The brain neurovascular epigenome and its association with dementia

adssx · November 4, 2025, 9:11am

Cerebral small vessel disease (SVD) is frequently comorbid with Alzheimer’s disease (AD), and brain endothelial cells (BECs) express genes associated with AD genetic risk. However, the epigenome of neurovascular cells and its intersection with genetic risk remain unexplored. Here, we generated gene regulomes for human BECs, mural cells, and other brain cell types and showed that AD heritability is primarily immune related, with a modest BEC enrichment. On the other hand, SVD heritability is enriched across the neurovascular unit, including astrocytes. Enhancer-to-gene interactomes implicated disease-distinct putative risk genes associated with amyloid and phospholipid processes in AD and senescence-associated pathways in SVD. Motifs for putative partners of lineage-determining transcription factors (TFs) in microglia and BECs were enriched for AD and SVD variants linked to disease-associated genes. In silico screening of compounds implicated vitamin D receptor agonists, mammalian target of rapamycin (mTOR), histone deacetylase (HDAC), and vascular endothelial growth factor receptor (VEGFR) inhibitors for AD. Our findings highlight regulatory mechanisms and therapeutic targets within the neurovascular system.

Gene prioritization and in silico screening identified compounds with modes of action that have previously been implicated in AD, including VDR agonists and inhibitors for mTOR, HDACs, and VEGFR. VDR is a nuclear receptor TF, which in microglia suppresses proinflammatory cytokine responses94,95 and has been linked to clearance of Aβ by macrophages.96,97 Vitamin D deficiency has been associated with increased risk of all-cause dementia and AD,98,99 although a causative role has remained unclear. A recent study showed that protective APOE2-expressing microglia have increased chromatin accessibility at VDR DNA-binding sites,100 supporting our genetics-based finding of a putative protective role in AD.
mTOR is a serine/threonine kinase associated with metabolism and mitochondrial function, cell growth, and proteostasis,101 and it is upregulated in early AD, including activated microglia.102,103 The mTOR inhibitor, rapamycin, has shown beneficial effects in cognition, amyloid and tau pathology, and synaptic plasticity in preclinical models and is being explored in clinical trials for dementia. However, activation of microglial mTOR was associated with elevated triggering receptor expressed on myeloid cells 2 (TREM2), increased spine number, and improved cognition, indicating that timing of drug administration may be critical.
HDAC inhibitors elevate histone acetylation, improve cognition, and reverse deficits in AD-associated models.106,107,108 Histone acetylation is dysregulated in AD brains,109,110,111,112 and inhibition of HDACs induced a disease-associated microglial signature and increased Aβ uptake,113 although broad isoform specificity remains a translational challenge.
Lastly, VEGF is an angiogenic factor, and increased angiogenesis and hypervascularization have been associated with AD, which may contribute to increased BBB permeability.114,115 The VEGF receptor is elevated in microglia and BECs in AD and has been associated with cognitive deficits and increased neuropathology.116,117 In addition, antibodies against VEGF and its receptor had protective cerebrovascular, cognitive, and neuroinflammatory effects in AD models.118,119 Although many of these drugs have been identified using epidemiology or in disease models, our approach provides a genetic rationale for drug prioritization. Importantly, our findings present targets and potential avenues for therapeutic intervention in AD and SVD, which have the potential to lower dementia risk.

Compounds that can target the above:

VDR: Vitamin D₃ (+ K2?)
mTOR: rapa
HDAC: Curcumin? Sulforaphane? Lithium (but indirect?)
VEGFR: EGCG? Curcumin?

CronosTempi · November 4, 2025, 12:08pm

Note, that some of these drugs may work at cross purposes or have undesirable interactions. Hasn’t there been raised concern about curcumin and rapamycin? While one drug may work in one system it may have the opposite effect on another. You may suppress mTOR with one while increase the expression with another. It’s hard to isolate effects and silo the impact. All these systems are biologically integrated. You may try to somewhat ameliorate these by timing, where you start a therapy with one and washout, then start another. That’s in essence what pulsing rapamycin is trying to do. You hit the system with a high dose suppression of mTOR to clear out older cells (as used in reproductive health), then give it a break so new stem cells can repopulate a healthier tissue, where you need a period of mTOR activation.

The promise and peril of polypharmacy. One reason I spend so much time looking for interaction between various drugs, supplements, diet and lifestyle, is because longer term effects can be quite unexpected. What is listed on a drug insert is just the start. The real challenge is that the body is a dynamic system with endless interactions.

So when I see a list of drugs as above, purporting to prevent dementia, I immediately think about the huge complexity of interactions. In some ways it’s simpler to focus on one drug like a SGLT2i, or rapamycin. By the time you’re throwing in 4, 5, 6 or more, you are looking at a lottery. YMMV.

Barnabas · November 4, 2025, 4:02pm

Good point. I found the Price and Hood book Scientific Wellness interesting, and one of the things I really liked was the observation / hypothesis that disease is simple in it’s earliest stages and more complex over time.

It’s an elegant conjecture with a common sense logic. Something is a little off. Other things respond downstream of that first thing being off. The more dominoes that fall, the more likely the affected person perceives symptoms and goes to the doctor. By that time there are a bunch of observable correlates of the disease state. Their idea was to track people longitudinally and then “rewind the tape” by looking at past samples preceding the disease diagnosis to try to isolate the earliest signs of transition to that disease state.

The thing about mucking around with drugs is that you might be correcting the initial rickety domino or you might be knocking over a domino that would otherwise have been standing tall. Sometimes you see this in e.g. taking a statin to reduce apoB and then having some worsening of glucose management and adding another drug to address that, etc.

I am going to start a flozin soon, and I’ll be on a statin, a sartan, a flozin, and rapamycin – all for preventative purposes with no diagnosed disease. I want to reduce the degradation rate of my arteries, my kidneys, and my whole system if possible. With a hat tip to Scientific Wellness, it would be great if there were a richer set of biomarkers we could track to have a more sophisticated model to guide treatment. Instead of do-no-harm meaning “Don’t take some drug off-label for prevention” it would be “Take that drug at a dose that moves these numbers toward optimal without overshooting or negatively impacting other numbers that are equally important.”