Nat Commun. 2018 Apr 6;9(1):1329 

De novo reconstruction of human adipose transcriptome reveals conserved lncRNAs as regulators of brown adipogenesis

Chunming Ding


摘要:

Obesity has emerged as an alarming health crisis due to its association with metabolic risk factors such as diabetes, dyslipidemia, and hypertension. Recent work has demonstrated the multifaceted roles of lncRNAs in regulating mouse adipose development, but their implication in human adipocytes remains largely unknown. Here we present a catalog of 3149 adipose active lncRNAs, of which 909 are specifically detected in brown adipose tissue (BAT) by performing deep RNA-seq on adult subcutaneous, omental white adipose tissue and fetal BATs. A total of 169 conserved human lncRNAs show positive correlation with their nearby mRNAs, and knockdown assay supports a role of lncRNAs in regulating their nearby mRNAs. The knockdown of one of those, lnc-dPrdm16, impairs brown adipocyte differentiation in vitro and a significant reduction of BAT-selective markers in in vivo. Together, our work provides a comprehensive human adipose catalog built from diverse fat depots and establishes a roadmap to facilitate the discovery of functional lncRNAs in adipocyte development.High-throughput sequencing coupled with computational pipelines have driven tremendous progress in the discovery of novel lncRNAs24,32,33. Nonetheless, the current lncRNA annotation is far from complete owing to its tissue-specific expression property. To better identify and characterize lncRNAs expressed in human adipose in vivo, we set out to de novo reconstruct the transcriptome by profiling human fetal BAT, oWAT, and sWAT (Fig. 1a, Supplementary Fig. 1A, Supplementary Table 1). We performed deep strand-specific, 100 bp paired-end sequencing on poly(A)-selected RNA and generated ~682.4 million reads. Reads were first mapped to Hg19 using Tophat34 and subsequently input to Cufflinks34 for transcriptome assembly and gene quantification. The precision of our RNA-seq data and de novo assembly were confirmed by examining the gene expression levels of pan, white, and brown fat markers (Fig. 1b) and predicted gene structures for Ucp1 and Leptin (Supplementary Fig. 1B).We applied a stringent filter to focus only on long (>200 bp) transcripts that do not overlap with mRNA exons on the same strand (against UCSC, refSeq, and Ensembl databases) and show no evidence of protein-coding capacity (Fig. 1c). To prevent artifacts introduced by single-exonic fragments with low expression, only multi-exon transcripts were considered. By implementing this strategy, we identified 3149 lncRNAs, which exhibit similar characteristics as previously reported24,29,32,33,35 such as lower gene expression (Fig. 1d), lower isoform number (Supplementary Fig. 1C), lower exon count (Supplementary Fig. 1D), shorter transcripts (Supplementary Fig. 1E), and shorter open reading frame (Supplementary Fig. 1F) than mRNAs. 1587 out of 3149 (50.40%) lncRNAs (Fig. 1e) while 1631 out of 14 383 (11.34%) mRNAs (Supplementary Fig. 1G) are detectable in only one of the three adipose subtypes, highlighting the tissue-specific expression nature of lncRNAs.Given the exponential progress of lncRNA annotations in well-established database such as Gencode, we proceeded to assess the contribution of our de novo-reconstructed catalog to the existing knowledge base. Referencing against Gencode v24 human lncRNA annotation, we identified 2129 previously unannotated lncRNAs, accounting for more than two-thirds of newly built catalog. Analysis of the novel transcripts against the annotated counterparts revealed that novel lncRNAs have significantly fewer exons (two-sided Mann–Whitney U-test p < 2.2 × 10−16, Supplementary Fig. 1H) and isoforms (two-sided Mann–Whitney U-test p = 1.909e-15, Supplementary Fig. 1I), shorter (Supplementary Fig. 1J) and lower expressed (Fig. 1F, Supplementary Fig. 1K) than annotated ones. Importantly, the proportion of tissue-unique lncRNAs were higher in novel (57.4%) than annotated category (36.5%; Fig. 1g, Supplementary Fig. 1L), confirming that current existing lncRNA database tends to miss tissue-specific lncRNAs. Taken together, we have provided a comprehensive catalog of human adipose lncRNAs, which are mostly unannotated.Human fetal BAT was obtained from Advanced Bioscience Resources (Alameda, CA) from deceased donors as approved under exemption 4 in the HHS regulations (45 CFR Part 46). ABR follows established procedures for written informed parental consent. Zen-Bio conducted basic research in accordance with NIH guidelines and the Federal Provisions Pertaining to Research Use of Human Fetal Tissue by NIH Investigators. Zen-Bio’s research related to human tissues is approved under its Institutional Review Board (IRB) through PearlIRB.RNA samples from human adipose tissue are the generous gifts from Zen-Bio Inc. This work was supported by Singapore NRF fellowship (NRF-2011NRF-NRFF 001-025), Tanoto Initiative in Diabetes Research to L.S., National Medical Research Council’s Cooperative Basic Research Grant (CBRG; NMRC/CBRG/0070/2014 and NMRC/CBRG/0101/2016), Open Fund-Individual Research (OF-IRG) Grant (NMRC/OFIRG/0062/2017), and Ministry of Education (MOE) Tier2 grant (MOE2017-T2-2-009). This work was supported by the RNA Biology Center at CSI Singapore, NUS, from funding by the Singapore Ministry of Education’s Tier 3 grants, grant number MOE2014-T3-1-006. This work was also supported by the Recruitment Program for Young Professionals (C.D.), Zhejiang Key Subject of Medical Science (C.D.), National Natural Science Foundation of China (81700770), and Zhejiang Provincial Natural Science Foundation of China (LY18C060006).The authors declare no competing interests.These authors contribute equally: Chunming Ding, Yen Ching Lim.Electronic supplementary materialSupplementary Information accompanies this paper at 10.1038/s41467-018-03754-3.Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
>>点击索取文献
  • 侧边栏广告 - 模式动物成功案例

在线支持/投诉建议

用户名是必须的

电话号码是必须的

用户 Email 是必须的

请输入有效的 Email.

内容是必须的