05/06/2026
Astrocytes are increasingly recognized as active drivers of neurodegeneration rather than passive bystanders, but our ability to interrogate which transcription factors actually shape their disease behavior has been limited. This Science paper from the Sloan and Bassik groups (Liu et al.) introduces an in vivo gain-of-function Perturb-seq platform that begins to fill that gap at scale.
The authors built a functional atlas of approximately 1,000 transcription factors in astrocytes by combining high-throughput overexpression with single-cell RNA sequencing readouts in vivo. They identified cofunctional TF modules, annotated previously uncharacterized regulators, and predicted disease-associated clusters. Most striking for our field: applying the platform in a neuroinflammation model identified a TF whose astrocyte-specific overexpression alleviated Alzheimer's disease symptoms in mice.
What I find conceptually important is the inversion. Most CRISPR screens in neurodegeneration are loss-of-function and identify dependencies. Gain-of-function screens like this identify sufficiency — which transcription factor, when reactivated, reorganizes astrocytes toward a protective state. That is a different therapeutic lens, more aligned with reprogramming approaches than knockdown approaches.
For translational purposes, the resource matters as much as the individual hit. A searchable atlas of astrocyte TF effects gives us a hypothesis engine for every neurodegenerative disease where astrocyte states matter — AD, ALS, MS, stroke. We should expect a wave of follow-up work testing whether the same TFs that protect in AD models can be redirected against tau, alpha-synuclein, or TDP-43 pathology.
https://doi.org/10.1126/science.adw2156
