Bring Spatial Techniques to Understanding Alzheimer’s Disease Pathology

February 23, 2022

By Allison Proffitt 

February 23, 2022 | At this week’s Molecular & Precision Medicine Tri-Con event, Simon Gregory, profession or neurology at the Duke University School of Medicine, outlined how his team has been using 10X Genomics’ Visium platform and Oxford Nanopore long read sequencing to understand the pathology of Alzheimer’s Disease development.

Alzheimer’s Disease accounts for 60-70% of dementia cases and affects millions of people worldwide. While we have known about APOE alleles and others genetic drivers of risk for many years, the disease progression is still not well understood. 

There are pathological hallmarks of the disease—amyloid-B plaques and neurofibrillary tangles—but the role of those features is not clear. However it does seem that the disease progresses in a temporospatial way: with the accumulation of amyloid plaques first, and then  neurofibrillary tangles, both increasing and spreading as the disease progresses.

“We’re going to take advantage of this temporospatial difference in accumulation and development of the pathology to try to understand the progression of Alzheimer’s Disease itself,” Gregory explained.

The team used the 10x Genomics Visium spatial genomics platform to examine changes in gene expression within four different tissues of the brain for one particular Alzheimer’s patient, paying particular attention how to genes change in proximity to amyloid plaques. “We believe that these four regions represent the trajectory of disease development,” he said.

Spatial Analysis

The team was able to do cell phenotyping from spatial data and see the progressive pathology of the disease in the four different tissues they sampled.

From there, they looked at cell-cell interaction. Because the current Visium arrays have a 55 micron spot size, multiple cells could fit into each spot, requiring researchers to de-convolute the data from each spot to separate the different cells, and identify which cells are likely to be next to each other.

“What we see, in general, is reduced interaction of astrocytes with the neuronal compartment, and increased interaction of astrocytes with oligodendrocytes,” he said. 

The team also looked at the microenvironment of the amyloid beta plaques and identified 243 genes that were increased within plaques. This finding was an unsupervised analysis, Gregory pointed out, just looking at differential expression, but it does have interesting overlap with findings in the REACTOME database of genes involved in collagen formation, degradation, and synthesis. “We’re following up on these candidate genes and how they’re relevant to the phenotype,” he said. 

Long Read Sequencing

The team is using the larger fragments of cDNA that the Visium delivered and sequencing them on an Oxford Nanopore MinION sequencer. “We’re less interested in the quantitation of what genes are present by looking at just the expression alignment of the sequence to a transcriptome from the data and the alignment of the transcripts to a particular region within the tissues. Here we’re using long read sequencing to look at alternative splicing,” he explained. 

Because the cDNA is going through Nanopore sequencing once it’s already been barcoded, the data generated can be related to the positional information from the original tissue. 

Together, these technologies are revealing conclusions that Gregory said his team is actively pursuing. Data from their first samples suggest strongly that a single individual can represent disease progression, and they are expanding their analysis to a second patient to confirm.

They plan to integrate the Visium and Nanopore splicing data. They are looking forward to an update to the Visium platform coming his summer, Gregory said: decreasing of the spot size from 50 microns to 4 microns, which he calls, “incredibly more informative or more useful.” He is also looking forward to two new technologies: in situ transcriptomics and spatial protein profiling from Akoya’s PhenoCycler.

“Instead of digitizing the expression of the genes from the tissue and then relating it back to the tissue, in situ transcriptomics approaches will be generating sequence of transcripts within the cells themselves. This will enable us to start looking at cell-to-cell interactions or cell neighborhood interactions, which will allow us to have a more informed contextual expression view of what’s happening.”