Brain's Protective Microglia Harness T Cell Gene CD28 to Curb Alzheimer's Damage

Microglia, the resident immune cells of the brain, can adopt protective or harmful identities during Alzheimer's disease (AD), but the molecular switches that determine which path they take have remained elusive. This landmark Nature study identifies PU.1—a master transcription factor governing myeloid and lymphoid cell identity—as a central regulator of neuroprotective microglial function, and uncovers an unexpected role for the T cell co-stimulatory receptor CD28 in restraining neuroinflammation.

Using the 5xFAD amyloid mouse model alongside human AD brain tissue, the team identified a distinct subpopulation of plaque-associated microglia characterized by low PU.1 (encoded by SPI1) expression. These PU.1-low microglia preferentially localized around amyloid plaques in both mice and human AD cases. Single-cell transcriptomic and epigenomic profiling revealed that reduced PU.1 activity in this subpopulation licenses the expression of a lymphoid gene program, most notably CD28—a surface co-stimulatory receptor canonically associated with T cell activation. This co-option of lymphoid receptor machinery by microglia is a novel and surprising finding.

Functional experiments demonstrated that reducing PU.1 expression specifically in microglia of 5xFAD mice attenuated amyloid pathology severity, consistent with the protective role suggested by human genetic data linking a SPI1 variant to delayed AD onset. Conversely, microglia-specific deletion of CD28 drove microglia toward a broad pro-inflammatory state and was associated with significantly increased amyloid plaque burden. This positions CD28-expressing PU.1-low microglia as suppressive cells that apply a brake on neuroinflammation, analogous to regulatory immune checkpoint mechanisms in peripheral immunity.

The study draws a compelling parallel to T cell biology: just as CD28 signals modulate T cell activation thresholds and regulatory T cell function, CD28 in this microglial subpopulation appears to tune the magnitude of the inflammatory response to amyloid. The authors speculate that other lymphoid co-stimulatory and co-inhibitory receptors may similarly be expressed by microglia under disease conditions, opening an entirely new avenue for understanding microglial immunoregulation.

From a therapeutic standpoint, these findings suggest that amplifying the PU.1-low, CD28-positive suppressive microglial state—or targeting the CD28 signaling pathway in microglia directly—could reduce neuroinflammation and slow AD progression. Given that CD28 and its ligands (B7 family) are already targets of approved immunotherapies in oncology and autoimmunity, repurposing or adapting these strategies for AD may be feasible. A key caveat is that most mechanistic data derive from mouse amyloid models, which do not fully recapitulate human AD, and the precise ligands and downstream signaling of microglial CD28 remain to be characterized.