Emerging mechanisms of long noncoding RNA function during normal and malignant hematopoiesis

Abstract

Long noncoding RNAs (lncRNAs) are increasingly recognized as vital components of gene programs controlling cell differentiation and function. Central to their functions is an ability to act as scaffolds or as decoys that recruit or sequester effector proteins from their DNA, RNA, or protein targets. lncRNA-modulated effectors include regulators of transcription, chromatin organization, RNA processing, and translation, such that lncRNAs can influence gene expression at multiple levels. Here we review the current understanding of how lncRNAs help coordinate gene expression to modulate cell fate in the hematopoietic system. We focus on a growing number of mechanistic studies to synthesize emerging principles of lncRNA function, emphasizing how they facilitate diversification of gene programming during development. We also survey how disrupted lncRNA function can contribute to malignant transformation, highlighting opportunities for therapeutic intervention in specific myeloid and lymphoid cancers. Finally, we discuss challenges and prospects for further elucidation of lncRNA mechanisms.

Figure 1.

Unique features enable lncRNA-mediated control of cell fate and function. (A) lncRNAs can bind nucleic acids via base-pairing or proteins through secondary structures. The same lncRNA can harbor multiple interaction modules, enabling it to spatially organize diverse effectors and their targets and to scaffold ribonucleoprotein complexes. (B) lncRNAs are highly tissue-, developmental-, and physiological state-specific. Shown are 2 loci encoding lncRNAs (purple and blue boxes); gray boxes encode other transcripts. During early development (left), 1 lncRNA (purple transcripts) can form distinct nuclear compartments by recruiting a regulatory protein (yellow circles) to genomic sites surrounding its own locus. That lncRNA may be turned off at a later developmental stage whereas a second (blue transcript) is induced and recruits the same yellow protein but to different chromosome loci. (C) lncRNA–protein complexes are dynamic and flexible. The same lncRNA can bind distinct proteins based on their local concentration or by adopting distinct structural conformations with distinct RNA–protein binding affinities.

Figure 2.

Many lncRNAs scaffold proteins to their targets. (A) Xist directs X chromosome inactivation across eutherian females. RNA fluorescence in situ hybridization (FISH) image shows focal retention of Xist transcripts (red; blue fluorescence demarks the nucleus). Transcript models from mouse and human are shown at bottom. The gray bar at the right marks the A-repeat region, with the 8 tandem A-repeats depicted above. Base pairing among the repeats is stochastic, and one of the specific conformations they adopt in vivo is shown at top. Xist remains cis-tethered to DNA (blue line) by binding the chromatin attachment factor HNRNPU. Binding to SPEN via the A-repeat domain followed by recruitment of histone deacetylase 3 (HDAC3), which catalyzes H3K4 demethylation leading to transcription machinery exclusion, enables transcription repression. Xist also binds HNRNPK, which facilitates scaffolding of epigenetic repressors such as PRC2, which deposits H3K27me3, leading to a transcription-repressive chromatin state. As Xist spreads in cis, it binds the lamin B receptor (LBR) to reposition its chromosome to the nuclear lamina, forming a compact repressive compartment. (B) lincRNA-EPS represses apoptotic/inflammatory response genes. RNA FISH image shows nuclear diffusion of lincRNA-EPS transcripts (red). Transcript models from mouse and human are shown at bottom. The gray bar marks a 3′ region with tandem CANACA repeats, depicted at top, which is not conserved in humans. Its predicted secondary structure, comprising the minimum free energy conformation identified by RNAfold, is shown at top. Under steady-state conditions, lincRNA-EPS binds HNRNPL via the CANACA-repeat domain and localizes to promoters of target genes, conferring a repressed chromatin state by promoting nucleosome occupancy upstream of their transcription start site. This basal transcription repression is lifted during infection/inflammatory responses. (C) LUNAR1 sustains oncogenic IGF1R activation in cis. Aberrantly activated NOTCH1 engages an enhancer element intronic to the IGF1R gene (orange) which activates LUNAR1, which in turn cooccupies this element and favors recruitment/retention of the transcription machinery, including Mediator, leading to enhanced activation of the IGF1R promoter and accumulation of additional IGF1R transcripts. (D) Bloodlinc potentiates the terminal erythropoiesis gene program. RNA FISH image shows nuclear diffusion of Bloodlinc molecules. Transcript models from mouse and human are shown at bottom. Bloodlinc diffuses from a super-enhancer domain to trans-loci encoding terminal erythropoiesis modulators, and binds HNRNPU as well as transcription coactivators/corepressors, mediating activation or repression of its targets.

Figure 3.

Many lncRNAs sequester proteins from their targets. (A) NRON sequesters phosphorylated NFAT in the cytoplasm. NRON scaffolds NFAT within a complex that includes the calmodulin-binding protein IQGAP1, the nuclear importer KPNB1, and the inhibitory kinase LRRK2, keeping NFAT in a phosphorylated inactive state. T-cell activation leads to increased Ca2⁺ levels, causing disassembly of this complex and NFAT dephosphorylation, which in turn favors NFAT nuclear import and subsequent activation of cytokine expression. (B) lnc-DC promotes the terminal dendritic maturation gene program. lnc-DC binds the C terminus of STAT3 and sequesters it away from the phosphatase SHP1, preventing its dephosphorylation and favoring nuclear import of phosphorylated STAT3 and subsequent activation of the terminal DC differentiation gene program. (C) PACER promotes COX2 activation in cis. PACER is induced as part of the inflammatory response in response to lipopolysaccharide (LPS) stimulation. PACER binds to and competes with the repressive NF-κB1 homodimer away from the adjacent COX2 promoter, favoring binding of the activating RELA/NF-κB1 heterodimer to the promoter and subsequent COX2 activation. (D) Lethe tunes inflammatory responses in trans. Lethe, a chromatin-retained lncRNA, is transcribed from the Rps15a-ps4 pseudogene and acts as an RNA decoy that titrates activating RELA complexes away from target gene promoters, restraining cytokine activation downstream of inflammatory triggers.