Back

Molecular and cellular signatures differentiate Parkinson's disease from Parkinson's disease with dementia

Fairbrother-Browne, A.; Grant-Peters, M.; Brenton, J. W.; Nelvagal, H.; Reynolds, R. H.; Lim, Y. M.; Gustavsson, E. K.; Macpherson, H.; Montgomery, K.; Evans, J. R.; Hicks, A. R.; Chiraki, N.; Curless, T.; Real, R.; Xenakis, T.; Houlden, H.; Morris, H. R.; Gandhi, S.; Wood, N. W.; Hardy, J.; Jaunmuktane, Z.; Ryten, M.

2025-03-04 neuroscience
10.1101/2025.03.04.641379 bioRxiv
Show abstract

Parkinsons disease (PD) affects millions of people worldwide, and up to 40% of these patients develop dementia, profoundly affecting their quality of life. Whether Parkinsons disease dementia (PDD) simply represents a late stage of PD or constitutes a distinct neurodegenerative process remains unresolved. To clarify this, we generated the largest single nuclear transcriptomic atlas of PD and PDD to date--almost one million nuclei derived from the anterior cingulate cortex and inferior parietal lobule of 64 post-mortem donors. By integrating these data with long-read RNA-seq, we found that the cellular compositions, biological pathways, and molecular profiles diverge substantially between PD and PDD, with limited overlap in differentially expressed genes and pathways. While PD was characterised by widespread upregulation of gene expression programs and robust regional signatures, PDD showed extensive pathway downregulation, loss of cortical regional identity, and significant shifts in transcript usage, including alterations in APP isoforms that may influence pathological amyloid beta accumulation. These findings reveal that PD and PDD represent fundamentally distinct disease states, offering important insights for understanding their underlying mechanisms and will guide the development of targeted therapies and more effective clinical trials.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.