Comprehensive tissue specificity analysis identifies (novel) retina-specific, long non-coding RNAs

Abstract

Purpose : Long non-coding RNAs (lncRNAs) are fascinating molecules that regulate gene expression at the right time and place. Thus far, little is known about the role of lncRNAs in the human retina and inherited retinal disease (IRD). Here, an extensive bioinformatic study was performed to identify both known and novel lncRNAs that are exclusively or predominantly expressed in human retina, followed by validation of lncRNAs potentially regulating IRD genes.

Methods : By re-analysis of RNA expression data of adult post-mortem retina (n=152) and 53 tissues from the GTEx project (n=7,294), a specificity analysis was performed using the Jensen-Shannon divergence (JSD) score. Next, transcripts were filtered on location within the topologically associating domains (TADs) of 218 IRD disease genes. For a selection of lncRNAs, the coding potential was predicted using in silico tools (PLEK, CPAT, and CPC2). Transcript structure and expression pattern were determined by direct cDNA long-read sequencing of human retina (ONT), RT-qPCR, re-analysis of single-cell RNAseq data, and single-molecule RNA in situ hybridization (RNAscope).

Results : In total, 785 lncRNA genes were located in IRD gene TADs, of which 25 displayed transcript(s) enriched in retina (JSD>0.35, TPM>1). To identify novel loci, filtering was narrowed to 30 kb up- and downstream of the IRD gene, resulting in 37 loci enriched in retina (JSD>0.35, TPM>1).
Four annotated and one novel lncRNA were selected for validation. In silico predictions indicated a high non-coding potential for all lncRNAs. Deciphering the transcript structure using long-read RNAseq revealed novel isoforms for four lncRNAs, which were validated by RT-qPCR. Remarkably, one annotated lncRNA, located in the TAD of NR2E3 (ENSG00000260037), appeared to be a readthrough of NR2E3, which was confirmed by Sanger sequencing. Single-cell RNAseq data pointed to photoreceptor-specific expression of three of the remaining lncRNAs, with rod-photoreceptor specificity for two. These expression patterns were confirmed for two lncRNAs using RNAscope (remaining two ongoing).

Conclusions : This extensive tissue-specificity analysis generated a comprehensive catalogue of transcripts predominantly expressed in adult human retina, and allowed the identification and validation of both known and novel retina-specific lncRNAs located in TADs of IRD disease genes.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.