From Axon Damage to Disease: Common Pathways in Neurodegeneration

DLK is an essential player in the axonal response to neuronal injury. The gene and others that act within the same pathway have independently come up in unbiased screens for regulators of neuronal regeneration and degeneration in different species. It is known that DLK can promote axon degeneration, neuronal cell death, and regeneration. However, the exact mechanisms by which it acts within neurons is incompletely understood. To identify regulators and substrates of DLK function, we study DLK localization, trafficking, and interactors in human iPSC (induced pluripotent stem cells)–derived neurons. We are also elucidating mechanisms of axon degeneration in human neurons [Reference 1], as well as downstream mechanisms of DLK signaling by studying the function of genes we identified as hits from CRISPR inhibition survival screens. We found that DLK can initiate an apoptotic cascade of mitochondrial fission in the axon after injury. Using live confocal imaging of human axons, we found that, upon an injury to the axon, DLK sets off a wave of apoptotic mitochondrial fission that travels from the damage site back to the cell body, leading to axon degeneration and neuronal cell death. The retrograde mitochondrial fission requires DLK–dependent phosphorylation of the GTPase dynamin–related peptide 1 DRP1 [Reference 2].

We and others have previously shown that axon damage causes transcriptional reprogramming of peripheral sensory neurons, including the upregulation of injury markers such as Atf3 (activating transcription factor 3). We used a mouse generated in my lab (Atf3-IRES-Cre) (IRES, internal ribosome entry site; Cre, a DNA recombinase) to gain genetic access to injured neurons. In a study published earlier this year, we examined whether neurons of the central nervous system (CNS) exhibit transcriptional reprogramming after traumatic injury. In this work, we characterized a model of concussion and investigated whether transcriptomic plasticity also occurs in the CNS. We found that cortical neurons do not undergo the broad transcriptional reprogramming observed in the peripheral nervous system, but that Atf3 is upregulated in a population of neurons, a subset of which undergoes DLK–dependent death [Reference 3].

It has long been known that specific subpopulations of motor neurons are more vulnerable than others to disease. We profile individual transcriptomes of spinal cord motor neurons in healthy mice and disease models to track the transcriptional changes occurring in these cells during disease progression. In 2021, we published a single-cell transcriptomic atlas of adult mouse spinal motor neurons [Reference 4]. Previously, very few spinal motor neurons (MNs) had been resolved at the single-cell level, both because they are relatively rare among all spinal cells, and because they do not survive single-cell isolation protocols. We comprehensively defined all subtypes of adult mouse MNs. The data can be browsed at http://www.spinalcordatlas.org and also at https://seqseek.ninds.nih.gov. We are now applying this method to interrogate transcriptional changes that arise in motor neurons during disease progression. Comparing differential expression between populations of resilient and vulnerable neurons, we are identifying candidate transcription factors that can promote resilience, and testing these in vivo.

In a recent collaboration, we provided in vivo proof of principle that it is possible to specifically target motor neurons that are undergoing pathological change, using a gene-therapy approach that remains silent in healthy cells [Reference 5]. This strategy takes advantage of the unique biology of TDP-43, a splicing repressor of many genes, which becomes dysregulated in a subset of neurons in amyotrophic lateral sclerosis (ALS) and in other neurodegenerative disorders. As the mis-localization of TDP-43 from the nucleus and concurrent aggregation in the cytoplasm are a conserved pathological hallmark of many cases neurodegenerative disease (virtually all cases of ALS, 50% of frontotemporal dementia, and subsets of Parkinson's and Alzheimer's disease), this approach could be broadly applicable across neurodegenerative conditions.