and genes with DES events.

Figure 5. RNA-seq reads coverage of the gene ADD3.

The RNA-Seq reads were mapping to the UCSC reference genome (hg19) of ADD3. The CRC tissue tracks were shown in red, the adjacent non-tumor in green and the normal tissue in blue. The counts of reads spanning the junction of exons were shown.

Figure 6. Illustration of the PTGFRN-NOTCH2 gene fusion in cancer tissue.

A: The inter-chromosomal gene fusion involves the PTGFRN (shown in yellow) and NOTCH2 (shown in green) loci; the distance between these two loci is about three Mbp. The fusion events were detected by pair-end reads that spanned the fusion region and reads that crossed the fusion region; B: The comparison of the reads mapping results for the fusion transcripts among the three samples. The structure of the fusion gene is at the bottom. The reads counts for “normal”, “adjacent non-tumor” and “cancer” tissue are denoted as “green”, “blue” and “red” bars, respectively. C: The RT-PCR with the sequencing results of the fusion transcript in the three samples. The PCR primer is marked in panel A. D: The prediction of ORF of and its function for PTGFRN-NOTCH2 by bioinformatics, The start codon was using the start codon of NOTCH2 and the stop codon (red “*”) located in the fusion sequence from PTGFRN. The fusion peptide contained the domains (predicted by CD-Search in NCBI) in its region from NOTCH2, like EGF domains, but some key domains in the protein NOTCH2, such as NOTCH domain and Ankyrin repeats, were missing.