이전 사이트 이동

The heterogeneity in various aspects of Parkinson's disease (PD) is increasingly recognized as an important aspect in understanding the condition, with ethnicity and race being one of the major causes of heterogeneity that affects the risk and symptoms of PD onset. While there have been numerous reports related to PD in East Asia, there has been a lack of contribution from single-cell (or nucleus) transcriptome studies, which have been making significant contributions to understanding PD. In this study, we profiled nuclei extracted from the substantia nigra (SN) of confirmed pathological PD and control patients in South Korea, revealing 8 different cell types through cluster analysis. Monocle-based pseudotime analysis identified two disease-associated trajectories for each astrocyte and microglia and identified genes that differentiate them. Interestingly, we uncovered the inflammatory intervention in the early PD-associated transition in microglia and identified the molecular features of this intermediate state of microglia. In addition, gene regulatory networks (GRNs) based on TENET analysis revealed the detrimental effect of an HSPA5-led module in microglia and MSRB3- and HDAC8- led modules specifying the two different astrocyte trajectories. In SN neurons, we observed high heterogeneity and population changes, a decrease in dopaminergic and glutamatergic neurons and a proportional increase in GABAergic neurons. By deconvolution in spatial transcriptome obtained the PD sample, we confirmed spatiotemporal heterogeneity of neuronal subpopulations and PD-associated progressive gliosis specific to dopaminergic nuclei, SN and ventral tegmental areas (VTAs). In conclusion, our approach has enabled us to identify the genetic and spatial characterization of neurons and to demonstrate different glial fates in PD. These findings advance our molecular understanding of cell type-specific changes in Korean PD progression, providing an important foundation for predicting and validating interventions or drug effects for future treatments.

case control design

Spatial transcriptomics by high-throughput sequencing

For midbrain preparation, fresh tissue was serially sectioned into approximately 2-mm-thick slices following routine brain bank protocols. To ensure anatomical consistency across donors, a standardized midbrain level was selected based on gross anatomical landmarks. Specifically, the slide at the level where oculomotor nerve fibers were macroscopically visible, corresponding to the level of the superior colliculus and red nucleus, was consistently used for all cases. Tissue blocks for spatial transcriptomics (10x Visium v2) and histological validation (RNAscope and immunohistochemistry) were obtained from the adjacent face of the same tissue block or from the contralateral side at the identical anatomical level (oculomotor nerve level). To preserve RNA integrity for spatial profiling, fixation in 10% neutral buffered formalin was strictly limited to less than 72 hours prior to paraffin embedding.

FFPE tissue quality control (QC) and sectioning were performed in accordance with the Visium CytAssist Tissue Preparation Guide (CG000518). Samples with an RNA DV200 value of 30% or higher were selected for analysis. The tissues were sectioned at a thickness of 5 um, followed by immunofluorescence (IF) staining according to the CG000519 protocol. Subsequently, probe hybridization was conducted following the CG000495 protocol.

Following the CG000495 protocol, the optimal number of PCR cycles for the subsequent indexing step was determined via qPCR. Following this determination, indexing PCR was performed to construct the final library.

Prepared libraries were sequenced on Illumina NovaSeqX.

Raw Visium spatial transcriptomics count data from multiple human midbrain substantia nigra sections were processed using standard quality control, normalization, and integration workflows. Low quality genes and spots, as well as spots outside tissue regions, were removed based on expression and histological criteria. Normalized expression matrices from individual slides were integrated to account for batch effects and used to define spatial domains through unsupervised clustering. Processed data include spatially resolved gene expression, spatial cluster annotations, cell type abundance estimates inferred using single nucleus RNA seq reference profiles, and downstream spatial analyses, including ligand receptor interaction analysis and annotation of spots relative to alpha synuclein positive regions.