Heterozygous transcriptional and nonsense decay signatures in blood outgrowth endothelial cells from patients with hereditary haemorrhagic telangiectasia

In order to identify cellular phenotypes resulting from nonsense (gain of stop/premature termination codon) variants, we devised a framework of analytic methods that minimised confounder contributions, and applied to blood outgrowth endothelial cells (BOECs) derived from controls and patients with heterozygous nonsense variants in ACVRL1 , ENG or SMAD4 causing hereditary haemorrhagic telangiectasia (HHT). Following validation of 48 pre-selected genes by single cell qRT-PCR, discovery RNASeq ranked HHT-differential alignments of 16,807 Ensembl transcripts. Consistent gene ontology (GO) processes enriched compared to randomly-selected gene lists included bone morphogenetic protein, transforming growth factor-β and angiogenesis GO processes already implicated in HHT, further validating methodologies. Additional terms/genes including for endoplasmic reticulum stress could be attributed to a generic process of inefficient nonsense mediated decay (NMD). NMD efficiency ranged from 78-92% (mean 87%) in different BOEC cultures, with misprocessed mutant protein production confirmed by pulse chase experiments. Genes in HHT-specific and generic nonsense decay (ND) lists displayed differing expression profiles in normal endothelial cells exposed to an additional stress of exogenous 10μmol/L iron which acutely upregulates multiple mRNAs: Despite differing donors and endothelial cell types, >50% of iron-induced variability could be explained by the magnitude of transcript downregulation in HHT BOECs with less efficient NMD. The Genotype Tissue Expression (GTEx) Project indicated ND list genes were usually most highly expressed in non-endothelial tissues. However, across 5 major tissues, although 18/486 nonsense and frameshift variants in highly expressed genes were captured in GTEx, none were sufficiently prevalent to obtain genome-wide significant p values for expression quantitative trait loci (GnomAD allele frequencies <0.0005). In conclusion, RNASeq analytics of rare genotype-selected, patient-derived endothelial cells facilitated identification of natural disease-specific and more generic transcriptional signatures. Future studies should evaluate wider relevance and whether injury from external agents is augmented in cells with already high burdens of defective protein production. ### Competing Interest Statement The authors have declared no competing interest.