Microarray expression profiling of dysregulated long non-coding RNAs in triple-negative breast cancer

Abstract

Triple-negative breast cancer (TNBC) represents a collection of malignant breast tumors that are often aggressive and have an increased risk of metastasis and relapse. Long non-coding RNAs are generally defined as RNA transcripts measuring 200 nucleotides or longer that do not encode for any protein. During the past decade, increasing evidence has shown that lncRNAs play important roles in oncogenesis and tumor suppression; however, the roles of lncRNAs in TNBC are poorly understood. To address this issue, we used Agilent human lncRNA microarray chips and bioinformatics tools, including Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG), to assess lncRNA expression in 3 pairs of TNBC tissues. A dysregulated lncRNA expression profile was identified by microarray and verified by qRT-PCR in 48 pairs of breast cancer subtype tissues. Metastasis is the major cause of cancer-related deaths, including those in TNBC, and the presence of dormant residual disseminated tumor cells (DTC) may be a key factor leading to metastasis. ANKRD30A, a potential target for breast cancer immunotherapy, is currently one of the most used DTC markers. Notably, we found the expression levels of the novel intergenic lncRNA LINC00993 to be associated with the expression levels of ANKRD30A. Furthermore, our qRT-PCR data indicated that the expression of LINC00993 was also associated with the expression of the estrogen receptor. In conclusion, our study identified a set of lncRNAs that were consistently aberrantly expressed in TNBC, and these dysregulated lncRNAs may be involved in the development and/or progression of TNBC.

Keywords: ANKRD30A; LINC00993; expression profile; long non-coding RNA; microarray; triple-negative breast cancer.

Figure 1.

(A) Summary of microarray results. Expression levels of 46,506 lncRNAs and 30,656 mRNAs were assessed in 3 pairs of TNBC tissues using Agilent Human lncRNA 4*180K microarrays. Compared with non-tumor tissues, 2925 lncRNAs (6.29%) and 2314 mRNAs (7.54%) had significant changes in expression levels (fold change >2, p < 0.05). A total of 2041 lncRNAs and 1375 mRNAs were excluded due to low expression levels. A total of 884 lncRNAs were then identified from the screen, with 626 up-regulated and 258 down-regulated. (B) Hierarchical clustering map. Hierarchical clustering analysis of the top 100 dysregulated lncRNAs and mRNAs. This clustering map revealed a set of lncRNAs that were often aberrantly expressed in TNBC compared with non-tumor tissues. Each row represents a single lncRNA or mRNA and each column represents one tissue sample. Expression levels of these transcripts are represented in red (elevated expression) or green (reduced expression), indicating expression above and below the median expression levels, respectively. C: carcinoma group, P: paired non-tumor group.

Figure 2.

(A) Chromosomal locations of variably expressed lncRNAs and mRNAs. The X-axis represents the ordinal of the chromosome, and the “NA” stands for non-annotation. The Y-axis represents the number of lncRNAs that expressed differently in TNBC tissues (fold change>2; p < 0.05). Breast cancer associated genes (ESR1, MKI67, PGR, ERBB2, TP53 and BRCA1) are presented according to their respective chromosomal locations. (B) Volcano plots. The negative log of the p value (base 10) was plotted on the Y-axis, and the log of the fold change (base 2) was plotted on the X-axis. The red points on this graph represent lncRNAs that were significantly differently expressed in TNBC tissues (fold change >2 and p <0.05); the gray points represent the remaining lncRNAs (fold change <2 or p >0.05). (C) Scatter plots demonstrating the heterogeneity between samples. A consistent set of lncRNAs was found to be frequently dysregulated in TNBC despite the heterogeneity of the samples. The X-axis represents the logarithmic (base 2) fluorescence signal values of the microarray probes in a specific sample, while the Y-axis represents the equivalent values in a second sample. The size of the red area corresponds to the number of differently expressed lncRNAs; the larger the red area, the more significant the heterogeneity between samples. C: carcinoma group, P: paired non-tumor group.

Figure 3.

qRT-PCR verification of 8 candidate lncRNAs in 48 pairs of BC tissues of different subtypes. Expression of the 8 candidate lncRNAs in 24 pairs of (A) ER-negative and (B) ER-positive samples and (C) 48 pairs of BC and corresponding non-tumor tissues. (D) Expression levels of the 8 candidate lncRNAs in TNBC tissues, as measured by microarray and qRT-PCR. The Y-axis represents the relative expression levels of lncRNAs assessed by the 2^−ΔCt method and the error bar represents standard deviation. Paired t-tests (2-tailed) were performed to compare the expression levels between carcinoma and non-tumor tissues, and a p value < 0.05 indicated statistical significance. (*p < 0.05, **p < 0.01 and ***p < 0.001).

Figure 4.

GO and KEGG pathway analysis. Functions of significantly differently expressed mRNAs (fold change>2; p < 0.05) were analyzed by GO and KEGG pathway annotations, and functions of novel lncRNAs were deduced by their co-expressed mRNAs. The GO database includes 3 parts (cellular component, biological process and molecular function) that together describe the genes and gene products across all species (http://www.geneontology.org). The KEGG pathway is a collection of manually drawn signal pathway maps and provides a valuable tool for mapping a specific gene to its corresponding pathway (http://www.genome.jp/kegg/).