SORL1 is a receptor for tau that promotes tau seeding

Abstract

Sortilin-related receptor 1 (SORL1) is an intracellular sorting receptor genetically implicated in Alzheimer's disease (AD) that impacts amyloid precursor protein trafficking. The objective of these studies was to test the hypothesis that SORL1 binds tau, modulates its cellular trafficking and impacts the aggregation of cytoplasmic tau induced by pathological forms of tau. Using surface plasmon resonance measurements, we observed high-affinity binding of tau to SORL1 and the vacuolar protein sorting 10 domain of SORL1. Interestingly, unlike LDL receptor-related protein 1, SORL1 binds tau at both pH 7.4 and pH 5.5, revealing its ability to bind tau at endosomal pH. Immunofluorescence studies confirmed that exogenously added tau colocalized with SORL1 in H4 neuroglioma cells, while overexpression of SORL1 in LDL receptor-related protein 1-deficient Chinese hamster ovary (CHO) cells resulted in a marked increase in the internalization of tau, indicating that SORL1 can bind and mediate the internalization of monomeric forms of tau. We further demonstrated that SORL1 mediates tau seeding when tau RD P301S FRET biosensor cells expressing SORL1 were incubated with high molecular weight forms of tau isolated from the brains of patients with AD. Seeding in H4 neuroglioma cells is significantly reduced when SORL1 is knocked down with siRNA. Finally, we demonstrate that the N1358S mutant of SORL1 significantly increases tau seeding when compared to WT SORL1, identifying for the first time a potential mechanism that connects this specific SORL1 mutation to Alzheimer's disease. Together, these studies identify SORL1 as a receptor that contributes to trafficking and seeding of pathogenic tau.

Keywords: SORL1; apolipoprotein E (ApoE); lipoprotein receptor-related protein (LRP); neurodegenerative disease; tau protein (tau); tau seeding; tauopathy.

Figure 1

Recombinant monomeric tau binds to SORL1 and the VPS10 domain of SORL1.A, equilibrium analysis of the binding of increasing concentrations of recombinant 2N4R tau to full-length SORL1 (blue circles) and the SORL1 VPS10 domain (orange triangles) coupled to a Biacore CM5 sensor chip. b and c, recombinant SORL1 (B) or VPS10 (C) was immobilized on the chip and binding of 25 nM tau ± 500 nM VPS10 was assessed. Shown are representative images. D, binding of tau isoforms 2N4R, 0N3R, 2N3R, and tau MBD to SORL1 VPS10 domain assessed by SPR equilibrium analysis. E and F, the binding of tau produced by Sf9 cells along with two mutant forms of tau to SORL1 VPS10 domain (E) or SORL1 (F) was measured by SPR; 6A (T181, S199, S202, S396, S400, and S404 are all converted to alanine) and 6E, in which all these residues are converted to glutamic acid. G and H, binding of mutant forms of tau to SORL1 VPS10 domain (G) or SORL1 (H): 3XKQ in which lysine residues 311, 317, and 321 were converted to glutamine residues and 9XKQ tau in which lysine residues 311, 217, 321, 340, 343, 347, 353, 369, and 375 are all converted to glutamine residues. The data were normalized to Rmax to correct for slight differences in coupling to the SPR surfaces. Each experiment was repeated at least three times. Shown are mean ± SEM. MBD, microtubule-binding domain; SORL1, sortilin-related receptor 1; SPR, surface plasmon resonance; VPS10, vacuolar protein sorting 10.

Figure 2

SORL1 mediates the internalization of tau and colocalizes with internalized tau. WT CHO and LRP1-deficient CHO 13-5-1 cells were transfected with SORL1 plasmid or empty vector (Mock). A, transfection efficiency was validated via immunoblot analysis. B, transfected cells were incubated with 20 nM ¹²⁵I-labeled tau ± 1 μM RAP for 2 h, and the amount of tau internalized by the cells was quantified. (Mean ± SEM, two-way ANOVA followed by Tukey multiple comparisons test). C, H4 cells were transfected with SORL1-GFP plasmid, then incubated with 40 nM tau labeled with Alexa Flour 594 for 2 h. Cells were fixed and imaged on a Nikon spinning disk confocal microscope at 60×. Cells expressing SORL1 (teal) internalize more tau (magenta) than neighboring cells not expressing SORL1-GFP (nuclei marked by DAPI staining, blue). Tau and SORL1 colocalization is represented in white in the merged image. The scale bar represents 10 μm. D, single cell analysis of the percent of tau that colocalizes with SORL1-GFP using Imaris Bitplane Software. Shown is mean ± SD, n = 4. CHO, Chinese hamster ovary; DAPI, 4′,6-diamidino-2-phenylindole; LDL receptor-related protein 1; RAP, receptor associated protein; LRP1; SORL1, sortilin-related receptor 1.

Figure 3

SORL1 knockdown does not impact tau uptake in H4 cells.A, time course for ¹²⁵I-tau internalization in H4 cells. Cells were incubated with 20 nM ¹²⁵I-labeled tau ± 1 μM RAP for designated times and internalized tau measured. B and C, H4 cells were treated with vehicle control, nonsense siRNA, or siRNA targeting SORL1, then incubated with 20 nM ¹²⁵I-labeled tau ± 1 μM RAP or 300 μg/ml R2629 for 2 h. B, SORL1 knockdown confirmed by immunoblot analysis. C, internalized tau was quantified. Shown are mean ± SEM; 2-way ANOVA, Tukey’s multiple comparison test, n = 3. RAP, receptor associated protein; SORL1, sortilin-related receptor 1.

Figure 4

Tau colocalizes with endogenous SORL1 in H4 cells.A, H4 cells were incubated with 40 nM tau labeled with Alexa Flour 594 for 2 h, then fixed and immunostained with anti-SORL1 antibody to label endogenous SORL1. The scale bar represents 10 μm. (B) Inset from panel (A) zoomed in to 200%, scale bar is 2 μm. All cells were fixed and imaged on Nikon spinning disk confocal microscope at 60×. Imaris Bitplane software was used to identify colocalization (white) between SORL1 (teal) and tau (magenta). Nuclei marked by DAPI staining, blue. C, single-cell analysis of the percent of tau that colocalizes with SORL1. Shown is mean ± SD, n = 7. DAPI, 4′,6-diamidino-2-phenylindole; SORL1, sortilin-related receptor 1.

Figure 5

Tau binds SORL1 at pH 7.4 as well as pH 5.5. SPR analysis of tau binding LRP1 or SORL1 at pH 7.4 or pH 5.5. Increasing concentrations of tau (11, 33, 100, 300, and 900 nM were titrated on CM5 sensor chips coated with LRP1 (A and B) or SORL1 (C and D). The experiment was performed at pH 7.4 (blue lines) or pH 5.5 (orange lines). B and C, equilibrium fits of the data in panels a and c. Shown are mean ± SEM, n = 3. LRP1, LDL receptor-related protein 1; SORL1, sortilin-related receptor 1; SPR, surface plasmon resonance.

Figure 6

SORL1 transfection reconstitutes pathogenic internalization and seeding in HEK293T reporter cells. HEK293T Tau RD P301S FRET Biosensor cells were transfected with SORL1, then incubated with (A) human brain homogenate from an Alzheimer’s patient (AD) or age-match healthy control (Ctrl) fraction or (B) HMW SEC fractions from AD patient brain (AD) or healthy control (Ctrl) and tau seeding was quantified. C, siRNA was used to knockdown SORL1 in H4 cells and was confirmed by immunoblot analysis, SORL1 band marked. Lane 1, nontransfected, lane 2, transfected with pcDNA SORL1, lane 3, control siRNA, and lane 4, SORL1 siRNA. D, H4 cells stably expressing the FRET reporter system with SORL1 knockdown were incubated with 300 ng/well AD brain derived HMW tau seeding material and tau seeding was quantified. (Mean ± SEM; 2-way ANOVA, Tukey’s multiple comparison test). SEC, size exclusion chromatography; HMW, high molecular weight; SORL1, sortilin-related receptor 1.

Figure 7

SORL1 harboring the N1358S mutation exhibits increased seeding capacity in HEK293T cells. HEK293T Tau RD P301S FRET Biosensor cells were transfected with plasmid containing WT SORL1 or SORL1 harboring the G511R or N1358S mutations. A, expression level of different variants was measured by Western blot (GAPDH was used as a loading control). These cells were incubated with (B) HMW or LMW SEC fractions from human AD patient brain or vehicle control and tau seeding was quantified. In a separate experiment, transfected cells expressing equal levels of the SORL1 variants (C) were incubated with (D) 20 nM ¹²⁵I-labeled tau in the presence or absence of 1 μM RAP for 2 h and then internalized tau was quantified. (Mean ± SEM; 2-way ANOVA, Tukey’s multiple comparison test). AD, Alzheimer’s disease; HMW, high molecular weight; LMW, low molecular weight; RAP, receptor associated protein; SORL1, sortilin-related receptor 1.

Figure 8

Model of LRP1 and SORL1-mediated tau trafficking. LRP1 mediates the internalization of monomeric forms of tau, resulting in rapid and efficient tau degradation. Internalized tau colocalizes with SORL1, in compartments that remain to be identified, and likely impacts tau trafficking through the endosomal pathways. Tau associates with SORL1 at endosomal pH of 5.5, but not LRP1, indicating a possible hand-off of tau from LRP1 to SORL1 as it traffics through endocytic pathways. LRP1 and SORL1 both mediate tau seeding in response to tau from human AD patient brains, and which could be through direct or indirect mechanisms. We postulate that AD pathogenic tau escapes the major degradative pathway to reach the extracellular space where it can induce cytoplasmic seeding. Furthermore, the AD-associated N1358S mutation in SORL1 increases the propensity of tau to seed aggregation, likely through impacting tau intracellular trafficking. Created with BioRender.com. AD, Alzheimer’s disease; LRP1, LDL receptor-related protein 1; SORL1, sortilin-related receptor 1.