A Cell study of 190 Alzheimer's brains finds that microglia carry somatic mutations in blood cancer driver genes — turning the existing oncology drug pipeline into a candidate shortlist for a disease that has spent two decades chasing amyloid and tau.
Alzheimer's is usually described as a slow protein problem. Amyloid plaques, tau tangles, dying neurons. A new study from Boston Children's Hospital, published in Cell on April 21, 2026 (doi:10.1016/j.cell.2026.03.040), inserts a different kind of biology into that picture: blood-cancer-style mutations quietly accumulating inside the brain's immune cells. The finding — based on deep sequencing of 311 brain samples from 190 Alzheimer's patients and 121 age-matched controls — reframes a slice of Alzheimer's as a clonal disease of microglia-like brain macrophages, and recasts the existing oncology pipeline as a candidate shortlist rather than a guess.
In the study, researchers led by Christopher Walsh, a Howard Hughes Medical Institute investigator and chief of genetics and genomics at Boston Children's, analyzed 149 genes known to drive blood cancers in brain tissue from 190 people who had Alzheimer's and 121 age-matched controls. The Alzheimer's samples contained significantly more single-letter DNA changes in those cancer-driver genes than the healthy tissue, according to the Cell paper abstract and its ScienceDaily press summary. Five of those cancer-driver genes recurred specifically in the Alzheimer's microglia-like cells and not in controls — a statistically significant enrichment (p < 0.05, one-tailed proportion test) in tumor suppressor genes, not oncogenes. The three genes highlighted as exhibiting significant positive selection in AD brains are TET2, ASXL1, and DNMT3A, all well-established drivers of clonal hematopoiesis. The five AD-enriched genes in total are named in Figure 2E of the paper (referenced via Table S3), and their precise identities should be confirmed against the full paper tables during fact-check.
The mechanism proposed by the team borrows from clonal hematopoiesis (CH), the documented phenomenon in which a single blood stem cell picks up a cancer-driving mutation and expands. In Alzheimer's patients, the team found those same mutations in circulating blood cells, and the mutant cells appear to cross a leaky blood-brain barrier and take up residence in the brain as microglia-like macrophages. Once there, they exhibit inflammatory and proliferative transcriptional signatures characteristic of disease-associated microglia, releasing signals that damage synapses and accelerate cognitive decline. The co-leads, Alice Eunjung Lee and August Yue Huang, frame the overlap with blood-cancer genes as a clue — not a diagnosis. Walsh has called the picture "a little like cancer," a framing the team uses as a research roadmap, not a clinical claim.
A companion preprint from the same group extends the logic to risk: Huang and Lee report that blood-cancer-driver mutations in peripheral blood predict later Alzheimer's diagnosis independently of APOE4, the most famous genetic risk factor for the disease. That result, currently posted as a preprint and not yet peer-reviewed, matters because APOE4 carriers account for only a fraction of cases, and the field has long needed a way to flag the rest of the at-risk population. If the finding holds, a routine blood test could flag at-risk patients years before symptoms appear — a much earlier screening path than today's cognitive testing.
For drug developers, the practical implication is to treat the existing oncology pipeline as a candidate list. Walsh has framed the cancer therapeutics already developed against the same driver genes — particularly TET2, ASXL1, and DNMT3A — as a repurposing pool to be evaluated for their ability to quiet inflammatory microglia. The Cell study gives those efforts a target list. Lee has pointed to blood-based genetic screening as a future diagnostic path.
The limits are real. The Cell paper is a single postmortem study of 190 AD and 121 control brains from the ROSMAP cohort; microglia-like brain macrophages are one driver among several in a disease that also involves amyloid, tau, vascular decline, and aging. The three genes functionally validated in iPSC-derived microglia — TET2, ASXL1, and DNMT3A — represent a subset of the broader five AD-enriched genes, and the full five-gene list needs confirmation from the paper's tables. Cancer-drug repurposing into neurodegeneration has a long and mixed history. The bioRxiv result on APOE4-independent risk is a preprint, not peer-reviewed, and its sample size and statistical metrics have not yet been independently confirmed. Walsh himself has been careful to describe the cancer framing as a research roadmap, not a therapy.
What changes is the search space. Alzheimer's research has spent two decades testing compounds against amyloid and tau, with limited clinical return. The Walsh, Lee, and Huang work argues that a third track — mutation-driven inflammation in the brain's immune cells — deserves systematic drug-by-drug interrogation. The oncology world has already built the compounds, the assays, and the patient registries. The next question is whether the neuroscience world can build the matching trials fast enough to test them.