Abstract
The mechanisms by which imprinted loci control activity of hematopoietic stem cells (HSCs) are not known. In this issue of Cell Stem Cell, Qian et al. (2016) demonstrate that non-coding RNAs expressed by the maternal-imprinted locus Dlk1-Gtl2 maintain HSC self-renewal through the inhibition of PI3K-mTOR signaling, mitochondrial biogenesis, and metabolic activity.
Regulation of hematopoietic stem cell (HSC) metabolism is the subject of active research. Adult HSC activity has been associated with quiescence, and adult HSC quiescence is inversely correlated with active oxidative metabolism. In contrast, fetal HSCs remain functional although they are continuously cycling. The question therefore arises as to whether there are ontogenically conserved repressive metabolic programs that maintain low metabolic activity in cycling fetal HSCs, or, alternatively, whether adult HSCs may activate completely different programs connecting cycling and metabolism.
Genomic imprinting is an epigenetic mechanism resulting in monoallelic expression of genes according to their parental origin. Imprinting of one of the two parental chromosomes is established during gametogenesis at specific differentially methylated regions (DMRs) (Barlow and Bartolomei, 2014). Forms of genomic imprinting have been demonstrated in fungi, plants, and invertebrates. However, genomic imprinting is a fairly rare phenomenon in mammals. A limited number of loci do remain imprinted during evolution, which suggests crucial regulatory roles for these loci in important cellular functions. Current evidence supports that imprinted loci frequently contain gene clusters for both coding transcripts and non-coding RNAs (ncRNAs) with specific program silencing functions (reviewed in Kung et al., 2013), and typically, the paternally inherited allele’s expression promotes growth, while the maternally inherited allele’s expression suppresses it (Barlow and Bartolomei, 2014).
Previous work by Goodell and colleagues showed that a network of imprinted genes (IGN) is predominantly expressed in adult somatic stem cells, including HSCs, relative to their differentiated progeny (Berg et al., 2011), and they identified the imprinted Gtl2 gene as one of the genes exclusively expressed in HSCs and their committed lineages, acting as a “fingerprint” reflecting ontogeny (Chambers et al., 2007). The Gtl2 gene is located in the maternal imprinted Dkl1/Gtl2 cluster on human chromosome 14q32 (murine chromosome 12qF) and is regulated by an intergenic DMR (IG-DMR) between the Dlk1 and Gtl2 genes and a DMR spanning the Gtl2 promoter (Gtl2 DMR). In humans, maternal and paternal UPD14 syndromes are associated with growth and neurodevelopment defects and are pathogenetically related to dysregulation of the Dlk1-Gtl2 locus (Plasschaert and Bartolomei, 2014).
In this issue of Cell Stem Cell, Li and colleagues used an unbiased approach to reveal the expression of unique long ncRNA (lncRNA) fingerprints in 17 populations of the hematopoietic hierarchy (Qian et al., 2016). They observed that a higher level of expression of 12 lncRNA transcripts distinguished fetal liver (FL)-HSCs and adult CD49blo long-term (LT)-HSCs (which have quiescence features) from other HSC subpopulations, progenitors, and mature hematopoietic cells. One of these lncRNA transcripts corresponded to the maternally imprinted H19 locus, which had been found by the same group to serve as a source of microRNA (miR)-675 and to restrict paternal Igf1r expression (Venkatraman et al., 2013). Conditional deletion of maternal, but not paternal, H19-DMR results in activation of the Igf2-Igfr1 pathway, translocation of phosphorylated (inactive) FoxO3 from nuclear to cytoplasmic location, and release of FoxO3-mediated cell cycle arrest, leading to increased HSC activation and exhaustion (Venkatraman et al., 2013). Out of the other 11 lncRNAs, 3 lncRNAs are located within the imprinted Dlk1-Gtl2 locus. The Dlk1-Gtl2 locus includes three paternally inherited protein-coding genes (Dlk1, Rtl1, and Dio3) and multiple maternally inherited ncRNAs, which are regulated by a common cis-element and epigenetic control that results in a large polycistronic transcription unit (da Rocha et al., 2008) (Figure 1). These ncRNAs include the lncRNA genes Gtl2, Rian (which also expresses 22 box C/D small nucleolar RNAs (snoRNAs) and has the ability to methylate other RNA molecules), and Gm26906. The Dlk1-Gtl2 locus also expresses the largest miRNA cluster in mammalian cells consisting of anti-Rtl1 (containing seven miRNAs within the miR-127/miR-136 cluster) and Mirg (containing 39 miRNAs within the miR-379/miR-410 cluster) (da Rocha et al., 2008) (Seitz et al., 2004) (Figure 1).
Figure 1. Non-coding RNAs Expressed by the Dlk1-Gtl2 Locus Prevent Metabolic Activation of HSCs.

Bottom panel: schema of the Dlk1-Gtl2 locus located in the murine chromosome 12qF1 including DMR (with open oval marks representing active transcription in the maternal allele and solid oval marks representing repressed transcription in the paternal allele), genes transcribing for the protein-coding genes Dlk1, Rtl1, and Dio3 in the paternal allele (blue), and intergenic transcripts for multiple ncRNAs regulated by a common cis-element and epigenetic control resulting in a large polycistronic transcription unit that encompasses lncRNA, snoRNA, and miRNA (dotted hairpins) in the maternal allele (pink). Arrows depict direction of transcription. Dotted lines depict post-transcriptional process of miRNA cistron. Upper panel: repressive role of ncRNA on the post-transcriptional process of mediators of the PI3K-mTOR signaling pathway, which controls ribosomal protein translation and mitochondrial biogenesis/metabolism resulting in maintenance of LT-HSC activity.
Li’s group studied the function of the Dlk1-Gtl2 locus in HSCs by using genetic mutant mice with an allele-specific DMR deficiency (Qian et al., 2016). Maternal-specific DMR deficiency of the Dlk1-Gtl2 locus induced lethality in E16 embryos. Loss of epigenetic marks at the maternal Dlk1-Gtl2 (mat ΔIG) locus or the Gtl2 gene (mat ΔGtl2) decreased the size of the Lin−/c-kit+/Sca-1+ (LSK) cell population in hematopoietic E15 FL. Maternal-specific FL mutants contained fewer competitive repopulation units (CRUs) and increased cycling of CD49blo (quiescent) adult HSCs, and donor-derived hematopoiesis displayed features of biased differentiation ability, including myeloid lineage bias, associated with impaired HSC activity. Gene ontology and enrichment analysis of the transcriptomes of wild-type and mutant FL LSK cells confirmed increased levels of proliferation, growth, development, and metabolic activation in mutant HSCs, which was associated with upregulation of cellular response to insulin stimulus and positive regulation of the PI3K cascade. The investigators hypothesized that the Dlk-Gtl2 locus is a negative regulator of the PI3K-mTOR pathway.
The PI3K/Akt/mTOR signaling pathway is considered to be a central activator of HSC cell cycle activity and controls the transition from the G0 to GAlert phase in quiescent HSCs (Rodgers et al., 2014). Using intracellular phospho-flow analysis, the investigators found that active pAKTS473, mTORC1S2448, mTORc2S2481, and pS6S235/S236 were upregulated in maternal-DMR-deficient HSCs, but not in their progenitor lineage counterparts. Moreover, RNA-sequencing analysis revealed enrichment of the small ncRNAs in the Gtl2 locus in FL LSK and adult CD49lo HSCs, which were downregulated in mat ΔIG FL HSC. By combining the ncRNAs downregulated in mat ΔIG and upregulated in CD49blo HSCs, the authors identified 15 miRNAs and 10 snoRNAs that are overexpressed in adult CD49blo HSCs, which have possible activity in the PI3K-mTOR pathway (Figure 1). A total of 21 genes distributed over the entire PI3K-mTOR pathway were found to contain one or more conserved binding sites of Gtl2-derived miRNAs in their respective 3ʹ untranslated region (UTR). Target sites for Gtl2-derived miRNAs were confirmed by luciferase reporter assays for IGF1, IGF1R, IRS1, PIK3R1, AKT1, RHEB, RPTOR, PGC-1A, RPS6KB1, and EIF4E. Loss of imprinting at the Dlk1/Gtl2 domain enhanced mitochondrial biogenesis oxidative phosphorylation pathways. The activation of both pathways is associated with loss of quiesence, which in turn resulted in elevated basal adenosine triphosphate and reactive oxygen species (ROS) levels in mat ΔIG HSCs, but not in lineage-committed progenitors. Finally, the authors demonstrated that pharmacological inhibition of the PI3K-mTOR signaling pathway by rapamycin, or pharmacological ROS neutralization by N-acetyl cysteine (NAC), partially restored the deficient activity of mat ΔIG HSCs.
This “tour de force” work has exposed a network of regulatory ncRNAs that acts as a crucial negative post-transcriptional regulator of the mediators of the PI3K-mTOR signaling pathway and HSC activation, and to date, it provides the best studied example of regulation by an imprinted locus of repression of HSC activation. This is one of the first studies, but surely not the last, that identifies the mechanisms of how maternal and paternal “imprintomes” affect HSC activity and fate, and it emphasizes the developmentally regulated nature of HSC function and illustrates an ontogenically conserved pathway of repressive metabolic programs that maintain low metabolic activity in HSCs.
ACKNOWLEDGMENTS
The authors want to thank Ms. Margaret O’Leary for her editorial assistance. This work was partly funded by NIH R01 GM110628.
REFERENCES
- Barlow DP, and Bartolomei MS (2014). Cold Spring Harb. Perspect. Biol 6, a018382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Berg JS, Lin KK, Sonnet C, Boles NC, Weksberg DC, Nguyen H, Holt LJ, Rickwood D, Daly RJ, and Goodell MA (2011). PLoS ONE 6, e26410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chambers SM, Boles NC, Lin KY, Tierney MP, Bowman TV, Bradfute SB, Chen AJ, Merchant AA, Sirin O, Weksberg DC, et al. (2007). Cell Stem Cell 1, 578–591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- da Rocha ST, Edwards CA, Ito M, Ogata T, and Ferguson-Smith AC (2008). Trends Genet. 24, 306–316. [DOI] [PubMed] [Google Scholar]
- Kung JT, Colognori D, and Lee JT (2013). Genetics 193, 651–669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Plasschaert RN, and Bartolomei MS (2014). Development 141, 1805–1813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Qian P, He XC, Paulson A, Li Z, Tao F, Perry JM, Guo F, Zhao M, Zhi L, Venkatraman A, et al. (2016). Cell Stem Cell 18, this issue, 214–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rodgers JT, King KY, Brett JO, Cromie MJ, Charville GW, Maguire KK, Brunson C, Mastey N, Liu L, Tsai CR, et al. (2014). Nature 510, 393–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Seitz H, Royo H, Bortolin ML, Lin SP, Ferguson-Smith AC, and Cavaillé J (2004). Genome Res. 14, 1741–1748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Venkatraman A, He XC, Thorvaldsen JL, Sugimura R, Perry JM, Tao F, Zhao M, Christenson MK, Sanchez R, Yu JY, et al. (2013). Nature 500, 345–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
