MicroRNAs and long non-coding RNAs as novel regulators of ribosome biogenesis

Ribosome biogenesis is the essential, complex, and tightly regulated process in which large and small ribosomal subunits are assembled. This process accounts for the majority of energy usage in a eukaryotic cell (reviewed in [1,2]). Throughout this process, the co-ordinated effort of all three RNA polymerases (RNAPI, II, III), ∼80 ribosomal proteins (RPs), and over 200 assembly factors (AFs) is required (Figure 1). Ribosome biogenesis starts in the nucleolus when RNAPI engages tandemly repeated ribosomal DNA (rDNA) arrays to transcribe the polycistronic precursor ribosomal RNA (rRNA). In humans, this 47S pre-rRNA is processed to generate the mature 18S, 5.8S, and 28S rRNAs, which are modified by C/D and H/ACA box small nucleolar ribonucleoproteins (snoRNPs) to undergo 2′-O-methylation and pseudouridylation, respectively. The mature rRNAs are assembled with the RNAPIII-transcribed 5S rRNA and RPs to form the 40S (18S and small RPs) and 60S ribosomal subunits (28S, 5.8S, 5S, and large RPs) that translate messenger RNAs (mRNAs) into polypeptides in the cytoplasm. More in-depth regulation of ribosome biogenesis continues to be discovered, including regulation by micro and long non-coding RNAs (Figure 1; Tables 1 and 2).

MicroRNAs (miRNAs or miRs) are a class of short (∼21 nucleotide, nt) non-coding RNAs (ncRNAs) that post-transcriptionally regulate mRNA stability and translatability. Canonical miRNA biogenesis is well-studied; typically, a primary miRNA (pri-miRNA) transcript is synthesized by RNAPII from a miRNA gene and then undergoes two processing steps before being loaded into the Argonaute protein (AGO) to form an active RNA-induced silencing complex (RISC) (Figure 2A). The active silencing complex binds mRNA transcripts at a site usually within the 3′UTR of the targets. Based on the degree of miRNA:mRNA complementarity, RISC can induce transcript destabilization or translational repression (low complementarity) or transcript cleavage (high complementarity) (Figure 2B), thereby down-regulating target expression post-transcriptionally [3,4].

While most miRNAs originate from independently transcribed miRNA genes, many noncanonical miRNA sources have been discovered. Mature miRNAs can be processed out of endogenously transcribed intron lariats (mirtrons), short hairpin RNAs (shRNAs), transfer RNAs (tRNAs), and small nuclear and nucleolar RNAs (sn- and snoRNAs) [3]. Notably, miRNA-like molecules interchangeably called small rDNA-derived RNAs or rRNA-hosted miRNA analogs (srRNAs or rmiRNAs) derive from precursor and mature rRNAs through poorly understood mechanisms (Figure 2C) [5,6]. srRNA/rmiRNA generation is likely not due to random degradation but rather controlled processing that creates stable srRNA products, and may be Drosha-dependent or independent [6–8]. Functionally, rmiRNAs have been shown to bind AGO proteins and are implicated in the regulation of various metabolic and developmental pathways. srRNAs in fly and human cells associate with AGO complexes [7]. Differential hepatic srRNA expression was observed in diabetic mice, and specific srRNAs were found to modulate transcription of regulators of gluconeogenesis in mouse hepatoma cells [7]. Several human rmiRNAs were predicted to target stress- and cancer-related pathways [6], and differential rmiRNA expression has been observed upon heat stress in rice [9] and wheat [10]. In zebrafish, rmiRNAs mapping to the 18S and 28S rRNAs were found to be critical for early embryogenesis [8] and blood vessel formation [11], respectively. Additionally, differential spatiotemporal transcription of rDNA sequence variants may also alter rmiRNA expression [8,12]. Novel rRNA-derived miRNAs exemplify the emerging connection between miRNAs and ribosome biogenesis, underscoring how diverse cellular processes can be modulated by miRNAs that interact with ribosome production.

Dozens of miRNAs are stably enriched in the nucleolus in a variety of human and mammalian cell lines, and their localization is resistant to RNAPI transcription inhibition or other cellular and nucleolar stresses [13–15]. These nucleolar miRNAs may have diverse noncanonical biological roles including direct regulation of ribosomal subunit formation [16] or rRNA synthesis [17], the protected formation of pre-silenced miRNA:mRNA pairs away from the crowded and competitive cytoplasm [18], mediation of a defensive response to exogenous genetic material [14], or even targeting for deactivation by nucleolar deaminase editing [13]. These initial observations warrant additional follow-up to better define the identities and roles of nucleolar miRNAs. For a more extensive discussion of miRNAs localized to the nucleolus, we refer the reader to a brief review by Catalanotto et al. [19].

A growing cadre of disease-associated miRNAs have been shown to control components, subprocesses, or regulators of ribosome biogenesis (Figure 1, Table 1). Dysregulation of miRNA expression correlates with the progression of many types of cancer [20,21], and the link between cancer and ribosome biogenesis is well-established [22]. We review recent publications that consider (1) direct interplay among cancer, ribosome biogenesis and miRNAs holistically and (2) miRNAs regulators of ribosome biogenesis in ribosomopathies and in chronic obstructive pulmonary disease (COPD).

The well-studied cancer switches TP53 and MYC sit as regulatory endpoints of several miRNAs [23] and are intimately connected with ribosome biogenesis [22,24]. A genetic circuit encompassing hsa-miR-504, the nucleolar protein isoform FGF13 1A, and genome guardian TP53 attenuates ribosome biogenesis in a manner promoting cell survival in models of oncogenic escape (Figures 1 and 3A) [25]. miR-504 is an FGF13 mirtron that targets TP53, although constitutive transcription of the FGF13/MIR504 locus itself is negatively regulated by TP53 via understudied mechanisms. Concerted up-regulation of FGF13 1A and miR-504 represses pre-rRNA transcription and TP53 translation, in turn attenuating global protein synthesis, oncogenic proteotoxic and oxidative stress, and tumor cell apoptosis.

Several RPs are necessary for miRNA-mediated regulation of the MYC oncogene, which itself controls transcription of rDNA, RPs, AFs, and translation initiation factors [24]. uL5 (RPL11), uL18 (RPL5), and uS11 (RPS14), which stabilize TP53 in the nucleolar stress response [26–28], have been shown to escort the armed RISC complex to MYC transcripts for silencing by hsa-miR-24 [29,30] or hsa-miR-145 (Figure 3B) [31]. UV irradiation also induces uL5-guided MYC repression by hsa-miR-130a [32]. Via LIN28B, eL22 (RPL22) indirectly controls maturation of hsa-miR-let-7 family paralogs [33] that repress MYC, KRAS, HMGA1 and HMGA2, and other oncogenes [34,35].

Additionally, miRNAs affecting cancer treatment outcomes or oncogene expression have links with ribosome biogenesis. hsa-miR-7641 has been shown to repress uS9 (RPS16) directly and eight other RPs indirectly, and miR-7641 depletion sensitized breast and colon cancer cells to doxorubicin-induced apoptosis (Figure 1) [36]. The tumor suppressor hsa-miR-7-5p, which has been reported to target oncogenic EGFR, BCL2 [37], RELA (p65) [38], among others [39–41], may also play a role in down-regulating ribosome biogenesis. Depletion of the miR-7 homolog in Chinese hamster ovary (CHO) cells unleashed cell proliferation and antibody production, correlating with derepression of ribosomal assembly factor BMS1 and Akt pathway activators STRN3 and Ezrin (Figure 1) [42]. Overall, miRNAs can fine-tune the balance between proliferation and oncogenesis by modulating upstream regulators of ribosome biogenesis such as TP53 and MYC, as well as other oncogenes, RPs, and AFs downstream of pre-rRNA transcription.

Ribosomopathies are diseases of impaired ribosome biogenesis that result in tissue-specific defects in affected patients. Interruption of particular ribosome biogenesis subprocesses have been shown to cause a wide range of clinical presentations including anemia, leukemia, neutropenia, craniofacial defects, cancer, and other diseases of major organs [43–45]. miRNAs are involved in two classical hematological ribosomopathies, myelodysplastic syndrome (MDS) and Diamond–Blackfan anemia (DBA). As part of the 5q- MDS deletion, hsa-miR-145 and hsa-miR-146a [46] are lost, while deletion of hsa-miR-595 occurs in –7/7q- MDS [47]. In murine hematopoietic progenitor cells, codepletion of miR-145 and miR-146a homologs relieved repression of TIRAP and TRAF6, activating the innate immune response via IL-6 and leading to 5q- syndrome-like thrombocytosis and megakaryopoiesis [46]; these features were mirrored in similar experiments on patient bone marrow. miR-595 silences RPL27A resulting in TP53 activation, disruption of nucleolar architecture independent of TP53, reduction of mature 60S subunits, and blockade of erythroid proliferation and differentiation (Figure 1) [47]. RPL27A derepression enhanced proliferation in cell lines and was observed in –7/7q- MDS patients, though the authors cite the need for further translational studies [47]. Finally, tens of miRNAs were found to be differentially expressed in RP-depleted zebrafish models of the ribosomopathy DBA [48]. Potential targets of the up-regulated miRNAs were enriched for functions in transcriptional regulation and neuronal and cellular development. It will be of interest to reexamine other ribosomopathies for miRNAs important for pathogenesis.

Other disease-associated miRNAs have been connected to ribosome biogenesis. COPD and intensive care unit patients with muscle wasting exhibit elevated expression of hsa-miR-424-5p, which targets the RNAPI pre-initiation complex factors POLR1A and UBTF (Figure 1) [49]. miR-424-5p was found to down-regulate mature rRNA levels in myoblasts and to cause muscle atrophy in mice [49]. We hypothesize that yet-undiscovered miRNA regulators of ribosome biogenesis may also play central roles in diseases arising from defects in growth-sensitive biological processes, such as development and angiogenesis, that are already known to be partially controlled by miRNAs [50,51].

Although miRNAs generally down-regulate the expression of their target genes, select cases have unveiled miRNA-mediated enhancement of transcript-specific or global translation (Table 1). Both endogenous and synthetic miRNAs can increase translation efficiency of target transcripts by a specialized microRNA ribonucleoprotein (miRNP) complex that lacks the normally present repressor protein GW182/TNRC6 but contains the RNA binder FXR1a [52–54]. hsa-miR-10a binds the 5′UTR of RP transcripts harboring a 5′TOP motif, up-regulating their translation, augmenting ribosome biogenesis, and enhancing proliferation (Figures 1 and 3C) [55]. Such noncanonical translational activation is also implicated in cellular quiescence and senescence, two paused growth states instigated by mTOR deactivation that feature diminished ribosome biogenesis [56,57]. Translation-activating miRNAs were originally discovered in quiescent human cells [52,53], and down-regulation of miRNA processing machinery initiates global translation reduction and induction of senescence (Figure 1) [55]. In summary, miRNAs can also enhance ribosome biogenesis and global protein synthesis via targeted RP up-regulation or mechanisms implicated in cellular aging.

Besides the essential non-coding rRNAs, other long non-coding RNAs (lncRNAs) have been increasingly found to possess significant roles in eukaryotic ribosome biogenesis, with some even dysregulated in disease and proposed as therapeutic targets or biomarkers. lncRNAs are RNAs over 200 nt and without protein-coding potential. lncRNAs are transcribed in a variety of genomic contexts: intergenic, antisense, intronic, and overlapping orientations with respect to protein-coding genes [58]. This leads to their cis functions, where they regulate the expression of nearby genes. In addition, lncRNAs have diverse trans functions to regulate distant genes. Recent work has established many diverse cellular roles for these non-coding portions of the genome.

lncRNAs utilize different biogenesis pathways and sequence properties to perform roles in ribosome biogenesis. Some transcripts are transcribed by RNAPI from the rDNA gene locus (Figure 4A), while others are independently transcribed by RNAPII (lincRNAs) or produced by alternative splicing of RNAPII transcripts (Figure 4B). Ribosome biogenesis-specific lncRNA roles may be direct, modulatory, or stress response-specific. These molecular mechanisms include protein binding or sequestration, rDNA chromatin modifications, snoRNP formation, and transcript-specific translation modulations (Table 2). Here, we discuss examples of lncRNAs that function within the nucleolus to modulate rDNA transcription, pre-rRNA processing, and nucleolar structure as well as lncRNAs that alter ribosome biogenesis levels and translation located outside the nucleolus.

LoNA is a 1.5 kb polyadenylated lncRNA that performs dual independent functions to regulate RNAPI transcription, pre-rRNA processing, and modification (Figures 1 and 5A). LoNA was discovered by Li et al. [59] as the 4th most abundant lncRNA present in mouse neuroblastoma cell nucleoli. Overexpression of LoNA reduces pre-rRNA levels, mature rRNA levels, and global translation; conversely, knockdown enhances ribosome production. Two 5′-end nucleolin binding motifs sequester nucleolin from establishing rDNA euchromatin modifications and enhancing RNAPI transcription (Figure 5A) (nucleolin functions reviewed in [60]). At the same time, LoNA reduces pre-rRNA processing and rRNA modification by competition binding using its two 3′-end C/D box snoRNA sequences that form noncanonical C/D box snoRNPs. Additionally, LoNA levels are increased in Alzheimer's disease model mice, connecting a lncRNA to the pathogenesis of a neurodegenerative disease known to have associations with nucleolar stress and reduced rRNA production [59,61,62] (reviewed in [63]).

Other lncRNAs have been implicated in the control of pre-rRNA transcription through direct roles in rDNA chromatin modifications. Although pre-rRNA is the most highly transcribed RNA, rDNA genes can adopt inactive, silent, or active chromatin states. Here, we focus on rDNA chromatin activity dynamics regulating rDNA transcription; however, it is important to note these dynamics also play an important role in genome stability and organization (reviewed in [64,65]).

Promoter associated RNA (pRNA), is transcribed by RNAPI from the intergenic spacer (IGS) region of the rDNA and degraded by the exosome to produce 150–300 nt molecules (Figure 4A) [66]. pRNA is required for rDNA silencing through the recruitment of the nucleolar remodeling complex (NoRC) via direct interaction (Figures 1 and 5B) [67]. Additionally, a 5′ portion of pRNA forms a DNA:RNA triplex with the T₀ regulatory element of the rDNA promoter. While triplex formation has only been observed in vitro, it suggests that pRNA recruits the DNA methyltransferase DNMT3b to further establish repressive chromatin marks by CpG methylation at this element (Figure 5B) [68]. pRNA's function is also necessary for exit from pluripotency in embryonic stem cells [69] and was speculated to be important for maintaining genomic stability [70].

There are many IGS transcripts that regulate rDNA chromatin modifications, including promoter and pre-rRNA antisense transcripts (PAPAS) (Figures 1 and 4A). These transcripts inactivate pre-rRNA transcription upon a variety of stressors, including density arrest and serum deprivation that lead to H4K20me3 modifications at the rDNA promoter [71]. Additionally, hypo-osmotic, hypotonic, and heat shock stressors promote nucleosome remodeling and deacetylase (NuRD) complex recruitment and H4ac inhibition by a similarly proposed DNA:RNA triplex formation as that as pRNA at the rDNA promoter [72–74].

Pre-rRNA transcription modulation also occurs beyond the chromatin level by a snoRNA-ended lncRNA that enhances pre-rRNA transcription (SLERT) (Figure 1). Both of its ends bear H/ACA snoRNA sequences that are essential for its nucleolar localization [75]. Additionally, there is a 143-nt internal sequence that interacts with DDX21, an RNA helicase that forms rings around multiple RNAPI complexes to inhibit pre-rRNA transcription. This interaction relieves suppression and increases pre-rRNA transcription [75]. DDX21 is also important for pre-rRNA processing steps [76], but SLERT's role in pre-rRNA processing has not yet been explored.

Only a few lncRNAs have been reported to modulate pre-rRNA processing, but similarly to LoNA and SLERT, these lncRNAs modulate pre-rRNA processing via sequestration mechanisms. circANRIL binds and sequesters PES1, a member of the PES1, BOP1, WDR12 (PeBoW) complex that is essential for large subunit processing through internal transcribed spacer 2 (ITS2) cleavage [77]. PES1 binds circANRIL at a 47S pre-rRNA homology domain bearing sequence identities across the pre-rRNA primary transcript (Figure 5C) [78]. This interaction most likely inhibits PeBoW complex formation and PES1 interaction with the pre-rRNA, leading to nucleolar stress (Figure 1). Identification of other lncRNAs with pre-rRNA homology could provide competition regulation like circANRIL, which is proposed to be present in sufficient amounts in cells to modulate ribosome production. This anti-proliferative function is supported by observations of lower circANRIL abundance in childhood systemic vasculitis (Kawasaki disease) patients [79] and association with a higher risk of atherosclerosis in a human cohort [78].

RNAPII transcription inhibition disrupts nucleolar structure, indicating a possible function of RNAPII transcripts within the nucleolus [80]. Caudron-Herger et al. [81] recently identified RNAPII alu-repeat transcripts that are partially responsible for maintaining nucleolar structure. These 100–300 nt RNAs are spliced from introns and enriched in the nucleolus. They interact with nucleolin and nucleophosmin to enhance RNAPI transcription and support nucleolar structure [81]. These RNAs provide a possible mechanism for the coordination of RNAPI and RNAPII transcription, particularly since alu elements represent a large portion (∼10%) of the human genome [82].

Some nucleolar lncRNAs are not directly involved in ribosome biogenesis, but arise under different stress conditions and are associated with the nucleolar stress response. This stress response elicits outputs to inhibit ribosome synthesis and induce apoptosis. These lncRNAs include multiple IGS RNAs that are transcribed by RNAPI (Figure 4A) and form foci within the nucleolus under stress conditions, termed ‘nucleolar detention centers.' These lncRNAs sequester and inactivate proteins within these detention centers. Establishment of this remodeled nucleolus is thought to be important for stress-induced transcriptional inactivation [83] (reviewed in [65]).

While transcription of the polycistronic 47S pre-rRNA by RNAPI occurs within the nucleolus, the 5S rRNA is transcribed in the nucleoplasm by RNAPIII. 5S overlapping transcript (5S-OT) regulates this important extra-nucleolar transcription step (Figure 1). 5S-OT is transcribed by RNAPII and is complementary to the 5S rRNA. Its depletion leads to decreases in the production of 5S rRNAs through undiscovered cis-acting mechanisms [84].

snoRNA sequences in lncRNAs can be utilized in other unique ways to modulate ribosome biogenesis levels. This includes an orphan C/D box snoRNA, snoRD86 within an intron of the NOP56 pre-mRNA. Upon increased levels of NOP56 and snoRNPs, this intronic snoRNA is able to bind and sequester C/D box snoRNPs to inhibit splicing of the NOP56 pre-mRNA. This ultimately ends in nonsense-mediated decay of the unspliced transcript, containing snoRD86, termed snoRD86 cytoplasmic 5′ snoRNA capped 3′ polyadenylated RNA (cSPA) (Figure 1) [85]. It will be of interest to determine the extent to which splicing feedback regulation occurs for intronic snoRNAs in other ribosome biogenesis factor pre-mRNAs.

While ribosome biogenesis is generally increased in cancer (reviewed in [86]), synchronization of this process with mitochondrial ribosome production is also necessary for cancer cell fitness [87,88]. Survival associated mitochondrial melanoma specific oncogenic non-coding RNA (SAMMSON) is another example of a lncRNA that binds ribosome biogenesis factors to fine-tune ribosome biogenesis in melanoma cells for synchronization of mitochondrial and nucleolar ribosome production (Figure 1). SAMMSON binds p32, a protein essential for mitoribosome formation [89], and increases its mitochondrial localization [88]. Furthermore, SAMMSON increases the nucleolar localization of XRN2, an exonuclease necessary for pre-rRNA processing [90]. It performs this latter task by binding CARF, a negative regulator of XRN2 [91], that normally increases its nucleoplasmic localization [92]. SAMMSON's presence increases nucleolar and mitochondrial ribosome synthesis, while its depletion leads to aberrant pre-rRNA processing intermediates. SAMMSON expression has only been reported in cancer cells; however, it is possibly expressed during embryonic development, since p32 is essential for development [89].

snoRD86 cSPA and SAMMSON utilize binding sequences to sequester RNA binding proteins from performing their functions in pre-rRNA processing. While it is not yet clear how these lncRNAs function in normal cellular metabolism, their study has revealed a mechanism for how ribosome biogenesis can be controlled from outside the nucleolus. Nevertheless, the presence of snoRNA sequences within lncRNAs might not always lead to direct roles in ribosome biogenesis modulation. Prader–Willi Syndrome deletion region (15q11-q13) transcribes several SPAs and snolncRNAs that have no current identified function in ribosome biogenesis (reviewed in [93]).

Although lncRNAs have been increasingly observed to be translated into micro peptides (reviewed in [94]), there is a lncRNA subcellular localization system (lncSLdb) that identified 300 lncRNAs associated with mature ribosomes in the cytoplasm through extensive literature mining [95]. This indicates possible lncRNA regulation of final ribosome maturation steps or translation efficiency if not translated into micro peptides. As of now, there are cases of the latter being studied (reviewed in [96]) (see examples in Figure 1 and Table 2).

Studies continue to emerge and implicate novel nucleolar miRNAs and lncRNAs in ribosome biogenesis regulation. Although canonical miRNA-mediated translational repression occurs in the cytoplasm, new roles have been revealed for miRNAs localized to the nucleolus including regulation of RNAPI and subunit assembly, defense against exogenous nucleic acids, and precise control of miRNA stability and target recognition. The lncRNA erythrocyte membrane protein band 4.1 Like 4A antisense 1 (EPB41L4A-AS1) was identified as a possible regulator of glycolysis and glutaminolysis in cancer. Yet, it interacts with HDAC2 and NPM1 to increase their nucleolar localization [97], raising the question: does EPB41L4A-AS1 expression down-regulate RNAPI transcription if HDAC2 is able to establish heterochromatin marks on the rDNA (Figure 1)? Additionally, lncRNAs expressed from the distal flanking junction of the rDNA (DISNOR 187 and 238) have been recently discovered. Depletion of these leads to decreased pre-rRNA transcription [98], suggesting a functional role of these lncRNAs in the transcription of the rDNA (Figure 1). Further study of nucleolar miRNAs and lncRNAs will illuminate how ncRNAs other than rRNA shape the dynamic landscape of the nucleolus.

Both miRNAs and lncRNAs play important roles in cancer and development, but to date only miRNAs have been associated with ribosomopathies, such as myelodysplasia and DBA. However, one study identified candidate miRNAs and lncRNAs that have altered expression in a DBA zebrafish model compared with wildtype [48]. Continued work will help elucidate new ncRNAs that play a role in ribosomopathy pathogenesis and disease-related nucleolar stress responses.

Published bioinformatic databases may contain more miRNAs and lncRNAs that are candidates for ribosome biogenesis regulation based on their localization and links to disease. RNALocate notes 79 nucleolar human miRNAs, plus 448 nucleolar and 190 ribosome-associated human lncRNAs on top of the lncRNAs in lncSLdb [99]. lncATLAS identified 28 lncRNAs that were enriched in the nucleolus compared with the nucleoplasm by subcellular fractionation followed by RNA-seq in K652 cells [100]. In addition to LoNA, Li et al. [59] also identified several other nucleolar localized lncRNAs by a similar RNA-seq method. It would be of interest to extend these experiments to look for conserved nucleolar miRNAs and lncRNAs across cell types and species. The mammal ncRNA-disease repository (MNDR) database, which catalogs ncRNAs linked to diseases, may provide clues about novel miRNA and lncRNA regulators of ribosome biogenesis based on their involvement in ribosomopathies or otherwise growth-sensitive diseases [101]. Together, bioinformatic databases bearing localization and disease linkage data about miRNAs and lncRNAs can help reveal novel ncRNA regulators of ribosome biogenesis.

• miRNAs and lncRNAs play diverse roles in the regulation of ribosome biogenesis, forming an additional dense layer of control over cellular growth and translational output. In diseases where ribosome biogenesis is implicated, such as cancer and various ribosomopathies, these ncRNAs comprise another conduit by which biochemical defects become the basis of medical pathogenesis.

• miRNAs and lncRNAs modulate core ribosome biogenesis processes including RNAPI pre-rRNA transcription, pre-rRNA processing, and ribosome assembly as well as nucleolar structural maintenance and global translation.

• Future discovery of novel ncRNA regulators of ribosome biogenesis and of expanded regulatory roles for known ncRNAs will be aided by (1) experimental inspection of specific steps in ribosome biogenesis and (2) bioinformatic databases containing localization and disease linkage data.

The authors declare that there are no competing interests associated with the manuscript.

S.J.B. acknowledges 1R35GM131687 from the National Institutes of Health (NIH), and the Breast Cancer Alliance for support of her laboratory. M.A.M. and C.J.B. acknowledge T32GM007223 from the NIH for funding.

C.J.B. wrote the miRNA portion and M.A.M. wrote the lncRNA portion. All authors read, commented on, discussed, edited, and approved the final manuscript.

Thank you to the members of the laboratory of S.J.B. for insightful information, questions, and comments throughout the manuscript writing process.

AFs

assembly factors

AGO

Argonaute protein

COPD

chronic obstructive pulmonary disease

cSPA

capped 3′ polyadenylated RNA

DBA

Diamond–Blackfan anemia

IGS

intergenic spacer

lncRNAs

long non-coding RNAs

LoNA

long nucleolar RNA

MDS

myelodysplastic syndrome

miRNAs

microRNAs

mRNA

messenger RNA

ncRNAs

non-coding RNAs

pRNA

promoter associated RNA

rDNA

ribosomal DNA

RISC

RNA-induced silencing complex

RPs

ribosomal proteins

rRNA

ribosomal RNA

snoRNPs

small nucleolar ribonucleoproteins