Loss of gene-to-gene transcriptional coordination is a general attribute of aging

  Guy Amit  ,  Orr Levy [1]  ,  Dana Vaknin [1]  ,  Tom Snir [2]  ,  Sol Efroni [2]  ,  Peter Castaldi [3]  ,  Yang-Yu Liu [3,4]  ,  Haim Y. Cohen [2]  ,  Amir Bashan [1]  
[1] Department of Physics, Bar-Ilan University, Ramat-Gan 52900, Israel
[2] The Mina and Everard Goodman Faculty of Life Science, Bar Ilan University, Ramat-Gan 52900, Israel
[3] Channing Division of Network Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts
[4] Center for Cancer Systems Biology, Dana-Farber Cancer Institute, Boston, Massachusetts, 02115, USA

Living cells function is determined primarily by a complex network of interactions among the genes, called the gene regulatory network. An increasing number of observations have shown that stochasticity plays a major role in the process of aging. The accumulation of random damage, such as genetic and epigenetic mutations, may adversely affect the gene regulatory networks of individual cells and eventually lead to declined tissue function. To date, however, scRNA-seq studies from diverse species and tissues have not found a general and consistent attribute that can reflect the stochastic aspects of aging in gene regulation, casting doubt on the generality of the accumulation of damage paradigm. Here, we have devised a novel approach that focuses on the interactions between the genes rather than the expression levels of individual genes. We have developed a top-down computational analysis tool that quantifies the “coordination level” between genes, avoiding the shortcomings of traditional methods for network inference or coexpression analysis. Consistently, across very different species: fruit fly, mouse and human, and very different cell types: hematopoietic stem cells (HSCs), immune cells, pancreatic cells, neurons and glial cells, we find a decrease in gene-to-gene coordination in cells from old organisms compared with those from young organisms. In addition, we find that high mutational load in human pancreatic cells from the same subject is strongly correlated with loss of gene-to-gene coordination, drawing an aging-independent link between random genetic damage and system-wide alterations of gene regulation. We also find a list of aging-susceptible KEGG pathways in mouse HSCs for which gene-to-gene coordination is significantly reduced during aging. These observations suggest that the general, potentially universal, transcriptional consequences of stochastic damage during aging is a disruption of the interactions between genes.