Comprehensive accounting of lncRNA dynamics within vascular smooth muscle cell pathological transitions

Abstract

Vascular smooth muscle cells (VSMCs) provide vital contractile force within blood vessel walls, yet also propagate widespread cardiovascular pathologies with high mortality rates through pathological activities. The targeting of such phenotypes in VSMCs has been a commonly-touted strategy for decades yet we still have no viable option to implement this. Recent studies have established that VSMC phenotypes are driven, in part, by the diverse effects of long non-coding RNAs (lncRNAs) on gene expression. This class of largely uncharacterised gene regulators may offer a wealth of novel targets to be used to target VSMCs. However, their characterisation in VSMCs in pathological states is hampered by incomplete lncRNA representation in reference annotation.

In this thesis, we address this by assembling non-reference transcripts in RNA sequencing datasets describing saphenous vein VSMCs stimulated in vitro with cytokines and growth factors or arterial VSMCs stimulated with mechanical stress. We also utilised VSMCs isolated from atherosclerotic plaques. All transcripts were subject to a rigorous lncRNA prediction pipeline to provide an expanded VSMC transcriptome with an unprecedented level of detail on the lncRNAs associated with VSMC pathological states.

We found substantially improved coverage of lncRNAs responding to pro-mitogenic stimuli, with non-reference lncRNAs contributing 21–32% per dataset. We also demonstrate non-reference lncRNAs were biased towards enriched expression within VSMCs, suggesting extra lncRNAs highlighted by our pipeline have particular relevance to VSMC-specific processes. They were also biased towards transcription from enhancer sites suggesting they coordinate the regulation of neighbouring protein-coding genes. Both VSMC-enriched and enhancer-transcribed lncRNAs were large components of lncRNAs responding to pathological stimuli, yet without novel transcript discovery 33–46% of these lncRNAs would remain hidden. In parallel to this analysis, we mined the expanded VSMC annotation to initially explore functionality in a small cohort of uncharacterised lncRNAs within the saphenous vein VSMC in vitro model.

In our final round of analysis, we hypothesised that many lncRNAs may be involved in directing early transcriptional changes leading up to proliferation – and so constitute targets that may be particularly high value through acting upstream of multiple mitogenic or pathogenic pathways. We therefore used our expanded VSMC annotation as a foundation to perform a deeper analysis of lncRNA activity within RNAseq samples obtained from the first 24 hours of stimulation in the saphenous vein VSMC in vitro model, aiming to identify lncRNAs influencing initial transcriptional changes prior to observable cell division.

We noted an enrichment of lncRNA induction – particularly those which were VSMC-enriched or enhancer-transcribed – within an early phase of SVSMC stimuli response prior to proliferation. Transcription factor mRNA dynamics also localised to earlier phases whilst cell cycle mRNAs were overwhelmingly induced after 8 hours. This suggests the involvement of lncRNAs in an early phase of gene regulation sets the VSMC on a path towards later proliferation. To predict lncRNAs with functional impact in the 4 hour regulatory phase, we looked for evidence of their cis-regulation of nearby genes. Genes located near differentially expressed intergenic lncRNAs were 1.51x more likely to be differentially expressed within the four hour window than those located elsewhere in the genome. This effect was also identified for transcription factors and was particularly potent for genes around enhancer-transcribed lncRNAs (2.26x) but weakened when considering over longer time periods (1.16x). Together this suggests a focus of lncRNA-dependent cis-regulation activity in the first four hours after stimulation that could lead to wider downstream impact on VSMC pathological states. We finish by identifying a cohort of uncharacterised lncRNAs regulated in the initial four hour phase that have strong correlations in expression with transcription factors or other genes that explicitly link to vasculoproliferative pathology.

Overall, our comprehensive VSMC lncRNA repertoire provides much needed clarity on the activity of lncRNAs within VSMC pathological states. The approach we outline allows proper prioritisation of candidates for characterisation and exemplifies a strategy to broaden our knowledge of lncRNA across a range of disease states.