The cholesterol 24-hydroxylase CYP46A1 promotes α-synuclein pathology in Parkinson's disease

Abstract

Parkinson's disease (PD) is a neurodegenerative disease characterized by the death of dopaminergic neurons in the substantia nigra and the formation of Lewy bodies that are composed of aggregated α-synuclein (α-Syn). However, the factors that regulate α-Syn pathology and nigrostriatal dopaminergic degeneration remain poorly understood. Previous studies demonstrate cholesterol 24-hydroxylase (CYP46A1) increases the risk for PD. Moreover, 24-hydroxycholesterol (24-OHC), a brain-specific oxysterol that is catalyzed by CYP46A1, is elevated in the cerebrospinal fluid of PD patients. Herein, we show that the levels of CYP46A1 and 24-OHC are elevated in PD patients and increase with age in a mouse model. Overexpression of CYP46A1 intensifies α-Syn pathology, whereas genetic removal of CYP46A1 attenuates α-Syn neurotoxicity and nigrostriatal dopaminergic degeneration in the brain. Moreover, supplementation with exogenous 24-OHC exacerbates the mitochondrial dysfunction induced by α-Syn fibrils. Intracerebral injection of 24-OHC enhances the spread of α-Syn pathology and dopaminergic neurodegeneration via elevated X-box binding protein 1 (XBP1) and lymphocyte-activation gene 3 (LAG3) levels. Thus, elevated CYP46A1 and 24-OHC promote neurotoxicity and the spread of α-Syn via the XBP1-LAG3 axis. Strategies aimed at inhibiting the CYP46A1-24-OHC axis and LAG3 could hold promise as disease-modifying therapies for PD.

Fig 1. CYP46A1 and 24-OHC are up-regulated in PD patients and PD model mice.

(A) Immunohistochemistry showing the levels of CYP46A1 in the SNpc of PD patients (n = 5 subjects in each group). Scale bar, 20 μm. (B) Levels of 24-OHC in the plasma of PD patients (n = 38) and control subjects (n = 19) determined by LC–MS. Student’s t test with Welch’s correction. **P < 0.01. (C) Representative immunohistochemistry images of CYP46A1 in the cortex and SN of α-Syn A53T transgenic mice at different ages. Scale bar, 20 μm. Bar graphs are the optical density (n = 5 mice per group). (D) Immunoblotting showing CYP46A1 levels in the striatal lysates of α-Syn A53T transgenic mice at different ages. GAPDH was used as the loading control (n = 6 mice per group). (E) LC–MS analysis of plasma 24-OHC levels in α-Syn A53T transgenic mice at different ages (n = 4–5 mice per group). (F) LC–MS analysis of 24-OHC levels in the striatal lysates of α-Syn A53T transgenic mice at different ages (n = 4–5 mice per group). All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. *P < 0.05, ****P < 0.001, ****P < 0.0001, and ns, not significant. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. PD, Parkinson’s disease; SN, substantia nigra; SNpc, substantia nigra pars compacta.

Fig 2. α-Syn pathology and its spread are significantly reduced after CYP46A1 removal in vivo.

WT, CYP46A1^+/−, and CYP46A1^−/− mice were injected with α-Syn PFFs (5 μg). (A) Ipsilateral pS129 immunohistochemistry of various brain regions at 180 dpi. Scale bar, 20 μm. STR: striatum. SN: substantia nigra. DMV: dorsal motor nucleus of the vagus nerve. (B) Representative images and quantification of TH-positive cells in the ipsilateral and contralateral SNpc. Scale bar, 100 μm. n = 6 mice per group. (C) Representative images and quantification of TH-positive cells in the ipsilateral and contralateral striatum. Scale bar, 100 μm. n = 6 mice per group. (D) Immunoblots and quantification of pS129 and TH in the ipsilateral striatum. n = 6 mice per group. (E) The contents of striatal DA and DOPAC in the ipsilateral striatum were measured by HPLC. n = 3–6 mice per group. (F) Behavioral tests, including the rotarod test, wire hang test, pole test, and balance beam test. n = 10 mice per group. All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001. ns, not significant. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. dpi, days postinjection; HPLC, high-performance liquid chromatography; PFF, preformed fibril; SNpc, substantia nigra pars compacta; WT, wild type.

Fig 3. Elevated CYP46A1 promotes α-Syn aggregation by 24-OHC in vitro.

(A, B) α-Syn-GFP HEK293 cells were transfected with His-CYP46A1 and then treated with α-Syn PFFs for 48 h. PBS was used as the transduction control for PFFs, while the His vector was used as a transfection control for His-CYP46A1. (A) Triton X-100 (1%) in PBS was used to permeabilize soluble proteins to observe insoluble α-Syn inclusions. Images of pS129 (red) and insoluble α-Syn inclusions (green). n = 6 independent experiments. AU, arbitrary units. (B) Western blot analysis of α-Syn, CYP46A1, and pS129. Scale bar, 20 μm. n = 6 independent experiments. (C, D) Primary neurons were infected with AAV-CYP46A1 or control AAVs and then treated with α-Syn PFFs for 9 days. pS129 (red) and MAP 2 (green) staining. Quantitative analysis of pS129. n = 6 independent experiments. (E, F) Primary neurons from WT or CYP46A1 KO mice were transduced with α-Syn PFFs for 9 days. Shown are MAP2 (green) and pS129 (red). Quantitative analysis of pS129. n = 6 independent experiments. (G) Kinetics of α-Syn fibrillation (2 mg/ml monomer) in the presence or absence of 30 μm 24-OHC or 25-OHC in the real-time ThT fluorescence assay (n = 6 independent experiments). AFU, arbitrary fluorescence unit. (H) Ultrastructural images of α-Syn PFFs and 24-OHC PFFs revealed by TEM. Scale bars, 200 nm. (I, J) Primary neurons were treated with α-Syn PFFs and 24-OHC PFFs for 9 days. (I) Double immunofluorescence of pS129 (red) and MAP2 (green) in primary neuron. Scale bar, 20 μm. (J) Cell viability of neurons was detected by Cell Counting Kit-8 (CCK8) after treatment with 24-OHC or 24-OHC PFFs (n = 6 independent experiments). All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. ****P < 0.0001. ns, not significant. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. PFF, preformed fibril; TEM, transmission electron microscopy; WT, wild type.

Fig 4. The seeding activity of 24-OHC PFFs was greater than that of α-Syn PFFs in vivo.

Three-month-old WT mice received a single intrastriatal injection of α-Syn PFFs or 24-OHC PFFs (5 μg). (A) Ipsilateral pS129 immunohistochemistry of various brain regions at 30 dpi. Scale bar, 20 μm. STR: striatum. SN: substantia nigra. DMV: dorsal motor nucleus of the vagus nerve. (B) Ipsilateral pS129 immunohistochemistry of various brain regions at 90 dpi. Scale bar, 20 μm. (C) Ipsilateral pS129 immunohistochemistry of various brain regions at 180 dpi. Scale bar, 20 μm. (D) Heatmap of pS129 pathology in mice at 30, 90, and 180 dpi. (E) Co-localization of pS129 (red) and TH (green) in the SNpc at 180 dpi. Scale bar, 20 μm. (F) Representative immunohistochemistry images and quantification of TH-positive neurons in the ipsilateral and contralateral SNpc at 180 dpi. Scale bar, 100 μm. n = 6 mice per group. (G) Immunoblots and quantification of pS129 and TH in the ipsilateral striatum at 180 dpi. n = 6 mice per group. (H) Behavioral tests, including the balance test, grip test, pole test, and rotarod test, were performed at 180 dpi. n = 12 mice per group. All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test (n = 6 independent experiments). ****P < 0.0001. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. dpi, days postinjection; PFF, preformed fibril; SNpc, substantia nigra pars compacta; WT, wild type.

Fig 5. 24-OHC facilitates the propagation of α-Syn pathology and mitochondrial dysfunction.

(A, B) α-Syn-GFP HEK293 cells were transduced with α-Syn PFFs in the presence or absence of 24-OHC (30 μm) for 48 h. (A) Triton X-100 (1%) in PBS was used to permeabilize soluble proteins to observe insoluble α-Syn inclusions. Shown are the representative immunostaining and quantification of the insoluble α-Syn inclusions (green) and pS129 (red). AU, arbitrary units. n = 6 independent experiments. (B) Immunoblots of phosphorylated α-Syn (pS129) and α-Syn. n = 5 independent experiments. (C) Neurons were exposed to α-Syn PFFs in the presence or absence of 24-OHC (30 μm) for 9 days. Shown are pS129 (red) and MAP2 (green). AU, arbitrary unit. n = 5 independent experiments. (D) α-Syn-GFP HEK293 cells were transduced with α-Syn PFFs in the presence or absence of 24-OHC (30 μm) for 48 h. Representative images showing the co-localization of mitochondria (MitoTracker, red) and α-Syn inclusions (green). Scale bar, 8 μm. n = 6 independent experiments. (E) SH-SY5Y cells were transduced with α-Syn PFFs in the presence or absence of 24-OHC for 48 h. TMRM staining of the mitochondrial membrane potential (red) and Fluo-4 Am (green) intracellular calcium concentration. Scale bar, 20 μm. n = 6 independent experiments. All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001, and ns, not significant. AU, arbitrary unit. (F) SH-SY5Y cells were transfected with Mitodendra2 and then transduced with α-Syn PFFs in the presence or absence of 24-OHC (30 μm) for 48 h. Cells were live-imagined using a confocal microscope. A subpopulation of mitochondria within a single cell was subjected to full-photon shifting by exposing the region of interest to 405-nm fluorescence (10% power) for 2 s. Representative images showing mitochondrial fusion. Red: shifted mitochondria; Green: original mitochondria; Yellow: fused mitochondria. Scale bar, 8 μm. The bar graph shows the quantification of mitochondrial fusion. One-way ANOVA with Tukey’s multiple comparisons test. n = 6 independent experiments. PFF + 24-OHC group vs. PFF + vehicle group: ^&P < 0.05 and ^&&&P < 0.001. PFFs + 24-OHC group vs. PBS + vehicle group: ***P < 0.001 and ****P < 0.0001. PFFs + vehicle group vs. PBS + vehicle group: ^#P < 0.05 and ^####P < 0.05. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. PFF, preformed fibril.

Fig 6. Intracerebroventricular injection of 24-OHC aggravates neurodegeneration.

(A) Schematic representation of the mouse experiments. (B) Schematic of the injection site. The mice were injected with α-Syn PFFs and then intracerebroventricularly injected with 24-OHC (100 μg/kg body weight). (C) Ipsilateral pS129 in various brain regions. Scale bar, 20 μm. STR: striatum. SN: substantia nigra. DMV: dorsal motor nucleus of the vagus nerve. (D) Representative immunohistochemistry images showing ipsilateral pS129-positive LB-like accumulation after PK digestion. Scale bar, 20 μm. (E) Representative immunohistochemistry and quantification of TH-positive cells in the ipsilateral and contralateral SNpc. Scale bar, 100 μm. n = 6 mice per group. (F) Immunoblots and quantification of pS129 in the ipsilateral striatum. n = 6 mice per group. (G) Behavioral tests, including the rotarod test, wire hang test, pole test, and balance beam test. n = 10 mice per group. All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001, and ns, not significant. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. PFF, preformed fibril; SNpc, substantia nigra pars compacta.

Fig 7. 24-OHC activates the XBP1/LAG3 axis both in vivo and in vitro.

(A) Immunoblots and quantification of XBP1 and LAG3 in the striatal tissues of WT mice injected with 24-OHC (100 μg/kg body weight) and α-Syn PFFs at 180 dpi (as in Fig 6). n = 6 mice per group. (B) Immunoblots and quantification of XBP1 and LAG3 in the striatum of CYP46A1 knockout mice injected with α-Syn PFFs at 180 dpi (as in Fig 2). n = 6 mice per group. (C) Representative images showing LAG3 (green) and pS129 (red) in the substantial nigra of mice injected with 24-OHC and PFFs (as in Fig 6). n = 6 mice per group. Scale bar, 20 μm. AU, arbitrary unit. (D) Immunoblots and quantification of LAG3 and XBP1 in α-Syn-GFP HEK293 cells transduced with α-Syn PFFs in the presence of 24-OHC (30 μm) for 48 h. n = 6 independent experiments. (E) Effect of the XBP1 inhibitor toyocamycin on the expression of LAG3 in α-Syn-GFP HEK293 cells treated with α-Syn PFFs and 24-OHC (30 μm). n = 6 independent experiments. **P < 0.01, ****P < 0.0001, and ns, not significant. Underlying data can be found in S1 Data. The uncropped blots are included in S1 Raw Images. dpi, days postinjection; LAG3, lymphocyte-activation gene 3; PFF, preformed fibril; WT, wild type; XBP1, X-box binding protein 1.

Fig 8. XBP1 and LAG3 mediate the effect of 24-OHC on the cell-to-cell transmission of α-syn.

(A) Representative images of insoluble inclusions and pS129 in α-Syn-GFP HEK293 cells transduced with α-Syn PFFs in the presence or absence of 24-OHC (30 μm) and toyocamycin. n = 6 independent experiments. (B) Representative images of insoluble inclusions and pS129 in α-Syn-GFP HEK293 cells transduced with α-Syn PFFs in the presence of LAG3 antibody. n = 6 independent experiments. (C) Schematic representation of chambers in which neurons were cultured in chamber 1 (C1) and chamber 2 (C2). (D) Transmission of pathologic p129 from C1 to C2 at 14 days after treatment with α-syn PFFs and 24-OHC in combination with toyocamycin or an anti-LAG3 antibody. n = 6 independent experiments. All data are means ± SEM. One-way ANOVA with Tukey’s multiple comparisons test. n = 6 independent experiments. ns, not significant, *P < 0.05, *P < 0.01 ***P < 0.001, and ****P < 0.0001. Scale bar, 20 μm. AU, arbitrary unit. Underlying data can be found in S1 Data. LAG3, lymphocyte-activation gene 3; PFF, preformed fibril; XBP1, X-box binding protein 1.

Fig 9. Schematic representation of CYP46A1 and 24-OHC in α-Syn pathology.

During aging, overactivated CYP461 catalyzes cholesterol to produce 24-OHC, which promotes XBP1s splicing to produce the XBP1u fragment. XBP1u is incorporated into the nucleus and increases LAG3 expression. The membrane protein LAG3 binds to α-Syn PFFs and promotes the cell-to-cell transmission of α-Syn seeds. In addition, 24-OHC facilitates α-Syn assembly, resulting in the formation of morphologically different 24-OHC PFFs, which are more neurotoxic than pure α-Syn PFFs. LAG3, lymphocyte-activation gene 3; PFF, preformed fibril; XBP1, X-box binding protein 1.