Integrative genomic analyses reveal clinically relevant long noncoding RNAs in human cancer

Abstract

Despite growing appreciation of the importance of long noncoding RNAs (lncRNAs) in normal physiology and disease, our knowledge of cancer-related lncRNAs remains limited. By repurposing microarray probes, we constructed expression profiles of 10,207 lncRNA genes in approximately 1,300 tumors over four different cancer types. Through integrative analysis of the lncRNA expression profiles with clinical outcome and somatic copy-number alterations, we identified lncRNAs that are associated with cancer subtypes and clinical prognosis and predicted those that are potential drivers of cancer progression. We validated our predictions by experimentally confirming prostate cancer cell growth dependence on two newly identified lncRNAs. Our analysis provides a resource of clinically relevant lncRNAs for the development of lncRNA biomarkers and the identification of lncRNA therapeutic targets. It also demonstrates the power of integrating publically available genomic data sets and clinical information for discovering disease-associated lncRNAs.

Figure 1. Human Exon array re-annotation and lncRNA classification

Affymetrix Human Exon array probe re-annotation pipeline for lncRNA was shown in (a). (b) Adopting the classification scheme from a previous study (Ref. 20), lncRNA were classified into four categories: intergenic, overlapping, intronic and exonic on the basis of their relationship with protein-coding genes. (c) Pie charts showed the number of lncRNA in each category for all collected lncRNA and for those with at least 4 uniquely mapped exon array probes.

Figure 2. The number and the expression profile of lncRNA that have disease-specific or subtype-specific expression in prostate cancer, GBM, OvCa and Lung SCC

(a) The expression level of lncRNA that showed significantly differential expression between cancer and normal prostate tissues were shown in heatmap across 29 normal prostate samples, 131 primary and 19 metastatic prostate tumor samples. Several known cancer-related lncRNA or lncRNA with established function in non-cancer context were highlighted. (b) Venn diagram represented the number of subtype-specific lncRNA in three cancers. The expression profile of top 100 lncRNA that exhibited significantly higher expression in one subtype than the others for (c) GBM, (d) OvCa and (e) Lung SCC were shown in heatmap (Note: the rank was based on the ascending order of the p-value). Tumor samples were hierarchically clustered within each subtype.

Figure 3. LncRNA associated with prognosis or in the genomic regions of SCNA

(a) Kaplan-Meier curve of two patient groups with higher (top 50%, n = 64) and lower expression (bottom 50%, n = 64) of ENSG00000261582 in Lung SCC and OvCa (Red: higher expression, blue: lower expression) was shown. The boxplot demonstrated that ENSG00000261582 expressed higher in the “differentiated” subtype of OvCa than the other subtypes. Both the p-value of multivariate Cox model for lncRNA expression and the p-value of log-rank test were shown. (b) Kaplan-Meier curve for overall and progression-free survival of two patient groups with higher (top 50%) and lower expression (bottom 50%) of ENSG00000225128 in OvCa was shown. (c) The number of lncRNA located in the SCNA (gain) and SCNA (loss) regions in different cancers was shown as Venn diagrams.

Figure 4. The genetic alteration and the expression profile of PCAN-R1 and PCAN-R2 in normal prostate tissues or prostate tumors and their transcript structure in cell line

(a) The heatmap showed the expression of PCAN-R1 and PCAN-R2 in normal prostate tissue, primary and metastatic prostate cancer. (b) The boxplot of PCAN-R1 and PCAN-R2 expression in tumors with genomic amplification (n = 7 and n = 9) and in the tumors without genomic amplification (n = 121 and n = 119) were compared. The boxplot showed the expression distribution of PCAN-R1 and PCAN-R2 in two groups and mann-Whitney U test was performed for the comparison. (c) The transcript structures of PCAN-R1 and PCAN-R2 from Ensembl annotation and determined by 5′ and 3′ RACE experiments in LNCaP cell were shown. In addition, the H3K4me3, and DNase I hypersensitive region profile in the same cell line were shown.

Figure 5. Functional validation of PCAN-R1 and PCAN-R2

(a) The Northern blot of PCAN-R1 and PCAN-R2 transcripts was shown (Mr: RNA marker). (b) The relative expression level of PCAN-R1 and PCAN-R2 upon knockdown by two different siRNA (orange and green) and upon control siRNA treatment (purple) was shown. (c) The growth curves of LNCaP cell with or without targeted siRNA-mediated knockdown of PCAN-R1 or PCAN-R2 were shown. The growth curves of control siRNA-treated cells and the growth curves of two targeted siRNA-treated cells were plotted in purple, orange, and green, respectively. Data were shown as Mean+S.D. n=3. (d) The number of soft-agar colony formation of LNCaP cell with or without targeted siRNA-mediated knockdown of PCAN-R1 or PCAN-R2 was shown.