Transcriptional expression patterns of the cortical morphometric similarity network in progressive supranuclear palsy

Abstract

BackgroundIt has been demonstrated that progressive supranuclear palsy (PSP) correlates with structural abnormalities in several distinct regions of the brain. However, whether there are changes in the morphological similarity network (MSN) and the relationship between changes in brain structure and gene expression remain largely unknown.MethodsWe used two independent cohorts (discovery dataset: PSP: 51, healthy controls (HC): 82; replication dataset: PSP: 53, HC: 55) for MSN analysis and comparing the longitudinal changes in the MSN of PSP. Then, we applied partial least squares regression to determine the relationships between changes in MSN and spatial transcriptional features and identified specific genes associated with MSN differences in PSP. We further investigated the biological processes enriched in PSP‐associated genes and the cellular characteristics of these genes, and finally, we performed an exploratory analysis of the relationship between MSN changes and neurotransmitter receptors.ResultsWe found that the MSN in PSP patients was mainly decreased in the frontal and temporal cortex but increased in the occipital cortical region. This difference is replicable. In longitudinal studies, MSN differences are mainly manifested in the frontal and parietal regions. Furthermore, the expression pattern associated with MSN changes in PSP involves genes implicated in astrocytes and excitatory and inhibitory neurons and is functionally enriched in neuron‐specific biological processes related to synaptic signaling. Finally, we found that the changes in MSN were mainly negatively correlated with the levels of serotonin, norepinephrine, and opioid receptors.ConclusionsThese results have enhanced our understanding of the microscale genetic and cellular mechanisms responsible for large‐scale morphological abnormalities in PSP patients, suggesting potential targets for future therapeutic trials.