Aging is a major risk factor for impaired cardiovascular health. The aging heart is characterized by vascular dysfunction, increased hypertrophy, fibrosis and electrophysiological alterations. The mechanism of age-associated pathophysiological alterations are incompletely understood. Previous studies showed impaired endothelial cell functions associated with senescence in aging. Since vessels align with nerves and this interplay is critical for tissue homeostasis in other organs, we investigated whether an impairment of the neuro-vascular interface in the aging heart may contribute to age-associated pathologies.To investigate the innervation of the aging mouse heart, we assessed the cardiac autonomic nervous system histologically in 12-week and 20-month old mice using the pan-neuronal marker Tuj1 and the sympathetic marker tyrosine hydroxylase. Interestingly, sympathetic innervation of the left ventricle, especially around arteries, was reduced by 0.66±0.07-fold with aging (n=9; p<0.05) indicating that aging reduces innervation in the left ventricle. Similar findings were observed in G4 telomerase-deficient mice, which develop a pre-mature senescent phenotype (reduced innervation by 0.48±0.09-fold, p<0.01, n=4). Electrophysiological studies confirmed an increased incidence of ventricular arrhythmias in old versus young hearts.

To determine a potential contribution of neurovascular cellular cross-talk, we analyzed the gene expression of isolated endothelial cells in the aging mouse heart by bulk RNA sequencing. Aging significantly induced genes associated with GO-terms related to neuronal cell death and axon injury. Sema3a, a known axon repelling factor, was significantly increased by 1.5±0.2-fold in endothelial cells of aged mice hearts, a finding which was confirmed by qPCR of isolated endothelial cells and by histology (p<0.05).

Mechanistically, expression of miR145, which targets Sema3a, was significantly reduced in old hearts, while miR145 knockout mice demonstrated increased vascular Sema3a expression paralleled by reduced innervation. Sema3a expression was significantly increased in senescent endothelial cells in vitro, thus suggesting a putative role for senescence-induced alterations of gene expression in impaired cardiac innervation. Indeed, we observed an increase in acidic β-galactosidase activity as a marker for cellular senescence in the aging heart, in particular around arteries. Importantly, in miR-145-deficient mice, senescence was exclusively detected around the arteries. Thus, we hypothesized that eliminating senescent cells may restore left ventricular axon density. Therefore, we treated 18-month old mice with a combination of the two senolytic drugs dasatinib and quercetin (n=5), which are known to deplete senescent cells in vivo. Indeed, after 2 months of senolytic treatment, cellular senescence was profoundly reduced by 0.4±0.1-fold (p<0.05) in the aging heart in parallel with a reduced expression of endothelial Sema3a and other repulsive genes. Finally, ventricular innervation was restored (by 1.8±0.4-fold), and   heart rate variability was normalized in senolytics compared to placebo treated mice