Sorting out Parkinson’s disease: one cell at a time

This scientific commentary refers to ‘Single-cell sequencing of human midbrain reveals glial activation and a Parkinson-specific neuronal state’ by Smajić et al. (https://doi.org/10.1093/brain/awab446).

Parkinson’s disease is the second most common neurodegenerative disorder and is characterized by bradykinesia, tremor, rigidity and postural instability. These motor signs are caused by loss of pigmented midbrain dopaminergic (DA) neurons that project to the dorsolateral putamen. Quantification of dopaminergic neuronal loss often relies on marker genes, such as those encoding tyrosine hydroxylase (TH) and the dopamine transporter (DAT). However, these markers may miss a population of ‘sick’ dopaminergic neurons that have lost TH and DAT expression but could potentially regain function with disease-modifying therapy.^1,2 In this issue of Brain, Smajić and co-workers³ identify CADPS2 (calcium dependent secretion activator 2) as a potential marker of these dysfunctional dopaminergic neurons, in the first single-cell sequencing study of Parkinson’s disease.

Single-cell or single-nucleus RNA sequencing (scRNA-seq or snRNA-seq) reveals the global expression profile of individual cells, and recent advances have enabled sequencing of thousands of cells at a time. By analysing these expression profiles, it is possible to subcategorize neurons expressing the same neurotransmitter into molecularly distinct groups. For instance, the subpopulation of dopaminergic neurons at risk in Parkinson’s disease is now known to have a characteristic molecular profile: CALB1−, SOX6+, ALDH1A1+.⁴ Other analyses have shown that a number of genes, including CADPS2, are preferentially expressed during intermediate stages of dopaminergic neuronal development, with lower expression in mature neurons. However, these single-cell studies of midbrain dopaminergic neurons have thus far all been limited to healthy mouse or embryonic human tissue, with none having examined the Parkinson’s disease brain.

To address this gap, Smajić and colleagues³ generated the first single-cell atlas of Parkinson’s disease, using nuclei isolated from the post-mortem ventral midbrains of six patients with idiopathic Parkinson’s disease and five age- and sex-matched control subjects (Fig. 1A and B). They determined the relative abundance of cell types in the midbrain, established the expression patterns of Parkinson’s disease-associated genes in each cell type, and then used these expression patterns to identify enrichment of Parkinson’s disease risk variants in individual cell types.

Figure 1

Identification of CADPS2 as a potential marker of ‘sick’ dopaminergic neurons in idiopathic Parkinson’s disease. (A) Smajić and colleagues³ performed nuclear isolation and single-nucleus sequencing from post-mortem ventral midbrain sections of six patients with idiopathic Parkinson’s disease (IPD) and five age- and sex-matched control subjects. (B) The single-nucleus transcriptomes displayed a cluster structure driven mostly by cell type. (C) A small neuronal population characterized by strong expression of CADPS2 and weak expression of TH originated almost exclusively from the midbrains of patients with idiopathic Parkinson’s disease. (D) CADPS2 is expressed in midbrain dopaminergic neurons during development and is expressed most strongly at an intermediate developmental stage.

In the first of three key findings, the authors discovered a small Parkinson’s disease-specific neuronal population characterized by high expression of CADPS2 and low expression of TH (Fig. 1C). In a follow-up microdissection experiment, these cells were identified as surviving pigmented midbrain neurons, suggesting that CADPS2 may mark the ‘sick’ dopaminergic neurons identified in prior studies.^1,2

CADPS2 is a calcium-binding protein that mediates activity-dependent exocytosis of dense core vesicles to promote release of neuropeptides, including brain-derived neurotrophic factor and neurotrophin-3. CADPS2 is important in brain development, and mutations in CAPDS2 have been implicated in autism.⁵ Notably, CADPS2 is most strongly expressed at an intermediate stage of dopaminergic neuron development, during which neurotrophic factors play a key role (Fig. 1D).^6,7 CADPS2 was recently found to be upregulated in induced pluripotent stem cell-derived dopaminergic neurons with the Parkinson’s disease-causing LRRK2 G2019S mutation, suggesting that this Parkinson’s disease-related response can be modelled in cell culture.⁸ Taken together, these findings point to upregulation of CADPS2 as a likely response to dopaminergic neuronal injury and as a potentially useful marker for dysfunctional neurons in Parkinson’s disease.

In a second key finding, Smajić and colleagues³ used their rich dataset to characterize the glial response to injury: consistent with prior studies, they found the midbrains of patients with Parkinson’s disease to be enriched in astrocytes and microglia, suggestive of neuroinflammation, and to contain reduced numbers of oligodendrocytes, consistent with injury. Microglia in the midbrains had an amoeboid morphology and could be divided into two groups, each exhibiting an expression pattern consistent with Parkinson’s disease-related activation: a GPNMB^high population that was also enriched for genes associated with cytokine secretion, and an HSP90AA1^high population that was enriched for genes in the unfolded protein response. Thus, two separate microglial populations were activated in Parkinson’s disease, demonstrating heterogeneity in the microglial response to injury.

By comparing the genes that were upregulated in astrocytes in patients versus control subjects, the authors found evidence consistent with a disease-specific response to misfolded proteins, with upregulation of chaperones, such as αB-crystallin. Upregulation of αB-crystallin and other chaperones in the substantia nigra pars compacta of patients with Parkinson’s disease has been shown previously at the protein level and may play a role in the clearance of misfolded α-synuclein.⁹

In the Parkinson’s disease midbrains, oligodendrocytes were decreased in number, but the remaining cells displayed an upregulation of S100B. S100B overexpression is associated with neurodegeneration and may be a response to enhanced cytokine release from the activated microglia and astrocytes. Thus, the findings suggest that inflammatory signalling in Parkinson’s disease may extend from microglia and astrocytes to oligodendrocytes, further reinforcing the importance of neuroinflammation in the disorder.

Finally, Smajić and colleagues³ used cell type enrichment analysis to determine which cell types preferentially express risk genes identified in the recent Parkinson’s disease genome-wide association study meta-analysis, META5.¹⁰ Previously, the only available references for midbrain cell type expression patterns were derived from healthy mouse or human tissue. However, Smajić et al.³ show differing patterns of risk gene expression in single nuclei datasets from Parkinson’s disease versus control brains, with risk genes most enriched in microglia and neurons. Disease-specific single-cell atlases may thus allow more precise association of risk variants with disease-relevant cell types.

The main limitation of this study is the small number of subjects examined. The discovery of differentially regulated genes in specific cell types was likely limited by the small number of neurons sequenced and variability in glial cell activation between subjects. Future studies with larger sample sizes will help establish which subtypes of DA neurons are most affected in Parkinson’s disease, and will reveal disease-specific transcriptional changes in DA neurons and glia.

Considered as a whole, this exciting new study demonstrates the potential of single cell sequencing in Parkinson’s disease. It identifies CADPS2 as a potential marker of ‘sick’ dopaminergic neurons, confirms the importance in neuroinflammation in Parkinson’s disease, and shows that microglia and neurons bear the brunt of Parkinson’s disease genetic risk. In future studies, it will be important to establish whether CADPS2 can be used to track ‘sick’ dopaminergic neurons in cellular and animal models, and, crucially, if the function of these neurons can be rescued by disease-modifying therapy.

Funding

This work was supported in part by the Intramural Research Program of the National Institute of Neurological Disorders, National Institutes of Health: project number 1ZIA-NS003169.

Competing interests

The authors report no competing interests.