Enhancer Dysregulation in Neurodegeneration: Lessons from Glial and Neuronal Models

Across Alzheimer's disease, Parkinson's disease, ALS, and multiple sclerosis, the majority of risk variants lie in non-coding sequence — and a disproportionate share fall in enhancers that are active in microglia rather than in neurons. Several of the best-characterized risk genes for Alzheimer's disease are expressed predominantly in myeloid cells.

That has two consequences for experimental design. First, the relevant cell type is often not the neuron; a neuronal model can be the wrong system for a risk variant that acts in microglia. Second, the phenotype is likely to be a change in an inflammatory or homeostatic programme rather than a change in neuronal excitability.

This article summarizes what glial and neuronal models have established, where they disagree, and what each can and cannot support. The broader per-field validation standards are set out in the pillar on validation paradigms; the choice of a specific model system is treated in our companion piece on models for neuronal enhancer dysregulation.

A risk variant's cell type can often be inferred before any experiment: intersect the credible set with cell-type-specific accessibility and factor occupancy. The inference is a prioritization, not a demonstration — but it determines which model to build.

Three findings recur across studies:

• • Risk variants can act cell-autonomously in microglia. Isogenic iPSC-derived microglia carrying risk alleles show altered inflammatory responses, phagocytic capacity, or survival, in a direction consistent with the genetic association.
• • Enhancer variants frequently alter factor binding rather than abolish the element. The effect is quantitative — a change in the magnitude or kinetics of an activation response — rather than binary.
• • The phenotype often appears only after an inflammatory stimulus. Baseline differences between genotypes can be small, and the effect emerges under challenge. Screens that measure only resting cells may find nothing.

That third point is the most practical: stimulation is part of the assay, not an optional extra.

Neuronal work has contributed a different set of lessons:

• • Activity-dependent enhancers respond to stimulation, and some risk variants alter the magnitude of that response rather than baseline expression.
• • Long-range regulation matters: enhancer–promoter pairs in neurons can span hundreds of kilobases, so a variant's target gene is often not the nearest one.
• • Maturation state determines the answer. A phenotype measured in immature neurons may not persist in mature ones, and vice versa — a recurring problem in iPSC-derived systems.

The second point connects directly to the question of how enhancers find their targets, which our pillar on cis-regulatory elements treats in detail.

Disagreement between studies is common and usually traceable to four variables:

None of these is a flaw in principle; all of them must be reported. A finding that depends on a specific stimulation protocol is a finding about that protocol, and should be described as such.

• • In vivo consequence. Cell-autonomous effects in culture do not establish a role in disease progression; that requires tissue or animal work.
• • Which target gene. A variant in an enhancer does not identify its target; contact data and perturbation are needed.
• • Direction of effect on disease. An allele may increase or decrease a protective programme; the direction must be measured, not assumed.
• • Cell non-autonomous effects. Microglial state changes alter neuronal survival through secreted factors; monoculture cannot capture that loop.

For a risk variant in a glial enhancer, the sequence that holds up:

1. Confirm the variant's cell-type context by intersecting the credible set with accessibility data.
2. Build isogenic iPSC-derived microglia (or the relevant glial type), both edit directions, multiple clones.
3. Measure baseline and stimulated phenotypes, with the stimulation paradigm pre-specified.
4. Show the enhancer effect at the level of the mark or accessibility, not only the transcript.
5. Identify the target gene by perturbation, not proximity.
6. State the limitation: no in vivo consequence was tested.

EDITGENE builds isogenic iPSC-derived microglia and neurons with characterized identity and maturity, enhancer-level readouts (accessibility, marks, eRNA), and stimulated functional assays — designed for variant-level comparisons rather than bulk phenotypes.

Tell us the variant and the cell type implicated. Contact EDITGENE to request a project consultation or quote.

Cell-type assignment of a risk variant is inferential and depends on the reference datasets used; conclusions should state the evidence behind the assignment.

iPSC-derived models approximate but do not reproduce human tissue. Phenotypes observed in culture require corroboration, and their dependence on stimulation protocols should be reported.

• • Nott A, et al. Brain cell type-specific enhancer–promoter interactome maps and disease-risk association. Science. 2019;366:1134–1139.
• • Corces MR, et al. Single-cell epigenomic analyses implicate candidate causal variants at inherited risk loci for Alzheimer's and Parkinson's diseases. Nature Genetics. 2020;52:1158–1168.
• • Dolmetsch R, Geschwind DH. The human brain in a dish: the promise of iPSC-derived neurons. Cell. 2011;145:831–834.
• • Jansen IE, et al. Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer's disease risk. Nature Genetics. 2019;51:404–413.