Human Brain Cell Triple Culture Reveals Astrocytes Drive Microglial Disease State
A new iPSC triple-culture model shows astrocytes push microglia toward a disease-associated state—and Alzheimer's neurons suppress it.
Summary
Researchers at Brigham and Women's Hospital developed a reproducible human iPSC-derived triple culture (triculture) system combining neurons, astrocytes, and microglia to study brain cell communication in health and Alzheimer's disease. Using single-cell RNA sequencing and protein analysis, they found that co-culturing all three cell types dramatically changes each cell's behavior. Most strikingly, astrocytes drive microglia into a disease-associated microglia (DAM) state—marked by elevated TREM2, APOE, SPP1, and GPNMB—even without overt pathological stimuli. Paradoxically, when neurons carrying familial Alzheimer's disease mutations were introduced, this astrocyte-driven DAM signature was significantly suppressed, despite triggering a separate inflammatory response. The model is operational within 20 days of thawing cryopreserved cells, making it broadly accessible for neurodegeneration research.
Detailed Summary
Neuroinflammation is a central feature of Alzheimer's disease (AD), yet the intercellular signals that govern human microglial activation states remain poorly understood. Mouse models and postmortem tissue studies have generated important hypotheses, but a reproducible, physiologically relevant human system for testing them has been lacking. This study introduces a human iPSC-derived triculture (TC) platform that addresses that gap with practical accessibility in mind.
The research team independently differentiated three cell types—excitatory neurons (iNs) via NGN2 overexpression, astrocytes (iAs) via SOX9/NFIB, and microglia (iMGs) via hematopoietic precursor intermediates—and then combined them after cryopreservation. By optimizing six candidate co-culture media formulations, they identified BrainPhys-based media (TCM5) as ideal for maintaining all three cell identities. The full triculture is operational within 20 days of thaw, with a final approximate ratio of 6:2:3 (neurons:astrocytes:microglia), and remains stable through at least six days of co-culture.
Single-cell RNA sequencing comparing monocultures (MCs) and tricultures revealed profound transcriptional remodeling across all three cell types. Neurons in TC showed increased dendritic spine density, elevated synaptic vesicle release, and enhanced electrophysiological activity. Astrocytes upregulated cholesterol efflux and cell adhesion pathways, while microglia shifted toward a more ramified morphology and upregulated lipoprotein catabolism genes. Critically, a subset of microglia in TC acquired a disease-associated microglia (DAM) transcriptional signature—including strong upregulation of TREM2, APOE, SPP1, and GPNMB at both mRNA and protein levels. Conditioned media and co-culture experiments demonstrated this DAM induction is driven specifically by astrocyte-secreted signals, not neuronal contact.
To probe disease relevance, the team introduced neurons carrying homozygous familial AD mutations (APP-Swedish; PSEN1-M146V) into the system. These fAD neurons triggered a prototypical proinflammatory microglial response. Paradoxically, they also substantially suppressed the astrocyte-induced DAM signature—reducing TREM2, APOE, SPP1, and GPNMB protein levels in microglia. This suggests that the high amyloid-beta secretion from fAD neurons disrupts normal astrocyte-microglia communication channels that ordinarily sustain the DAM state, revealing a potentially important early disease mechanism.
The triculture model was validated across two donor genetic backgrounds and multiple independent differentiations, demonstrating strong reproducibility. The cryopreservation-based workflow lowers the barrier to adoption compared to continuous long-term differentiation protocols. As a proof of utility, the system has already been applied in a companion study demonstrating that microglia mediate CLU-dependent synapse loss and tau phosphorylation. Collectively, the findings reframe the DAM state as regulated by ongoing glial crosstalk rather than purely pathological triggers, with significant implications for understanding early AD pathogenesis.
Key Findings
- Astrocytes alone are sufficient to drive a disease-associated microglia (DAM) state marked by TREM2, APOE, SPP1, and GPNMB upregulation.
- Familial AD neurons suppress astrocyte-induced DAM signature in microglia despite triggering separate proinflammatory responses.
- Neurons in triculture show increased dendritic spine density, synaptic vesicle release, and electrophysiological activity.
- The triculture system is fully operational within 20 days post-thaw using cryopreserved cells, enabling consistent reproducibility.
- Co-culture transcriptional states recapitulate gene expression patterns observed in human postmortem brain tissue.
Methodology
Human iPSC-derived neurons, astrocytes, and microglia from two cognitively normal donors were independently differentiated, cryopreserved, and combined into a triculture within 20 days of thaw. Characterization included single-cell RNA sequencing, western blotting, immunostaining, electrophysiology, and dendritic spine analysis in monoculture versus triculture across multiple differentiations and genetic backgrounds.
Study Limitations
The triculture uses acutely co-cultured cells over only 6 days, which may not capture longer-term adaptive changes relevant to chronic neurodegeneration. Astrocytes were underrepresented in single-cell RNA sequencing due to known technical biases, limiting the depth of astrocyte transcriptomic analysis. All neurons are excitatory (NGN2-derived) and do not recapitulate the full diversity of neuronal subtypes affected in Alzheimer's disease.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.