iPSC-derived neural precursor cells engineering GBA1 recovers acid β-glucosidase deficiency and diminishes α-synuclein and neuropathology

Abstract

Mutations in GBA1, encoding the lysosomal acid β-glucosidase (GCase), cause neuronopathic Gaucher disease (nGD) and promote Parkinson disease (PD). The mutations on GBA1 include deletion and missense mutations that are pathological and lead to GCase deficiency in Gaucher disease. Both nGD and PD lack disease-modifying treatments and are critical unmet medical needs. In this study, we evaluated a cell therapy treatment using mouse iPSC-derived neural precursor cells (NPCs) engineered to overexpress GCase (termed hGBA1-NPCs). The hGBA1-NPCs secreted GCase that was taken up by adjacent mouse Gba^−/− neurons and improved GCase activity, reduced GCase substrate accumulation, and improved mitochondrial function. Short-term in vivo effects were evaluated in 9H/PS-NA mice, an nGD mouse model exhibiting neuropathology and α-synuclein aggregation, the typical PD phenotypes. Intravenously administrated hGBA1-NPCs were engrafted throughout the brain and differentiated into neural lineages. GCase activity was increased in various brain regions of treated 9H/PS-NA mice. Compared with vehicle, hGBA1-NPC-transplanted mice showed ∼50% reduction of α-synuclein aggregates in the substantia nigra, significant reduction of neuroinflammation and neurodegeneration in the regions of NPC migration, and increased expression of neurotrophic factors that support neural cell function. Together, these results support the therapeutic benefit of intravenous delivery of iPSC-derived NPCs overexpressing GCase in mitigating nGD and PD phenotypes and establish the feasibility of combined cell and gene therapy for GBA1-associated PD.

Graphical abstract

The study provides in vivo proof-of-concept that intravenous injection of iPSC-derived neural progenitor cells engineered to overexpress hGBA1 reduces neuropathology and toxic insoluble α-Synuclein in a mouse model of neuronopathic Gaucher and Parkinson diseases, suggesting a potential novel non-invasive disease-modifying approach for treatment of these diseases.

Introduction

GBA1 encodes for the lysosomal acid β-glucosidase (GCase), a lysosomal enzyme responsible for glycosphingolipid degradation. GBA1 mutations cause Gaucher disease (GD) and are also the major genetic risk factors for developing Parkinson disease (PD).^1,2 In GD, insufficient GCase activity leads to progressive cellular accumulation of glucosylceramide (GC) and its deacylated form, glucosylsphingosine (GS). Both GC and GS are pathological and account for a continuum of clinical phenotypes. GD is classified as visceral (GD1) or neuronopathic (GD2 and GD3) diseases. Prevalence of GD in the general population is 1 per 60,000, and ∼10% of GD patients in the United States and Europe and ∼75% in Asian countries are diagnosed with neuronopathic disease.^3,4,5,6 Neuronopathic GD (nGD) affects the central nervous system (CNS) through progressive neurodegeneration and inflammation, leading to significant functional deficits and mortality.^6,7,8,9 Approximately 5%–10% of the PD population carries GBA1 mutations, which accounts for ∼700,000 patients.¹ Decreased GCase activity has been documented in the PD brain, even in patients without GBA1 mutations.^10,11 Therefore, enhancing GCase activity in the CNS could be a disease-modifying therapy for nGD and PD.

Induced pluripotent stem cells (iPSCs) represent a source of unlimited patient-specific neural precursor cells (NPCs). A subclass of NPCs express VLA4 (Integrin alpha4beta1, Very Late Antigen-4) that allows systemically delivered NPCs to cross the blood-brain-barrier (BBB) via interaction with endothelial VCAM 1 (vascular cell adhesion molecule 1), and enter the CNS.^12,13 The successful delivery of NPCs and resulting therapeutic efficacy of this approach has been previously validated in models of autoimmune disease and motor neuron diseases.^12,13,14 The therapeutic potential of this cell transplantation approach for treating nGD has also been demonstrated in an acute nGD mouse model.¹⁵ Therefore, non-invasive intravenous delivery of iPSC-derived VLA4+NPCs represents a promising therapeutic approach for neurodegenerative disease, such as PD. Transplantation of NPCs to the CNS may have therapeutic benefits in PD by promoting neuroprotection and direct cell replacement through differentiation to functional cell types in the CNS. Furthermore, because overexpressed GCase is secreted and can cross-correct adjacent cells,¹⁶ increasing GCase activity may also rescue GCase function and promote reduction of α-synuclein (α-Syn) aggregates in PD.

In this study, we developed an iPSC-derived NPC-based cell and gene therapy for nGD and PD. Wild-type (WT) mouse iPSC-derived VLA4+NPCs expressing transgenic human GCase (hereafter termed hGBA1-NPC) provided both self- and cross-correction of GCase deficiency in cell and mouse models of nGD. Intravenous (i.v.) injection of hGBA1-NPC in the 9H/PS-NA nGD mouse model showed engraftment of hGBA1-NPC and increased GCase activity in the brain, improved neuropathology, and reduced PD marker α-Syn pathology. This study provides in vivo proof-of-concept of using this novel cell-based therapy, i.e., iPSC-derived NPCs with GCase overexpression, for nGD and PD treatment.

Results

Generation of hGBA1-NPCs

To increase GCase expression and enable secreted GCase to cross-correct GCase deficiency in surrounding cells, GFP⁺VLA4⁺ WT mouse iPSC-derived NPCs were generated as previously described¹⁵ and transduced with lentiviral vectors constitutively overexpressing human GCase and mCherry (hGBA1-NPCs). Transduced cells expressing both GFP and mCherry were enriched by flow cytometric cell sorting (Figure 1A). Control NPCs were generated in parallel by transfection with the same lentiviral vector but expressing mCherry alone without hGBA1 (hereafter termed vector-NPCs). Both hGBA1-NPCs and vector-NPCs were replated and expanded in NPC culture media and expressed neural precursor cell markers Nestin and Sox2, and the BBB targeting marker VLA4 (Figure 1B), demonstrating retention of neural precursor identity. Those NPCs were verified by immunofluorescence (IF) staining of Oct4+ pluripotent iPSCs and showed complete lack of detectable Oct4+ cells, confirming the absence of iPSCs in our NPC cultures (Figure S1). hGBA1-NPCs and vector-NPCs also retained the ability to differentiate into astrocytes (GFAP), neurons (Tuj1), and oligodendrocytes (O4) (Figure 1D), indicating the modified NPCs retained neural potency. Approximately 19% of NPCs differentiate into GFAP+ astrocytes, 31% to 35% differentiate into Tuj1+ neurons, and ∼13% to 15% differentiate into O4+ oligodendrocytes from vector-NPCs and hGBA1-NPCs (Figure 1E). There was no significant difference in cell differentiation between vector-NPCs and hGBA1-NPCs. MTT (3-(e,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay confirmed that lentiviral transfection did not affect NPC proliferation (Figure 1C). These results suggest that human GCase expression does not affect NPC differentiation and growth.

Figure 1.

Generation and validation of hGBA1-NPCs

(A) VLA4+ GFP+ NPCs were transduced with lentiviral vector of EF1a-hGBA1-mCherry to generate hGBA1-NPCs. Immunofluorescence (IF) signals of mCherry (red) in those cells were acquired by conventional fluorescence microscopy at 48 h post transduction. The cells with mCherry signal indicate hGBA1 expression. The transduced cells were sorted by flow cytometry to enrich GFP+ and mCherry+ NPCs at day 7 post transduction. (B) IF confirmed the sorted hGBA1-NPCs are VLA4+, Nestin+, and Sox2+ neural precursor cells. (C) LV vector transduction did not affect NPC proliferation analyzed by MTT assay, a colorimetric assay for assessing cell metabolic activity. MTT results did not show significant difference in hGBA1-NPCs and vector-NPCs compared with non-transfected NPCs. (D) hGBA1-NPCs and vector-NPCs had potency in differentiation into astrocytes (GFAP), neurons (Tuj1), and oligodendrocytes (O4). DAPI (blue) stained cell nuclei. Scale bar, 50 μm in (B) and (D). (E) Quantitation data showed percentage of GFAP+, Tuj1+, and O4+ cells in total cells (DAPI+) in (D). n = 5 images/group. (F) GCase activity in hGBA1-NPCs were significantly increased (1.45-fold) compared with untransduced NPCs. GCase activity was detected in the medium of hGBA1-NPCs compared with basal level of GCase in the untransduced NPC medium. Student’s t test, ∗p < 0.05, ∗∗p < 0.01 (n = 6 samples). (G) Human (h) GCase protein was detected in hGBA1-NPCs by immunoblot with human-specific anti-GCase antibody. GAPDH is the internal control. Experiments were repeated three times. Data are presented as means ± SEM.

Human GCase expression in hGBA1-NPCs

GCase expression from hGBA1-NPCs was measured by assessing protein expression and activity. GCase activity in hGBA1-NPCs was above the level (1.45-fold) of endogenous GCase activity in vector-NPCs (Figure 1F). In addition, GCase activity was ∼10-fold higher in medium harvested from hGBA1-NPCs vs. medium from vector-NPCs. Because endogenous GCase is barely secreted,^16,17 this result demonstrates the human GCase secretion from hGBA1-NPCs into the culture medium (Figure 1F). hGBA1-NPCs expressed GCase protein was detected by immunoblot using human-specific anti-GCase antibody,¹⁸ but was not present in the negative control of vector-NPCs (Figure 1G).

The secreted human GCase restored GCase deficiency in Gba^−/− cells

To determine whether secreted human GCase could cross-correct activity in GCase-deficient cells, Gba knockout mouse (Gba^−/−) primary fibroblasts were cultured in medium conditioned by hGBA1-NPCs for 0, 3, and 24 h. GCase activity assay demonstrated increased GCase activity dependent on conditioning time (Figure 2A). Cellular GCase protein level was also increased in a time-dependent fashion after culturing in conditioned medium for 2, 4, and 24 h (Figure 2B). As GCase is a lysosomal enzyme, we next assessed lysosomal accumulation of human GCase taken up by Gba^−/− cells. Co-IF analysis demonstrated co-localization of human GCase and the lysosome marker Lamp1 in Gba^−/− fibroblasts cultured for 3 h in conditioned medium (Figure 2C). Quantification of the percentage of cells that take up GCase was performed by counting GCase+ and Lamp1+ cells from total cells (DAPI+). The result showed that 80% Gba^−/− fibroblasts take up the human GCase into lysosomes from the conditioned medium (Figure 2C graph).

Figure 2.

Secreted human GCase from hGBA1-NPCs restored GCase deficiency in mouse Gba^−/− fibroblasts and neurons

(A) GCase activity was increased in Gba^−/− fibroblasts co-cultured in the medium conditioned by hGBA1-NPCs for 3 and 24 h. (B) The human GCase protein was detected in Gba^−/− fibroblasts co-cultured in the conditioned medium for 2, 4, and 24 h. (C) IF data showed that uptake of human GCase (green) was trafficked into lysosomes of Gba^−/− fibroblasts cultured in the conditioned medium for 3 h and co-stained (yellow, arrows) with lysosome marker (Lamp1, red). No GCase detected in the untreated Gba^−/− cells. The quantitation data showed the percentages of hGCase+Lamp1+ fibroblasts in total cells (DAPI+), indicating ∼80% Gba^−/− fibroblast uptake of the hGCase into lysosomes. n = 8 images/group. (D) GCase activity was significantly increased in the Gba^−/− neurons co-cultured with the conditioned medium for 4, 24, and 48 h. (E) Substrate (GC and GS) levels were significantly reduced after uptake of hGCase from the conditioned medium at 4 to 15 days. (F) Oxygen consumption rate (OCR) was significantly increased in Gba^−/− neurons co-cultured with the conditioned medium at 2 and 6 days. One-way ANOVA test (n = 6–8 samples). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. ns, not significant. Experiments were repeated twice. Data are presented as means ± SEM.

Cross-correction of GCase function by secreted GCase was also validated using cultured immortalized neurons derived from Gba^−/− mice.¹⁹ In Gba^−/− neurons, GCase activity was significantly increased by ∼24-fold after 48 h culture in conditioned medium (Figure 2D). Culture in conditioned medium for 4 to 15 days also significantly decreased GC and GS substrate accumulation in Gba^−/− neurons in a time-dependent fashion (Figure 2E). Finally, to assess cross-correction of GCase in restoring mitochondrial dysfunction that is known to play a role in nGD and PD pathogenesis,^20,21,22 mitochondrial function was measured by Seahorse assay of oxygen consumption rates in Gba^−/− neurons cultured in conditioned medium for 2 and 6 days. Oxygen consumption rates on ATP production, maximal respiration, space capacity, and proton leak were significantly improved, especially after 6 days of culture in conditioned medium (Figure 2F).

These results demonstrate that functional GCase is expressed and secreted from hGBA1-NPCs, resulting in cross-correction of GCase activity, substrate degradation, and improvement of mitochondrial function in adjacent cells. Therefore, we next evaluated the potential therapeutic activity of hGBA1-NPCs on CNS phenotypes in a mouse model of nGD with key PD pathology (accumulated α-Syn).

Transplantation of hGBA1-NPCs into a nGD mouse model

The 9H/PS-NA is a chronic nGD mouse model exhibiting neurological phenotypes and α-Syn pathology,^23,24 and was used for evaluation of short-term in vivo effects of hGBA1-NPCs. 9H/PS-NA mice were administered GFP⁺ VLA4⁺ hGBA1-NPCs at doses of 1 × 10⁶ cells/injection via tail vein administration. Control groups consisted of non-injected WT mice and vehicle-9H/PS-NA mice injected with saline. Intravenous injections were started at 6 weeks of age, at which time there is pathogenic onset of brain inflammation in 9H/PS-NA mice, determined by immunohistochemistry (IHC) for CD68 and GFAP (Figures S2A and S2B). Body weights of mice injected with hGBA1-NPCs were not different from controls during the treatment (Figure S3), demonstrating that injections were well-tolerated. Mice injected 3X per week for 8 weeks were analyzed for neuropathological phenotypes and GCase function. The difference of in vivo effects between hGBA1-NPCs and vector-NPCs was assessed in a group of 9H/PS-NA mice with cell injections of 2X per week for 4 weeks.

Engraftment and differentiation of transplanted hGBA1-NPCs

Engraftment of transplanted GFP⁺ VLA4⁺ hGBA1-NPCs was detected by IF using anti-GFP antibody in the 9H/PS-NA mouse brain collected 1 day after the last i.v. injection from the 8-week treatment (Figure 3). The hGBA1-NPCs were distributed primarily in brainstem, cortex, midbrain, and thalamus. A small number of GFP-positive cells were observed in hippocampus, hypothalamus, and cerebellar peduncle and white matter (Figure 3A). GFP signals were also detected in the same regions in GFP⁺ VLA4⁺ vector-NPC transplanted 9H/PS-NA mice (Figure S4A). Transplanted GFP⁺ hGBA1-NPCs were observed adjacent to the VCAM1⁺ endothelial cell layer in the brain (Figure 3B). Engrafted hGBA1-NPCs survived in the brain and remained multipotent, evidenced by differentiation into astrocytes (Figure 3C), neurons (Figure 3D), and oligodendrocytes (Figure 3E). Quantification of transplanted NPC-differentiated astrocytes, neurons, and oligodendrocytes showed that ∼8% to 12% astrocytes, 11% to 20% neurons, and 2% to 7% oligodendrocytes were in various brain regions including brainstem, cortex, midbrain, and thalamus (Figure 3F). The survival of transplanted hGBA1-NPCs in the cortex was also measured by quantitative GFP⁺ cells at 1 day, 7 days, and 28 days post last i.v. injection. GFP⁺ cells were reduced ∼50% after 7 days compared with 1 day post injection, and barely detectable after 28 days (Figure S5A). Based on these data, using nonlinear regression (curve fit) modeling with Prism software, the estimated half-life of transplanted cells in the brain was 9.5 days (Figure S5B), supporting a weekly cell treatment regimen. These data indicate that the transplanted hGBA1-NPCs enable neural cell differentiation.

Figure 3.

Intravenous administered hGBA1-NPCs engrafted and proliferated in 9H/PS-NA mice brains

hGBA1-NPCs (1 × 10⁶ cells) were transplanted by i.v. injection, 3X/week for 8 weeks, into 9H/PS-NA mice starting at 6 weeks of age. Tissues were collected 1 week post last injection. (A) Distribution of GFP+ hGBA1-NPCs cells in treated mouse brain was determined by IF and mapped (stars) to brain regions. The GFP+ cells were distributed mainly in thalamus, brainstem, midbrain, and cortex regions and some cells were detected in hippocampus, hypothalamus, and cerebellar peduncle and white matter. (B) Transplanted hGBA1-NPCs (green) were detected (white arrow) in the VCAM1+ endothelial cell layer (red) in brainstem. (C–E) Engrafted hGBA1-NPCs (green) differentiated to (C) astrocytes (GFAP, orange), (D) neurons (NeuN, orange), and (E) oligodendrocytes (O4, orange). White arrows indicate GFP+ neural cells. Scale bar, 50 μm. Representative images are shown. (F) Quantitation data of (C) to (E) show the percentages of GFP+ astrocytes (GFAP+), neurons (NeuN+), and oligodendrocytes (O4+) of total cells (DAPI+) in each brain region image. N = 3–6 images of three mice/group. Data are presented as means ± SEM.

Human GCase is expressed in the CNS of hGBA1-NPC-treated mice

To explore whether the engrafted hGBA1-NPCs expressed human GCase in the brain of 9H/PS-NA mice with 8 weeks of treatment, brain tissue lysates were analyzed by immunoblot using human-specific anti-GCase antibody that does not react with mouse GCase. Compared with mice treated with vehicle, human GCase protein was expressed and detected in hGBA1-NPC-treated mice 7 days post injection (Figure 4A). GCase expressing transplanted cells (GCase+GFP+, indicated by pink arrows) were detected in the brainstem, cortex, midbrain, thalamus, and spinal cord (Figure 4B). The surrounding cells (GCase+GFP−, indicated by white arrows) that took up the secreted GCase did not show a GFP+ signal (Figure 4B). Quantification showed that GCase expressing cells (GCase+GFP+) were ∼15% in brainstem, 15% in cortex, 15% in midbrain, and 24% in thalamus, whereas the surrounding endogenous cells that had taken up GCase (GCase+GFP−) were ∼10% in brainstem, 16% in cortex, 15% in midbrain, and 16% in thalamus (Figure 4C). Accordingly, GCase activity in hGBA1-NPC-treated mice showed the trend of increase in brainstem (110.6%), cortex (157.4%), and midbrain (118.8%), as well as in cerebellum (130%), and reached significance, compared with vehicle-treated mice, with the exception of the brainstem (Figure 4D). GCase activity was also significantly increased in whole brain lysates from hGBA1-NPC-treated mice compared with vector-NPC- and vehicle-treated mice (Figure S4B). These results demonstrate that human GCase was expressed from transplanted hGBA1-NPCs in the CNS of 9H/PS-NA mice. Interestingly, despite significant increases in enzyme activity, we failed to detect concomitant decreases in substrate accumulation (Figure S4C), suggesting more active GCase is needed for correction of substrate accumulation in this model (see discussion).

Figure 4.

GCase expression in hGBA-NPC-treated 9H/PS-NA mouse brains

(A) Human GCase protein was detected in hGBA1-NPCs treated 9H/PS-NA mice brains by immunoprecipitation/immunoblot. rhGCase: recombinant human GCase (20 ng) as positive loading control. Experiments were repeated twice. (B) Human GCase protein (orange) and GFP (green) were detected by IF in brain regions and spinal cord of treated (hGBA1-NPC) and untreated (Vehicle) 9H/PS-NA mice. Pink arrows indicate hGCase expressing transplant cells (GCase+GFP+). White arrows indicate surrounding endogenous cells that have taken up GCase (GCase+GFP−). Scale bar, 50 μm. (C) Quantitation data of (B) showed the percentage of GCase+GFP+ and GCase+GFP− cells of total cells (DAPI+) in each brain region image. N = 6 images of three mice/group. (D) GCase activity was increased in brainstem, cortex, midbrain, and cerebellum of hGBA1-NPC-treated 9H/PS-NA mouse brains compared with vehicle (saline) control mice. Student’s t test, ∗∗∗p < 0.001 (n = 6 mice/group). Cell treatment is described in the Figure 3 legend. Data are presented as means ± SEM.

Decreased α-Syn in hGBA1-NPCs treated mice brain

The presynaptic protein α-Syn is linked genetically and neuropathologically to PD.^25,26 9H/PS-NA mice develop α-Syn aggregates in the brain, an important pathology that supports the GBA1 mutation and GD association with PD.²⁴ The effect of hGBA1-NPC treatment on α-Syn aggregates was evaluated by IF and immunoblot for insoluble α-Syn aggregates using anti-α-Syn phosphor (Ser129) antibodies.²⁷ Compared with vehicle control, α-Syn aggregation was significantly decreased in sections of the mice brains with 8 weeks of hGBA1-NPC treatment (Figures 5A and 5B). Furthermore, immunoblot of Triton X-100 (TX)-soluble and -insoluble fractions of brain lysates showed a significant reduction in the insoluble form of α-Syn in hGBA1-NPC treated mice (Figures 5C–5E). Insoluble form of α-Syn was also significantly reduced in hGBA1-NPC-treated mice compared with vector-NPC with 4 weeks of treatment (Figures S4D and S4G). These results indicate that hGBA1-NPC treatment enabled prevention or reduction of α-Syn pathology in nGD mice.

Figure 5.

Evaluation of α-Syn pathology in treated 9H/PS-NA mice brains

(A) α-Syn was detected by IF staining using anti-α-Syn phosphor (Ser129) antibody. (B) Quantitation of (A). Average IF pixels of α-Syn aggregations at >2 μm size were quantified. α-Syn aggregates were reduced by about 52% in hGBA1-NPC-treated mice brains. N = 5–10 images of three mice/group. (C) Soluble and insoluble α-Syn was analyzed by immunoblot in mice brains from vehicle, hGBA1-NPC, and WT groups. (D and E) Quantitation of (C). (D) Insoluble α-Syn levels were reduced 60% by hGBA1-NPC treatment. (E) Soluble α-Syn levels were not changed. One-way ANOVA test, ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant (n = 3 mice/group). Cell treatment is described in the Figure 3 legend. Experiments were repeated three times. Data are presented as means ± SEM.

CNS phenotype improvement by hGBA1-NPC treatment

9H/PS-NA brains develop neurodegeneration and neuroinflammation.²³ Neurodegeneration was analyzed by staining the brain sections from WT and vehicle- and hGBA1-NPC-treated mice with Fluoro-Jade C (FJC), a fluorescent dye that has been used to label degenerating neurons in the brain.¹⁵ Significantly reduced FJC signal was observed in hGBA1-NPC-treated 9H/PS-NA brains (Figures 6A and 6B). Quantitative RT-PCR analysis of neurotrophic factors that possess neuroprotective properties (Bdnf, Cntf, and Nt3) revealed increased mRNA expression levels with treatment (Figures 6C–6E).^28,29 These results suggest a neuroprotective effect by hGBA1-NPCs, possibly by modulating the expression of neurotrophic factors. The effect of hGBA1-NPC treatment on neuroinflammation in 9H/PS-NA mice was assessed by IHC using markers of activated microglia (CD68) and astrogliosis (GFAP). hGBA1-NPC treatment reduced both CD68 (Figures 7A and 7B) and GFAP signals (Figures 7C and 7D) in brainstem, cortex, midbrain, and thalamus. Reduction of CD68 and GFAP was also detected by IF in spinal cord (Figure 7E). Finally, positive effects of hGBA1-NPC treatment on neuroinflammation were supported by a significant reduction in mRNA expression of the major inflammatory markers Tnf and Il6 (Figure 7F). Furthermore, immunoblot analysis was performed to compare in vivo effect of hGBA1-NPC with vector-NPC in 9H/PS-NA mice with 4 weeks of treatment. hGBA1-NPC treatment showed significantly improved CNS pathology compared with vector-NPC treatment on neurodegeneration measured by neuron marker NeuN (Figures S4E and S4H) and brain inflammation determined by astrogliosis marker GFAP (Figures S4F and S4I). Together, these analyses demonstrate that i.v. infusion of hGBA1-NPCs results in significant reductions of markers of both neurodegeneration and inflammation in the CNS of 9H/PS-NA mice.

Figure 6.

Evaluation of neurodegeneration and neurotrophic factors in treated 9H/PS-NA mice brains

(A) Neurodegeneration was measured by FJC (Fluoro-Jade C, a green fluorescent dye that has been used to label degenerating neurons in the brain). Scale bar, 50 μm. (B) Quantitation of (A). The FJC signals were significantly reduced in hGBA1-NPC-treated 9H/PS-NA mice brains compared with vehicle control. Student’s t test, ∗p < 0.05; ∗∗p < 0.01. N = 3 images/mouse, 3 mice/group. (C–E) qRT-PCR of neurotrophic factors (Bdnf, Cntf, and Nt3) mRNA levels in treated 9H/PS-NA mice cortex tissues compared with vehicle-9H/PS-NA or untreated WT mice (CTL). One-way ANOVA test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns, not significant. (n = 2–5 mice/group, replicate assays). Cell treatment is described in the Figure 3 legend. Experiments were repeated three times. Data are presented as means ± SEM.

Figure 7.

Reduced neuroinflammation in hGBA1-NPC-treated 9H/PS-NA mice

(A–D) The mice brain sections were stained with inflammatory markers, anti-CD68 (A, brown) for microgliosis and anti-GFAP antibody (C, brown) for astrogliosis. (B and D) Quantitation of (A) and (C), respectively. Compared with vehicle control, hGBA1-NPC-treated mice brains showed significant reduction of CD68 (B) and GFAP (D) signals in brainstem, cortex, midbrain, and thalamus, indicating decreased neuroinflammation. Scale bar, 50 μm. Cell treatment is described in the Figure 3 legend. Student’s t test, ∗p < 0.05; ∗∗p < 0.01. N = 3 images/mouse, 3 mice/group. (E) CD68 and GFAP signals by IF were reduced in spinal cords of hGBA1-NPC-treated mice (2X/week for 4 weeks). Scale bar, 100 μM.

(F) qRT-PCR of proinflammatory factors (Tnf and Il6) mRNA levels in treated 9H/PS-NA mice cortex tissues were significantly reduced compared with vehicle-9H/PS-NA. Cell treatment is described in the Figure 3 legend. Student’s t test, ∗p < 0.05; ∗∗∗p < 0.001 (n = 3–5 mice/group, replicate assays). Experiments were repeated three times. Data are presented as means ± SEM.

No tumor formation observed in hGBA1-NPC-treated mice

Finally, we assessed whether transplanted NPCs could be detected in visceral tissues (livers and spleens) and whether teratomas formed. In contrast to brain sections, IF staining showed no detection of GFP⁺ hGBA1-NPCs adjacent to endothelium and other regions of livers and spleens of infused mice (Figure S6A). No tumor formation was detected by gross examination of treated mice 9 weeks after first injection (Figures 6, 7, and S6B). In addition, GCase activity in 9H/PS-NA mouse livers was not significantly changed after hGBA1-NPC treatment compared with vehicle group (Figure S6C). These data confirmed hGBA1-NPCs were not engrafted in visceral tissues (liver and spleen) and the cell treatment did not result in detectable tumors using this regimen.

Discussion

The CNS component in nGD patients is currently untreatable. While some symptoms of PD can be managed by medications, effective disease-modifying therapy is not available to treat PD. There is a pressing unmet need to develop effective medical treatment for nGD and PD patients. Cell replacement therapy has been proposed as one approach for PD and other neurodegenerative diseases.^30,31 However, transplantation of therapeutic cells to the CNS involves highly invasive procedures and is limited by the availability of suitable donor cells. The iPSCs represent a source of unlimited patient-specific NPCs. These cells can be derived from any individual, making it possible to generate autologous NPCs that can be used in cell therapies. Intravenous administration of NPCs that express VLA4 allows systemically delivered NPCs to cross the BBB via interaction with endothelial VCAM 1 and enter the CNS.^12,13 This non-invasive approach has an advantage over intracranial cell delivery to prevent medical complications from brain surgery. The therapeutic value of VLA4+NPC transplantation approach is supported by previous animal model studies, in which i.v. infusion of iPSC-derived VLA4+NPCs led to improvements in disease phenotype and survival in a mouse model of nGD and amyotrophic lateral sclerosis.^12,13

GCase deficiency in CNS causes nGD and can also prompt PD,^1,6 therefore, the iPSC-NPC approach can be extended to GBA1 mutation associated PD to deliver normal GCase into the CNS. Because endogenous GCase is generally not secreted from the cells except in overexpression systems, unlike many lysosomal enzymes,^16,32 in this study, the NPCs were transduced with a lentiviral vector for constitutive human GCase expression to enhance GCase levels. Overexpressed GCase was secreted from the NPCs, which was taken up by Gba^−/− cells and provided cross-correction of enzyme deficiency in surrounding cells (Figures 1 and 2). GCase uptake requires the mannose receptor. This is the preferred uptake mechanism.^33,34 We and others have shown that both fibroblasts and neurons use this mannose receptor-mediated pathway.^33,35 Critically, the take up of GCase was functional in Gba^−/− neurons evidenced by reduced GCase substrate accumulation (Figure 2E). Impaired mitochondrial function is a well-recognized neuronopathic phenotype in nGD and PD.^15,22,36 Our study in Gba^−/− neurons also reinforced the association of mitochondrial dysfunction with GCase deficiency.²¹ Importantly, mitochondrial function was improved in GCase rescued Gba^−/− neurons (Figure 2F). Thus, cross-correction of GCase in the cells not only restored GCase function in substrate degradation, but also rescued mitochondria function, a GD and PD pathological marker, highlighting the therapeutic benefit of hGBA1-NPC. In the ex vivo experiments, GCase activity in fibroblast was measured after 3 and 24 h and in neurons was assayed after 4-, 24-, and 48-h incubation in the conditioned medium. GCase activity showed time-dependent increase until 48 h (Figure 2E), suggesting the secreted GCase activity lasts >48 h in Gba^−/− cells. These ex vivo data support the feasibility of hGBA1-NPCs in rescuing GCase deficiency and associated disease phenotypes. Thus, successful overexpression and secretion of human GCase from hGBA1-NPCs allows replacement and self- and cross-correction of enzyme deficiency not only in transplanted cells, but also in adjacent host cells to maximize the therapeutic efficacy.

The in vivo efficacy of hGBA1-NPCs was evaluated in the 9H/PS-NA nGD mouse line, a viable GBA1-associated PD mouse model.^22,24 This model develops chronic progressive neuroinflammation, neurodegeneration, and α-Syn aggregates, hallmark PD pathology. Its ∼22-week life span is suitable for cell therapy study.^22,23,24,37 The hGBA1-NPCs treatment was started at disease onset, as indicated by neuroinflammation, at 6 weeks of age to evaluate potential prevention of disease progression. Intravenous infused hGBA1-NPCs were successfully engrafted and differentiated to neural cells in the CNS and expressed human GCase broadly throughout the brain and in the spinal cord (Figures 3 and 4). GCase is a ubiquitously expressed protein. All neural cells including neurons,²¹ astrocytes,³⁸ and microglia³⁹ accumulate GC and GS if lacking GCase. Therefore, all neural cells would need GCase enhancement. In fact, our results showed that all major neural cells (neuron, astrocyte, oligodendrocyte) expressed human GCase (Figure 3). Effectively, cell treatment reduced CNS inflammation and neurodegeneration in brain and spinal cord, a positive outcome of NPC therapy in mitigating GD and PD (Figures 6 and 7). Increased expression of neurotrophic factors (Bdnf, Cntf, and Nt3) in treated mice may provide neuroprotection and promote neural differentiation (Figures 6C–6E),^15,28,40 a unique advantage of NPC therapy. Furthermore, hGBA1-NPCs with only 4-week treatment resulted in significantly reduced neurodegeneration, inflammation, and α-Syn pathology compared with vector-NPC treatment (Figures S4D–S4I), supporting a positive effect of GCase overexpressed hGBA1-NPCs in GD and PD treatment. These therapeutic effects in 9H/PS-NA mice reiterate the positive outcomes from the iPSC-NPC study in a nGD mouse model with acute and severe neurological disease¹⁵ and support the in vivo benefits of hGBA1-NPC in recovering GCase and mitigating neuropathology in the CNS.

GBA1 mutations are the most common genetic risk factors for developing PD.^1,2 Reduction of GCase is associated with increased levels of α-Syn and documented in the PD brain, even in patients without GBA1 mutations.^10,11 There is a reciprocal interaction between GCase and α-Syn levels.²⁶ This interaction supports a therapeutic approach to increase GCase and consequently reduce α-Syn misfolding to modify the course of PD. The proposed mechanisms of genetic cause of PD by mutated GCase include a "gain-of-function" mechanism that misfolded mutant GCase plays a part in α-Syn aggregate formation through stress to the Ubiquitin-Proteasome System, and to the development of PD, and the "loss-of-function" mechanism that one normal GBA1 allele is not sufficient to provide functional GCase and deficient GCase could hinder α-Syn metabolism.⁴¹ Thus, cross-correction of GCase deficiency by expressing GCase from hGBA1-NPC, thereby, is expected to improve α-Syn metabolism to alleviate PD. Indeed, our results showed that hGBA1-NPC treatment reduced α-Syn aggregation (Figures 5, S4D, and S4G), despite no significant reduction in substrates GC or GS. In hGBA1-NPC-treated mice, GCase activity was significantly increased in the brain. Remarkably, with a short-term treatment, α-Syn aggregates were significantly reduced by ∼50% (8-week treatment) or ∼30% (4-week treatment) in 9H/PS-NA mice in the substantia nigra and cortex (Figures 5, S4D, and S4G), major areas where α-Syn aggregates develop.²⁴ This result holds promise and supports the approach of restoring GCase in CNS to mitigate PD pathology. There is a debate on the confounding factors that promote PD, including defective GCase or GCase substrate accumulation.^26,42 Mutant GCase has been shown to trigger α-Syn aggregation by blocking chaperone-mediated autophagy.⁴³ The intra-cerebral gene delivery of GBA1 has been used as gene therapy to prevent α-Syn accumulation in animal models of PD.^44,45 In contrast, reduction of GCase substrate level by a small molecular drug that inhibits glucosylceramide production did not affect symptoms in PD patients in a clinical trial.⁴⁶ In this study, our results showed that hGBA1-NPC treatment reduced toxic α-Syn aggregation (Figures 5, S4D, and S4G), despite no significant CNS substrate reduction (Figure S4C). Those findings and the results from this study all support the CNS delivery of healthy GCase to be a therapeutically effective treatment and beneficial for PD. The main hypothesis tested in this study is increased GCase activity, through delivery of NPCs engineered to overexpress GCase, will be therapeutic for treating GBA1-mutation associated PD with GCase deficiency. Currently, the specific mechanism of GCase activity reduction in PD without GBA1 mutation is not known. It has been suggested that decreased GCase activity in these cases is not due to changes in GBA1 mRNA expression, but rather is associated with impaired lysosomal function.^11,47 While the NPC cell therapy in the present study is primarily targeting PD with GBA1 mutation, its therapeutic benefit may also extend to PD without GBA1 mutation by delivering GCase to compensate for lost GCase activity.

While the NPC treatment showed a robust therapeutic effect in the CNS and α-Syn pathology, it did not result in changes in GCase substrate levels (Figure S4C). The 9H/PS-NA model used in this study accumulates multiple glycosphingolipids due to deficient prosaposin and derived saposins (A, B, C, and D).²³ Those saposins are sphingolipid activators involved in degradation of multiple glycosphingolipids, i.e., Saposin B and Saposin C for lactosylceramide, Saposin C for GC, and Saposin D for ceramide.⁴⁸ In the glycosphingolipids degradation pathway, lactosylceramide is upstream from GC. Lactosylceramide accumulation could influence the level of GC, which may offset the GC reduction by hGBA1-NPCs treatment. Compared with our prior results of NPC therapy on acute nGD model (4L; C∗ mice) that present rapidly developed disease parameters and short life span (<8 weeks),¹⁵ the 9H/PS-NA model in the current manuscript is the chronic nGD model.²³ 9H/PS-NA mice exhibit slower disease progression and longer life span with different age at start and received different duration of treatment. The 9H/PS-NA mice present chronic neurological disease over 22 weeks and current treatment was administered for 2 months. Although this regimen led to an increase of CNS GCase and prevention of neuropathology, a longer course of treatment and earlier intervention might be even more effective to reverse the disease. Moreover, GCase is an unstable lysosomal enzyme with a very short half-life (20 min) in body fluids at neutral pH, therefore, improving its half-life by pharmacological agents, e.g., small molecule GCase chaperone that can protect GCase stability^49,50,51 could be used together with hGBA1-NPCs to enhance the in vivo efficacy.

In this study, NPC cell transplantation accompanied by stable expression of human GCase has many advantages. First, the novel therapeutic approach provides the potential to use a GD patient’s iPSC-derived NPC cells for personalized therapy, which will likely reduce the immune response. We have functionally assessed the presence of undifferentiated iPSCs in our VLA4+ NPC population by teratoma formation (Figure S6) in this study and a prior study.¹⁵ Both analyses demonstrated no-teratoma formation, supporting the safety of the iPSC-NPC therapy. Second, LV-mediated human GCase provides long-lasting transgene expression in transplanted cells. Most importantly, the GBA1 transgene was contained in the transduced NPC cells without affecting genome integrity of other brain cells, which contributes to therapeutic safety. Although GCase is an unstable lysosomal enzyme with relatively short half-life, its robust expression in NPC cells provides more GCase than empty vector-transduced WT NPCs and benefits surrounding brain tissue. Recently, adeno-associated virus (AAV)-mediated gene therapy has been developed for GD treatment.^52,53 AAV gene therapy could provide efficient gene transfer to target monogenetic disease and has the ability to provide long-term therapy, which makes this vector system very attractive. Although different AAV capsids possess unique tissue tropism, off-target, innate, and adaptive immune responses to AAV vectors and their transgene products, and toxicity of viral genome to host genome stability pose substantial hurdles to clinical development and wide use in patients, which demands further improvement. Additional modifications, e.g., combined expression cassette, microRNA regulation,⁵⁴ tolerance induction,⁵⁵ and reduction of AAV dose may be necessary to overcome those obstacles. The current approach of NPCs with expressed VLA4 allows systemically delivered NPCs to cross the BBB via interaction with endothelial VCAM1 to enter the CNS, but restricting the access to other organs. Such directed cell delivery to the CNS via a systematic administration route by i.v. administration provides a novel approach and greatly avoids off-tissue targeting simultaneously. However, the rapid turnover of GCase and NPCs in brain and necessity of repeated cell injection may be drawbacks of this NPC transplantation, which would require further optimization.

In summary, these studies demonstrate that iPSC-derived NPCs in combination with transgenic overexpression of GCase mitigated nGD and PD phenotypes in cell and mouse models. Our non-invasive approach by i.v. infusion of hGBA1-NPCs enables reduction of insoluble α-Syn, which supports the therapeutic benefit of increasing brain GCase to treat PD.^56,57,58 hGBA1-NPCs offer a new therapeutic option, in addition to gene therapy and chaperone therapy, for modulating GCase as a disease-modifying treatment for PD.^50,51,59 This hGBA1-NPC approach may exert therapeutic activity through several mechanisms: functional cell replacement and improvement of neural phenotypes (e.g., mitochondrial function); secretion of trophic factors that confer neuroprotection and suppress inflammation; and cross-correction of GCase deficiency, thereby reducing α-Syn aggregation. These studies demonstrate beneficial CNS effects from non-surgical, i.v. transplantation of neural cells combined with therapeutic expression of GCase, and provide in vivo proof-of-concept of cell and gene therapy in amelioration of nGD and GBA1 mutation associated PD.

Materials and methods

Generation of GBA1 lentivirus and hGBA1-NPCs

Human WT GBA1 cDNA was from Origene Technologies (GBA [untagged]-Human glucosidase, beta, acid [GBA], transcript variant 3, NM_001005742.1). The full-length cDNA was cloned into pLVX-EF1α-IRES-mCherry Vector (Clontech, 631987) by using In-Fusion Cloning kit (Clontech). Briefly, GBA1 PCR fragment and 15-base pair vector sequence were fused with linearized vector by In-Fusion Enzyme. This construct and empty vector were packaged into viral particles using HEK293T cells. The lentiviruses were concentrated by ultracentrifugation at Cincinnati Children’s Hospital Medical Center (CCHMC) vector core.

Previously generated GFP⁺VLA4⁺mNPCs that were derived from C57BL6 mice were cultured in STEMdiff Neural Progenitor Medium.¹⁵ For transfection, GFP⁺VLA4⁺ mNPCs were plated at a concentration of 7 × 10⁴ cells/well in a 6-well plate. The cells were transfected with either GBA1 lentiviral vectors (hGBA1-NPC) or empty lentiviral vectors (vector-NPC) in serum-free medium for 6 h, and then 1 mL culture medium was added to the well and incubated overnight. Polybrene (4 μg/mL) (an enhancer reagent of retrovirus-mediated gene transduction) was added to increase the infection efficiency. Two weeks after transfection, transduced NPCs were sorted by flow cytometry and GFP⁺ and mCherry⁺ NPCs were selected. hGBA1-transduced NPCs were designated hGBA1-NPCs and empty vector-transduced NPCs were designated vector-NPCs.

NPCs multipotency

The hGBA1-NPCs and vector-NPCs were cultured in STEMdiff Neural Progenitor Medium (Stem Cell Technologies) and confirmed with neural stem cell markers (Nestin and Sox2).¹⁵ Neural potency of the NPCs was measured by neural differentiation. hGBA1-NPCs and vector-NPCs were plated on a polyornithine and laminin-coated culture dish in complete StemPro NSC SFM media at 2.5 to 5 × 10⁴ cells/cm² for 2 days and replaced with each differentiation medium: neuron differentiation medium (Neurobasal Medium, B-27 Serum-Free Supplement, GlutaMAX-I Supplement), astrocyte differentiation medium (D-MEM with N-2, GlutaMAX-I, and FBS), and oligodendrocyte differentiation medium (Neurobasal medium with B-27, GlutaMAX-I, and T3). The medium was refreshed every 3 to 4 days.^15,60,61 Cell types were determined by IF staining of neuronal markers, neuron (Tuj1), astrocyte (GFAP), or oligodendrocyte (O4) as described.⁶¹

Human GCase expression and secretion

The culture medium and cells were collected from 80% to 90% confluence of hGBA1-NPCs to measure human GCase protein by immunoblot using human-specific anti-GCase antibody (1; 1,000) and GCase activity.³⁵ For co-culture experiments, the medium was collected daily from 80% to 90% confluence of hGBA1-NPCs culture and loaded as conditioned medium to mouse Gba^−/− primary fibroblasts or immortalized Gba^−/− neurons culture. The conditioned medium was changed every 24 h. Cells were co-cultured with the conditioned medium for 2 to 48 h or 2 to 15 days.

Mouse model and treatment

9H/PS-NA mice carry the Gba1 D409H/D409H (9H) mutation and express low levels (4%–45% of WT) of prosaposin and saposins (PS-NA) in the background of C57BL/6J/129SvEV/FVB.²³ 9H/PS-NA mice display deficient GCase function and accumulation of substrate GC and GS in the brain and visceral tissues and develop neurological phenotypes and brain α-Syn aggregation.^22,24 This model has sufficient life span (∼22 weeks) to allow for preclinical studies of therapies and is thus particularly suitable for the evaluation of PD phenotypic improvements.

For efficacy evaluation, 9H/PS-NA mice at 6 weeks of age were transplanted with hGBA1-NPCs (1 × 10⁶ cells/100 μL saline) or 100 μL saline as vehicle by i.v. tail vein injection with three injections (3X)/week for 8 weeks. In vivo effect of vector-NPCs in comparison with hGBA1-NPC was assessed in 9H/PS-NA mice (6 weeks) i.v. administered vector-NPCs or hGBA1-NPC (1 × 10⁶ cells/100 μL saline), two injections (2X)/week, for 4 weeks. Spinal cord histology was analyzed in 9H/PS-NA mice (6 weeks) treated with hGBA1-NPCs (1 × 10⁶ cells/100 μL saline), 2X/week for 4 weeks. The experiments were not designed to address sex differences, but male and female mice were used for all animal experiments. All mouse housing (under pathogen-free conditions in the animal facility) and cell treatment was performed according to Institutional Animal Care and Use Committee-approved protocol (2015-0050) at CCHMC. Mouse tissues were collected after transcardial perfusion with saline at 7 days after last injection or as indicated. Dissected tissues were either fixed in 4% paraformaldehyde (PFA) or snap-frozen for further analyses.

Immunofluorescence

The mouse brain and spinal cord tissues or cells were fixed in 4% PFA and permeabilized in 0.3% Triton X-100 (TX) for 30 min. The samples were treated with 50 mM NH₄Cl in 1xPBS for 15 min followed by 1xPBS wash, and then blocked for 1 h at room temperature (RT) in blocking buffer (10% goat serum and 0.4% TX in PBS). Following primary antibodies diluted in 5% goat serum were applied to the cells or tissue sections: rat anti-CD106 (VCAM1) (1:100, ebioscience, MR106), rat anti-Lamp1 (1:100, RDI, MCD107A-D4B), rabbit anti-GFP (1:200, Cell Signaling Technology, 2956), rabbit anti-hGCase (1:50, made at CCHMC), mouse anti-Sox2 (1:200, Millipore, AB5603), mouse anti-Nestin (1:100, Millipore, MAB353), mouse anti-integrin alpha 4 (VLA4) (1:100, Cell Signaling Technology, 8440S), mouse anti-GFAP (1:100, Millipore, MAB360), mouse anti-TUJ1 (1:200, StemCell Technologies, 60052), mouse anti-O4 (1:100, Millipore, MAB365), anti-Oct4 (1:250, Millipore, MAB345), and mouse anti-NeuN antibody (1:500, Millipore, MAB377). After washing in PBS (3 × 10 min), the secondary antibodies: goat anti-mouse conjugated with Alexa Fluor 647 (1:1000) for Sox2, Nestin, VLA4, TUJ1, GFAP, and O4 detection, goat anti-rabbit conjugated with Alexa Fluor 488 (1:1000) for GFP detection, goat anti-mouse immunoglobulin G (H + L) cross-adsorbed conjugated with Alexa Fluor 594 (1:500) for Oct4, and goat anti-rat conjugated with Alexa Fluor 647 (1:1000) for VCAM1 detection, were applied. The cell and tissue sections were counter-stained for nuclei with DAPI in mounting medium. Immunofluorescent signals were evaluated by conventional fluorescence microscopy (Zeiss Axiophot; Oberkochen, Germany). Cell quantification was performed by manually counting specific marker-stained cells in total DAPI+ cells. The percentage of specific marker-stained cells from multiple images and multiple mice per group were graphed using Prism.

MTT cell proliferation assay

Cell proliferation and viability were assessed by MTT (3-(e,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay (Molecular Probes, Vybrant^@ MTT cell proliferation assay kit, V-13154) as described.²¹ NPCs, hGBA1-NPCs, and vector-NPCs at the logarithmic growth phase were collected and seeded in 96-well culture plates (5 × 10³ cells/well) and allowed to adhere overnight. After culturing the cells in STEMdiff Neural Progenitor Medium for 1, 2, 3, 4, 5, or 6 days, 20 μL of MTT (5 mg/mL) was added to each well and incubated at 37°C for 4 h followed by incubation in 150 μL of dimethyl sulfoxide for 1 h at 37°C. The absorbance at 570 nm was measured using a Synergy HT plate reader (Bio-Tek Instruments, Winooski, VT, USA).²¹

GCase activity

The cells and mouse tissues were homogenized in 1% Na-taurocholate/1% Triton X-100 (TC/TX). GCase activity was determined fluorometrically with 4MU-Glc as substrate in 0.25% TC/TX diluted in 0.1M citrate phosphate buffer (pH 5.6).^32,62 Protein concentrations of samples were determined by BCA assay using BSA as standard.

Immunoprecipitation and immunoblotting of GCase

Immunoprecipitation (IP) was performed using the Dynabeads protein G kit to determine human GCase protein in mouse brains. Protein G was cross-linked with the trapping antibody, rabbit anti-human GCase, using bis (sulfosuccinimidyl) suberate (BS3, Life Technologies. Carlsbad, CA) following the manufacturer’s instructions. Each milligram of brain lysate protein from hGBA1-NPC- or vehicle-treated mice was mixed with 80 μL of protein G beads cross-linked with rabbit anti-human GCase antibody and incubated at 4°C overnight allowing human GCase to bind to the antibody. The bound GCase proteins were eluted from beads with elution buffer and treated with loading buffer for electrophoresis. For immunoblot, cell lysate, brain lysate and immunoprecipitated human GCase from mouse brains were resolved on 4%–12% NuPAGE gel (Invitrogen) and then transferred to PVDF membrane using iBlot 2 gel transfer device (Life Technologies) following the manufacturer’s instruction. The blotted membranes were incubated with rabbit anti-hGCase antibody (1/1,000). mouse anti-NeuN (Millipore, 1:1,000) or mouse anti-GFAP antibody (Stemcell, 1:1,000) in 1.5% BSA/1.5% milk/PBS buffer overnight at 4°C. The signals were detected using a LI-COR detection system according to the manufacturer’s instruction.

Glycosphingolipids analysis

Frozen brain tissues and neuron cells were homogenized in water/chloroform/methanol and glycosphingolipids were extracted.⁶³ Cell lipid extracts were processed for GC and GS quantification by liquid chromatography-tandem mass spectrometry (LC/MS-MS) at the Clinical Mass Spectrometry Laboratory at CCHMC. The GC and GS levels in cells were normalized to milligrams of cell protein determined by BCA assay.²¹ GC and GS in tissues were analyzed at the Medical University of South Carolina Lipidomics Shared Resource: Analytical Unit. The concentration of tissue GC and GS was normalized to milligrams of tissue weight.³⁵

Seahorse assay

Immortalized Gba^+/+ and Gba^−/− neuron cells were cultured in conditioned medium from hGBA-NPC culture for 0, 2, and 6 days, separately. The cells were replated in an XFe 96-well plate overnight for oxygen consumption rate (OCR) measurement.⁶⁴ The data were analyzed using the XFe Wave software. Basal respiration, ATP production, maximal respiration, space capacity, and proton leak rate were normalized to cell number.⁶⁵

α-Syn measurement

For immunoblot, mouse brain tissues were homogenized in tissue homogenization buffer (50 mM Tris-buffer pH 7.6, 150 mM NaCl, 1% TX, 2 mM EDTA) containing protease and phosphatase inhibitors. Tissue lysate was sonicated on ice for 1 min × 3 times (Branson Sonifier 450, 100 mm Cup probe). The supernatant was collected as “TX soluble fraction” by centrifugation at 120,000 x g for 60 min at 4°C. The remaining pellet was washed three times with 1X PBS/1% TX, centrifuged at 13,000 x g for 15 min, and re-suspended in SDS extraction buffer (50 mM Tris pH 7.6, 150 mM NaCl, 1% TX, 0.5% Na-deoxycholate, 1% SDS). The SDS suspension was sonicated and centrifuged at 120,000 x g for 60 min at 4°C. The supernatant from SDS suspension was collected as “TX insoluble fraction.” The α-Syn in the TX soluble and TX insoluble fractions were subjected to electrophoresis by SDS-PAGE gels (Invitrogen NuPage 4%–12%) and blotted on PVDF membrane (Invitrogen iBlot). The α-Syn was detected using mouse monoclonal antibody (Biolegend, MMS529-R, Syn 202, 1:500) for TX soluble α-Syn fraction and rabbit monoclonal anti-phospho-α-syn antibody (Ser129) (Abcam, ab51253, 1:500) for TX insoluble fraction. IRDye 680RD Goat anti-Mouse (1:5000) was used for signal detection. β-actin was a loading control and detected by anti-β-actin antibody (Invitrogen, MA5-15739, monoclonal BA3R).

IF staining of α-Syn was performed on mouse brain cryostat sections (40 μm). The sections were free-floating fixed in 4% PFA and washed with 1X PBS (3X, 10 min). The tissue sections were permeabilized in 0.3% TX/PBS followed by 1X PBS wash (3X, 10 min) and treated with 5 μg/mL of proteinase K for 5 min at 37°C. Following wash with 1X PBS (3X, 10 min) and blocking with 1.5%BSA/0.5%geletin/PBS for 60 min, the sections were incubated with anti-phospho-α-syn antibody (Ser129) (Abcam ab51253, 1:200 to 1:500) followed by secondary antibody of goat-anti-mouse Texas Red 595 (Invitrogen, 1:500) for detection. Tissue sections were counter-stained for nuclei with DAPI in mounting medium.

Fluoro-Jade C staining

Fluoro-Jade C (FJC) staining was performed on the frozen brain sections as described.¹⁵ FJC images were evaluated by Nikon wide-field microscope. The FJC signals were acquired from multiple images per brain region and analyzed by Fiji software.

Immunohistochemistry

Frozen tissue (brain, spinal cord) sections were fixed with 4% PFA. The fixed sections were incubated with rat anti-mouse CD68 monoclonal antibodies (1:2000, Serotec, MCA1957) and mouse anti-GFAP monoclonal antibody (1:200, Roche, #7604345) as described.⁶⁴ Visceral tissues (liver and spleen) were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with H&E. The morphology was analyzed by Axioskop light microscopy equipped with SPOT Advance Image Software (Zeiss, Germany).⁶⁶ CD68 and GFAP signals were analyzed by Fiji software. For CD68 and GFAP analysis, the signal was acquired from multiple images per brain region.

Quantitative real-time PCR

Total RNA from the mouse cortex was isolated using an RNeasy Micro Kit (QIAGEN) and quantified by spectrophotometric analysis (ND-100; NanoDrop, Thermo Scientific).⁶⁵ The RNA was treated with DNase to remove potential genomic DNA contamination. Total RNA was reverse transcribed into complementary DNA using random hexamers and Transcriptor Reverse Transcriptase (Roche Diagnostics). Real-time PCR was performed according to the manufacturer’s protocol using TaqMan Gene Expression Assays and an ABI Prism 7300 Sequence Detection System (Applied Biosystems). The expression of the following genes was determined: Nt3 (Gene ID: 18205; neurotrophin 3), Bdnf (Gene ID: 12064; brain derived neurotrophic factor), Cntf (Gene ID: 12803; ciliary neurotrophic factor), Tnf (Gene ID: 21926; tumor necrosis factor), and Il6 (Gene ID: 16193; interleukin 6). Primer sequences are listed in Table S1. The reactions were performed in triplicate per cortex sample per mouse, three mice for each group. Gene expression Ct values were corrected for β-actin Ct values in the same sample using the ΔΔCt method and presented as the relative mRNA expression of treated-9H/PS-NA or WT mice compared with vehicle-9H/PS-NA mice.

Statistical analyses

Results are presented as means ± SEM. The data were analyzed by Student’s t test or ANOVA test using GraphPad Prism 9. p < 0.05 was considered statistically significant.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplemental information.