Age-dependent neurodegeneration and organelle transport deficiencies in mutant TDP43 patient-derived neurons are independent of TDP43 aggregation

Highlights

• TDP43 mutations do not affect motor neuron differentiation capacitiy

• TDP43 mutations increase neuron death without signs of protein aggregation during aging

• TDP43 mutations induce age-related disturbance of distal axonal organelle morphology including a dramatic loss of organelle motility

• TDP43 mutant neurons show dramatic axo-skeletal degeneration but lack obvious signs of nuclear stress

• D-Sorbitol rescues mutant motor neuron organelle morphology and motility without affecting TDP43 aggregation

Abstract

TAR DNA-binding protein 43 (TDP43) plays a significant role in familiar and sporadic amyotrophic lateral sclerosis (ALS). The diverse postulated mechanisms by which TDP43 mutations cause the disease are not fully understood. Human wildtype and TDP43 S393L and G294V mutant spinal motor neuron cultures were differentiated from patient-derived iPSCs. Mutant hTDP43 and wildtype motor neuron cultures did not differ in neuron differentiation capacity during early maturation stage. During aging we detected a dramatic neurodegeneration including neuron loss and pathological neurofilament abnormalities only in TDP43 mutant cultures. Additionally mitochondria and lysosomes of aging spinal motor neurons revealed robust TDP43 mutation dependent abnormal phenotypes in size, shape, speed and motility which all appeared without TDP43 mislocalization or aggregation formation. Furthermore, D-sorbitol – known to induce stress granules and cytoplasmic mislocalization of TDP43 – rescued axonal trafficking phenotypes without signs of TDP43 mislocalization or aggregation formation. Our data indicate TDP43 mutation-dependent but cytosolic aggregation-independent mechanisms of motor neuron degeneration in TDP43 ALS.

Introduction

Worldwide 2–4 of 100.000 people per year suffer from amyotrophic lateral sclerosis (ALS) which is the most frequent variant of motor neuron disease with a survival time of 1–5 years after symptom onset. Motor neurons are the mainly affected cell type that undergoes degeneration and death in ALS (Ludolph & Sperfeld, 2005; Johnston et al., 2006). Several gene mutations were found to cause familiar forms of ALS (fALS) of which the first identified were SOD1 (superoxide dismutase 1) gene mutations in 1993 (Rosen et al., 1993; Robberecht & Philips, 2013). In 2006 mutations in TDP43 (TAR DNA binding protein, TARDBP) were identified to cause fALS as well as sporadic ALS but the pathophysiology caused by mutant TDP43 is still not fully understood (Neumann et al., 2006). Most importantly, TDP43 is the main aggregating protein in sporadic ALS also being aggregated in some fALS forms (e.g. TDP43, C9ORF72) (Scotter et al., 2015). However, TDP43-mediated pathophysiology appears to be mechanistically distinct from mutant SOD1 action (Mackenzie et al., 2007).

Different mechanisms for TDP43 mutant pathology have been suggested. These include nuclear loss of function leading to transcription and splicing defects, toxic properties in the cytoplasm sequestering RNAs and RNA binding proteins and finally loss of axonal transport function with RNA granule transport deficiencies (Jovicic & Gitler, 2014). The main focus has been drawn on pathological accumulation of cytosolic TDP43 (Han et al., 2013; Shan et al., 2010; Zhang et al., 2013). Mutant TDP43 human induced pluripotent stem cell (hiPSC)-derived neurons were reported to show elevated levels of soluble and detergent- resistant TDP43 protein and decreased survival in long-term differentiation (Bilican et al., 2012). Such lines showed a two-fold increase in cytosolic TDP43 compared to the controls which could be – at least partially – reduced by 30% by allele-specific knockdown (Nishimura et al., 2014). However, on a closer look, also in models reporting cytosolic mislocalization of TDP43, only up to one third of cells showed this phenotype while the remaining cells still showed physiological nuclear localization of TDP43 protein (Shan et al., 2010; Zhang et al., 2013; Fallini et al., 2012).

TDP43 was shown to be actively transported within the axon in both directions (Fallini et al., 2012). Thereby TDP43 forms cytoplasmic mRNP granules that undergo bidirectional, microtubule-dependent transport in neurons and facilitate the delivery of target mRNA to distal neuronal compartments (Alami et al., 2014). Mutations of TDP43 interfere with its mRNA transport function (Fallini et al., 2012; Alami et al., 2014). Neurofilament light chain (NEFL) mRNA was shown to be transported in such TDP43 granules and the amount of NEFL mRNA was reduced in TDP43 mutant motor neurons (Shan et al., 2010; Alami et al., 2014).

ALS-linked mutations in TDP43 can impair stress granule dynamics (Liu-Yesucevitz et al., 2010). TDP43 had previously been shown to associate with multiple proteins that are part of mRNP granules (e.g., staufen, FMRP, SMN, and HuD), including neuronal transport granules (Freibaum et al., 2010). TDP43 is known to be directed to stress granules by sorbitol, a novel physiological osmotic and oxidative stressor (Dewey et al., 2011). In addition, TDP43 was found to regulate FOXO-dependent protein quality control in stress response (Zhang et al., 2014).

Mutations in TDP43 led to abnormal (Han et al., 2013) or reduced neurite outgrowth (Duan et al., 2011; Wächter et al., 2015). TDP43 mutant mice show lack of mitochondria in motor axon terminals (Shan et al., 2010). Furthermore, TDP43 was reported to impair mitochondria and lysosome morphology and function in cell models of ALS. Interestingly, both overexpression and knockdown caused perturbations in primary murine neurons suggesting a tight regulation of TDP43 (Wang et al., 2013; Xia et al., 2016). Overexpression of wildtype or mutant TDP43 caused mitochondrial shortening in dendrites, not in axons, due to increased fission and decreased mitochondrial movement in axons and dendrites. In contrast, knockdown increased mitochondrial length but also decreased dendritic and axonal mitochondrial motility. Mutations in TDP43 caused co-localization with mitochondria, which was rescued by Mfn2 expression.

However, the relation between (stress induced) cytoplasmic TDP43 aggregation and neuronal dysfunction and degeneration remains enigmatic. Furthermore, data on organelle trafficking and neurite morphology in human patient-derived motor neurons are still lacking.

This prompted us to use patient-derived induced pluripotent stem cells (hiPSCs)-derived motor neurons to investigate (i) the sequential appearance of neuronal dysfunction and neurodegeneration, (ii) aggregate formation and (iii) how these events are mechanistically connected. Surprisingly, we found significant neuronal dysfunction and degeneration during cellular aging, but did not detect relevant signs of pathological TDP43 aggregation, arguing against an upstream function of TDP43 aggregation in TDP43-ALS.