Mouse atlas identifies 92 protective proteins crossing brain barrier

By Tom Whitfield

Of 861 tagged mouse blood proteins tested for entry into the brain, 92 crossed the blood-brain barrier, producing what the authors describe as the first atlas of neuroprotective proteins imported into the brain.

Researchers at the Gladstone Institutes, the University of California and other institutions published the work in Cell under the title “The blood-brain barrier imports diverse neuroprotective proteins through pathways impaired early in Alzheimer’s disease models.”

Haoyue Zhou, PhD, a postdoctoral scholar in Andrew Yang’s lab, is first author. Yang, a Gladstone investigator, is a co-author.

What did the researchers measure?

The team developed a bioorthogonal proteomics strategy, described in the paper as an in vivo click-enrichment screen of labeled plasma proteins entering the brain. Its labeling chemistry followed proteins from organs including the liver and kidney without disrupting their biology.

They tagged various mouse blood proteins, injected them back into the bloodstream and examined brain tissue for entry. Of the 861 proteins tracked, 92 reached the brain; most were produced primarily in the liver, with adipose tissue and the intestine also contributing.

After entry, the proteins reached specific brain cell types for specialized functions. Yang described the barrier as an intelligent interface that selects useful blood proteins from distant parts of the body and routes them to the right brain cells.

The method’s benchmark against a common biotin-labeling approach produced consistently stronger signals in plasma and brain vessels. It also enabled what the paper calls robust detection of plasma proteins in neurons, astrocytes and microglia.

For lab workflow, that matters because the chemistry offered stronger reads across blood, vessels and three brain cell types. The source provides no quantitative sensitivity, specificity, throughput or cost comparison, so performance claims need replication before teams replace an established workflow.

What does the Alzheimer’s model show?

Apolipoprotein AI, or APOA1, provides the clearest example. The liver-made protein activates a pathway needed to maintain healthy astrocyte connections once it reaches the brain; the source cites studies linking its absence to greater Alzheimer’s risk, without giving their designs or cohort sizes.

As Alzheimer’s disease progresses, the barrier breaks down and permits harmful proteins to leak into the brain. The labeling method detected another change first: even before overt breakdown, model mice stopped importing protective proteins such as APOA1. The team linked that loss to eroded neuronal connections, which contribute to behavioral changes and other early disease signs.

Zhou, the first author, said: “We think of blood proteins as messengers carrying signals from distant body parts, and the cells of the blood-brain barrier import only the ones the brain really needs.”

The team then tracked proteins in mice engineered to produce extra APOA1 in the liver. As more APOA1 entered the brain, the animals showed fewer signs of disease, an association reported alongside the impaired-import findings.

That result offers a hypothesis, not a treatment. The source account contains no drug, dose or human trial, and it gives no group sizes, effect sizes, follow-up duration or statistical tests for the overproduction experiment.

Yang framed the therapeutic hypothesis: “Alzheimer’s disease may not start entirely inside the brain … If that’s right, we may be able to fight this disease from outside the brain, by restoring what the brain has lost.”

Which treatment routes remain proposals?

• Increase production of beneficial proteins such as APOA1. The mouse experiment tested liver-driven overproduction; translation into a drug strategy remains untested here.
• Use the brain’s natural “ZIP code” system. Researchers propose using that targeting mechanism to deliver therapies directly to specific brain cell types.
• Harness barrier transport machinery. Yang and colleagues are exploring whether the blood-brain barrier’s import pathways can help drugs reach the brain more efficiently.

These are therapeutic proposals, not validated treatments. The study reports no human tests, therapeutic candidates, dosing regimens or clinical endpoints.

What should R&D teams check before acting on the atlas?

The source names the Gladstone Institutes, the University of California and other institutions but does not identify a funding sponsor. It also omits group sizes and statistical uncertainty for key experiments, so portfolio committees should check the full paper and its disclosures before assigning development value to the findings.

Replication should quantify whether protective import falls before barrier leakage, whether restoring APOA1 improves connections and behavior, and whether engineered liver overproduction changes brain exposure. Those questions separate a delivery mechanism from a therapeutic effect.

For R&D leaders, the next funding decision is which route to test first: restore the liver supply line, target brain cell types through the natural delivery code, or borrow transport machinery for drugs. Until mouse-model findings are replicated with group sizes and quantified outcomes, the atlas supports target discovery and assay planning, not a clinical forecast; the immediate test is whether restoring the body-to-brain protein supply improves brain endpoints.