Stress-Dependent SUMOylation Protects Neurons from Age Related TDP-43 Toxicity with Implications for ALS/FTD and Aging in the Mammalian Nervous System

Abstract

To maintain longevity, neurons must appropriately respond to and effectively recover from environmental stressors throughout their lifespan. Disruption of these stress response pathways can lead to complex neurodegenerative diseases such as Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Dementia (FTD). The RNA binding protein TDP-43 helps facilitate appropriate stress responses through many essential roles within the nucleus. However, in ALS/FTD and conditions of prolonged stress, TDP-43 can be found aberrantly mislocalized forming large cytoplasmic aggregates causing neurotoxicity. It is critical to determine the molecular mechanisms natively regulating TDP-43 upon stress to uncover pathways that protect neurons from neurodegeneration. Covalent conjugation of SUMOs through SUMOylation has long been characterized to occur in response to cellular stress. Due to the strong overlap between roles of TDP-43 and those regulated by SUMOylation, we hypothesized that TDP-43 is likely a substrate of SUMOylation. We found that TDP-43 is SUMOylated by SUMO2/3 in a stress dependent manner at lysine 408 (K408) and is subsequently cleared through the ubiquitin-proteasome pathway upon recovery. To determine the consequences of blocking SUMOylation of TDP-43, we generated a knock-in mouse line with a TDP-43 K408R point mutation thereby preventing SUMOylation. In primary neuron cultures, we found that blocking TDP-43 SUMOylation impairs the ability for neurons to recover from stress leading to the formation of nuclear TDP-43 foci. We posited that aging may act as a stressor in vivo to elicit stress-dependent phenotypes and observed that TDP-43 K408R mice develop age-related cognitive impairment and motor neuron pathology. In human temporal lobe samples, we found that global SUMOylation is significantly upregulated during aging suggesting potentially important roles for SUMOylation beyond TDP-43 in protecting neurons in elderly individuals. Due to this global response, we question whether other targets of SUMOylation may be involved in protecting different neuron populations during aging. However, uncovering roles for SUMOylation in vivo has been a longstanding challenge due to lack of effective tools to directly compare SUMO paralogs and unbiasedly identify native substrates. Thus, we generated a His6-HA-SUMO2 knock-in mouse to complement a previously established His6-HA-SUMO1 model and demonstrated using whole brain imaging and immunoprecipitation coupled with mass spectrometry that these models facilitate the direct comparison of SUMO1 and SUMO2 distribution and paralog-specific substrates in vivo. These models will facilitate future aging studies to uncover how global SUMOylation and specific substrates maintain neuronal longevity in health and disease.