The Neuroscience Journal Club: Connecting vascular dysfunction to Huntington’s disease
Publications are critical to communicating advancements in the scientific community, but do not cover the interesting backstories of the research project: why did they take this approach? Where are they going next? What else did they discover? The answers to these questions can influence your own research, which is why we started the virtual Neuroscience Journal Club—a forum where you can ask questions directly to the authors of new, high-impact publications.
In the latest meeting of the Neuroscience Journal Club, Francisco Garcia (Heiman Lab, Massachusetts Institute of Technology) led a discussion on his recent Nature publication, where he characterized transcriptional dysregulation in the neurovasculature of Huntington’s disease (HD). In this discussion, he touched on HD's transcriptional and cellular signatures, answered questions about whether his findings are recapitulated in other neurodegenerative disorders, and covered his novel approach toward single nuclei sequencing of neurovascular cells.
Linking cerebrovascular dysfunction to Huntington’s disease
In his talk, Francisco highlighted how the neurovasculature is not only dysregulated in the context of HD and other neurodegenerative diseases but that this dysregulation manifests earlier in disease than many other pathologies. However, single cell analysis of these cell types has proven difficult due to their relative rarity and “stickiness” with other cells.
To address this challenge, Francisco’s group developed a novel blood vessel enrichment protocol and coupled it with single nuclei RNA sequencing (snRNA-seq) to transcriptionally characterize human vascular cells from healthy controls and patients with HD. The protocol showed remarkable consistency of cell types and marker genes between both frozen and fresh human brain tissue, as well as in data generated from mice. Intriguingly, while some human vascular markers were present in mouse models, they did not show cell-type specificity in mice.
Using this information, they examined possible functional pathways affected by this cell type–specific transcriptional dysregulation and revealed signatures consistent with innate immune activation. Turning to mouse models, they used immunofluorescence to reveal that, in HD, astrocytic processes colocalize and engulf blood vessels, eventually leaving only a glial scar and no vessel.
Over the course of his discussion, attendees asked questions and were encouraged to submit additional questions for discussion at the end of the panel. We’ve included a subset of the questions (and Francisco’s answers to them) below; read what he had to say, then watch the full on-demand webinar here.
Sample prep and protocol
You mentioned the enriched vasculature is not pure. Can you assess how variable the purity and proportion of cells is across samples?
This is actually something we do with the quality checks for everything. […] Right now, the protocol that I use is about a 14–15 fold enrichment from previous samples without enrichment. Now, we also have improved protocols in the lab, and we’re getting a higher percentage of enrichment so that, out of the total population, we’re getting…about 40% vascular cell types. Every time we do these experiments, in each sample, we try to characterize what proportion of them are vascular cell types and what portions are neuronal and glial.
Cell type–specific considerations
Any thoughts on where the pericytes are located and what their function is?
One idea that I’ve been having is, potentially, that, because we saw this transcriptional gradient between the subtypes of fibroblasts and one type of pericyte, I think, perhaps, that subtype of pericyte is coming from the lineage of the perivascular fibroblasts that are being differentiated. It’d make sense since it’s been shown that perivascular fibroblasts are more localized to the larger vessels, particularly for arterioles. So, perhaps, these “pericyte 2” subtypes are closer to arterioles because [they are] in a transcriptional continuum with the perivascular fibroblasts; whereas [the “pericyte 1” population], which is not in the continuum, may be deeper in the capillary bed. As to function, I’m not entirely sure. It’s a great debate in the field. Given that there’s a lot of heterogeneity in the pericytes, I think a lot of work needs to be done, and, now that we have these marker genes, we can better try to define these functions.
Are the transcriptomic changes in the vasculature thought to be cell autonomous or caused by neuronal changes?
This is an amazing question because it’s always the chicken and egg question of what comes first: is it the vascular dysfunction or neuronal dysfunction? As I mentioned, vascular changes are some of the first to occur. In other neurodegenerative diseases, we know these occur before the onset of other pathologies. For example, in Alzheimer’s, we know there’s vascular dysfunction before amyloid beta or tau pathologies. Now, there’s a hypothetical model [in which] people are thinking we should put a vascular curve before any of this. You might be inclined to think vascular changes are leading to this cell-autonomous change. I think it’s very much intertwined.
I think that it’s not just cell autonomous or just caused by neuronal changes; I think it’s caused by a mix of endothelial cells trying to respond and losing their integrity, but also there’s a lot of communication with the neurons that are actually causing this dysregulation to happen.
Relevance to other neurodegenerative diseases
Are any of the human-specific genes that you found in these brain vascular cell types also seen in other human diseases?
Yes, actually. One thing that came up as pretty interesting is that this gene called Slc20a2 is actually implicated in what was formerly known as Fahr’s disease—now known as primary familial brain calcification. One of the genes we validated by IF [and] showed is very much enriched in the vasculature in people but not in mice. It was interesting to us because you have mutations in human populations of this gene that lead to vascular pathologies, so when we’re looking at/thinking about making mouse models, we’re interested in whether the gene is present in the right cell population.
As far as the transcriptional changes in HD, do you think they are similar to other neurodegenerative diseases like Alzheimer’s, Parkinson’s, ALS, etc.?
Absolutely, yes. […] We’re beginning to see a lot of agreement between the datasets. Not just for a baseline profile, but also in the disease context, particularly Alzheimer’s. We’re seeing a lot of common overlapping genes that are dysregulated in brain vasculature across HD and Alzheimer’s. As I mentioned at the end of the talk, it’s exciting because now we can think about approaches targeting the vasculature that might be not just beneficial for HD, but for other diseases [as well].
We’d like to thank Francisco for sharing his work, leading the discussion, and being both willing and enthusiastic in answering all the questions submitted by participants! Watch the entire webinar on-demand, and make sure you don’t miss the next Neuroscience Journal Club.