Aplp1 interacts with Lag3 to facilitate transmission of pathologic α-synuclein

Abstract

Pathologic α-synuclein (α-syn) spreads from cell-to-cell, in part, through binding to the lymphocyte-activation gene 3 (Lag3). Here we report that amyloid β precursor-like protein 1 (Aplp1) interacts with Lag3 that facilitates the binding, internalization, transmission, and toxicity of pathologic α-syn. Deletion of both Aplp1 and Lag3 eliminates the loss of dopaminergic neurons and the accompanying behavioral deficits induced by α-syn preformed fibrils (PFF). Anti-Lag3 prevents the internalization of α-syn PFF by disrupting the interaction of Aplp1 and Lag3, and blocks the neurodegeneration induced by α-syn PFF in vivo. The identification of Aplp1 and the interplay with Lag3 for α-syn PFF induced pathology deepens our insight about molecular mechanisms of cell-to-cell transmission of pathologic α-syn and provides additional targets for therapeutic strategies aimed at preventing neurodegeneration in Parkinson's disease and related α-synucleinopathies.

Fig. 1. α-Syn PFF binds to Aplp1.

a Schematic diagram of Aplp1 deletions mutants: the ΔE1, ΔAcD+E2, ΔJMR (juxtamembrane region), ΔTM (transmembrane region), ΔICD (intracellular domain), two subdomain deletion mutants of ectodomain E1 (ΔGFLD and ΔCuBD), and the deletion mutant (Δ7aa) and the substitutional mutant (7aa-A × 7) of a seven-amino acid (7aa) motif. b Quantification of binding signals of deletion mutants of Aplp1 with α-syn-biotin PFF by normalization with the expression of indicated deletion mutants of Aplp1. (p-values: FL vs. ΔE1 < 0.0001, FL vs. ΔGFLD 0.0267, FL vs. ΔCuBD 0.2771, FL vs. 7aa-A × 7 < 0.0001, FL vs. Δ7aa 0.0001, FL vs. ΔAcD + E2 0.5265, FL vs. ΔJMR < 0.0001, FL vs. ΔICD 0.0076, FL vs. GFP < 0.0001, FL vs. None <0.0001). c, d Schematic diagram of Aplp1 chimeras. Quantification of binding signals of Aplp1(E1)-Aplp2 and Aplp1(E1)-App chimeras with α-syn-biotin PFF by normalization with the expression of indicated Aplp1 chimeras. (p-values: Aplp1 vs. Aplp1(E1)-Aplp2 < 0.0001, Aplp1 vs. Aplp1(E1)-App <0.0001, Aplp1 vs. Aplp2 < 0.0001, Aplp1 vs. App <0.0001). e Schematic diagram of Lag3 and deletions mutants. f Quantification of binding signals of deletion mutants of Lag3 with α-syn-biotin PFF by normalization with the expression of indicated deletion mutants of Lag3. (p-values: FL vs. ΔD1 < 0.0001, FL vs. Δ7aa < 0.0001, FL vs. Δ7aa-A × 7 < 0.0001, FL vs. ΔICD < 0.0001, FL vs. None <0.0001, FL vs. GFP < 0.0001) b, d, f ***P < 0.001, n.s. not significant. Data are the means ± SEM, from 3 individual experiments, one-way ANOVA followed by Dunnett’s correction. Source data are provided as a Source Data file.

Fig. 2. Aplp1 mediates the endocytosis of α-syn PFF and subsequent pathology.

a Quantification of α-syn-pHrodo PFF endocytosis in WT and Aplp1^–/– neurons control lentivirus (WT: 38 cells, Aplp1^−/−: 40 cells) and Aplp1-lentivirus transduction (WT+Aplp1: 29 cells, Aplp1^−/− + Aplp1: 35 cells). n = 4 (WT, Aplp1^–/–, Aplp1^−/− + Aplp1); n = 3 (WT + Aplp1) independent experiments. Two-way ANOVA with Tukey’s correction. F (DFn, DFd): F (Time x Group) = F (18, 66) = 7.777; F (Time) = F (6, 66) = 143.2; F (Group) = F (3, 11) 46.15. b, c Immunoblot and quantification of α-syn-biotin PFF in the endolysosome fraction. n = 3 independent experiments. One-way ANOVA. d Quantification of the co-localization of α-syn-biotin PFF with Rab7 in WT and Aplp1^–/– neurons. WT (38 cells), WT + Aplp1 (12 cells), Aplp1^–/– (49 cells), Aplp1^–/– + Aplp1 (13 cells) from 3 independent experiments. (p-values: WT vs. WT + Aplp1 0.0059, WT vs. Aplp1^−/− 0.0230, Aplp1^−/− vs. Aplp1^−/− + Aplp1 0.0024). e Quantification of co-localization of α-syn-biotin PFF and Rab7 in WT neurons transiently expressing full-length Aplp1 (FL: 25 cells) and deletion mutants: ΔE1(10 cells), ΔAcD + E2 (12 cells), ΔICD (15 cells), App (5 cells) and Aplp2 (5 cells), and control (16 cells) from 3 independent experiments. (p-values: FL vs. Control <0.0001, FL vs. ΔE1 < 0.0001, FL vs. ΔAcD + E2 0.0178, FL vs. ΔICD 0.0831, FL vs. App < 0.0001, FL vs. Aplp2 < 0.0001). f Quantification of co-localization signal of α-syn-biotin PFF in WT neurons transiently expressing Aplp1, chimeric Aplp1(E1)-Aplp2 and Aplp1(E1)-App. n = 3 independent experiments. g Immunostaining of anti-pS129 in WT and Aplp1^–/– neurons expressing Aplp1 and control lentivirus, with α-syn PFF and PBS administration. Scale bar, 100 μm. h Quantification of (g). n = 3 (WT); n = 4 (WT + Aplp, Aplp1^–/–, Aplp1^−/− + Aplp1) independent experiments. i–k, Immunoblots for insoluble α-syn, pS129, soluble α-syn, and β-actin in WT and Aplp1^–/– neuronal lysates post 15 days α-syn PFF treatment. n = 3 independent experiments. d, e One-way ANOVA with Dunnett’s correction. f, h, j, k One-way ANOVA with Tukey’s correction; a, c, d, e, f, h, j, k *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant, Data are as means ± SEM. Source data are provided as a Source Data file.

Fig. 3. Aplp1 and Lag3 bind to each other.

a Lag3 pulls down Aplp1 by anti-Lag3 410C9 immunoprecipitation in WT mouse brain lysates, but not in Lag3^–/– lysates. b Aplp1 pulls down Lag3 by anti-Aplp1 CT11 immunoprecipitation in WT mouse brain lysates, but not in Aplp1^–/– lysates. c, d Mapping of the Lag3-binding domains in Aplp1. HEK293FT cells were transfected with full-length (FL) or deletion mutants of FLAG-Aplp1, and Myc-Lag3 for co-IP experiments. The GFLD subdomain in the E1 domain of Aplp1 is the major subdomain responsible for the Lag3 interaction. e Mapping of the Aplp1-binding domains in Lag3. HEK293FT cells were transfected with FL, deletion mutants of Myc-Lag3, and FLAG-Aplp1 for co-IP experiments. Experiments in (a–e) were repeated three times independently with similar results. f–h Identification of the interface of A1E1 (E1 domain of APLP1) binding to L3D2 (D2 domain of LAG3). f Overlay of the 2D ¹H-¹⁵N HSQC spectra of A1E1 alone (gray) and in the presence of 2 molar folds of L3D2 (blue). Four residues with significant calculated chemical shift deviations (CSDs) (> 0.03 ppm) are highlighted and enlarged in the black boxes. g Histogram of the CSDs of A1E1 in the presence of L3D2 at a molar ratio of 1:2 (A1E1/L3D2). The domain organization of A1E1 is indicated on the top, with blue boxes indicating the β-strands and the red box indicating the α-helix. A dashed line was drawn to highlight the residues with CSDs >0.01 ppm. h The 37 residues with large CSDs (>0.01 ppm) upon L3D2 titration are highlighted in blue on the ribbon diagram of the A1E1 modeled structure. i Validation of the NMR results. We substituted nine residues of APLP1 with alanine to generate FLAG-APLP1(mut9) and performed the co-IP experiment to assess the APLP1-Lag3 interaction. n = 3 independent experiments. Data are the means ± SEM, Two-tailed Student’s t-test; p-value = 0.0009. j The scheme for the interaction among Aplp1, Lag3, and α-syn PFF. Source data are provided as a Source Data file.

Fig. 4. The role of Aplp1-Lag3 interaction in mediating the binding, endocytosis of α-syn PFF, and subsequent pathology.

a Aplp1 and Lag3 account for more than 40% α-syn-biotin PFF binding to cortical neurons analyzed in supplementary Fig. 6a. Data are the means ± SD. b Co-expressing Aplp1 and Lag3 in Aplp1^–/–/Lag3^–/– cortical neurons substantially increases binding that is greater than the sum of α-syn-biotin PFF binding to Aplp1 and Lag3. Refer to supplementary Fig. 6b. (p-values: +Aplp1 vs. +Aplp1 + Lag3 0.0001, +Lag3 vs. +Aplp1 + Lag3 < 0.0001). c, d Immunoblots of endosomal fractions from WT, Aplp1^–/–, Lag3^–/–, and Aplp1^–/–/Lag3^–/– cortical neurons. Aplp1^–/–/Lag3^–/– significantly decreases (~70%) internalized α-syn-biotin PFF compared to WT, Aplp1^–/–, Lag3^–/–, neurons. (p-values: WT vs. Aplp1^−/− 0.0019, WT vs. Lag3^−/− 0.0039, WT vs. Aplp1^−/−/Lag3^−/− <0.0001, Aplp1^−/− vs. Aplp1^−/−/Lag3^−/− 0.0208, Lag3^−/− vs. Aplp1^−/−/Lag3^−/− 0.0091) e, APLP1 and Lag3 co-expression by transfection increases α-syn-biotin uptake in SH-SY5Y cells compared to APLP1(mut9) expression. The biotin signal was normalized using the intensity of Lag3. Lag3 (15 cells), Lag3 + Aplp1 (14 cells), Lag3 + Aplp1(mut9) (11 cells). (p-values: Lag3 vs. Lag3 + Aplp1 < 0.0001, Lag3 vs. Lag3 + Aplp1(mut9) 0.1720, Lag3 + Aplp1 vs. Lag3+Aplp1(mut9) 0.0003). f, g Anti-Lag3 410C9 significantly disrupts the co-IP of FLAG-Aplp1 and Myc-Lag3 in HEK293FT cells. (p-values: PBS vs. mIgG 0.1304, mIgG vs. 410C9 0.0159). h, i Anti-Lag3 410C9 (330 nM) reduced the internalization of α-syn-biotin PFF (1 μM) in WT neurons compared to Lag3^−/− neurons. Quantification of the intensity of internalized α-syn-biotin PFF normalized by Rab7. (p-values: WT vs. Lag3^−/− < 0.0057, mIgG vs. 410C9 < 0.0277, 410C9 vs. Lag3^−/− 0.0152). j–l Both the levels of the endosomal Aplp1 and Lag3 were significantly decreased by 410C9 in WT cortical neuron cultures. n = 3 independent experiments. Two-tailed Student’s t-test; p-value = 0.0019 (k), 0.0019 (l). For all experiments, n = 3 independent experiments. For graphs, b, d, e, g, i, k, l Data are the means ± SEM; b, d, e, g, i One-way ANOVA with Tukey’s correction; *P < 0.05, **P < 0.01, ***P < 0.001; n.s. not significant. Source data are provided as a Source Data file.

Fig. 5. The role of Aplp1-Lag3 interaction in α-syn pathology propagation, transmission, and neurotoxicity in vitro.

a, b Aplp1-Lag3 double deletion significantly decreased (70%) pS129 immunostaining induced by α-syn PFF, compared to Aplp1^–/–, Lag3^–/–, or WT neurons. Scale bar, 20 μm. (p-values: WT vs. Aplp1^−/− 0.0016, WT vs. Lag3^−/− < 0.0001, WT vs. Aplp1^−/−/Lag3^−/− <0.0001, Aplp1^−/− vs. Lag3^−/− 0.0296, Aplp1^−/− vs. Aplp1^−/−/Lag3^−/− < 0.0001, Lag3^−/− vs. Aplp1^−/−/Lag3^−/− 0.0004). c, d Immunostaining of anti-pS129 in Aplp1^–/–/Lag3^–/– neurons, treated with ^αα-syn PFF, transduced with Aplp1, Lag3, or Aplp1 + Lag3. Scale bar, 20 μm. (p-values: +Aplp1 vs. +Aplp1 + Lag3 < 0.0001, +Lag3 vs. +Aplp1 + Lag3 < 0.0001). e Schematic of a microfluidic device with three chambers. α-Syn transmission from chamber 1 (C1) to chamber 2 (C2), to chamber 3 (C3) 14 days α-syn PFF treatment in C1. The different combinations of neurons tested in C2, listed as C1-(C2)-C3, are WT-(WT)-WT, WT-(Aplp1^–/–)-WT, WT-(Lag3^–/–)-WT, WT-(Aplp1^–/–/Lag3^–/–)-WT. f–i Immunostaining images and quantification of pS129 signals in the transmission. Scale bar, 100 μm. (P-values: Chamber 2(h), WT-WT-WT vs. WT-Aplp1^−/−-WT < 0.0001, WT-WT-WT vs. WT-Lag3^−/−-WT < 0.0001, WT-WT-WT vs. WT-Aplp1^−/−/Lag3^−/−-WT < 0.0001, WT-Aplp1^−/−-WT vs. WT-Aplp1^−/−/Lag3^−/−-WT 0.0032, WT-Lag3^−/−-WT vs. WT-Aplp1^−/−/Lag3^−/−-WT 0.0427; Chamber 3(i) WT-WT-WT vs. WT-Aplp1^−/−-WT 0.0111, WT-WT-WT vs. WT-Lag3^−/−-WT 0.0141, WT-WT-WT vs. WT-Aplp1^−/−/Lag3^−/−-WT 0.0154). j, k PI/Hoechst staining for cell death in WT, Aplp1^–/–, Lag3^–/–, and Aplp1^–/–/Lag3^–/– cortical neuron cultures, treated with α-syn PFF. Scale bar, 100 μm. Aplp1-Lag3 deletion exhibited significantly less cell death than in Aplp1^–/–, Lag3^–/–, or WT cultures, treated with α-syn PFF. (p-values: PBS WT vs. PFF WT < 0.0001, PFF group: WT vs. Aplp1^−/− < 0.0001, WT vs. Lag3^−/− < 0.0001, WT vs. Aplp1^−/−/Lag3^−/− < 0.0001, Aplp1^−/− vs. Aplp1^−/−/Lag3^−/− <0.0001, Lag3^−/− vs. Aplp1^−/−/Lag3^−/− < 0.0001). For all experiments, n = 3 independent experiments. For graphs, Data are the means ± SEM. One-way ANOVA with Turkey’s correction. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant. Source data are provided as a Source Data file.

Fig. 6. Roles of Aplp1 and the Aplp1-Lag3 interaction in mediating α-syn PFF-induced neurodegeneration in vivo.

a Representative TH (tyrosine hydroxylase) immunohistochemistry and Nissl staining images of dopamine (DA) neurons in the SNpc of α-syn PFF-injected hemisphere in the WT, Aplp1^–/–, and Aplp1^–/–/Lag3^–/– mice. b Stereological counting of TH- and Nissl-positive neurons in the substantia nigra after 6 months of α-syn PFF injection (WT: n = 7; Aplp1^–/–: n = 7; Aplp1^–/–/Lag3^–/–: n = 7). c DA concentrations in the striatum of PBS and α-syn PFF-injected mice measured at 180 days by HPLC. n = 8 mice per group, two-way ANOVA with Sidak’s correction. d, e Representative images, and quantification of pS129 positive inclusions in the substantia nigra of WT, Aplp1^–/–, and Aplp1^–/–/Lag3^–/– mice. (n = 6, each group). f Distribution of LB/LN-like pathology in the brain sections of α-syn PFF-injected WT, Aplp1^–/–, and Aplp1^–/–/Lag3^–/– mice (pS129 positive neuron, red dots; pS129 positive neurites, red lines). g, h Representative images of LB/LN-like pathology (the black box in h) and the quantification of pS129 intensity (green) from each coronal section (1–4) stained with pS129. (n = 6, each group). i, j Assessments of the behavioral deficits measured by the pole test (WT: n = 13; Aplp1^–/–: n = 13; Aplp1^–/–/Lag3^–/–: n = 7) and cylinder test (n = 10, each group). (p-values: for pole test (i), PBS WT vs. PFF WT < 0.0001, PFF WT vs. PFF Aplp1^−/− < 0.0010, PFF WT vs. PFF Aplp1^−/−Lag3^−/− < 0.0001, PFF Aplp1^−/− vs. PFF Aplp1^−/−Lag3^−/− < 0.0312; For cylinder test (j): PBS WT vs. PFF WT < 0.0001, PFF WT vs. PFF Aplp1^−/− 0.0004, PFF WT vs. PFF Aplp1^−/−/Lag3^−/− < 0.0001, PFF Aplp1^−/− vs. PFF Aplp1^−/−/Lag3^−/− 0.0304). For graphs, e, h One-way ANOVA with Tukey’s correction; b, j Two-way ANOVA with Tukey’s correction; Data are the means ± SEM, *P < 0.05. **P < 0.01. ***P < 0.001. n.s. not significant. Source data are provided as a Source Data file.

Fig. 7. The role of anti-Lag3 in blocking α-syn PFF-induced neurodegeneration in vivo.

a Representative images of TH and Nissl staining of SNpc DA neurons of WT mice treated with mIgG or 410C9 at 6 months after intrastriatal injection of α-syn PFF. b Stereology counts of data are the means ± SEM, (mIgG-PBS: n = 5; 410C9-PBS: n = 5; mIgG-PFF: n = 4; 410C9-PFF: n = 5). One-way ANOVA with Turkey’s correction. (p-values: for Nissl: PBS mIgG vs. PFF mIgG 0.0022, PFF mIgG vs. PFF 410C9 0.0301, for TH: PBS mIgG vs. PFF mIgG <0.0001, PFF mIgG vs. PFF 410C9 0.0311) *P < 0.05, **P < 0.01, ***P < 0.001. c DA concentrations in the striatum of α-syn PFF or PBS-injected mice treated with 410C9 or mIgG measured at 6 months by HPLC-ECD (high-performance liquid chromatography-electrochemical detection). (mIgG-PBS: n = 4; 410C9-PBS: n = 4; mIgG-PFF: n = 3; 410C9-PFF: n = 4), Data are the means ± SEM, Two-way ANOVA with Tukey’s correction; (p-values: PBS mIgG vs. PFF mIgG 0.0050, PFF mIgG vs. PFF 410C9 0.0111). d–g Behavioral deficits were ameliorated by 410C9. Six months after α-syn PFF injection, the grip strength test and pole test were performed. Behavioral abnormalities in the grip strength test and pole test induced by α-syn PFF were ameliorated by 410C9 treatment (one day after stereotaxic injection). (mIgG-PBS: n = 8; 410C9-PBS: n = 8; mIgG-PFF: n = 8; 410C9-PFF: n = 9). Data are the means ± SEM, one-way ANOVA with Dunnett’s correction; (p-values: d PBS mIgG vs. PBS 410C9 0.9970 PBS, mIgG vs. PFF mIgG 0.0001, PFF mIgG vs. PFF 410C9 0.0001; e PBS mIgG vs. PBS 410C9 0.9655 PBS, mIgG vs. PFF mIgG 0.0001, PFF mIgG vs. PFF 410C9 0.0001; f PBS mIgG vs. PBS 410C9 0.9986 PBS, mIgG vs. PFF mIgG 0.0089, PFF mIgG vs. PFF 410C9 0.0179; g PBS mIgG vs. PBS 410C9 0.5869 PBS, mIgG vs. PFF mIgG 0.0057, PFF mIgG vs. PFF 410C9 0.0026). *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file.