We report the first comprehensive biochemical analysis of the effects of GBA
mutations in PD brains. There is widespread deficiency of GCase activity and protein levels in PD+GBA brains, with the most severe defect located in the SN (58%) and putamen (48%). This loss of GCase is unlikely to simply represent neurodegeneration, as there was a 47% defect in cerebellum, an area not involved in the degenerative process of PD. GCase activity was also unaffected in the amygdala of AD brains, a region associated with marked neurodegeneration. There was no decrease in GCase mRNA levels in the putamen, indicating that the reduction in protein levels, in this region at least, was not a result of downregulation of gene expression, but rather a post-translational effect on protein levels. Three markers of UPR/ERAD were increased and might play a role in reduced protein levels, although further studies are required to confirm this. Given the normal cathepsin D levels and hexosaminidase activity, the loss of GCase protein was not due to a general reduction in lysosomal content or activity. There was a trend for LC3-II levels to be increased in PD+GBA putamen, suggesting an increase in autophagosome number, a feature previously reported in PD SN and amygdala.10, 11
As autophagy flux cannot be measured in postmortem tissue, it is unclear whether this is due to an increase in macroautophagy or an inhibition of lysosomal degradation of autophagosomes.
We also report for the first time a significant deficiency of GCase activity in PD SN (33%) and cerebellum (24%), with reduced protein levels, in sporadic PD brains. Enzyme activities were comparable to control in other PD brain regions. The patterns of lysosomal proteins and UPR/ERAD markers were similar in PD and GBA+PD brains.
The SN is the site of greatest pathology in PD brain and biochemical abnormalities thought relevant to PD pathogenesis, including α-synuclein deposition, mitochondrial dysfunction, and oxidative stress. Although the loss of GCase activity in PD+GBA brains is in part related to GBA
mutations, this cannot be the explanation in PD brains. Thus, we hypothesized that mitochondrial dysfunction and/or oxidative stress, or increased α-synuclein levels might cause this defect in PD and an exacerbation of the GCase deficiency in PD+GBA. Mitochondrial dysfunction/oxidative stress as a result of PINK1 KD28, 29
caused a loss of GCase activity and protein in cells. Increased α-synuclein expression also significantly decreased GCase activity and protein levels.
Recent studies have highlighted the reciprocal relation between GCase activity and α-synuclein. Decreased GCase activity caused increased α-synuclein levels in toxin and transgenic GBA mouse models, and in neuronal cultures.14–18
Accumulation of the GCase substrate glucosylceramide can stabilize soluble α-synuclein oligomers in vitro,18
and oligomeric α-synuclein has been reported in patients with homozygous or heterozygous GBA
mutations with Lewy body dementia, although not in PD.18, 31
No statistical differences in Lewy body pathology were reported between sporadic PD brains and PD+GBA brains studied here.8
We also found no noticeable difference in the solubility or amount of monomeric α-synuclein between PD+GBA brains and sporadic PD. Differences in α-synuclein conformation between these groups warrants further investigation.
GCase has also been shown to bind α-synuclein at lysosomal pH, and increased α-synuclein levels in the cortex can result in the depletion of lysosomal GCase.18, 19
Our in vitro data suggest that increased sensitivity of GCase to endo-H in cells with increased α-synuclein results in the aberrant trafficking of GCase through the ER. α-Synuclein can affect ER/Golgi apparatus trafficking,32, 33
but the mechanism by which GCase transport is reduced remains unclear. LIMP-2, the protein required for GCase transport to the lysosome,30
does not bind α-synuclein. However, less GCase is bound to LIMP-2 in α-synuclein–overexpressing cells, and therefore the amount of enzyme delivered to the lysosome is decreased. LIMP-2 levels were unaffected in the SN of PD brains. The decrease in steady-state GCase mRNA levels observed in vitro needs further investigation to determine whether this is a decrease in transcription or increased degradation of mRNA, and the extent to which this contributes to GCase deficiency.
Accumulation of proteins in the ER can induce UPR/ERAD. The 2 most common heterozygote GBA
mutations associated with PD (N370S and L444P)7
have been reported to undergo UPR/ERAD in cultured cells.24, 25, 34
However, 2 Gaucher mouse studies could not find evidence of UPR/ERAD in the brain.17, 35
UPR/ERAD markers were increased in both PD brains with GBA
mutations and sporadic PD. Aberrant trafficking of GCase might contribute to this increase in UPR/ERAD. However, it could also be a result of perturbed calcium homeostasis, redox status, or proteostasis or mitochondrial dysfunction, all of which are linked with PD pathogenesis.26, 36
Surprisingly, GCase activity was decreased in the cerebellum of PD and PD+GBA brains, although this is not a site associated with neurodegeneration, α-synuclein accumulation,37
mitochondrial dysfunction, or oxidative stress in PD.38
It is unclear why GCase activity is decreased in this region. We speculate that deficiency in the cerebellum is caused by a mechanism separate from that occurring in recognized pathogenic areas of PD such as the SN, putamen, and amygdala.
In conclusion, our studies confirm a widespread deficiency of GCase activity in the brains of PD patients carrying GBA mutations. We also demonstrate that PD patients without GBA mutations exhibit deficiency of GCase in SN. Based upon the relation between GCase, α-synuclein, and mitochondrial function shown here and by others, we propose that PD pathology is exacerbated and accelerated, but not necessarily initiated, by GBA mutations (). This would explain why not all GBA mutation carriers develop PD, and why those who do tend to do so at an earlier age.