Selective loss of glucocerebrosidase activity in sporadic Parkinson’s disease and dementia with Lewy bodies

The identification of the underlying causes of Parkinson’s disease (PD) is a major challenge, since in most instances, cases present with sporadic disease. The discovery of the genetic determinants of PD would be a major step forward in describing the underlying etiology of the disorder and also in identifying possible therapeutic strategies [1]. Mutations in the GBA gene, encoding for the lysosomal hydrolase β-glucocerebrosidase (GCase, EC = 3.2.1.45) cause Gaucher disease (GD) a rare lysosomal storage disorder, and represent a common risk factor for PD [2-4]. Patients carrying GBA loss of function mutations have a five-fold increased risk of developing PD with respect to non-carriers [3], and may show a similar phenotype to idiopathic PD. The GCase substrate, glucosylceramide, stabilizes soluble oligomeric α-syn species [5]. As a consequence, GCase deficiency may contribute to α-syn aggregation and accumulation. GCase deficiency has been identified in post mortem brain from patients diagnosed with PD either with or without GBA mutations, the decrease being most evident in the substantia nigra, cerebellum and cortex of PD patients [6,7]. This non-selective loss of GCase activity irrespective of mutation status may either represent a global defect in lysosomal enzymes in PD, or may simply represent the presence of other pathologies such as neuronal loss leading to deficiency.

The activity of GCase and other lysosomal enzymes can also be reliably measured in cerebrospinal fluid (CSF) [8]. The reduction of GCase activity has been found also in the cerebrospinal fluid (CSF) of PD patients when compared to neurological controls [9-11] and in fibroblasts from patients with GD or PD carrying GBA mutations [12]. Interestingly, other lysosomal enzymes involved in different degradation pathways, were also found to be altered in PD patients diagnosed with PD. An increased β-hexosaminidase activity, a lysosomal enzyme able to hydrolyze the GM2 ganglioside, has been found in CSF and fibroblasts of PD patients [11,12]. Another study showed increased activity of cathepsin E and β-galactosidase in CSF of de-novo PD patients while α-fucosidase activity was significantly decreased [13]. Together, these findings suggest that there may be a more widespread dysfunction of lysosomal enzymes in PD and related disorders.

Here we determined the specific activity of several endo-lysosomal enzymes, namely β-hexosaminidase, α-fucosidase, β-mannosidase, α-mannosidase, β-galactosidase, β-glucocerebrosidase and cathepsin E, in different brain areas of PD patients, age-matched controls and, for some brain areas, of patients diagnosed with dementia with Lewy bodies (DLB). Our aims were to map the activity of these lysosomal enzymes in different human brain regions and also to evaluate the occurrence of lysosomal dysfunction not only in PD but also in another synucleinopathy, DLB. This would provide the necessary information to determine if lysosomal dysfunction is widespread in PD, or if there is a relatively selective loss of GCase.

We utilized a large series of clinically and neuropathologically verified cases of PD and DLB and age matched control cases. In Table 1 the demographic characteristics of the patients included in the study are reported. Post-mortem brain tissue was analyzed for lysosomal enzyme activity, GBA genotype and GBA mRNA expression. GBA sequencing was performed to verify the presence of pathological mutations on GBA gene. The results showed that our cohort was composed mainly of patients and controls without GBA mutations. Only two out of 26 PD patients were heterozygous for pathogenic GBA mutations. One patient carried the L444P mutation while the other one carried the IVS2 + 1G > A mutation. The remaining PD and DLB patients and control subjects were wild type for the GBA gene (CTRL =9, PD = 23, DLB = 15). GBA genotype was not available in 4 controls, 1 PD and 1 DLB. Using hierarchical clustering analysis, we compared the lysosomal enzyme activities across different brain areas (Figure 1a). The substantia nigra and the hippocampus clustered together in the analysis, globally showing a high level of almost all the enzyme activities. GCase was the exception, showing low levels in substantia nigra but only in the pathological groups, while its levels were high in frontal cortex. The caudate clustered independently from the other areas and showed the lowest levels of the measured activities. Notably, the differences among the brain areas were greater than the differences between PD, DLB patients and control subjects.

The comparison of the mean levels of lysosomal enzyme activities in the three groups is reported in Additional file 1: Table S1. We first evaluated the effect of post mortem delay as a confounder in the different comparisons. No significant association was found between postmortem delay and lysosomal enzymes activity when it was considered as a potential confounder (data not shown). GCase specific activity in substantia nigra and caudate was significantly lower in PD patients when compared to control subjects (p < 0.01 and p < 0.05 respectively, Figure 1b). In DLB patients, GCase activity was also reduced, but the difference was not significant (Additional file 1: Table S1). GCase activity did not change significantly in the other brain regions in any of the diagnostic groups (Figure 1b), suggesting a selective loss in the nigrostriatal system. The two patients carrying GBA mutations showed detectable GCase activity, with values within the range of the PD and DLB patients without mutations. Exclusion of the two patients from the analysis did not change the significance of the final results (data not shown).

To evaluate if the decrease in GCase activity was due to a change in gene expression we performed qPCR on GBA mRNA in the substantia nigra. mRNA levels of GBA gene, normalized against GAPDH, were significantly lower in PD and DLB compared to controls (p < 0.05). The same result was found with respect to the subunit A of the succinate dehydrogenase complex (SDHA) (Figure 1c).

Other enzymatic activities were changed in different brain areas. In particular, α-fucosidase activity in frontal cortex was significantly lower in PD compared to controls (p < 0.01), while its activity in DLB patients did not change significantly (Additional file 1: Figure S1). Interestingly, α-mannosidase activity was increased in frontal cortex only in DLB patients (Additional file 1: Figure S1).

We also performed correlation analysis between the enzyme activities and Alzheimer’s neurofibrillary pathology expressed as Braak stage. No significant associations were found for any of the enzyme activities tested (data not shown).