Exploring the transcriptional hallmarks of aging using single-cell RNA sequencing

Abstract

Aging is associated with a series of phenotypic changes at the cellular level, including changes in gene expression. In recent years, aged tissues have been reported to present an increase in transcriptional noise or cell-to-cell variability. In order to test this, a new method (Scallop) for the quantification of transcriptional noise from single-cell RNA sequencing (scRNA-seq) data has been developed, which is based on cluster membership. In addition, a compilation of methods used in the literature has been implemented in the Python toolkit Decibel. A systematic analysis of published scRNA-seq datasets of murine and human aged tissues using Scallop and Decibel revealed that increased transcriptional noise is not a universal hallmark of aging. A deeper analysis of four scRNA-seq datasets of the murine aging lung showed that, even though there are no consistent cell type-specific changes in transcriptional noise levels, the aging lung is characterized by a reproducible enrichment in immune cell types, which was also seen in two human datasets. Interestingly, an exploratory analysis of gene expression datasets of murine and human aging revealed a previously undetected hallmark of aging: aged tissues present a generalized length-dependent transcriptional decay. This length-dependent underexpression (LDU) was also observed in premature aging models caused by exposure to environmental genotoxicity (UV radiation and tobacco smoke) and in two progeroid syndromes caused by defects in the transcription-coupled nucleotide excision repair pathway (TC-NER): Cockayne syndrome and trichothiodystrophy. These results suggest that accumulation of genotoxicity in long genes could lead to reduced RNA polymerase II procesivity.