New brain maps reveal hidden cellular vulnerabilities in Alzheimer's disease

A recent study published in the journal NatureResearchers have published a transcriptome atlas of multiple brain regions in patients with and without Alzheimer's disease (AD).

Postmortem diagnosis of Alzheimer's disease is staged according to the distribution and severity of features characterized by intraneuronal intracellular neurofibrillary tangles (NFTs) and extracellular amyloid-β (Aβ) deposits. Neurofibrillary tangles are observed sequentially in the entorhinal cortex (EC), hippocampus (HC), thalamus, and neocortex, a sequence consistent with cognitive decline.

Understanding the cytoarchitecture of affected regions may elucidate the underlying mechanisms of disease progression and impact early therapeutic intervention.Although some of the involved brain regions have been investigated, extensive molecular characterization of regional differences in AD is lacking.

study: Single-cell, multi-region analysis of Alzheimer's diseaseImage credit: Marko Aliaksandr / Shutterstock

Research and Results

In this study, we present a transcriptome atlas of six different brain regions in patients with and without AD. The researchers analyzed single-nuclear nuclear ribonucleic acid (RNA) sequences from 283 postmortem brain samples from 48 participants, 26 of whom had a pathological diagnosis of AD.

Six anatomical brain regions were profiled: HC, EC, anterior thalamus (AT), prefrontal cortex (PFC), middle temporal cortex (MT), and angular gyrus (AG). 76 high-resolution cell types across 14 major cell type groups were annotated. Cell types were characterized based on proliferation state and transcriptome size and compared to previously published atlases.

OneSummary of snRNA-seq profiling across 283 samples across six brain regions from 48 ROSMAP participants showing global pathology, Braak stage, and pathological diagnosis of AD (26 AD, 22 non-AD) or clinical diagnosis (16 AD dementia (dem.) and 32 no dementia). b,cjoint uniform manifold approximation and projection (UMAP), colored by major cell types (b) and place of origin (c). dRegional composition of major cell types. eRelative enrichment of major cell types between regions by quasi-binomial regression. False discovery rate (FDR)-corrected P values ​​are indicated with asterisks: ***P < 0.001, **P < 0.01, *P < 0.05. debt,Gthe overall composition, regional composition, enrichment and number of excitatory nuclei (debt) and inhibitory (G) Neuronal subtype. hGene expression analysis of the top four markers per inhibition subclass, averaged at the sample subclass level (columns). IValidation of FOXP2 and MEIS2 as markers of unique thalamic subtypes with RNAscope. Quantification (left) was performed using Student's t-test and representative images (right) are shown. Blue dots represent MEIS2 (top) or FOXP2 (bottom) transcripts, red dots represent GAD2 transcripts. FOXP2: n = 19 (PFC) and n = 22 (TH) cells. MEIS2: n = 35 (PFC) and n = 26 (TH) cells. Each dot represents an individual cell pooled from eight samples (four individuals, one PFC and one thalamic sample each). gunglutamatergic and GABAergic scores for all neuronal subtypes. Dotted lines represent 95% confidence intervals for linear fits. P values ​​were calculated using two-tailed F tests. Ast., astrocytes; exc., excitatory neurons; inh., inhibitory neurons; mic., microglia/immune cells; olig., oligodendrocytes; vasc., vascular/epithelial cells.

The proportion of neurons increased from TH (14.4%) to ectocortical HC (32.2%) and neocortical (PFC, MT, and AG) regions (58.9%). Glia, including microglia/immune cells, astrocytes, oligodendrocyte precursor cells (OPCs), and oligodendrocytes, were less abundant in neocortical samples. Notably, differences in the composition of major cell types were consistent across participants, regardless of disease status.

Furthermore, excitatory neuron subtypes were either highly region-specific or largely shared across neocortical regions. Conversely, inhibitory neuron subtypes were observed in all but one of the five cortical regions. Furthermore, the researchers observed numerous transcriptionally distinct subsets of principal neurons. Glial cells Astrocytes showed the highest regional heterogeneity, with both thalamic and neocortical enriched subtypes present.

Thalamic astrocytes were enriched for focal adhesion-related genes, while cortical astrocytes were enriched for genes involved in glutamate processing and transport.The team then investigated how AD affects cellular composition.Specifically in late-stage AD, differences in the PFC, EC, and HC regions showed a slight decrease in excitatory and inhibitory neurons and an increase in oligodendrocytes and vascular cells.

Among excitatory neurons, one HC-specific and four EC-specific subtypes were underrepresented in individuals with a pathological AD diagnosis. Individuals with lower levels of these vulnerable subtypes performed worse on cognitive tasks. Furthermore, 391 genes showed significantly elevated baseline expression in the vulnerable subtypes, including reelin signaling pathway, heparan sulfate proteoglycan biosynthesis, and kinase-related genes.

Vulnerable and non-vulnerable inhibitory neurons differentially expressed genes involved in heparan sulfate proteoglycan biosynthesis, enzyme-linked receptor protein signaling, and neuronal projection morphogenesis. Furthermore, vulnerable inhibitory neurons showed increased expression of Reelin signaling pathway genes.

Additionally, the researchers calculated differentially expressed genes (DEGs) for each excitatory subtype. Non-vulnerable subtype-associated DEGs were enriched in diverse functions, including heat shock family chaperones, ubiquitin ligase binding, and neuronal death mediators. In contrast, vulnerable subtype-associated DEGs were enriched only in mitochondrial oxidative phosphorylation.

We next assessed DEGs for each major cell type in each region and across regions. Excitatory and inhibitory neurons, and astrocytes showed the highest DEGs in all regions, but highest in ECs. Neuronal DEGs showed little overlap across regions, suggesting that neuronal differences in AD are primarily determined by region of origin or subtype. OPC and microglia DEGs overlapped within non-neocortical regions.

Finally, the researchers investigated transcriptional changes associated with cognitive resilience (CR) in AD, defined as the difference between observed and predicted cognition based on the level of pathology, or the absence of cognitive impairment despite pathological AD. Astrocytes were the only cell type that consistently had a high number of genes associated with CR across all nine measures of CR tested.

Various astrocyte genes that are consistently associated with multiple indicators of CR, termed “CR-associated genes,” promote or have antioxidant activity.Goal 1), glutathione peroxidase 3 (GPX3), and ornithine decarboxylase 1 (ODC1) was positively correlated with cognitive function, and its expression was highest in astrocytes from individuals with the least cognitive decline.

Conclusion

The study compiled a transcriptome atlas of six brain regions from AD and non-AD patients. The researchers annotated cellular diversity by region, explored gene expression and differences in AD across cell types, and highlighted region-specific AD-vulnerable cell populations. Additionally, the researchers discovered astrocyte genes associated with CR in AD. These genes converged in the choline metabolism and polyamine biosynthesis pathways, supporting them as attractive targets for promoting CR.