Antisense oligonucleotide-mediated perturbation of long non-coding RNA reveals functional features in stem cells and across cell types

Abstract

Long non-coding RNAs (lncRNAs) were shown to regulate pluripotency and differentiation, however, a comprehensive assessment of their regulatory roles in stem cells is currently lacking. By performing a large-scale knockdown of lncRNAs (n = 390) in human induced pluripotent stem (iPS) cells, we systematically characterized their roles in self-renewal and pluripotency. Cell growth and transcriptomic assays revealed respectively, 18.3% and 29.3% of these lncRNAs in altering cellular and molecular phenotypes upon knockdown. By integrating the molecular phenotypes with chromatin-interaction assays, we identified their cis- and trans- interacting partners as potential primary targets. In addition, cell-type enrichment revealed 16 lncRNAs associated with pluripotency and we highlighted LINC02595, CATG00000090305.1 and RP11-148B6.2. Next, we integrated fibroblasts phenotyping data (Ramilowski et al., 2020; Genome Research) and reported that only 2.9% lncRNAs exhibited consistent cell growth phenotype, while molecular phenotype showed 66.7% agreement. This highlights their consistent primary functional roles across cell-types and the distinct secondary effects under different cellular environments.

Functional identification of cis-regulatory long noncoding RNAs at controlled false-discovery rates

“…The usefulness of these methods is further supported by their internal and external consistency. Together, the TransCistor approaches correctly identify previously-described cis-activators H19 (33), RP11-398K22.12 (20,23), JPX (34), Evx1os (35), LINC01615 (36) and DA125942 (37) amongst the top-ranked cis-activators, while both independent experiments for XIST are amongst the top repressors (38).…”

“…Similar to previous reports, we observed a strong cis-regulation of nearby histone genes (Figure 2I, Supplementary Figure S2A). (20,23). Therefore, TransCistor correctly identifies previously-reported cis-lncRNAs.…”

“…The FANTOM perturbation datasets were downloaded from the Core FANTOM6 repository (20,23). The differential expression results were transformed into regulation files by applying an adjusted p-value threshold of 0.05 and using custom bash scripts.…”

“…We further define targets as activated or repressed, where they decrease / increase in response to lncRNA loss-of-function (LOF), respectively. Overall we collected 488 lncRNA LOF experiments targeting 268 human lncRNAs from a mixture of sources, including the recently published datasets of ASO knockdowns in human dermal fibroblasts and induced pluripotent stem cells (iPSCs) from the FANTOM consortium (20,23). To this we added 140 experiments for 134 lncRNAs from mouse (Figure 1B).…”