Construction of a cell-based aggregation and seeding model for the Tau protein

Abstract

A pathological hallmark of Alzheimer’s disease (AD), the most common neurodegenerative disease in elderly people, is the formation of neurofibrillary tangles (NFTs), which are mainly composed of bundles of amyloid fibrils formed by abnormal deposition of hyperphosphorylated full-length human Tau protein [1–3]. Recent studies have shown that AD-related cognitive decline and brain atrophy are closely correlated with Tau PET signal, further supporting the link between Tau pathology and AD symptomatology [4,5]. Despite the high incidence and severe burden to patients, caregivers, and health systems caused by AD, there are few disease-modifying therapies available. Normal functional human Tau protein binds to tubulin heterodimers through its microtubule-binding repeats and stabilizes microtubules. Hyperphosphorylated Tau detaches from microtubules and exposes the microtubule-binding domain, thereby leading to Tau self-oligomerization and aggregation [6]. Accumulating evidence suggests that filamentous Tau inclusions form first in a small number of brain cells, from which they are released and taken up by neighboring cells via endocytosis; these filamentous Tau inclusions act as templates for their own replication through monomeric Tau addition and propagate to other regions of the cells [7,8]. Propagation of neuropathology is called prion-like, which refers to the capacity of an abnormally assembled protein to induce the same pathological conformation in the same protein, initiating a self-amplifying cascade. Transcellular propagation and prion-like phenomena are thought to contribute to the progression of pathology in AD, suggesting that inhibiting Tau aggregation and seeding could slow disease progression [7,8]. Accordingly, different experimental aggregation models for the Tau protein have been developed. A cell-based model offers a physiological assay environment with controllable costs and reproducible results, making it the most widely used model for the development of Tau-targeted therapies. In most cellular models, self-assembly of naive monomeric Tau is promoted by the addition of an exogenous ‘seed’ template of synthetic or patient-derived pre-aggregated Tau [9]. Pre-prepared fibrillar seeds of Tau are added to the cell culture medium, taken up by cells, and act as templates to induce the aggregation of monomeric Tau. In addition to homotypic seeding, heterotypic seeding has also been demonstrated for Tau. Direct cross-seeding between pre-aggregated Aβ and Tau is supported by direct binding between Aβ peptides and Tau and direct induction of Tau fibrillization by pre-aggregated Aβ seeds [10]. Human-derived seeds are the most relevant source of pathological Tau protein; however, clinical material is not straightforward to obtain and work with, and it is difficult to guarantee the quality and stability of aggregated seeds. In addition, aggregate seeds could damage the integrity of the cell membrane and lead to cytotoxicity. In the present study, we reported the construction of a cell-based model for the aggregation of endogenous Tau protein in cells without pre-prepared seeds. By introducing an aggregation-driven pathological mutant, ΔK280, to the aggregation-prone truncated core fragment of Tau (Tau244‒372, K18), we constructed a stable cell line over-expressing K18-ΔK280. Over-expressed K18-ΔK280 spontaneously aggregated in SH-SY5Y cells, forming amyloid fibrils positive for thioflavin S (ThS) (Figure 1), a fluorescent dye with β-sheet binding properties, which is widely employed to observe amyloid plaque accumulation [10]. Based on the present cellular model, the properties of Tau aggregation after seeding can be further observed. The aggregates formed by K18-ΔK280 induce co-aggregation and phosphorylation of endogenous Tau in SH-SY5Y cells, which can be recognized by AT8 (phosphorylation at Ser202/Thr205) and pS396 (phosphorylation at Ser396) (Figures 2 and 3) because phosphorylation at Ser202, Thr205, and Ser396 occurred in endogenous Tau but not at K18-ΔK280. This model is easy to use and avoids the potential cytotoxicity caused by fibrillar seeds.

1. Introduction

A pathological hallmark of Alzheimer’s disease (AD), the most common neurodegenerative disease in elderly people, is the formation of neurofibrillary tangles (NFTs), which are mainly composed of bundles of amyloid fibrils formed by abnormal deposition of hyperphosphorylated full-length human Tau protein [1–3]. Recent studies have shown that AD-related cognitive decline and brain atrophy are closely correlated with Tau PET signal, further supporting the link between Tau pathology and AD symptomatology [4,5]. Despite the high incidence and severe burden to patients, caregivers, and health systems caused by AD, there are few disease-modifying therapies available.

Normal functional human Tau protein binds to tubulin heterodimers through its microtubule-binding repeats and stabilizes microtubules. Hyperphosphorylated Tau detaches from microtubules and exposes the microtubule-binding domain, thereby leading to Tau self-oligomerization and aggregation [6]. Accumulating evidence suggests that filamentous Tau inclusions form first in a small number of brain cells, from which they are released and taken up by neighboring cells via endocytosis; these filamentous Tau inclusions act as templates for their own replication through monomeric Tau addition and propagate to other regions of the cells [7,8]. Propagation of neuropathology is called prion-like, which refers to the capacity of an abnormally assembled protein to induce the same pathological conformation in the same protein, initiating a self-amplifying cascade. Transcellular propagation and prion-like phenomena are thought to contribute to the progression of pathology in AD, suggesting that inhibiting Tau aggregation and seeding could slow disease progression [7,8].