Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Jul 15;199(2):547-558.
doi: 10.4049/jimmunol.1700232. Epub 2017 Jun 9.

Profiles of Long Noncoding RNAs in Human Naive and Memory T Cells

Affiliations
Free PMC article

Profiles of Long Noncoding RNAs in Human Naive and Memory T Cells

Charles F Spurlock 3rd et al. J Immunol. .
Free PMC article

Abstract

We employed whole-genome RNA-sequencing to profile mRNAs and both annotated and novel long noncoding RNAs (lncRNAs) in human naive, central memory, and effector memory CD4+ T cells. Loci transcribing both lineage-specific annotated and novel lncRNA are adjacent to lineage-specific protein-coding genes in the genome. Lineage-specific novel lncRNA loci are transcribed from lineage-specific typical- and supertranscriptional enhancers and are not multiexonic, thus are more similar to enhancer RNAs. Novel enhancer-associated lncRNAs transcribed from the IFNG locus bind the transcription factor NF-κB and enhance binding of NF-κB to the IFNG genomic locus. Depletion of the annotated lncRNA, IFNG-AS1, or one IFNG enhancer-associated lncRNA abrogates IFNG expression by memory T cells, indicating these lncRNAs have biologic function.

Conflict of interest statement

Disclosures

The authors declare no conflict of interest.

Figures

FIGURE 1
FIGURE 1
Cell-type specific expression of mRNAs and lncRNAs. (A) Left panel, total numbers of mRNAs (GRCh38 release, total protein-coding genes=19950), annotated lncRNAs (GRCh38 release, long non-coding genes=15767) and novel lncRNAs detected here by de novo assembly with average FPKM > 1 in at least one cell lineage; middle panel, average expression levels of the different RNA classes in TN, TCM and TEM cells; right panel, total expression levels of the indicated RNA classes in TN, TCM and TEM cells (N=3). (B) Volcano plots identifying lineage specific RNA classes after FDR correction, P < 0.05. X-axis represents the indicated expression ratios, log2, and the Y-axis is P value, -log10. (C) Total numbers of mRNAs, annotated lncRNAs and novel lncRNAs that show lineage specific expression (from (B). (D) Hierarchical clustering showing expression levels of mRNAs expressed by human CD4+ T cells determined by whole genome RNA-seq aligned to hg38 (N=3). A scale bar is adjacent to the heat map. MSigDB analysis of lineage-specific mRNAs (identified in the heatmap) as input, outputs from the MSigDB analysis are shown with corresponding P-values. (E) Individual variation in lncRNA expression. Expression levels in FPKM of four annotated lncRNAs are shown in TN, TCM and TEM subsets isolated from three individual subjects.
FIGURE 2
FIGURE 2
Relationships between genomic positions of co-expressed lineage specific protein-coding genes and annotated lncRNA genes. (A) Results are expressed as the % of lineage-specific mRNA genes (○, TN; □, TCM; Δ, TEM) co-expressed with lineage-specific annotated lncRNA genes relative to total mRNA genes within the indicated distance in kb from a lineage-specific annotated lncRNA gene. (B) As in (A) except results are expressed relative to the number of genes in the genome away from the lineage-specific annotated lncRNA gene. (C) Percentage of co-expressed lineage-specific mRNA genes transcribed in antisense or sense orientations relative to adjacent co-expressed lineage-specific annotated lncRNA genes. (D) Genes encoding lineage-specific mRNAs are enriched in the genome near lineage-specific annotated lncRNA genes. The y axis is the percent of total lineage-specific TN, TCM and TEM mRNAs (Fig 1, heatmaps). The x axis is the distance of the gene encoding a lineage-specific mRNA from the nearest gene encoding a lineage-specific annotated lncRNA. P values determined by c2 analyses, TN; P = 10−22, TCM; P=10−12, TEM; P=10−14.
FIGURE 3
FIGURE 3
Discovery of lineage-specific novel lncRNAs. (A) Hierarchical clustering showing expression levels of novel lncRNAs expressed by human CD4+ TN, TCM and TEM cells determined by whole genome RNA-seq aligned to hg19 (N=3). (B) Co-localization of genes encoding lineage-specific mRNAs and novel lncRNAs in the genome. Left panel: % of lineage-specific protein-coding genes that are within the indicated distance in kb from another lineage-specific protein-coding gene, right panel: % of lineage-specific protein-coding genes that are within the indicated distance in kb from a co-expressed lineage-specific novel lncRNA-coding gene. (C) Genes encoding lineage-specific mRNAs are enriched in the genome near lineage-specific novel lncRNA loci. The y axis is the percent of total lineage-specific TN, TCM and TEM mRNAs (Fig 1, heatmaps). The x axis is the distance of the gene encoding a lineage-specific mRNA from the nearest loci encoding a lineage-specific novel lncRNA, P values determined by c2 analyses, TN, P = 10−21; TCM, P=10−14; and TEM P=10−19.
FIGURE 4
FIGURE 4
Relationships among lineage-specific lncRNAs and transcriptional enhancers. (A) Circosplot depicting genome-wide positions of CD4 naïve typical enhancers and super enhancers (purple) along with CD4 memory typical enhancers (orange) and super enhancers (blue) in the outer rings (from ref. 20). Bar graphs colored in black in the inner concentric circles show lineage specific novel lncRNAs in naïve (outermost), central memory, and effector memory (innermost) cells (B) Lineage specific mRNAs ranked by number of lineage-specific novel lncRNAs per gene loci. Ranks of certain gene loci are noted in the graphs. (C) Lineage-specific SE are ranked by number of novel lncRNAs per genomic locus. (D) Lineage-specific TEs are ranked by number of novel lncRNAs per genomic locus.
FIGURE 5
FIGURE 5
Genomic distributions of novel lncRNAs. (A) Distances between genomic loci transcribing unique novel lncRNAs. The X-axis ranks novel lncRNA loci according to distance to the next novel lncRNA loci and the Y-axis shows the actual distance in bp between neighboring lncRNA loci. (B) Characteristics of genomic ‘bins’ transcribing multiple novel lncRNAs. (C, D, & E) Examples of genomic bins selectively transcribing multiple novel lncRNAs in (C) TN relative to TM, D) TCM relative to TN and TEM, and (E) TEM relative to TN and TCM. X-axes show genomic positions from which novel lncRNAs are transcribed; lines illustrate sizes of the transcribed novel lncRNA. Y-axes are mean expression ratios, log2, between the indicated T cell lineages. (F) Genomic bins transcribing novel lncRNAs co-localize with TN and TM typical and super transcriptional enhancers (from ref. 20). We considered that the genomic bins may not exactly overlap with enhancers so we performed the analysis to include the proportion of bins that co-localize with an enhancer (0), and the proportion of bins that co-localized with enhancers if we extended bin size by 5, 10, 15, 20 or 30 kb in both 5’ and 3’ directions, X-axis. The Y-axis is the proportion of total bins with a TN or TM enhancer. P values, -log10, were determined by c2 analysis. (G) Number of bins within TM SE or TE (CD4MSE, CD4MTE) structures or TN SE or TE structures. X-axis is as in (F), Y axis is the number of bins with enhancers. (H&I) Re-analysis of RNA-seq data using ‘bins’ as definition lists. (H) After FDR correction, bins were identified that were over-expressed (blue lines) or under-expressed (brown lines) in TCM and TEM cells compared to TN cells. (I) Expression levels of bins differentially expressed in TCM relative to TN and TEM after FDR correction. Y-axes are expression ratios, log2, X-axes identify ratios; Tn=TN/TN, Tcm=TCM/TN, Tem=TEM/TN.
FIGURE 6
FIGURE 6
Novel lncRNAs at the IFNG locus. (A) RNA-seq tracks spanning ~300 kb from IFNG-AS1 past IL26 in primary and effector Th1, Th2 and Th17 cells, naïve CD4+ T cells and TCM and TEM cells. Below are positions of genomic loci encoding novel lncRNAs determined by de novo assembly, genomic positions of typical and super enhancers (TE & SE, respectively) identified in total CD4+ TM cells are also shown (from ref. 20). (B) PCR validation of RNA-seq results normalized to GAPDH. Each novel lncRNA is named relative to the IFNG transcriptional start site, - indicates toward the p-terminus, + toward the q-terminus, differences between either TCM and TN or TEM and TN were statistically significant, N=4, P < 0.05. (C) JQ1 treatment of memory T cells abrogates RNA polymerase II binding but not H3K27Ac across the IFNG locus. Indicated concentrations of JQ1 were added to memory T cells cultures. Cells were stimulated with plate-bound anti-CD3 for 24 hours and levels of H3K27-Ac (left panel) and RNA polymerase II (right panel) were determined by ChIP. For this analysis, we designed a tiling array. 100 PCR primer pairs were designed at ~3 kb intervals to span the 300-kb IFNG locus. Data are expressed as mean fraction of input (n=3) P < 0.05, ANOVA. (D) Total memory CD4+ T cells were cultured for 24 hr with JQ1 or I-BET to displace BRD-containing proteins from genomic loci. RNA was purified and levels of the indicated IFNG locus novel lncRNAs, IFNG-R-#, determined by PCR. Results are expressed relative to levels of GAPDH. P < 0.05, treated versus untreated (N=4). (E) JQ1 or I-BET inhibit IFNG-AS1 and IFNG expression. Total memory CD4+ T cells were treated with and without JQ1 or I-BET for 24 hr and stimulated with anti-CD3 for an additional 24 hr. Levels of three unique isoforms of IFNG-AS1 designated 1, 2, 3 (Fig. 6A), and IFNG were determined by PCR and normalized to levels of GAPDH, *=P < 0.05 (N=4). (F) Individual siRNAs targeting IFNG-AS1 or IFNG-R-49 were transfected into total memory CD4+ T cells prior to stimulation with anti-CD3 for 24 hours. Transfected cells were isolated by fluorescence activated cell sorting prior to mRNA analysis or to measure intracellular levels of IFN-g. P values were determined by unpaired T-test. Error bars are mean +/− S.D. * = p < 0.05.
FIGURE 7
FIGURE 7
IFNG-associated lncRNAs bind NF-κB. (A&B) Cells were cultured with or without JQ1 (150 nM, 4 hr) and lysed. After lysis, (A) anti-T-bet and (B) anti-NF-κB, p65 subunit, immunoprecipitations (IP) were performed. Levels of each IFNG-R-lncRNA in the anti-T-bet, anti-NF-κB, p65 subunit, and isotype control immunoprecipitates were determined by PCR. Results are expressed as fraction of input of the indicated IFNG-associated novel lncRNAs relative to totals of each IFNG-R-lncRNA. Levels of IFNG-R-associated lncRNAs in the isotype control immunoprecipitates were negligible and therefore not included in the calculations. Error bars are S.D of 8 independent experiments, *=P < 0.05. (C) As in (B) except ChIP assays were performed to measure NF-κB, p65, binding to IFNG genomic locus sites. Results are expressed as fraction of input relative to an isotype control. Error bars are S.D of 5 independent experiments, *=P < 0.05. (D) Chromatin was isolated from TM cells by lysis in non-ionic detergents and treated with RNase, 10 min, RT. Chromatin was cross linked with formaldehyde and processed for ChIP assays. NF-κB, p65 binding is expressed as fraction of input, N=3, *=P < 0.05 comparing treated and untreated samples. (E) After RNAse treatment, chromatin was pelleted by centrifugation and supernatant fluids harvested and analyzed for NF-κB protein by western blotting. The arrow indicates the p65 subunit of NF-κB. (F) Cells were transfected with the siRNA targeting IFNG-R-49 (+) or a scrambled control siRNA (−) and cultured for 24 hr. Cultures were fixed with paraformaldehyde, harvested and processed for ChIP assays using an NF-κB, p65 subunit, antibody. The Y-axis shows NF-κB, p65 subunit, binding as fraction of input at the indicated IFNG genomic loci, X-axis.
FIGURE 8
FIGURE 8
Novel lncRNAs at the BACH2 locus. (A) RNA sequencing tracks of the BACH2 gene in TN, TCM and TEM cells. Positions of known TEs and SEs in total CD4+ T cells are shown below the sequencing tracks (from ref. 20). Exon-intron structure of BACH2 is the bottom track. (B) Expression levels of the 13 BACH2-associated novel lncRNAs discovered by de novo assembly in TN, and TEM and TCM expressed as average FPKM, N=3, *=P < 0.05. (C) PCR verification of results in (B) expressed relative to GAPDH in TN and TM cells, N=4, *=P < 0.05. (D) TN cells were cultured for 24 hr with JQ1 or I-BET to displace BRD-containing proteins from genomic loci. RNA was purified and levels of 13 BACH2 associated novel lncRNAs were determined by PCR. Results are expressed relative to levels of GAPDH. N=4, *=P < 0.05. (E) As in (D), except levels of BACH2 mRNA were determined by PCR and are expressed relative to GAPDH. P values were determined using an unpaired T-test. Error bars are mean +/− S.D. * = p < 0.05

Similar articles

See all similar articles

Cited by 14 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback