Molecular differences in Alzheimer's disease between male and female patients determined by integrative network analysis

Abstract

Alzheimer's disease (AD) is a complex neurodegenerative disease and the most common cause of dementia among the elderly. There has been increasing recognition of sex differences in AD prevalence, clinical manifestation, disease course and prognosis. However, there have been few studies on the molecular mechanism underlying these differences. To address this issue, we carried out global gene expression and integrative network analyses based on expression profiles (GSE84422) across 17 cortical regions of 125 individuals with AD. There were few genes that were differentially expressed across the 17 regions between the two sexes, with only four (encoding glutamate metabotropic receptor 2, oestrogen-related receptor beta, kinesin family member 26B, and aspartoacylase) that were differentially expressed in three regions. A pan-cortical brain region co-expression network analysis identified pathways and genes (eg, glycogen synthase kinase 3β) that were significantly associated with clinical characteristics of AD (such as neurofibrillary score) in males only. Similarity analyses between region-specific networks indicated that male patients exhibited greater variability, especially in the superior parietal lobule, dorsolateral prefrontal cortex and occipital visual cortex. A network module analysis revealed an association between clinical traits and crosstalk of sex-specific modules. An examination of temporal and spatial patterns of sex differences in AD showed that molecular networks were more conserved in females than in males in different cortical regions and at different AD stages. These findings provide insight into critical molecular pathways governing sex differences in AD pathology.

Keywords: Alzheimer's disease; integrative network; sex differences; temporal and spatial patterns.

Figure 1

A, Distribution of gene expression values in male (blue) and female (orange) samples. B, differentially expressed genes (DEGs) in 17 cortical regions

Figure 2

Co‐expression network analysis. A, From left to right: male‐ and female‐specific networks and overlap network. B, Significantly enriched pathways for the networks in (A)

Figure 3

Kaplan‐Meier test of GSK3β based on (A) average neuritic plaque density; B, Braak neurofibrillary tangle score; C, sum of Consortium to Establish a Registry for Alzheimer's disease rating scores in multiple brain regions; and D, sum of neurofibrillary tangle density in multiple brain regions

Figure 4

A, Numbers of co‐expressing network edges and Jaccard coefficients of 17 cortical regions between males and females. B,C, Jaccard coefficients across the 17 regions in males (B) and females (C)

Figure 5

Crosstalk between sex‐specific modules. A, Jaccard coefficient between male and female modules. B, Male‐specific module. C, Female‐specific module

Figure 6

A, Dynamic scores in males and females. B,C, Conserved networks of males (B) and females (C) across 17 brain regions in definite AD. D,E, Significantly enriched pathways for conserved networks in males (D) and females (E)