Long non-coding RNA KIKAT/LINC01061 as a novel epigenetic regulator that relocates KDM4A on chromatin and modulates viral reactivation

Abstract

KDM4A is a histone lysine demethylase that has been described as an oncogene in various types of cancer. The importance of KDM4A-mediated epigenetic regulation in tumorigenesis is just emerging. Here, by using Kaposi's sarcoma associated herpesvirus (KSHV) as a screening model, we identified 6 oncogenic virus-induced long non-coding RNAs (lncRNAs) with the potential to open chromatin. RNA immunoprecipitation revealed KSHV-induced KDM4A-associated transcript (KIKAT)/LINC01061 as a binding partner of KDM4A. Integrated ChIP-seq and RNA-seq analysis showed that the KIKAT/LINC01061 interaction may mediate relocalization of KDM4A from the transcription start site (TSS) of the AMOT promoter region and transactivation of AMOT, an angiostatin binding protein that regulates endothelial cell migration. Knockdown of AMOT diminished the migration ability of uninfected SLK and iSLK-BAC16 cells in response to KIKAT/LINC01061 overexpression. Thus, we conclude that KIKAT/LINC01061 triggered shifting of KDM4A as a potential epigenetic mechanism regulating gene transactivation. Dysregulation of KIKAT/LINC01061 expression may represent a novel pathological mechanism contributing to KDM4A oncogenicity.

Fig 1. Transcriptome analysis and siRNA knockdown screening of KSHV reactivation-induced lncRNAs essential for KSHV latent-to-lytic switch.

(A) Workflow showing the identification of KSHV reactivation-induced long intergenic non-coding RNAs (LincRNAs). RNA-seq was performed using total RNA isolated from iSLK-BAC16 cells treated with and without Dox for 24 hours. Paired-end raw reads were aligned to human reference genome hg19 using CLC Genomics Workbench (Qiagen Germantown, MD, USA) and annotated to Refseq using Partek Genomics Suite (Partek, MO, USA). Reads per kilobase of transcript per million mapped read (RPKM) higher than 0.05 in at least one condition was considered as expressed and used for analysis. Pie chart (Top) showing the numbers (percentages) of long non-coding RNAs (LncRNAs) that were up- or down-regulated more than 1.5-fold after Dox induction for viral reactivation. Pie chart (Bottom) showing the numbers of different types of up-regulated lncRNAs. (B) Histogram demonstrating the expression (RPKM, log10 scale) of 82 KSHV reactivation up-regulated lincRNAs present in the siRNA Library (Dharmacon, GU-301000). (C) Schematic representing the siRNA screening procedure of lincRNAs that mediate KSHV reactivation. The recombinant KSHV, rKSHV.219, produced from Vero-rKSHV.219 cells were used to infect iSLK cells. After 24 hours, iSLK-rKSHV.219 cells were re-seeded in 96-well-plates at a density of 1000 cells/well. Cells were attached overnight and transfected with siRNAs. Eighteen hours after transfection, cells were treated with 2 μg/ml Dox for another 72 hours, fixed by 4% paraformaldehyde, stained with Hochest 33258. The GFP and RFP positive cells were observed with an inverted fluorescence microscope. Representative images (100x magnification) of iSLK-rKSHV.219 cells transfected with siLINC01061 was shown. Cells without siRNA transfection was used as positive control. Scale bar: 300 μm. (D) RT-qPCR analysis of the expression of 14 lincRNAs identified in Fig 1C in iSLK-BAC16 cells treated with 1 μg/ml Dox for 24 hours. The expression of lincRNAs was normalized to GAPDH and the relative expression of LINC00313 in iSLK-BAC16 cells (Ctrl) was considered to be 1. (E) rKSHV.219 was used to infect HMEC1 cells, followed by transient puromycin selection for a week. After selection, HMEC1-rKSHV.219 cells were re-seeded in 6-well-plates. Cells were attached overnight and transduced with lentiviruses harboring pLenti4-Flag-K-Rta or empty vector control (VC). 72 hours after transduction, total RNA was collected, followed by RT-qPCR analysis of the expression of 6 lincRNAs that were up-regulated in iSLK-BAC16 after viral reactivation. The expression of lincRNAs was normalized to GAPDH and reported as the relative expression of vector control (VC = 1). (F) RT-qPCR analysis of the expression of 6 lincRNAs identified in Fig 1D in TREx-F3H3-K-Rta BCBL-1 cells treated with 0.2 μg/ml Dox for 72 hours. The expression of lincRNAs was normalized to GAPDH and plotted as the relative expression versus non-induced cell Ctrl (Ctrl = 1).

Fig 2. KSHV-Induced KDM4A-Associated Transcript (KIKAT)/LINC01061 is associated with histone lysine demethylase 4A (KDM4A).

(A) Total cell lysates (TCLs) prepared from SLK, HMEC1 and HUVEC were examined for KDM4A by immunoblotting. Anti-α-Tubulin antibody was included as loading control. (B) SLK cell lysate was used for RNA immunoprecipitation (RIP) using KDM4A antibody or rabbit IgG as control. Immunoprecipitation (IP) specificity was confirmed by immunoblotting. (C) Copurified RNA from KDM4A and IgG IPs in (B) was assessed for LncRNAs enrichment by RT-qPCR. (D) Schematic diagram represented the structure of KIKAT/LINC01061 and the nearby genes Phosphodiesterase 5A (PDE5A) and Fatty acid-binding protein 2 (FABP2). Exons and introns are showed as blue boxes and thin lines, respectively. Arrows labeled transcription start site (TSS) and transcription direction. The numbers below the boxes and above the lines indicate the length of transcripts and of intergenic regions, respectively. (E) Cell lysate from TREx-F3H3-K-Rta BCBL-1 cells was used for RIP using KDM4A antibody or rabbit IgG as control. IP specificity was confirmed by immunoblotting. (F) Copurified RNA from KDM4A and IgG IPs in (E) was assessed for LncRNAs enrichment by RT-qPCR. Significance was determined by student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig 3. RNA-Seq analysis revealed that KIKAT/LINC01061 may function as an onco-lncRNA.

(A) RT-qPCR analysis of KIKAT/LINC01061 expression in SLK cells transiently transduced with lentivirus expressing KIKAT/LINC01061 or with no cDNA (mock). (B) RT-qPCR analysis of KIKAT/LINC01061 expression in a stable SLK cell line overexpressing KIKAT/LINC01061, SLK-KIKAT/LINC01061, and its vector control cells, SLK-VC. (C) RNA-seq was performed using total RNA isolated from cells in (A) and (B). Paired-end raw reads were aligned and annotated as described in Fig 1A. RPKM ≥ 0.05 in at least one condition was considered as expressed and used for further analysis. Pie chart showing the numbers (percentages) of mRNAs that were up- or down-regulated more than 2-fold in transient transduced cells (Left) and stable cell lines (Right). (D) Expression heatmap of top 10 up-regulated genes of RNA-seq data from both transient (left) and stable (right) KIKAT/LINC01061 expression SLK cells. (E) RT-qPCR analysis of the expression of top 10 up-regulated genes identified in (D) in SLK-KIKAT/LINC01061 and SLK-VC cells. (F) KIKAT/LINC01061 levels in skin cancer and normal tissues were obtained from the human cancer metastasis database (HCMDB). The statistical significance was calculated using student’s t-test. (G and H) GSEA analysis reveal the correlation of KIKAT/LINC01061 down-regulated genes (≥ 5-fold) with skin cancer (GSE7553) (G) and Kaposi’s sarcoma (KS) lesion (GSE147704) (H). Significance was determined by student’s t-test. *p < 0.05, **p < 0.01, n.s. non-significant.

Fig 4. Genome-wide ChIP-seq analysis revealed the regulatory role of KIKAT/LINC01061 in KDM4A occupancy in SLK cells.

(A) Workflow showing the identification of KDM4A ChIP-seq peaks. ChIP-seq was performed using chromatin DNA immunoprecipitated (IPed) from SLK-VC, SLK-KIKAT/LINC01061 and SLK-KDM4A KO-VC cells by KDM4A antibody. Paired-end raw reads were trimmed and aligned to human reference genome hg19 using CLC Genomics Workbench (Qiagen Germantown, MD, USA). KDM4A ChIP-seq peaks were called using Partek Genomics Suite (Partek, MO, USA) with KDM4A KO ChIP DNA as control. Peaks with FDR ≤ 0.05 were identified as significant KDM4A binding regions. KDM4A peaks were annotated to promoter regions (TSS ± 2 kb) based on RefSeq. Venn diagrams represent the KDM4A peaks (Middle) and potential target genes (bottom) coincide in SLK-VC and SLK-KIKAT/LINC01061. (B) Distance distribution of KDM4A peaks. (C) Venn diagram depicts the overlap of genes among RNA-seq (Fig 3C, right panel) and ChIP-seq (Fig 4A, bottom panel). Pie chart showing the numbers (percentages) of mRNAs that were up- or down-regulated more than 2-fold in SLK-KIKAT/LINC01061 when compared with SLK-VC.

Fig 5

Identification of AMOT as a potential KDM4A-KIKAT/LINC01061 target (A) Histograms of ChIP-seq profiles for KDM4A binding at AMOT loci in SLK-VC and SLK-KIKAT/LINC01061 cells. Red arrow in the enlarged view shows the direction of KDM4A relocation. (B) ChIP-qPCR assay revealed the binding of KDM4A (upper panel) and the modification of H3K4me3 (middle panel) and H3K9me3 (lower panel) to the promoter of AMOT in SLK-VC and SLK-KIKAT/LINC01061 cells. (C and D) RT-qPCR analysis verification of AMOT expression in transient KIKAT/LINC01061 transduced SLK cells (C) and in SLK-KIKAT/LINC01061 cell lines (D). (E) RT-qPCR analysis revealed the expression of KIKAT/LINC01061 and AMOT in SLK cells after transfected with siRNA targeting KIKAT/LINC01061 and with control siGLO. (F) ChIP-qPCR assay revealed the binding of KDM4A to the promoter of AMOT in SLK cells transient transfected with siKIKAT/LINC01061. (G) RT-qPCR analysis revealed the expression of AMOT in SLK and SLK-KDM4A KO cells. (H) ChIP-qPCR assay revealed the modification of H3K9me3 on the AMOT promoter in SLK-KDM4A KO cells. Significance was determined by student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, n.s. non-significant.

Fig 6. KIKAT/LINC01061 enhances SLK migration in an AMOT-dependent manner.

(A) Representative images (100x magnification) of in vitro transwell migration analysis of SLK-KIKAT/LINC01061 and SLK-VC cells at 6 hours (left). Scale bar: 100 μm. Quantification of cell migration (right). (B) Representative images (100x magnification) of in vitro wound healing migration assay of SLK-KIKAT/LINC01061 and SLK-VC cells at 8 hours (left panel). Scale bar: 300 μm. Quantification of wound healing (right panel). (C and D) Representative images (100x magnification) of migration analysis of HMEC1 (C) and HUVEC (D) cells transfected with siKIKAT/LINC01061 or with control siGLO. (upper panel). Scale bar: 100 μm. Quantification of cell migration (bottom panel). (E) Representative images (100x magnification) of migration analysis of KIKAT/LINC01061-transduced SLK cells with or without transduction of lentivirus expressing shAMOT. Scale bar: 100 μm. (F) Quantification of cell migration in (E). (G) Representative images (100x magnification) of migration analysis of KIKAT/LINC01061-transduced iSLK-BAC16 cells with or without transduction of lentivirus expressing shAMOT. Scale bar: 100 μm. (H) Quantification of cell migration in (G). Scale bar: 100 μm. Significance was determined by student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig 7. KIKAT/LINC01061 is important for KSHV reactivation.

(A and B) Immunoblotting (A) and RT-qPCR (B) analysis of the expression of K-bZIP, Orf45, Orf57, K8.1 in iSLK-BAC16 cells treated with 1 μg/ml Dox for 0, 24, 48 and 72 hours. (C) KSHV virion-associated DNA was determined by TaqMan qPCR. (D) Expression of KIKAT/LINC01061 was determined by RT-qPCR. (E) Immunoblotting of K-Rta in iSLK cells treated with or without Dox for 48 hours. (F) RT-qPCR analysis of KIKAT/LINC01061 in iSLK cells treated as described in (E). (G) The luciferase reporter plasmid containing KIKAT/LINC01061 promoter (TSS ± 500 bp) was co-transfected with pcDNA3-HA-K-Rta or with empty vector into 293T cells. After 48 hours, cells were collected for immunoblotting with anti-HA and anti-α-Tubulin antibodies. (H) Luciferase reporter assays were performed in 293T cells treated as described in (G). The activity of firefly luciferase was normalized to that of Renilla luciferase in the same assayed sample and reported as the relative activity of vector control. (I) iSLK-BAC16 cells were transfected with siKIKAT/LINC01061 or with control siGLO. After 6 hours, cells were re-seeded in 6-well plates and treated with 1 μg/ml Dox for another 48 hours. Relative expression of KIKAT/LINC01061 (upper panel) and KSHV virion-associated DNA (lower panel) was determined by RT-qPCR and TaqMan qPCR, respectively. (J) iSLK-BAC16 cells were transduced with lentivirus expressing full-length KIKAT/LINC01061 or with no cDNA (Mock). After 16 hours, cells were re-seeded in 6-well plates and treated with 1 μg/ml Dox for another 72 hours. Relative expression of KIKAT/LINC01061 (upper panel) and KSHV virion-associated DNA (lower panel) was determined as described in (I). (K) iSLK-BAC16 cells were transduced with lentivirus expressing full-length KIKAT/LINC01061 or with no cDNA (Mock). After 72 hours, relative expression of KIKAT/LINC01061 (upper panel) and KSHV virion-associated DNA (lower panel) was determined as described in (I).

Fig 8. KIKAT/LINC01061 and KDM4A association in iSLK-BAC16 cells.

(A) TCLs from control and Dox induced iSLK-BAC16 cells were used for RIP using KDM4A antibody or rabbit IgG as control. Copurified RNA from KDM4A and IgG IPs was assessed for KIKAT/LINC01061 enrichment by RT-qPCR. (B) Cells as described in Fig 7I were subjected to ChIP assay and the binding of KDM4A to the promoter of AMOT were analyzed by qPCR. (C) The enrichment of KDM4A on the promoter of several viral genes in cells described in Fig 7I were analyzed by qPCR. (D) iSLK-BAC16 cells were transfected with control siGLO or siKIKAT/LINC01061. After 6 hours, cells were re-seeded in 6-well plate and treated with 1 μg/ml Dox for another 48 hours. in vitro transwell migration analysis was determined. Representative images (100x magnification) of transwell migration analysis (left panel). Scale bar: 100 μm. Quantification of cell migration (right panel). Significance was determined by student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001.