CRISPRi screen uncovers lncRNA regulators of human monocyte growth

Abstract

Long noncoding RNAs are emerging as critical regulators of biological processes. While there are over 36,000 lncRNAs annotated in the human genome, we do not know the function of the majority. Here, we performed a high-throughput CRISPRi screen to identify those lncRNAs that are important for viability in human monocytes using the cell line THP1. We identified a total of 38 hits from the screen and validated and characterized two of the top intergenic hits. The first is a lncRNA neighboring the macrophage viability transcription factor IRF8 (RP11-542M13.2 hereafter referred to as long noncoding RNA regulator of monocyte proliferation, LNCRMP), and the second is a lncRNA called OLMALINC (oligodendrocyte maturation-associated long intervening non-coding RNA) that was previously found to be important in oligodendrocyte maturation. Transcriptional repression of LNCRMP and OLMALINC from monocytes severely limited their proliferation capabilities. RNA-seq analysis of knockdown lines showed that LNCRMP regulated proapoptotic pathways, while knockdown of OLMALINC impacted genes associated with the cell cycle. Data support both LNCRMP and OLMALINC functioning in cis to regulate their neighboring proteins that are also essential for THP1 cell growth. This research highlights the importance of high-throughput screening as a powerful tool for quickly discovering functional long non-coding RNAs (lncRNAs) that play a vital role in regulating monocyte viability.

Keywords: CRISPR; CRISPR inhibition; LNCRMP; Long noncoding RNAs; OLMALINC; cis regulation; high throughput pooled screen; monocytes.

Figure 1

CRISPRi screen identifies lncRNAs that impact macrophage cell growth.A, schematic of drop out CRISPRi screen pipeline. sgRNAs were designed to target the transcription start sites of over 2000 Gencode hg19 annotated lncRNAs. Transcription start sites were predicted using data from FANTOM and ENCODE. The THP1 lncRNA expression was estimated from THP1 RNAseq data. NFkB-EGFP-CRISPRi-THP1 cells were infected with pooled sgRNA libraries. The cells were selected with puromycin, and samples were collected at day 21 to identify the sgrnas that dropped out from the original library. sgRNAs from the library and day 21 were PCR amplified and sequenced. B, screen hits. Mann Whitney U analysis was used to determine the gene ranking from the RNA seq data. C, top-performing sgRNAs. Top three guides for each significant hit are represented and color coded to match their gene category in (D). D, ncRNA hit categories. Pie chart for the different categories of lncrnas in the CRISPRi screen. Intergenic lncrnas are lncrnas localized between two genes and they have their own transcription start site (TSS). Antisense lncrnas are lncrnas that transcribed from the same gene locus in the opposite direction. Intronic lncrnas are localized within an intron of a gene. Bidirectional lncRNAs are 1kb away from the gene locus and they are transcribed in the opposite direction. E, table summarizing top candidate lncRNAs, OLMALINC and LNCRMP and a positive control hit CASP8AP2.

Figure 2

Validation of OLMALINC and LNCRMP as hits involved in macrophage growth CRISPRi knockdown of OLMALINC and LNCRMP.A and B, sgRNAs (sgRNAs 1, 2, 3) were designed to target top lncRNA candidates OLMALINC and LNCRMP. qPCR measurement of both genes across three replicate experiments shows knockdown of OLMALINC (A) and LNCRMP (B) by all three sgRNAs (∗p-values<0.05, ∗∗p-values<0.01 and ∗∗∗p-values<0.005) vs a non-targeting control sgRNA. C, schematic of macrophage growth validation assay. Mcherry positive cells sgRNA containing cells were mixed with mcherry negative cells in a 1:1 ratio and plated in triplicate for each sgrna for each gene. The cells were flowed every other day for 21 days. Mcherry enrichment was then calculated to determine the cell growth. D, mixed cell assay results for OLMALINC and LNCRMP. We combined cherry-positive cells (containing sgRNAs) with unedited cherry-negative cells at a 1:1 ratio and monitored the growth of sgRNA-infected cells (cherry) relative to uninfected reference cells in a mixed-cell growth assay for 21 days. The experiment was repeated 3 times, and a representative experiment was displayed.

Figure 3

Functional characterization of OLMALINC’s regulation of monocyte growth.A–C, CRISPRi knockdown of OLMALINC impacts SCD expression. A, UCSC genome browser track displays RNA-Seq reads from three control (Ctl) and 3 OLMALINC knockdown (KD) cell lines at the OLMALINC genomic locus. B–C, Normalized read counts from DEseq analysis for OLMALINC (B) and its neighboring protein SCD (C) error bars represent standard deviations of the three biological replicates, ∗∗p < 0.01. D, OLMALINC and SCD share a TAD. OLMALINC and SCD are part of the same topologically associated domain (TAD). Indicating a close genomic proximity and cis-regulation (data analyzed from GSE60857). E, differentially expressed genes following OLMALINC knockdown. Volcano plot displaying significantly upregulated and downregulated genes in OLMALINC KD cells. F–G, heatmaps of top 20 upregulated or downregulated genes in OMALINC KD cells and control cells. H, OLMALINC is expressed in the cytosol. RT-qPCR analysis of OLMALINC expression in cytoplasmic and nuclear fractions from THP1 cells. The experiment was performed in triplicate, and error bars represent standard deviations of the replicates. I, OLMALINC possibly encodes short peptides. Riboseq data for OLMALINC shows Riboseq aggregates at open reading frame 1 (ORF1) and conservation between different species. Track one shows initiating Riboseq aggregates in the different ORFs. Track two shows Riboseq footprint aggregates. Track three shows Riboseq coverage aggregate. Track four shows Riboseq coverage in THP1 cells. J, OLMALINC Peptide conservation. Consensus conserved sequence is observed in different species such as human, gorilla, and green monkey. K, proposed working model Briefly, OLMALINC acts as a cis-regulator of SCD (viability coding gene) and regulates its expression. Under basal conditions, OLMALINC is expressed at basal levels, and regular levels of SCD are expressed, which results in a normal cell cycle. In the absence of OLMALINC, SCD levels are lower, and cell cycle is altered impacting cell growth.

Figure 4

Characterization of LNCRMP’s role in monocytic growth and function.A, UCSC genome browser track displays RNA-Seq reads from three control (Ctl) non-targeting sgRNAs and 2 LNCRMP knockdown (KD) cell lines at the LNCRMP genomic locus. B–C, expression of LNCRMP (B) and its neighboring protein SCD (C) from RNA-seq data comparing 3 Ctl sgRNAS to 2 LNCRMP sgRNA. Error bars represent standard deviations of the biological replicates, ∗p < 0.05. D, LNCRMP and IRF8 are part of the same topological associated domain (TAD). Indicating a close genomic proximity and possible cis regulation (data analyzed from GSE60857. E, volcano plot displaying significantly upregulated and downregulated genes in LNCRMP KD cells. F–G, heatmaps of top upregulated LNCRMP KD cells and controls across two categories outlined. H, RT-qPCR analysis of LNCRMP expression in cytoplasmic and nuclear fractions from THP1 cells. The experiment was performed in triplicate and error bars represent standard deviations of the replicates. I, proposed working model of how LNCRMP acts as a regulator of THP1 viability. Briefly, LNCRMP acts in cis to regulate IRF8.