Abstract
Somatodendritic missorting of the axonal protein TAU is a hallmark of Alzheimer's disease and related tauopathies. Rodent primary neurons and iPSC-derived neurons are used for studying mechanisms of neuronal polarity, including TAU trafficking. However, these models are expensive, time-consuming, and/or require the killing of animals. In this study, we tested four differentiation procedures to generate mature neuron cultures from human SH-SY5Y neuroblastoma cells and assessed the TAU sorting capacity. We show that SH-SY5Y-derived neurons, differentiated with sequential RA/BDNF treatment, are suitable for investigating axonal TAU sorting. These human neurons show pronounced neuronal polarity, axodendritic outgrowth, expression of the neuronal maturation markers TAU and MAP2, and, importantly, efficient axonal sorting of endogenous and transfected human wild-type TAU, similar to mouse primary neurons. We demonstrate that the N-terminal half of TAU is not sufficient for axonal targeting, as a C-terminus-lacking construct (N-term-TAUHA) is not axonally enriched in both neuronal cell models. Importantly, SH-SY5Y-derived neurons do not show the formation of a classical axon initial segment (AIS), indicated by the lack of ankyrin G (ANKG) and tripartite motif-containing protein 46 (TRIM46) at the proximal axon, which suggests that successful axonal TAU sorting is independent of classical AIS formation. Taken together, our results provide evidence that (i) SH-SY5Y-derived neurons are a valuable human neuronal cell model for studying TAU sorting readily accessible at low cost and without animal need, and that (ii) efficient axonal TAU targeting is independent of ANKG or TRIM46 enrichment at the proximal axon in these neurons.