Cofilin hyperphosphorylation triggers TDP-43 pathology in sporadic amyotrophic lateral sclerosis

Pathological forms of TAR-binding protein 43 (TDP-43), involving its aberrant mislocalization to the cytoplasm, inclusion formation, hyperphosphorylation and fragmentation, are present in ∼45-50% frontotemporal dementia (FTD) and Alzheimer's disease individuals, and most (97%) amyotrophic lateral sclerosis (ALS) cases. Hence, identifying mechanisms that induce TDP-43 pathology are central to neurodegeneration and developing new therapeutic targets in these conditions. Cofilin is a multi-functional protein with a crucial role in regulating the actin cytoskeleton. Actin has important neuronal-specific activities in dendritic spines, axonal growth cones and synapses and it is in constant equilibrium between two forms: monomeric globular actin (G-actin) and polymeric filamentous actin (F-actin). Cofilin controls actin dynamics by depolymerising and severing actin filaments. When cofilin is phosphorylated (at Serine-3) by LIM kinase1 (LIMK1), it becomes inactive, leading to production of more F-actin. Defects in cofilin are well described in other neurodegenerative disorders, unlike in ALS. We examined phosphorylation of cofilin and actin dynamics in post-mortem spinal cord tissue from sporadic ALS (SALS) patients, the TDP-43 rNLS8 transgenic mouse model, and NSC34 motor neuronal cells expressing cytoplasmic TDP-43. F-actin was pharmacologically stabilized to mimic cofilin hyperphosphorylation, and TDP-43 pathology was assessed. Neuronal cells were treated with a non-phosphorylatable cofilin S3A peptide (MAAGVAVSDGVIKVFN), and TDP-43 pathology and apoptosis were evaluated. Here, we show that cofilin is hyper-phosphorylated in human ALS and disease models compared to controls. This was detected in spinal motor neurons from sporadic ALS (SALS) patients and a TDP-43 mouse model (rNLS8) displaying key ALS phenotypes, and in motor neuronal NSC34-cells expressing cytoplasmic TDP-43. Supporting this observation, more F-actin relative to G-actin was present in cortical/spinal cord lysates from SALS patients and TDP-43 rNLS8 mice, and NSC34-cells expressing TDP-43. We also show that mimicking cofilin hyperphosphorylation by pharmacological stabilization of F-actin induced TDP-43 pathology: cytoplasmic mislocalization, inclusion formation, hyperphosphorylation, and fragmentation, and promoted its recruitment into stress granules (SGs). Furthermore, we detected increased levels of LIMK1 phosphorylation and tropomyosin isoforms 4.1 and 4.2 in SALS patients. These findings reveal aberrant cofilin hyperphosphorylation disrupts actin dynamics, triggering TDP-43 pathology and SG recruitment in SALS. They imply that preventing cofilin phosphorylation is a novel therapeutic strategy applicable to most ALS cases. Treatment of neuronal cells with the S3A peptide prevented features of TDP-43 pathology and apoptosis compared to control peptides. These findings thus describe a novel pathogenic mechanism producing TDP-43 pathology, applicable to most ALS cases and other neurodegenerative diseases.