Prosaposin maintains lipid homeostasis in dopamine neurons and counteracts experimental parkinsonism in rodents

Abstract

Prosaposin (PSAP) modulates glycosphingolipid metabolism and variants have been linked to Parkinson's disease (PD). Here, we find altered PSAP levels in the plasma, CSF and post-mortem brain of PD patients. Altered plasma and CSF PSAP levels correlate with PD-related motor impairments. Dopaminergic PSAP-deficient (cPSAP^DAT) mice display hypolocomotion and depression/anxiety-like symptoms with mildly impaired dopaminergic neurotransmission, while serotonergic PSAP-deficient (cPSAP^SERT) mice behave normally. Spatial lipidomics revealed an accumulation of highly unsaturated and shortened lipids and reduction of sphingolipids throughout the brains of cPSAP^DAT mice. The overexpression of α-synuclein via AAV lead to more severe dopaminergic degeneration and higher p-Ser129 α-synuclein levels in cPSAP^DAT mice compared to WT mice. Overexpression of PSAP via AAV and encapsulated cell biodelivery protected against 6-OHDA and α-synuclein toxicity in wild-type rodents. Thus, these findings suggest PSAP may maintain dopaminergic lipid homeostasis, which is dysregulated in PD, and counteract experimental parkinsonism.

Fig. 1. PSAP and PGRN in PD patients are divergently regulated and associated with different PD symptoms.

A Representative PSAP and PGRN immunofluorescent staining in postmortem substantia nigra sections from four PD patients and four controls. BF bright field, RN red nucleus, CP cerebral peduncle, NM neuromelanin. Scale bars, 100 μm. B, C Quantification of mean immunofluorescence intensity (MFI) of PSAP and PGRN staining in TH positive neurons (n = 959 and 430 neurons from four controls and four PD patients, respectively). Data are presented as mean ± S.E.M. Student’s t-test (B), or Welch’s t-test (C) is applied. Non-significant p value is not labeled, ****p < 0.0001. D, E Scatter plots representing the associations of CSF PSAP with plasma PSAP (D) and CSF PGRN with plasma PGRN (E). Each point depicts a CSF PSAP or PGRN value and the corresponding plasma PSAP or PGRN value of one PD patient, respectively. N = 19, 20 in (D, E), respectively. F–I Scatter plots representing associations of CSF or plasma PSAP with scores of UPDRS-III or BDI-II. Each point depicts CSF or plasma PSAP values and the corresponding score of UPDRS-III or BDI-II of a PD patient. N = 19, 54, 20, 50 in (F–I), respectively. J–M Scatter plots representing associations of CSF or plasma PGRN with scores of UPDRS-III or BDI-II. Each point depicts the CSF or plasma PGRN value and the corresponding score of UPDRS-III or BDI-II of one PD patient. N = 20, 55, 19, 50 in (J–M), respectively. Pearson correlation coefficients (r) and p-values are calculated in (D, E, G, H, J, L), and nonparametric Spearman correlation r and p-values are calculated in (F, I, K, M). The solid and dashed lines indicate the simple linear regression line and the 95% confidential interval (CI), respectively. UPDRS-III Unified Parkinson’s disease Rating Scale-III, BDI-II Beck Depression Inventory-II.

Fig. 2. cPSAP^DAT mice show reduced levels of dopaminergic markers and behavioral deficiencies, while cPSAP^SERT mice behave normally.

A Fluorescence in situ hybridization (FISH) images of TH (orange) and PSAP (green) mRNA in substantia nigra pars compacta/ventral tegmental area (SNc/VTA) of wild-type (WT) and cPSAP^DAT mice. B Line charts showing HPLC measurements of DA, HVA, and 3-MT in the striatum of WT and cPSAP^DAT mice of 4m-, 8m- and 16m-old. N_WT = 7, 7, 9, N_cPSAP^DAT = 10, 10, 15, respectively. C TH immunohistochemical staining in striatal sections of 2m- and 16m-old WT and cPSAP^DAT mice. D Graph showing densitometry analysis of TH staining in striatal sections of 2m-, 4m-, 8m- and 16m-old WT and cPSAP^DAT mice. N_WT = 7, 7, 7, 9, N_cPSAP^DAT = 6, 10, 10, 16, respectively. E Representative FISH images of tryptophan hydroxylase (Tph) (magenta) and PSAP (green) mRNA in dorsal raphe nucleus (DRN) of WT and cPSAP^SERT mice. F, G Dot plots showing HPLC measurements of 5-HT (F), 5-HIAA (G) in the hippocampus of WT and cPSAP^SERT mice of 8m- and 13m-old. N_WT = 8, N_cPSAP^SERT = 12. H TH (orange), PSAP (magenta), PGRN (cyan) immunofluorescent staining in substantia nigra of WT and cPSAP^DAT mice of 2m- and 16m-old. Left panels are low-magnification images; right panels are high-magnification images. I, J Graphs showing MFI quantification of PSAP (I) and PGRN (J) in TH positive neurons in the SNc of WT and cPSAP^DAT mice of 2m-, 4m-, 8m-, and 16m-old. N = 48, 70, 46, 51 cells from N = 4, 5, 3, 5 WT mice respectively, and N = 56, 67, 85, 63 cells from N = 4, 5, 5, 5 cPSAP^DAT mice respectively. K Timeline of behavioral tests in cPSAP^DAT and control mice. L Graph representing distance traveled in open field test by WT and cPSAP^DAT mice of 4m-, 8m-, and 16m-old. N_WT = 14, 16, 9, N_cPSAP^DAT = 20, 25, 15, respectively. M Graph showing T-turn time in pole test by WT and cPSAP^DAT mice of 4m-old. N_WT = 14 and N_cPSAP^DAT = 19. N Bar graph representing traversal time in beam traversal test by WT and cPSAP^DAT mice of 16m-old. N_WT = 8 and N_cPSAP^DAT = 16. O–R Graphs representing time in the center zone in the open field test (O), distance and time in light box in light-dark transition test (P, Q), and immobility time in forced swim test (R) by WT and cPSAP^DAT mice of 4m-, 8m-, and 16m-old. N_WT = 14, 16, 9 and N_cPSAP^DAT = 20, 23-24, 15-16 in (O–Q), N_WT = 6, 9, 8 and N_cPSAP^DAT = 10, 11, 16 in (R). S–W Graphs showing distance traveled and time in the center zone in open field test (S, T), distance and time in light box in light-dark transition test (U, V), and immobility time in forced swim test (W) by WT and cPSAP^SERT mice of 8m-old. N_WT = 13, N_cPSAP^SERT = 11 in (S–W). Scale bars, 200 μm (A, E), 1 mm (C),and 20 μm (H). Data are presented as mean ± S.E.M. Student’s t-test (B, S–U, W), Mann–Whitney test (M, N, V) or two-way ANOVA with Bonferroni’s (F, G, I, J, L, O–R) or Fisher’s LSD (D) post hoc test was applied appropriately. * Compared to WT, ^# compared to 2 m cPSAP^DAT; non-significant p values are not labeled, *^/#p < 0.05, **^/##p < 0.01, ^***/###p < 0.001, ****^/####p < 0.0001.

Fig. 3. cPSAP^DAT mice display increased highly unsaturated and shortened lipids along with reduction of sphingolipids throughout the brain, while cPSAP^SERT mice present confined accumulation of gangliosides and increased tryptophan metabolism in the dorsal raphe nucleus.

A Score plots presenting the first and second principal components (PC1 and PC2) generated by the principal component analysis (PCA) of all annotated lipids in the caudate-putamen (CPu) and cortex (Ctx) of 16m-old WT (gray) and cPSAP^DAT (orange) mice. Each point depicts one biological replicate. B Volcano plots showing fold differences and the minus logarithm of q value (−log₁₀ (q value)) of all detected lipids in the CPu and Ctx of two genotypes. The red dash line represents FDR (q) = 5%. Lipids regulated with a false discovery rate (FDR, q) <5% are highlighted with colors. Each dot depicts one lipid. C Images showing the analyzed striatal sections in bright field with dash lines delineating CPu. D, E Ion images (16 m) of cardiolipin (CL) peaks, sorted by numbers of double bonds (D) or chain length (E). F, G Graphs showing CLs arranged by double bonds (F) or chain length (G) with fold changes (FC) The red dashed arrows denote highly unsaturated (F) and shortened CLs (G)., N_cPSAP^DAT = 11 for 16m-old mice. H Ion images of sphingolipids and their precursors. The white arrows denote the metabolic pathways. I Graph showing ceramide-1-phosphate (CerP), palmitoyl-carnitine, and gangliosides in the CPu of 16m-old WT and cPSAP^DAT mice. N_WT = 6, N_cPSAP^DAT = 11. J Volcano plot showing the indicated fold differences and -log₁₀ (p-value) of all annotated lipids in dorsal raphe nucleus (DRN) of 8m-old WT and cPSAP^SERT mice. Non-corrected multiple t tests. The red dash line represents p = 0.05. Lipids regulated with p < 0.05 are labeled. All gangliosides are in black. K Ion images of ganglioside peaks in the DRN level of the midbrain of two genotypes. The top two images (left, WT; right, cPSAP^SERT) show the analyzed sections in bright field with DRN marked in magenta. The white arrows denote the metabolic pathways. Magenta arrows are metabolic pathways modulated by saposins. L Graph showing gangliosides in DRN of two genotypes. Student’s t-test. Each circle represents one mouse. N_WT = 5, N_cPSAP^SERT = 5. M Ion images of tryptophan and metabolite peaks in DRN level of midbrain of two genotypes. N Graph showing tryptophan and metabolites in DRN of two genotypes. N_WT = 5, N_cPSAP^SERT = 5. Scale bars, 1 mm; MALDI-MS ion images are shown using rainbow scale (scaled to 100% of max ion intensity scale) for visualization; data are presented as mean ± S.E.M. Student’s t-test (I, L, N). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. PCs phosphatidylcholines, PAs phosphatidic acids, SMs sphingomyelins, LPCs lyso-PCs, PIs phosphatidylinositols, PSs phosphatidylserines, PEs phosphatidylethanolamines, GMs monosialogangliosides.

Fig. 4. cPSAP^DAT mice are more vulnerable to AAV-α-synuclein-induced toxicity, while AAV-PSAP intranigral injection counteracts the toxicity by reducing p-Ser129 α-syn levels.

A Schematic representation of unilateral stereotaxic surgery (upper left) and timeline of experiments. B Quantification of apomorphine-induced net contralateral rotation (contralateral-ipsilateral) of AAV-α-syn injected mice with or without AAV-PSAP at 6w (gray), 10w (orange), and 14w (black). Each dot depicts one mouse. N = 8, 7, 8, 6, mice in four groups, respectively. Repeated measures (RM) two-way ANOVA with Bonferroni’s post hoc test was applied; *compared to 6w, ^# interactions. C Representative images of TH (orange), PSAP (magenta), and α-syn (cyan) immunofluorescent staining on postmortem substantia nigra sections of AAV-α-syn-injected mice with or without AAV-PSAP. Top panel, low-magnification (scale bar=1000 μm) images of the whole substantia nigra. Bottom panels, high-magnification (scale bar=50 μm) images of TH neurons. D Quantification of mean fluorescence intensity (MFI) of PSAP staining in TH-positive neurons of AAV-α-syn-injected WT and cPSAP^DAT mice. Black and red circles depict contralateral and ipsilateral PSAP immunoreactivity, respectively. N_WT = 8, N_cPSAP^DAT = 7. RM two-way ANOVA with Bonferroni’s post hoc test; *compared to contralateral, ^# compared to cPSAP^DAT. E Representative images of α-syn (red dots indicate the injection side) and TH immunohistochemical staining in striatal and substantia nigra sections, and DAT staining in striatal sections of AAV-α-syn-injected mice. Scale bar, 1 mm. F Densitometry analysis of ipsilateral TH immunoreactivity in the striatum and SNc, and ipsilateral DAT immunoreactivity in the striatum of all groups of mice. Values are normalized to the mean of their corresponding contralateral immunoreactivity. Each circle represents one mouse. N = 8, 8, 7, 7, 6, 6, 6, 6 mice in eightgroups, respectively. Two-way ANOVA with Fisher’s LSD post hoc test. G Representative images of p-Ser129 α-syn immunohistochemical staining (enhanced by nickel) in striatal sections of AAV-α-syn-injected mice. Scale bar, 25 μm. H Quantification of number of p-Ser129 α-syn accumulations (area ≥ 0.64 μm²) in striatal sections of AAV-α-syn-injected mice; N = 8, 8, 7, 7mice in four groups, respectively; One-way ANOVA with Bonferroni’s post hoc test. Data are presented as mean ± S.E.M. *^/#p < 0.05, **p < 0.01, ***p < 0.001, ****^/####p < 0.0001.

Fig. 5. Viral overexpression of PSAP protects mice from 6-OHDA induced dopaminergic degeneration.

A Timeline of AAV injection, 6-OHDA striatal lesion, and apomorphine rotation test. B Quantification of apomorphine-induced net contralateral rotation of AAV-GFP or AAV-PSAP injected WT mice. N = 10, 99-10 mice in two groups, respectively. C Representative images of TH (orange), PSAP (magenta), and GFP (cyan) immunofluorescent staining on postmortem substantia nigra sections of 6-OHDA-lesioned mice injected with AAV-GFP or AAV-PSAP. Top panel, low-magnification (scale bar = 1000 μm) images of the whole substantia nigra. Bottom panels, high-magnification (scale bar = 50 μm) images of TH neurons. D Representative images of TH immunohistochemical staining in striatal and substantia nigra sections, and DAT staining in striatal sections of 6-OHDA-lesioned mice injected with AAV-GFP or AAV-PSAP. Scale bar, 1 mm. E Densitometry analysis of ipsilateral TH immunoreactivity in the striatum and SNc, and ipsilateral DAT immunoreactivity in the striatum of all groups of mice; Values are normalized to their corresponding contralateral immunoreactivity; Each circle represents one mouse; N = 10, 9 mice in two groups, respectively. Student’s t-test. Data are presented as mean ± S.E.M. *p < 0.05, **p < 0.01.

Fig. 6. Encapsulated-cell biodelivery of PSAP protects against AAV-α-synuclein-induced parkinsonism in rats.

A Schematic representation of simultaneous unilateral intranigral AAV-α-syn injection and ipsilateral striatal ECB device implantation. B Timeline of experiments. C Representative tracking images of all groups of rats (AAV-α-syn, AAV-α-syn & ECB, AAV-α-syn & ECB-PSAP, AAV-α-syn & ECB-PGRN, and AAV-α-syn & ECB-PSAP-PGRN) in the open field test at 2w, 8w, and 12w. D Quantification of distance traveled in the open field test of all groups of rats at 2w (gray), 8w (orange), and 12w (black). N = 6, 8, 8, 8, 8 rats in five groups, respectively. Repeated measures two-way ANOVA with Bonferroni’s post hoc test was applied; *compared to 2w, ^# interactions. E Quantification of apomorphine-induced net contralateral rotation of all groups of rats. N = 6, 8, 7, 8, 8 rats in five groups, respectively. Kruskal–Wallis test followed by Dunn’s post-hoc test was applied. F Representative images of α-syn (top), DAT (middle), and VMAT2 (bottom) immunohistochemical staining in striatal sections of all groups of rats. Scale bar, 1 mm. G, H Densitometry analysis of ipsilateral DAT (G) and VMAT2 (H) immunoreactivity in the striatum of all groups of rats. Values are normalized to the mean value of their corresponding contralateral immunoreactivity. N = 6, 8, 7, 7, 7 rats in five groups, respectively. One-way ANOVA with Bonferroni’s post hoc test was applied in (G), and Kruskal–Wallis test followed by Dunn’s post hoc test was applied in (H). Data in (E, G, H) are presented as mean ± S.E.M. *^/#p < 0.05, **p < 0.01, ***p < 0.001.