|Home | About | Journals | Submit | Contact Us | Français|
Gaucher disease, a prevalent lysosomal storage disease, is caused by insufficient activity of acid β-glucosidase (GCase) and resultant glucosylceramide accumulation. Recently in Parkinson disease (PD) patients, heterozygous mutations in GCase have been associated with earlier onset and more progressive PD. To understand the pathogenic relationships between GCase variants and Parkinsonism, α-synuclein and ubiquitin distributions and levels in the brains of several mouse models containing GCase variants were evaluated by immunohistochemistry. Progressive α-synuclein and ubiquitin aggregate accumulations were observed in the cortex, hippocampus, basal ganglia, brainstem, and some cerebellar regions between 4-24 wks in mice that were homozygous for GCase [D409H (9H) or V394L (4L)] variants and also had a prosaposin hypomorphic (PS-NA) transgene. In 4L/PS-NA and 9H/PS-NA mice, this was coincident with progressive neurological manifestations and brain glucosylceramide accumulation. Ultrastructural studies showed electron dense inclusion bodies in neurons and axons of 9H/PS-NA brains. α-Synuclein aggregates were also observed in ventricular, brainstem, and cerebellar regions of older mice (>42-wk) with the GCase variant (D409H/D409H) without overt neurological disease. In a chemically induced GCase deficiency, α-synuclein aggregates and glucosylceramide accumulation also occurred. These studies demonstrate a relationship between glucosylceramide accumulation and α-synuclein aggregates, and implicate glucosylceramide accumulation as risk factor for the α-synucleinopathies.
Gaucher disease, an autosomal recessive disorder, is a common lysosomal storage disease [1, 2] that results from insufficient activity of acid β-glucosidase (GCase, encoded by the GBA1 gene) and resultant accumulation of its substrates glucosylceramide and glucosylsphingosine. Accumulations of glucosylceramide and glucosylsphingosine produce the visceral and CNS manifestations by as yet ill-defined mechanisms. Classically, the three clinical phenotypes include the “non-neuronopathic” (type 1), and “neuronopathic” (type 2 and 3) variants [1, 3, 4]. In all variants, glucosylceramide engorged visceral macrophages or “Gaucher cells” are a hallmark of the disease. In the “neuronopathic “variants, the CNS pathology includes neuronal cell death.
The distinction between the “nonneuronopathic” and “neuronopathic” variants has become somewhat blurred. Recent studies showed that Parkinson disease patients have a 3- to 7-fold increased risk of being heterozygous for GBA1 variants [5-13]. Such associations with Parkinsonism are not related to specific GBA1 mutations . Parkinsonism also occurs in Gaucher disease type 1 or 3 patients and in heterozygotes for GBA1 mutations [6-8, 11], but the risk of GBA1 heterozygotes of developing Parkinsonism is unknown. Parkinson signs and symptoms include memory loss, resting tremor, uncontrolled movements, kinetic rigidity syndrome, asymmetric onset, horizontal myoclonus, supranuclear gaze palsy, typical progression rigidity, difficulty ambulating, and bradykinesia [5-13, 15]. This spectrum of manifestations is similar in persons with or without GBA1 mutations, but can be more severe in their presence. These observations indicate that mutant GBA1 even in heterozygotes, is a significant risk factor for potentiating the effects of Parkinsonism. Neither the basis for these effects or the general pathology and their relationships to glucosylceramide accumulation are known.
These findings contrast with the neuropathology of Gaucher Disease types 2 and 3 in which neuronal loss and degeneration are the most consistent findings, particularly in the basal ganglia, nuclei of the midbrain, pons and medulla, cerebellum, dentate nucleus and hypothalamus [16-19]. Cerebral cortical laminar necrosis [16, 19] and neuronal loss with astrogliosis [20, 21] also have been reported, but only in some type 2 patients. Importantly, α-synuclein inclusion-associated neurodegenerative lesions (α-synucleinopathies) were reported in similar brain regions from some PD patients who also had Gaucher disease type 1 [7, 15, 22].
α-Synuclein is a small presynaptic cytosolic protein that is abundant in nerve terminals of dopaminergic system. Its normal function is incompletely defined, but it has been implicated in dopamine metabolism and synaptic vesicle homeostasis [23-25]. Mutations in the gene for α-synuclein (e.g., Ala30Pro or Ala53Thr) have been implicated directly in the pathogenesis of Parkinson disease [26, 27], as has the over-expression of a human wild-type α-synuclein [28-30]. The presence of α-synuclein insoluble intracellular aggregates (Lewy bodies) is a feature of Parkinson disease and other neurodegenerative disorders [31, 32]. Although α-synuclein aggregates have been observed in several lysosomal diseases and their animal analogues, only GBA1 mutations show a clear and, potentially direct risk association with α-synucleinopathies and Parkinson disease for this reason these shared clinical and neuropathologic findings suggested that GBA1 mutations or glucosylceramide excess act as contributory risk factors that interfere with the clearance of or promote the aggregation of α-synuclein in some patients.
Here, Gba1 point-mutated mice bearing a prosaposin hypomorph (4L/PS-NA and 9H/PS-NA)  or having CBE induced-GCase deficiency  were used as models of Gaucher disease. Prosaposin is a precursor of the four saposins A, B, C, and D that are essential proteins for the optimal activity of selected glycosphingolipid hydrolyases . Saposin C optimizes GCase hydrolysis of glucosylceramide and other substrates, as well as protecting GCase from proteolytic digestion [35, 36]. Saposin C's protective function accounts for the decreased GCase protein and activity with excess glucosylceramide in the CNS and visceral organs of 4L/PS-NA and 9H/PS-NA mice . Extensive histological and immunohistological analyses of such mice showed a particular pattern of α-synuclein accumulation that implicates mutant GCase and/or excess glucosylceramide in the development of α-synuclein accumulation.
The following were from commercial sources: Conduritol B epoxide (CBE, Calbiochem, San Diego, La Jolla, CA). 4-methyl-umbelliferyl-β-D-glucopyranoside (4MU-Glc; Biosynth AG, Switzerland). Sodium taurocholate and Protease Inhibitor Cocktail (Calbiochem, La Jolla, CA). Triton X-100 (Sigma, St. Louis, MO). Antibody sources are as follows: mouse monoclonal anti-α-synuclein, rabbit polyclonal anti-mouse α-synuclein, rabbit polyclonal anti-ubiquitin and tyrosine hydroxylase (Abcam, Inc. Cambridge, MA). Rat monoclonal to CD68 (Serotec, Raleigh, NC). Mouse monoclonal anti-GFAP (Sigma, St. Louis, MO). Mouse monoclonal anti-β-actin antibody (Sigma, St. Louis, MO). Goat anti-rabbit/rat (Alexa-488, FITC), or Goat anti-mouse-biotinylated antibody with streptavidin-Alexa-610 (Molecular Probes, Irvine, CA). Hybond ECL Nitrocellulose Membrane (Amersham, Piscataway, NJ). 10% Bis-Tris Gel (Invitrogen, Carlsbad, CA). BCA Protein Assay Kit, Peroxidase-Conjugated Goat anti-mouse IgG, M-Per Mammalian Protein Extraction Reagent and ECL Kit (PIERCE, Rockford, IL).
Several GCase, prosaposin, and other lysosomal disease mouse models were analyzed (Table 1). The homozygous point-mutated Gba1 mice, V394L (4L), D409H (9H), and D409V (9V), have normal life spans and no overt CNS disease . Hypomorphic prosaposin mice (PS-NA) and prosaposin knockout mice (PS-/-) had glucosylceramide accumulation in the CNS with CNS degenerative disease [33, 38, 39]. PS-NA mice that were also homozygous for V394L or D409H (4L/PS-NA or 9H/PS-NA, respectively), exhibit greater excesses of glucosylceramide accumulation in the CNS than PS-NA mice  at least in part due to additive deficiencies of GCase activity. These mice were used here as models for chronic effects of GCase deficiency and glucosylceramide accumulation since no other viable models for the CNS variants of GCase deficiency available. Other mice with lysosomal storage diseases were used for comparative purposes and included those with individual saposin deficiencies (saposin A, B, C or D), Niemann-Pick type C1 (NPC1), mucopolysaccharidosis type I (MPS I), and lysosomal acid lipase (LAL) deficiency. NPC1 is characterized by intracellular accumulations of free cholesterol and glycosphingolipids, including glucosylceramide, in the endosomal/lysosomal compartment . MPS I, α-L-iduronidase deficiency, leads to defective catabolism of the glycosaminoglycans heparan and dermatan, sulphate, and have secondary accumulations of GM2 in neurons . Lysosomal acid lipase is essential for the hydrolysis of the triglycerides and cholesteryl esters in lysosomes and deficient mice do not exhibit overt CNS disease during their 8-10 month lifespan . Age-matched wild-type mice in the FVB background were from Jackson Laboratory (FVB/NJ, Stock No 001800). The mice were maintained in microisolators in accordance with institutional guidelines under IACUC approval at the animal facility of Cincinnati Children's Hospital Research Foundation. Brains were collected from each genotype at the designated times for histological and biochemical analyses.
For histological studies, 3 to 4 mice from each genotype at the indicated ages, a minimum of 2 sections from each tissue were examined. Brains were dissected after perfusion with 1 × PBS, pH 7.4 and 4% paraformaldehyde in 1 × PBS, pH 7.4. The brains were post fixed overnight in 10% formalin or 4% paraformaldehyde, and then processed to paraffin or frozen block, respectively. Karnovsky's fixative was used for ultrastructural studies. Serial coronal and sagittal sections of 8 μm thickness were processed for H&E, silver staining, and immunohistochemistry analyses. For immunofluorescence, paraffin sections were deparaffinized in xylene and frozen sections were fixed in 100% acetone at -20°C for 10 min. The sections were blocked with 10% goat serum/PBS plus 0.4% Triton X-100, and then incubated with primary antibody against α-synuclein (1:200), ubiquitin (1:500), glial fibrillary acidic protein (GFAP, 1:500), CD68 (1:300), or tyrosine hydroxylase (1:300) followed by successive incubations with the appropriate biotin-conjugated anti-rabbit or anti-mouse antibodies. The signals were detected by Alexa-488, FITC, or streptavidin-Alexa-610. Antibodies used for double staining were tested individually and showed similar signal intensities. Images were captured with a Zeiss Apotome microscope (AxioV200) at excitation of 506 nm (for Alexa-488, FITC) or 590 nm (for Alexa-610, Rhodamine). In general, α-synuclein signals (Alexa-610, red) were captured with a green background (with FITC filter) to distinguish the specific signals from autofluorescence, except in dual antibody studies. Cell nuclei were stained with DAPI (blue). For negative controls, respective sections were processed for immunoreactivity without primary antibody or with pre-immune (mouse or rabbit) serum as the primary antibodies. The signal locations in brain sections were according to The Mouse Brain in Stereotaxic Coordinates . For immunohistochemistry study, brain sections were incubated with rat anti-mouse CD68 monoclonal antibody and detected with ABC Vectastain and DAB peroxidase substrate .
The cerebral cortex tissues from 4L/PS-NA, 9H/PS-NA and WT brains were microdissected and homogenized (1:5, wt/vol) in Mammalian Protein Extraction Reagent (M-Per) with a douncer (10-12 strokes). Cortex lysate protein (100 μg) was subjected to 10% SDS-PAGE. Immunoblots were pre-treated with 3% BSA in PBS-0.05% Tween-20 for 1 h and then incubated with the primary antibody, mouse monoclonal anti-α-synuclein, (1:1,000) in PBS. The blot was then incubated with peroxidase-conjugated goat anti-mouse IgG, in 1% BSA-PBS (1:1,500) and developed using ECL kit. Pure α-synuclein protein (~ 14 KD) was used as molecular weight reference. The developed immunoblots were stripped, and probed with anti-mouse β-actin as a loading control.
Gba1 point mutated 4L, 9H and 9V mice or WT mice were intraperitoneally injected with 100 mg/kg/day of CBE [34, 44]. In short-term experiments, daily injections were initiated at postnatal day 5 and continued for 6 daily doses. The mice were sacrificed on day 12 or 2 months after last injection. In long-term experiments, 4L mice were injected daily beginning at postnatal day 15 for 24 or 36 daily doses and sacrificed the day after last injection. Mice were perfused with PBS and organs were harvested for enzyme activity, lipid and histological analyses.
For GCase activities, tissues (~50 mg) were homogenized [1:10 (mg: μl) in 1% sodium taurocholate/Triton-X100 (TC/TX), sonicated (4°C, 30s × 3), and clarified (10,000×g; 5 min, 4°C) . The assay mixtures were pre-incubated in the presence or absence of CBE (1 mM; 30 min at room temperature), and then substrate (4MU-Glc) was added and incubated (1 h, 37°C). Protein concentrations were determined using the BCA kit according to manufacturer's instructions.
Anesthetized mice were perfused with 3 blood volumes of normal saline. The brains were immediately microdissected to isolate the cerebral cortices, cerebella, hippocampi, and brainstems. Brain tissues were homogenized in 1% TC/TX and protein concentrations were determined using the BCA kit. Glucosylceramide and glucosylsphingosine in the tissue extracts were analyzed by LC/MS (Lipidomic Core, University of South Carolina). The glucosylceramide and glucosylsphingosine amounts were normalized to the extract protein in each sample and expressed as pmol/mg protein. Three mice for each genotype were included in the analyses.
The homozygotes point-mutated Gba1 (4L, 9H, and 9V) have 22-27% of WT GCase activity in the CNS as assayed with the 4-MU-Glc substrate, and no CNS glucosylceramide accumulation by TLC, or gross CNS phenotypes (Table 1) . Crossing the 9H and 4L mice into hypomorphic prosaposin (PS-NA) mice led to severe neurological phenotypes and foreshortened life spans (~22 wks). The resultant 9H/PS-NA and 4L/PS-NA mice appeared phenotypically normal until ~10 wks, and then developed progressive ataxia, generalized tremor, and gross shaking to the point of falling over. Lack of rear limb control progressed to nearly complete paralysis by ~20 wks. The lack of body control eventually resulted in the inability to eat and drink, with consequent malnutrition, dehydration, and death (~22 wks). Histological studies of 9H/PS-NA brains showed degenerating neurons in the most regions of the brain including cerebral cortex (Fig. 1), which was highly similar to the changes in the PS-NA and 4L/PS-NA mice . Ultrastructural analyses revealed electron dense amorphous inclusions in the cytoplasm of neurons and in axons of 9H/PS-NA cortex (Fig. 1D and F, arrows). Brain glucosylceramide and glucosylsphingosine levels in 4L/PS-NA and 9H/PS-NA mice were higher than those in age-matched WT or PS-NA mice (Fig. 2A). In cerebellum, glucosylceramide levels were ~4-, 10-, or 6-fold higher in PS-NA, 4L/PS-NA, or 9H/PS-NA mice, respectively, than those in WT mice (Fig. 2A). In the hippocampus and brainstem, glucosylceramide levels were ~2 (9H/PS-NA mice) to 4-fold (4L/PS-NA mice) higher than those in WT mice; PS-NA mice showed almost WT levels (Fig. 2A). Glucosylsphingosine levels were increased from 6 to 28- fold in 9H/PS-NA brains and ~20 to 117-fold in 4L/PS-NA brains (Fig. 2B), whereas levels in PS-NA mice were comparable to WT levels. These biochemical and neurological findings in the 9H/PS-NA and 4L/PS-NA mice provided data for correlation of pathologic findings and potential insights of relationships between Gaucher disease and Parkinsonism.
Serial sagittal brain sections from 10 to 12-wk 9H/PS-NA, 4L/PS-NA, and PS-NA mice were examined by immunofluorescence microscopy to provide an overview of brain α-synuclein distribution. Such brain sections from 9H/PS-NA mice showed α-synuclein signals (Alexa-610, red) concentrated in two areas: 1) Deep cerebral cortical layer: the area overlaying the corpus callosum (cc) from dorsal tenia tecta (DTT) and ventral orbital cortex (VO) through dorsal peduncular cortex (DP), prelimbic cortex (PrL), cingulate cortex area 1 (Cg1), and in the secondary motor cortex (M2) to the retrosplenial agranular cortex (RSA) and retrosplenial granular cortex (RSGa) (Fig. 3 and supplemental Fig. 1). α-Synuclein signals were in ~5-20 μm aggregates within these regions in 9H/PS-NA (Fig. 3A), A FITC filter was used to make the background green allowing visualization of the brain fields, and to distinguish the authentic signals (red) from auto-fluorescence (yellow or green) caused by lipofuscin. In the hippocampal region small (2-5 μm) dense α-synuclein particles were observed in CA3 region of 9H/PS-NA brains (Fig. 3B), in contrast to larger α-synuclein particulates in the cortex. Only weak and very small α-synuclein particles (<2 μm) were seen in these regions from age matched WT mice (Fig. 3C and D).
In comparison to 9H/PS-NA mice, large α-synuclein particles (≥5 μm) were observed in cerebral cortex of 10-12 wk PS-NA and 4L/PS-NA mice (Fig. 4, arrows), but the signals were much weaker in PS-NA brains. Similar results were observed in the 18-wk age group (data not shown), No α-synuclein aggregates were observed in the same regions of age-matched WT mice (Fig. 4). Immunoblots for α-synuclein showed the presence of monomers in the cerebral cortex of 18-wk 4L/PS-NA, 9H/PS-NA and WT mice. However, two forms of α-synuclein oligomers were present in the cortices of 4L/PS-NA and 9H/PS-NA mice (Fig. 5).
Serial coronal sections of 9H/PS-NA and 4L/PS-NA brains were examined by immunofluorescence staining from Interaural 1.98 mm/Bregma -1.82 mm to Interaural 0.40 mm/Bregma -3.40 mm . This analysis included the basal ganglia region and brainstem where the neuronal lesions had been found in Gaucher disease type 2 and Parkinson disease [46, 47]. In the coronal sections, the regions containing α-synuclein-positive aggregates are indicated with text-abbreviations in Figure 6A (Interaural 1.98 mm/Bregma -1.82 mm) or 6B (Interaural 0.40 mm/Bregma -3.40 mm) that were constructed from 18 to 20 serial black-white images taken prior to processing for immunofluorescence. In the cortex, many α-synuclein aggregates were observed near the external capsule (ec) and near the temporal association cortex (TeA), Piriform cortex (Pir) (Fig. 6A), and retrosplenial agranular cortex (RSA), retrosplenial granular cortex (RSG), secondary visual cortex mediolateral (V2ML), primary visual cortex (V1), secondary visual cortex lateral (V2L), primary auditory cortex (Au1), cingulum (cg), perirhinal cortex (PRh), lateral entorhinal cortex (LEnt) (Fig. 6B). The α-synuclein-positive particles were also observed in the amydaloid nuclear complex, i.e., the lateral amydaloid nucleus, dorsolateral (LaDL) and lateral amydaloid nucleus, ventromedial (LaVM), basomedial amydaloid nucleus posterior (BMP), medial amydaloid nucleus posterodorsal (MepD), and medial amydaloid nucleus posteroventral (MepV) (Fig. 6A). In brainstem, α-synuclein-positive aggregates were seen at the area of cerebral peduncle basal part (cp), subthalamic nucleus (STH), ventral posteromedial thalamic nucleus (VPM), posterior thalamic nuclear group (po), paraventricular thalamic nucleus (PV), central medial thalamic nucleus (CM) (Fig. 6A) and substantia nigra (SN), medial geniculate nucleus ventral part (MGV), interstitial nucleus of Cajal (InC), red nucleus parvicellular part (RPC) interpeduncular nucleus rostral subnucleus (IPR), interpeduncular nucleus caudal subnucleus (IPc) (Fig. 6B). In the hippocampal region the signals were localized to lateral habenular nucleus medial part (LHbM), fasciculus retroflexus (fr) (Fig. 6A), lacunosum molecular layer (LMo), stratum radiatum (Rad), molecular layer of dentate gyrus (MoL), pyramidal cell layer (Py) and subiculum (Fig. 6B).
The α-synuclein aggregates in the brain were evaluated by silver staining since this has been used to identify pathological deposits in studies of Parkinson disease [48, 49]. In general, brain sections from 9H/PS-NA and 4L/PS-NA mice stained darker than those from control mice. The silver stained neurons were observed at multiple regions in 9H/PS-NA brains, but were less intense in the brains of 4L/PS-NA mice. Dense dark particles were found in the cytoplasm around the nuclei of neurons (Fig. 7A), suggesting aggregated proteins in the diseased mouse brain. To further characterize the silver stained aggregates, consecutive sections were probed with anti-α-synuclein and anti-ubiquitin antibodies. Some of these silver stained neurons showed α-synuclein (Fig. 7B) or ubiquitin (Fig. 7C) immunoreactivity positivity indicating intracellular aggregation of these proteins.
To characterize the developmental stages of α-synuclein accumulation, brain sections from 9H/PS-NA, 4L/PS-NA, PS-NA and WT mice at 4, 10-12, and 18-20 wks were examined. In general, fine α-synuclein signals (~1 μm) were seen in brain of WT and diseased mice at all age groups. α-Synuclein positive particles were observed in 4-wk 9H/PS-NA brain, [e.g., 1-2 μm in OB, Cb and dentate gyrus; 3-5 μm in CPu (data not shown)]; and a few larger (6-7 μm) α-synuclein particles were found in cortex at that age (not shown). The aggregated α-synuclein particles became more widely distributed with increasing age, especially in 9H/PS-NA mice (Fig. 8). Larger amounts of α-synuclein particles (5 - 30 μm) were seen in 9H/PS-NA brains at 10-or 20-wks (Fig. 8), and the particle sizes increased with age. In PS-NA or 4L/PS-NA mice, a few of α-synuclein positive particles (2-4 μm or 4-6 μm) were seen in the cortex at 4 to 20 wks (data not shown).
The relationships between α-synuclein and ubiquitin were studied in 9H/PS-NA (Fig. 9), 4L/PS-NA, PS-NA and WT mice. α-Synuclein aggregates appeared in brains at 4-wks, but the ubiquitin aggregates were detected at ~10-12 wks. In the brain of 10-wk 9H/PS-NA, 4L/PS-NA or PS-NA mice, sporadic ubiquitin signals were observed in cortex, brainstem and midbrain region (data not shown). Larger sized (≤18 μm) ubiquitin signals were present in the brainstem, medulla, midbrain and cortex of 20-wk 9H/PS-NA mice (Fig 9D); the overlap of ubiquitin and α-synuclein signals was partial (Fig. 9F, arrows). A few small ubiquitin signals were observed in PS-NA and 4L/PS-NA brains (data not shown).
Strong and dense GFAP staining, indicating astrogliosis, was in the cortex and ventricular area of 10-wk 9H/PS-NA mice (Fig. 10 top panels), but without colocalization with α-synuclein. α-Synuclein signals did not co-localize with CD68 signals (Fig. 10 lower panels), but some were embedded at the center of CD68 positive microglial cells, which suggested the phagocytosis of α-synuclein by microglial cells. These immunofluorescence studies indicated that α-synuclein was not present in astrocytes or microglial cells. Tyrosine hydroxylase-positive signals were mostly seen in basal ganglia and brainstem regions, but no significant differences were found between diseased and WT brains (not shown). In the 9H/PS-NA, 4L/PS-NA and PS-NA mice, no obvious relationships of tyrosine hydroxylase and α-synuclein signals were observed.
α-Synuclein aggregation was examined in selected Gba1 point-mutated mice (4L, 9H and 9V). Aggregated α-synuclein signals were only found in cerebellar and brainstem regions of 42-46 wk 9H mice, but not in age-matched 4L, 9V, or WT brains (Fig. 11). A GCase irreversible inhibitor, CBE, was used to generate mice with neuronopathic disease  in an effort to distinguish the potential for gain (malfolding) or loss (activity absence) of GCase. Short-term (6 day) CBE treatment of 4L, 9H, 9V and WT mice (100 mg/kg/day, intraperitoneal injection) from postnatal day 5 to 11 did not induce α-synuclein aggregates, either when harvested one day or 60 days after the last CBE injection. These mice did develop varying degrees of neuronopathic disease, e.g., shaking and paralysis after 6 injections . Long-term daily CBE treatment of 4L mice (100 mg/kg/day of CBE from postnatal day 15 for 24 or 36 doses) led to hind limb paralysis and small amounts of α-synuclein accumulation in the olfactory bulb, brainstem, and PVP (paraventricular th nu, posterior) near D3V (dorsal 3rd ventricle) (Fig. 12). In these mice, glucosylceramide accumulation in brain can be appreciated by TLC analyses as compared to untreated 4L or WT mice (data not shown). Interestingly, by LC/MS brain glucosylsphingosine levels in these mice were 20- to 30-fold higher than those in untreated 4L mice (not shown).
Several lysosomal storage disease models were used as controls for these studies as summarized in Table 1 and below. Ubiquitin was found in the cortex and midbrain of 4-wk PS-/- mice, but α-synuclein aggregates were not observed. Sporadic 4-8 μm α-synuclein aggregates were present in the cortical, cerebellar, and brainstem areas of NPC1. Occasional larger (≤ 15 μm) α-synuclein and ubiquitin aggregates were present in the CPu area of NPC1 mice (data not shown). In MPS1 mice, ubiquitin aggregates >10 μm were in the cortex, SNC, OB, and brainstem areas; most α-synuclein aggregates were in SNC (≤10 μm), and cerebellum and brainstem (~5 μm) (data not shown). In brains from LAL knockout or saposin (A, B, C, or D) deficient mice, α-synuclein and ubiquitin aggregates were not found (data not shown).
Here, α-synuclein expression was examined in various Gba1 mutated mice with or without CNS phenotypes including: Gba1 point mutated mice, and hypomorphic prosaposin mice (PS-NA), with or without 9H or 4L, and in CBE - treated Gba1 mutated mice. The 9H/PS-NA mice had the most significant α-synuclein accumulation, particularly in the deeper layers of the cortex. Brain lipid analyses showed moderately increased levels of glucosylceramide in the cortex and other regions compared with PS-NA and point mutated Gba1 models; this suggested a correlation of glucosylceramide accumulation with significant α-synuclein aggregation. In comparison with 9H/PS-NA mice, similar glucosylceramide levels and neuronopathology, but less significant α-synucleinopathy were observed in 4L/PS-NA mice. This result indicates that glucosylceramide levels are one factor in the development of α-synucleinopathy and that additional pathophysiologic mechanisms are needed, e.g., the in vivo properties of mutant GCases, (4L or 9H). Interestingly, glucosylsphingosine accumulation in 4L/PS-NA is much higher (10 times) than that in 9H/PS-NA, but with greater α-synuclein accumulations in the latter. Thus, glucosylsphingosine is not likely to play a major role in the α-synucleinopathy, but rather it contributes to neuronal and cell death. The rapid neuronal degeneration and death induced by glucosylsphingosine may not support the time needed for α-synuclein aggregate deposition. A complex and chronic mechanism is implied for α-synucleinopathy development.
The chronicity of the α-synuclein accumulation is supported by its age-dependency in the 9H/PS-NA, 4L/PS-NA, and Gba1 point-mutated mice. Low levels of α-synuclein accumulation were detected as early as 4-wks in 9H/PS-NA mice with progression to substantial aggregates by 12-20 wks. Immunoblot analysis confirmed the formation of α-synuclein oligomers in 9H/PS-NA and 4L/PS-NA cortices that contribute to the aggregates. The short-lived kn-9H mice (14 days)  that were GCase null in the CNS and had glucosylceramide accumulation in the brain, did not have α-synuclein aggregates (data not shown). In CBE induced GCase deficiency in adult mice, α-synuclein aggregates only developed after long-term treatment. These observations indicate that α-synuclein aggregation is of a chronic progressive nature, consistent with the time-dependent formation of α-synuclein aggregates following an in vitro initial nucleation event [50, 51]. α-Synuclein aggregates could be formed under a number of conditions including simple solutions, incubation with lipids, in cell cultures, or in animal models of disease . However, in vivo systems are complicated by interactions among multiple proteins/biomolecules involved in the formation of inclusions [52-54]. Here, ultrastructural analyses showed numerous inclusion bodies in neurons and the structure of these inclusions suggested aggregated proteins.
Various patterns of α-synuclein aggregation were observed in brains from the Gba1 mouse variants in these studies, particularly in 9H/PS-NA. Numerous large α-synuclein aggregates were present in the deep layers of cortex overlaying corpus callosum, whereas fine α-synuclein particles were in the hippocampal CA3 and dentate gyrus. The patterns of α-synuclein aggregates are similar to pathologic findings in Gaucher disease patients who had Parkinsonism. However, such patients characteristically exhibit Lewy bodylike α-synuclein inclusions with gliosis in cortical layers 3 and 5 (deeper layers) and in the hippocampal neurons (CA2-4 regions) . Similar aggregates were detected in 4L/PS-NA and 9H/PS-NA brain, but Lewy bodies were not observed in the current studies.
The relationships of α-synuclein accumulation and ubiquitin, astrocyte (GFAP), dopaminergic neuron (tyrosine hydroxylase), and microglial cells (CD68) were examined in brain sections. In 9H/PS-NA or 4L/PS-NA and PS-NA mice, ubiquitin partially colocalized with α-synuclein signals. In human samples, α-synuclein positive Lewy bodies were generally ubiquitin positive . The differences in co-localization of α-synuclein and ubiquitin between human and mouse suggest variation in the initiation mechanisms between these species. However, the end-stage nature of autopsy specimens from patients with both Gaucher disease and PD patients does not provide insights into the time course of this process. In comparison, 9H/PS-NA mice showed a progressive α-synucleinopathy that was detected as early as 4-wks, whereas ubiquitin aggregates were seen almost exclusively after 10 wks. The precise role of α-synuclein in the Lewy body formation or in the neurodegenerative processes is unclear. Lewy bodies are complex structures and contain multiple proteins [52, 53]. However, GCase has recently been found in Lewy bodies in patients heterozygous for GBA1 mutations suggesting a role for the GCase in their formation . In 9H/PS-NA or 4L/PS-NA brains, massive astrogliosis, and activated microglial cells were present, but signals for GFAP (astrocytes) or CD68 (microglial cells) colocalized with α-synuclein only occasionally. Interestingly, a few α-synuclein signals localize near the center of CD68 positive microglial cells (Fig. 10, CD68 positive), suggesting a microglial cell response to aggregated proteins in the brain. These observations indicate that α-synuclein was not expressed to a great extent in either astrocytes nor in microglial cells, but rather in neurons. Tyrosine hydroxylase expression in the mice used here was not significantly different from WT control mice.
Screening of several other LSD mouse models confirmed the presence of small α-synuclein aggregates in brains from NPC1 and MPS1 mice (data not shown), but not in LAL or individual saposin deficiency mice. NPC1 mice have a defect in a cholesterol efflux protein in the membrane of the lysosome [56-58]. However, glycosphingolipids, including glucosylceramide and gangliosides, also accumulate in NPC1 brain. MPS1 mice accumulate glycosaminoglycans heparan and dermatan sulfate, as well as GM2 and GM3 . These observations support the contention that glycosphingolipids and gangoliosides contribute to the pathogenesis of these diseases and may increase the risk for developing α-synuclein accumulations. In Sandhoff mice, neuronal accumulations of α-synuclein and GM2 ganglioside were present throughout the brain . α-Synuclein aggregates are also found in cathepsin D deficient mice, sheep, and humans. This is potentially due to the role of cathepsin D in α-synuclein processing. α-Synuclein and ubiquitin signals in NPC1 and MPS1 brains were significantly less than those in 9H/or 4L/PS-NA brains, some specificity for the α-synucleinopathy. Specific saposin A, B, C or D deficiencies in mice produced no α-synuclein and ubiquitin accumulation. The accumulated GSLs in each of these models are different, e.g., galactosylceramide and galactosylsphingosine with saposin A deficiency , sulfatide with saposin B deficiency , lactosylceramide with saposin C deficiency , and ceramide with saposin D deficiency . In total saposin deficient mice, (i.e., prosaposin knockout) moderate increases of glucosylceramide and positive anti-ubiquitin antibody staining were observed in neurons . These studies suggest that the respective sphingolipids or other saposin effects in the saposin deficient mice are not causally related to α-synuclein and ubiquitin accumulation. With the hypomorphic prosaposin-deficiency combined with low activity GCase mutations (4L/PS-NA or 9H/PS-NA), α-synuclein accumulation was observed as early as 4-wks. These studies suggest a more complicated relationship between glucosylceramide, GCase mutations, GSL accumulations, and the risk for α-synuclein/ubiquitin accumulations.
Many LSDs are associated with progressive and severe neurodegeneration. To date, the diverse biological pathways from lysosomal enzyme deficiency to neuron dysfunction and death have not been clarified . In the mouse models of multiple sulfatase deficiency (MSD) and mucopolysaccharidosis type IIIA (MPSIIIA), defective autophagosome-lysosome fusion was observed with severe neurodegeneration . In Huntington disease or familial of Parkinson disease, impaired autophagocytic pathways led to aggregation-prone proteins, which cause late-onset neurodegenerative conditions . However, knockout of selected autophagy genes caused abnormal protein accumulation in ubiquitinated inclusions and neurodegeneration in mice [69, 70]. These phenotypic similarities suggest the presence of common pathogenic mechanisms for LSDs with neurodegeneration .
Clinically, earlier onset and more progressive PD patients had heterozygous mutations in GBA1 [5-13]. However, Parkinsonism manifestations [6-8, 11] and α-synuclein/ubiquitin aggregates [7, 15, 22] were found in only some GD type 1 or 3 patients. Thus, the connections between GCase mutations (gain or loss of function or both) and the development of Parkinsonism or PD remain unclear. The current findings in mice provide a biological system and potential clues to the pathogenesis of neuronopathic Gaucher disease including the α-synuclein distribution patterns in Gba1 mutant mice. Ongoing studies should provide insight into the underlying neuronopathic mechanisms in Gaucher disease and the relationship between GCase deficiency and/or GBA1 heterozygosity and the development of Parkinsonian manifestations.
Supplemental Fig 1: Distribution of α-synuclein in sagittal sections of brain cortex from 9H/PS-NA mice. The figure was composed from ~70 images using 20×. Serial sagittal images were from 10-wk 9H/PS-NA mice and examined by immunofluorescence staining using anti-α-synuclein antibody conjugated with Alexa-610 (red). The abbreviations from left to right are as follows: VO, ventral orbital cortex; DTT, dorsal tenia tecta; Cg1, cingulate cortex area 1; PrL, prelimbic cortex; DP, dorsal peduncular cortex; gcc, genu of corpus callosum; Cpu, caudate putamen; M2, secondary motor cortex; LV, lateral ventricle; fi, fimbria of Hippocampus; cc, corpus callosum; CA2-3, hippocampus CA2-3 fields; RSA, retrosplenial agranular cortex; dentate gyrus (DG); LHb, lateral habenular nu; scc, splenium corpus callosum; RSGa, and retrosplenial granular cortex .
The authors thank Michelle Cooley for her clerical expertise, and Venette Inskeep for her excellent technical assistance; Drs. Hong Du and Dao Pan for providing LAL-/- and MPSI mouse brains, respectively. Lisa McMillin, Meredith Farmer and Sabina Sylvest for skilled tissue preparation; Irene Hofmann for EM preparation; Chris Woods for photograph processing. This work was supported by a grant from the NIH (DK 36729) to GAG.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.