Non-coding RNAs that regulate development

We focus on identifying and characterizing lncRNAs involved in modulation of cell differentiation along specific lineages. We have generated catalogs of lineage- specific lnc RNAs that are essential for the differentiation and function of erythroid cells, as well as others essential for formation of white and/or brown adipose cells. We are identifying their mRNA and protein targets, and studying their roles and mechanisms during cell development and disease.


Long non-coding RNAs (lncRNAs) are transcripts longer than 200nt that do not function through encoded protein products. Many are capped, polyadenylated, and often spliced, and transcribed by RNA Polymerase 2 (Pol2). lncRNAs constitute a significant fraction of the mammalian transcriptome. Compared to mRNAs, lncRNAs tend to be shorter and less well conserved at the primary sequence level. Expression of lncRNAs is often restricted to specific tissues and developmental stages, suggesting that many may regulate cell fate specification .A few dozen intergenic lncRNAs (lincRNAs) have been functionally characterized in mammals, and they have been associated with important developmental processes such as apoptosis, proliferation, and lineage commitment. However, the biological functions of the majority of these genes and their potential roles in development and disease still remain uncharacterized.

An erythroid-specific long non-coding RNA prevents apoptosis of erythroid progenitors and promotes terminal proliferation.

Wenqian Hu identified one erythroid-specific lncRNA, LincRNA-EPS, with potent anti-apoptotic activity. Expression of LincRNA-EPS is largely confined to terminally differentiating fetal erythroid cells and its expression is induced in CFU-E progenitors by Epo. Inhibition of this lncRNA blocks erythroid differentiation and promotes apoptosis. Ectopic expression of this lncRNA in CFU-E progenitors prevents erythroid progenitor cells from the apoptosis that is normally induced by erythropoietin deprivation. This lncRNA represses expression of several proapoptotic genes including the one encoding Pycard, an activator of caspases, explaining in part the inhibition of programmed cell death. These findings revealed a novel layer of regulation of cell differentiation and apoptosis by a lncRNA.

LincRNA-EPS acts as a transcriptional brake to restrain inflammation

In collaboration with the laboratory of Katherine Fitzgerald of the University of Massachusetts Medical School, Wenqian showed that lincRNA-EPS is also expressed in macrophages and is suppressed in macrophages exposed to diverse microbial products. To gain insight into its immune functions, Wenqian generated lincRNA-EPS-deficient mice. Transcriptome analysis of macrophages from these lincRNA-EPS-deficient mice, combined with both rescue and gain of function experiments, revealed a specific role for this lincRNA in restraining immune response gene (IRG) expression. lincRNA-EPS-deficient mice manifested enhanced inflammation and lethality following endotoxin challenge in vivo. lincRNA-EPS associates with chromatin at regulatory regions of IRGs to repress their transcription. Further, lincRNA-EPS mediates these effects by interacting with heterogeneous nuclear ribonucleoprotein L. The identification of lincRNA-EPS as a repressor of IRGs underscores the importance of long noncoding RNAs as crucial regulators of immune cell function.

Multiple types of long non-coding RNAs regulate red blood cell development

To obtain a comprehensive view of how lncRNAs contribute to erythropoiesis, Wenqian Hu and Juan R. Alvarez-Dominguez, together with two undergraduates, Staphany Park and Austin Gromatzky, generated and analyzed >1 billion RNA-seq reads of both Poly(A)+ and Poly(A)- RNA from mouse fetal liver erythroid progenitor cells and terminal differentiating erythroblasts. Using de novo transcript reconstruction, they identified 655 lncRNA genes including not only intergenic, antisense, and intronic RNAs but also transcripts of pseudogenes and enhancer loci. Over 100 of these genes were previously unrecognized and are highly erythroid specific. They then combined genome-wide surveys of expression levels, chromatin states, and transcription factor occupancy in these cells with computational analyses to systematically characterize all lncRNA subclasses by a spectrum of >30 features covering structural, conservation, regulation and expression traits. They uncovered global features of the biogenesis and the coordination of chromatin and transcription dynamics of lncRNAs during erythropoiesis, as well as subclass-specific patterns in conservation and tissue and developmental stage specificity.

They then focused on differentiation-induced lncRNAs, including novel erythroid-specific lncRNAs conserved in humans that are nuclear-localized. They selected 12 erythroid-specific lncRNAs that, like lincRNA-EPS, are greatly induced during erythroid terminal differentiation and are targeted at their promoters by the key erythroid transcription factors GATA1, TAL1 and KLF1. Remarkably, shRNA-mediated loss-of-function assays revealed that all 12, including alncRNA-EC7, now renamed Bloodlinc and discussed below, are essential for this developmental process.

In mammals, definitive erythropoiesis occurs sequentially in the fetal liver followed by the bone marrow. To explore the differences in definitive erythropoiesis in these two distinct niches, Wenqian performed transcriptomic profiling of terminal differentiating erythroblasts isolated from these two organs. In collaboration with Juan R. Alvarez-Dominguez, Wenqian is annotating and characterizing differentially expressed lncRNAs in these two hematopoietic organs and characterizing their functions in erythropoiesis. Much of this work is continuing in Wenqian’s own laboratory in the Mayo Clinic.

Jide Ezike, a technical assistant who works closely with Sherry and Xiaofei, is computationally analyzing public RNA-seq datasets to identify lncRNAs that are expressed in BFU-E cells and that may be important for BFU-E self-renewal. To date he has identified hundreds of lnc RNA’s that are differentially expressed between BFU-E and CFU-E cells. He has focused on a few interesting candidates; for each of these he has confirmed via quantitative-PCR their differential expression patterns. He is currently performing knockdown and over-expression experiments in the K562 erythroid cell-line as well as in primary human CD34+ cells, to determine the phenotypic role of these candidate lncRNAs. Identifying these differentially expressed candidates, Jide seeks to unearth lncRNAs that are important for BFU-E self-renewal promoted by corticosteroids and PPARα agonists, and uncover the underlying molecular mechanisms.

Bloodlinc, a super-enhancer derived trans-acting lncRNA that potentiates red blood cell development

Juan R. Alvarez-Dominguez, together with Marko Knoll, initially focused on enhancer -derived non-coding RNAs (ncRNAs) that are thought to contribute locally to the functioning of their parent enhancers. But their interest turned to understanding whether large domains of clustered enhancers (super-enhancers) also produce cis-acting noncoding RNAs that mediate their function.

To begin they globally profiled super-enhancers and their associated noncoding RNAs in primary erythroblasts, and showed that Bloodlinc is transcribed from an erythroid-specific super enhancer upstream of the gene encoding Band 3, a major erythrocyte membrane protein. Knockdown of Bloodlinc showed that it was essential, not only for Band3 gene activation, but also for establishing the entire red cell gene expression program during terminal differentiation. Inhibiting Bloodlinc inhibited erythroblast survival and blocked terminal cell proliferation and differentiation, whereas ectopically expressed Bloodlinc promoted erythroblast proliferation and enucleation in the absence of differentiation stimuli. Bloodlinc diffuses beyond its insulated superenhancer domain and mediates activation of the Gata1/Tal1/Klf1-regulated program of terminal erythropoiesis, including but not limited to Band3 induction. Thus, Bloodlinc represents a novel type of trans-acting, super-enhancer lncRNA essential for gene control during cell type-specific differentiation.

LincRNAs in fat cell development and function.

Many protein coding genes, mRNAs, and microRNAs have been implicated in regulating adipocyte development; however, the global expression patterns and functional contributions of long intergenic noncoding RNAs (lincRNAs) during adipogenesis have not been explored. Lei Sun and Ryan Alexander, collaborating with John Rinn’s group at the Broad Institute, examined the roles of lincRNAs in adipogenesis. To begin, they profiled the transcriptome of primary brown and white adipocytes, pre- brown and white adipocytes, and cultured adipocytes and identified 175 lincRNAs that are specifically regulated during both brown and white adipogenesis. Many lincRNAs are adipose-enriched, strongly induced during adipogenesis, and bound at their promoters by key adipogenic transcription factors such as PPARγ and CEBPα. RNAi-mediated loss of function screens identified 9 functional lincRNAs required for adipogenesis; mRNA analyses showed that each of these lincRNAs is essential for normal induction of a discrete set of adipocyte- induced mRNAs and for down regulation of a discrete set of mRNAs expressed in adipocyte progenitor cells.

They further focused on one of them, Firre, an X-linked lincRNA required for proper adipogenesis. Firre is exclusively nuclear and interacts with the nuclear matrix factor hnRNP-U through numerous copies of an RNA sequence motif conserved between human and mouse, an association that is required to mediate trans-chromosomal interactions between loci encoding known adipogenic factors. Thus, numerous lincRNAs comprise a critical transcriptional regulatory layer that is functionally required for proper differentiation of both brown and white adipocytes.

Marko Knoll has chosen 9 functional lincRNAs to determine the mechanism by which they influence the adipocyte developmental program. All 9 are upregulated during adipogenesis in vitro. Using a candidate approach Marko immunoprecipitated hnRNPU, Suz12, Ezh2 and CoREST proteins and tested if these 9 lincRNAs were potential interaction partners; three potential candidates were identified, additional to linc FIRRE, that are associated with hnRNPU. All three are strictly cytoplasmic lncRNAs, raising the question whether lincRNAs shuttle between the nucleus and cytoplasm because hnRNPU is strictly nuclear. Using RNA immunoprecipitation of hnRNPU in preadipocytes and differentiated mature adipocytes followed by deep sequencing, Marko aims to identify possible interaction partners. Further, using the CRISPR/Cas system, Marko aims to knock out or mutate these lincRNAs in the preadipocytic 3T3-L1 cell line.

Lei Sun is now continuing work on other lncRNAs in his own laboratory in the Duke- NUS Medical School in Singapore. Together with Juan R. Alvarez-Dominguez, they have used RNA-seq to reconstruct de novo transcriptomes across different fat depots, identifying ~1,500 adipose lncRNAs, including 127 brown fat-restricted lncRNAs that are often targeted by key transcriptional regulators PPARγ, C/EBPα, and C/EBPβ. They showed that one +of them, lnc-BATE1, is required for establishment and maintenance of BAT identity and thermogenic capacity. lnc-BATE1 inhibition impairs concurrent activation of brown fat and repression of white fat genes and is partially rescued by exogenous lnc-BATE1 with mutated siRNA-targeting sites, demonstrating a function of this lnc RNA in trans. Like Firre, lnc-BATE1 binds heterogeneous nuclear ribonucleoprotein U and both are required for brown adipogenesis. They have thus established lnc-BATE1 as a prototypical fat depot-selective lncRNA regulator of BAT development and physiology.

Mengxi Jiang, a new postdoctoral fellow, is investigating the role of lncRNAs in the development of two types of white adipose tissue – visceral (around internal organs) and subcutaneous fat (underneath the skin) – that differ in metabolic functions and are regulated by shared yet distinct transcriptional cascades. It is excess visceral fat that is the major risk factor for Type II diabetes. Our previous RNA-seq analysis identified several visceral, subcutaneous, and BAT- specific noncoding RNAs. She ranked these lncRNAs based on their abundance, tissue specificity, and relative expression between visceral and subcutaneous adipose tissue, and chose 5 top candidates from each depot for further experimental analysis. She showed that depletion of one visceral preadipocyte-specific lncRNA enhances visceral adipogenesis, but reduces subcutaneous adipogenesis in vitro. She is probing the role and mechanism of this and other non-coding RNAs in regulating the function and development of these two types of WAT using several molecular and cell biology techniques, and in genetically modified mice.

MicroRNAs in fat cell development and obesity

Marko Knoll, building on work of former lab members Lei Sun, Huangming Xie, and Ryan Alexander, is investigating the role of miRNAs in brown fat adipogenesis. Mammals have two principal types of fat: white adipose tissue (WAT) primarily serves to store extra energy as triglycerides, while brown adipose tissue (BAT) is specialized to burn lipids for heat generation and energy expenditure as a defense against cold and obesity.

Recent studies demonstrated that brown adipocytes arise in vivo from a Myf5-positive, bipotential myoblastic progenitor by the action of the Prdm16 (PR domain containing 16) transcription factor. Lei and colleagues identified a brown fat-enriched miRNA cluster, miR-193b-365, as a key regulator of brown fat development. Blocking miR-193b and/or miR-365 in primary brown preadipocytes dramatically impaired brown adipocyte adipogenesis by enhancing expression of Runx1t1 (runt-related transcription factor 1; translocated to 1) whereas myogenic markers were significantly induced. In contrast, forced expression of miR-193b and/or miR-365 in C2C12 myoblasts blocked the entire program of myogenesis, and ectopic miR-193b expression induced myoblasts to differentiate into brown adipocytes. MiR-193b-365 was upregulated by Prdm16 partially through the action of the transcription factor PPARγ. Taken together, these results underlie the importance of tissue enriched miRNAs 193b-365 in regulating lineage specification between brown fat and muscle, and also suggest that these or other miRNAs may have therapeutic potential in inducing expression of brown fat-specific genes.

Another miRNA, miR-203 was also enriched in brown fat. Marko Knoll is following up the work by Ryan Alexander. Ryan showed that knock down of miR-203 results in a block of adipogenesis in primary brown pre-adipocytes. Further, forced expression of miR-203 in C2C12 myoblasts blocked the myogenic program and induced differentiation into adipocytes. Marko aims to search for the target of miR-203. With the development of the CRISPR/Cas system it is possible to insert multiple mutations of up to 25 base pairs in a gene. Marko will use the CRISPR/Cas approach to generate mice that have a mutated seed sequence in miR-203 and analyze the effect of miR-203 seed region mutation on the development of white and brown adipose tissue.