Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis

Abstract

Large intervening non-coding RNAs (lincRNAs) are pervasively transcribed in the genome yet their potential involvement in human disease is not well understood. Recent studies of dosage compensation, imprinting, and homeotic gene expression suggest that individual lincRNAs can function as the interface between DNA and specific chromatin remodelling activities. Here we show that lincRNAs in the HOX loci become systematically dysregulated during breast cancer progression. The lincRNA termed HOTAIR is increased in expression in primary breast tumours and metastases, and HOTAIR expression level in primary tumours is a powerful predictor of eventual metastasis and death. Enforced expression of HOTAIR in epithelial cancer cells induced genome-wide re-targeting of Polycomb repressive complex 2 (PRC2) to an occupancy pattern more resembling embryonic fibroblasts, leading to altered histone H3 lysine 27 methylation, gene expression, and increased cancer invasiveness and metastasis in a manner dependent on PRC2. Conversely, loss of HOTAIR can inhibit cancer invasiveness, particularly in cells that possess excessive PRC2 activity. These findings indicate that lincRNAs have active roles in modulating the cancer epigenome and may be important targets for cancer diagnosis and therapy.

Figure 1. HOX lincRNAs are systematically dysregulated in breast carcinoma and have prognostic value for metastasis and survival

a, Heat map representing unsupervised hierarchical clustering of expression values of a panel of primary and metastatic breast cancers relative to normal breast epithelial cells (pooled from 5 breast organoids). An ultra-high density HOX tiling array was interrogated with either normal breast organoid RNA (Cy3 channel) or RNA derived from primary or metastatic breast tumors (Cy5 channel). Each column represents the indicated clinical sample. Each row indicates a transcribed region, either a HOX coding exon or HOX ncRNA. Expression values are depicted as a ratio relative to pooled normal and represented as a red-green color scale. b, Higher resolution of subset iii identifying transcripts that show higher relative expression in metastatic as compared to primary tumors and normal epithelia. HOTAIR is one such transcript (p= 0.03, Student t-test). c, Quantitative reverse transcription (qRT)-PCR validation of the expression tiling array results measuring HOTAIR abundance in a panel of normal breast epithelial-enriched organoids, primary breast tumors, and metastatic breast tumors. Metastatic tumors had at a minimum 125-fold higher level of HOTAIR than normal breast epithelia. Error bars represent s.d. (n=3). d, qRT-PCR analysis of HOTAIR in 132 primary breast tumors (stage I or II). Approximately one third of primary breast tumors had >125 fold overexpression of HOTAIR over normal (HOTAIR high, indicated in red), while roughly two third of tumors did not (HOTAIR low, indicated in blue). Error bars represent s.d. (n=3). e and f, Kaplan-Meier curves for (e) metastasis free survival or (f) overall survival of the same 132 primary breast tumors measured in panel (d).

Figure 2. HOTAIR promotes invasion of breast carcinoma cells

a, Relative fold increase in matrix invasion in three breast carcinoma cell lines after enforced expression of HOTAIR. Mean±s.d. are shown (n=3). b. Matrix invasion in the MCF-7 breast carcinoma cell line transfected with individual or pooled siRNAs targeting HOTAIR (error bars = s.d., n=3). c, HOTAIR expression in MDA-MB-231 cells enhances orthotopic growth in mammary fat pads and metastasis to lung (n=5 per arm; error bars = s.e.m.). d, HOTAIR promotes transient lung colonization of SK-BR3 after tail vein xenograft (n=15 per arm). e, HOTAIR promotes lung colonization of MDA-MB-231 cells after tail vein xenograft. f, Histologic analysis of the number of lung metastasis in vector- or HOTAIR- expressing MDA-MB-231 cells 8–9 weeks after tail vein xenograft (n=18 per arm, error bars = s.e.m.). Similar results were obtained using luciferase marked cells. *, p<0.05 between control cells and cells manipulated for HOTAIR.

Figure 3. HOTAIR promotes selective re-targeting of PRC2 and H3K27me3 genome-wide

a, Heat map representing genes with a significant relative change in chromatin occupancy of EZH2, SUZ12, and H3K27 following HOTAIR expression. MDA-MB-231 vector or HOTAIR cells were subjected to ChIP using anti-EZH2, H3K27me3, and SUZ12 antibodies followed by interrogation on a genome-wide promoter array. Values are depicted as relative ratio of HOTAIR over vector cells and represented as an orange-blue scale. b, Average SUZ12 occupancy of >800 PRC-2 target genes in HOTAIR or vector expressing cells across the length of gene promoter and gene body. All target genes are aligned by their transcriptional start sites (TSS). c, Module map of the 854 genes with a gain in PRC2 occupancy following HOTAR overexpression. (Left panel) Heat map of genes (column) showing a gain in PRC2 occupancy following HOTAR expression in breast carcinoma cells [see panel (a)] compared with PRC-2 occupancy patterns from the indicated cell or tissue type (rows). Binary scale is brown (match) or white (no match). (Right panel) Quantification of significance of pattern matching between gene sets.

Figure 4. HOTAIR-induced matrix invasion and global gene expression changes requires PRC2

a, Immunoblot of SUZ12 and EZH2 protein levels following transduction of MDA-MB-231 vector or HOTAIR cells with retrovirus expressing a shRNA targeting either GFP, EZH2, or SUZ12. b, Matrix invasion in vector or HOTAIR cells expressing the indicated shRNA. Mean±s.d. are shown (n=3). c, (left panel) Heat map of gene with significant induction (red) or repression (green) following HOTAIR expression in the MDA-MB-231 cells. (right panel) The relative expression of the same gene list in MDA-MB-231 HOTAIR cells expressing shEZH2 or shSUZ12 (expressed as a ratio to HOTAIR cells expressing shGFP). d, qRT-PCR of a representative panel of genes in MDA-MB-231 vector or HOTAIR cells also expressing the indicated shRNA (error bars = s.d., n=3). e, Matrix invasion in the immortalized H16N2 breast epithelial line expressing vector or EZH2 as well as EZH2-expressing cells transfected with siRNAs targeting GFP or HOTAIR (error bars = s.d., n=3). f, Working model of the role of HOTAIR in breast cancer progression. Selection for increased HOTAIR expression in a subset of breast primary tumors leads to a genome-wide retargeting of the PRC2 and H3K27me3 patterns, resulting in gene expression changes that promote tumor metastasis.