Introduction and objectives To improve our understanding of the dynamic interplay between non-coding and coding RNA species, we developed solution-based methods to capture the entire endothelial transcriptome. In contrast to microarrays and CHIP-based methodologies, there was no pre-specification of RNA target sequences, and the identity of the DNA strand of origin was preserved.

Methods Triplicate confluent primary human microvascular endothelial cells (Promocell GmbH) were cultured in the presence and absence of TGF-beta stimuli relevant to pulmonary arteriovenous malformations and hereditary haemorrhagic telangiectasia (HHT). Following rRNA-depletion of total RNA, seven libraries were prepared using Illumina reagents, and 8 pM of each library used for cluster generation and sequencing on Genome Analyser II. Algorithms for aligning reads to NCBI36/hg18 and GRCh37/hg19 included Bowtie, TopHat and Seqmap. Validations were performed using quantitative rt-PCR.

Results More than 2 Gigabases of sequence was generated. Transcriptome-wide profiles were similar between libraries, with sixteen types of RNA species detected including 146 micro (mi)RNA families (47 broadly conserved), and 10,749 protein-encoding mRNAs representing ~5.5% of mapped reads. Alignments to endothelial mRNAs/miRNAs were substantially higher than to gene loci for non-endothelial mRNAs/miRNAs. mRNA exon alignments demonstrated sharp exon boundary delineation, but replicate alignments to non-repetitive intronic regions involved in multi-exon deletions in HHT patients. There was an inverse relationship between alignments depths and qt-PCT cycle thresholds (Ct), where single alignments were detectable, and Ct values of 20 generated by 0.02 nM spiked RNA. Across all experiments in replicate donor/treatment RNAs, for a panel of single open reading frame miRNA genes, RNASeq alignments (gene strand read counts normalised to the total number of valid reads and exon/locus size) explained 72% of the variance of qt-PCR cycle threshold (p < 0.0001). Dynamic whole transcriptome profiling is in progress.

Conclusions These novel directional next generation RNA sequencing methods provide new insights for mutational mechanisms, and a systems approach to dissection of regulatory and target RNA networks relevant to human disease.