Analysis of alternative splicing uncovers a vastly expanded transcriptomic response to hypoxia in human vascular endothelial cells
Hypoxia is a fundamental pathophysiological stimulus that plays important roles in multiple cardiopulmonary diseases. During hypoxic stress, cells adapt by undergoing widespread transcriptional reprogramming. The conventional approach to investigating this response has largely involved RNA-seq analysis which typically quantifies transcriptional output at the level of individual genes. However, as most mammalian genes encode multiple transcripts that are divergently regulated by alternative splicing, consolidating these discrete features to the gene level can obscure critical changes within the transcriptome that are largely unappreciated. To more accurately define the role of individual transcript usage during hypoxia, herein we employed three different analytical strategies that collectively identified thousands of instances of alternative splicing in human endothelial cells undergoing hypoxic stress. Notably, the set of differentially utilized transcripts displayed minimal overlap with the genes identified to be differentially expressed by conventional RNA-seq analysis, indicating divergent usage of individual transcripts does not reliably culminate in a detectable change in overall expression. This outcome is particularly clear at the earliest time-point of hypoxia where we found the transcriptional response was mediated almost exclusively by alternative splicing. A subset of these acutely responsive genes was variably detected as differentially expressed or alternatively spliced at later time points, demonstrating hypoxic transcription is highly dynamic and temporally complex. Further, the number of genes in pathways incriminated in the hypoxic response was also expanded considerably when alternatively spliced transcripts were included in the analysis, suggesting distinct attributes of the hypoxic transcriptome are unveiled by transcript level assays that would otherwise be obfuscated by standard RNA-seq analyses that summarize expression to the gene level. Collectively, these results strongly point to alternative splicing being a significant, albeit understudied, component of the transcriptional response to hypoxic stress which is vastly more complicated than previously thought.