by Christos Evangelou, MSc, PhD – Medical Writer and Editor 

A recent spatial imaging study has provided new insights into the role of microglia in Alzheimer’s disease (AD). The study showed that microglial cells exist along a spectrum of activation states rather than in discrete subtypes, and that AD is associated with widespread microglial dysfunction across this spectrum, not just around plaques and tangles. According to the authors, these findings suggest that microglial dysfunction may be far more pervasive in AD than previously recognized. 

The study was published in Nature Immunology. 

Moving Beyond Animal Models

Microglia have traditionally been categorized into M1 and M2 states, and clusters of disease-associated microglia (DAMs) have been identified around amyloid plaques. However, our understanding of the role of microglia in neurodegeneration has relied primarily on studies in rodents, which do not fully capture the complexity of the aging human brain. In addition, the technical challenges associated with studying post-mortem samples of human brains have limited our understanding of microglial diversity in human tissue in health and disease. 

“Animals don’t live as long as humans or have as large and complex brains,” explained senior author Sean Bendall, PhD, Associate Professor at Stanford University School of Medicine. “We also know the human blood and immune system significantly remodels with age. For these reasons, we sought to study the diversity of these cells directly in aged and diseased human brains.” 

Methodology

Researchers at Stanford University used multiplexed ion beam imaging (MIBI), which employs elemental isotope reporters and nanometer-scale mass spectrometry, to simultaneously measure 38 different proteins in individual cells while preserving their spatial context within brain tissue. The team analyzed over 135,000 microglia across five brain regions from 23 individuals, including both cognitively normal older adults and patients with AD dementia.  

“MIBI enabled high-dimensional mapping of microglial states in situ at single-cell resolution by capturing millions of cells, including many hundreds of thousands of microglia, across different brain regions from dozens of individuals,” Bendall noted. 

He added that MIBI works particularly well with formalin-fixed, paraffin-embedded (FFPE) brain tissues, which are commonly used in human neurodegeneration studies. 

Microglial State Continuum

Rather than finding the classical M1 and M2 microglial subtypes, the researchers found that microglial cells exist in a “microglial state continuum,” which spans from low to high levels of immune activation. This spectrum reflects varying expression of proteins involved in immune surveillance, phagocytosis, antigen presentation, and cellular metabolism. 

“The continuous model allowed us to show a significantly higher portion of microglial cells in AD, up to 30% in some individuals, suggesting a bigger role for these cells in AD,” Bendall explained. 

The continuum of microglial state varied across brain regions. The hippocampus and substantia nigra harbored more highly activated microglia, whereas the middle frontal gyrus and caudate showed predominantly lower activation states. The researchers also found that the local tissue environment influenced the phenotypes of microglia. Cells in synapse-dense areas showed lower activation markers, whereas those near myelin tracts or blood vessels displayed higher immune signatures. 

Widespread Microglial Dysfunction in AD

DAMs are highly activated microglia that directly interact with plaques and tangles in AD, but these cells represent less than 1% of the total brain microglia. The research team found that AD is associated with much more widespread changes across the entire microglial population. 

In AD brains, particularly in the hippocampal CA1 region, microglia showed a shift toward the high end of the activation spectrum. However, rather than showing enhanced immune function, these cells displayed a dysfunctional state. They exhibited reduced expression of proteins involved in antigen presentation (HLA-DR), homeostatic surveillance (P2RY12), and lipid metabolism (ApoE), while showing increased expression of inflammatory markers, including CD44 and CD33. 

“The big shifts we see in AD-associated microglia are at the end of the spectrum, most associated with immune activation and phagocytosis. Except all of the shifts are in the dysfunctional direction,” Bendall said. “They have lower indicators of phagocytosis, lower metabolism, and less internalized synaptic proteins. Simply put, they seem to be doing less, yet they are concentrated in the regions of the hippocampus we know are prone to AD-related degeneration.” 

The spatial analysis also provided insights into how dysfunctional microglia interact with their local environment. In AD, high-activation microglia were predominantly located near parvalbumin-positive inhibitory interneurons rather than excitatory synapses. Furthermore, dysfunctional microglia were associated with reactive astrocytes showing high GFAP expression and low glutamine synthetase, indicating altered astrocytic metabolism. This suggests that microglial dysfunction may contribute to broader glial network abnormalities in AD. 

Potential Clinical Implications

According to the authors, their findings have implications for the development of AD therapeutics and patient stratification.  

“We believe that our findings add to the argument that microglia dysfunction, as opposed to over-activation, is a key part of the AD pathological process,” Bendall stated. “If so, detecting these cells early and reversing this microglial dysfunction, or replacing these cells altogether, could be a key strategy for AD intervention.” 

The continuous spectrum model also provides a framework for developing biomarkers to track disease progression and therapeutic response. The inter-individual variability observed in microglial states suggests that personalized approaches may be necessary, Bendall emphasized. 

Future Directions

The analysis primarily focused on older adults and used post-mortem tissue, which may not fully capture the dynamic changes that occur during disease progression. In addition, the protein-based approach represents only one layer of cellular complexity. The research team is currently validating their findings in a broader collection of human brain tissues. 

“This study was focused on looking at AD-related cells in individuals with an APOE3/3 genotype. We’ve already started to look at high-risk AD genotypes, like APOE4/4, which are believed to have a role in microglial function, as well as looking at AD-resilient individuals,” Bendall revealed. 

The study received financial support from the National Institutes of Health.  

References

1. MrdjenD, Cannon BJ, Amouzgar M, et al. Spatial proteomics of Alzheimer’s disease-specific human microglial states. Nat Immunol. 2025;26(8):1397-1410. doi:10.1038/s41590-025-02203-w