Dr. Hung-Ju Chueh, first author of the study, said: "What is striking is the precision of this response. The microglia were not simply engulfing damaged neurons but instead removing parts of the neuron containing aggregated alpha-synuclein, suggesting that, at certain stages of disease, microglia help neurons dispose of potentially harmful material."

What keeps the nibbling under control?

The response did not run unchecked. The researchers found it was fine-tuned by several molecular signals between neurons and microglia:

• P2RY12, a sensing signal that helps microglia detect the problem
• CD22, an inhibitory signal that reins the response in
• Interleukin-10 (IL-10), a cytokine that acts as a "self-regulated brake" on microglia, stopping them from damaging healthy tissue

Using single-cell RNA sequencing — a technique that reads the gene profiles of individual cells — the team also identified a specific subpopulation of activated microglia responsible for this beneficial clearance.

What role does GPNMB play?

The study went a step further. Professor George Tofaris, senior author, and his colleagues found that a protein called glycoprotein nonmetastatic melanoma protein B (GPNMB) is an important part of the protective process. Genome-wide association studies had previously linked genetic variation at the GPNMB locus to Parkinson's disease, but the protein's functional role remained unclear.

In the lab models, GPNMB levels rose when microglia encountered neurons loaded with aggregates, and the protein interacted with pathological alpha-synuclein inside the microglia. The researchers also found more GPNMB in microglia within the substantia nigra — the brain region most affected in Parkinson's — in tissue from people who had died with either incidental alpha-synuclein pathology or diagnosed Parkinson's disease.

The clinching evidence came from a genetic experiment. When the team selectively dialled down GPNMB expression in microglia using CRISPR interference (CRISPRi), the cells became measurably worse at clearing alpha-synuclein aggregates from neurons. That places GPNMB as an active player in the protective mechanism, not a bystander.

Could this lead to new treatments?

Professor Tofaris said: "Our findings highlight that the immune response in Parkinson's disease is more nuanced than simply being beneficial or harmful."

He added: "We have identified a population of human microglia that can actively remove pathological alpha-synuclein from neurons. Understanding how to enhance and monitor such beneficial microglial functions, without triggering damaging inflammation, could open up new avenues for developing disease-modifying treatments."

That balance is the central challenge. Microglia sit on a knife-edge: too little activity and toxic aggregates accumulate; too much and inflammation damages the very neurons they protect. Any future therapy would need to boost the nibbling response while leaving the IL-10 brake and other control signals intact.

There are caveats to keep in mind. The trogocytosis findings come from stem cell models in a dish, which, while human in origin, do not fully reproduce the complexity of a living brain. The post-mortem tissue findings on GPNMB support the relevance to real disease, but the work has not yet been tested in patients. Whether enhancing this mechanism can actually slow Parkinson's in people remains an open question for future research.

Still, the study reframes how scientists think about immunity in Parkinson's. Instead of casting microglia as villains of inflammation, it reveals a precise, regulated and potentially harnessable cleaning crew — one that, with the right guidance, might someday help clear the protein clumps that drive the disease.

The paper, by Hung-Ju Chueh et al., appears in Science Translational Medicine (DOI: 10.1126/scitranslmed.adz9258).