Natasha Hill’s evidence came out of a microscope: three images per brain region, per animal, from three male mice per group. None of those mice was put in a maze, set to find a hidden platform, or tested on anything resembling memory. That is the study behind the headline ScienceDaily ran on November 2: “Scientists discover new Alzheimer’s drug that repairs DNA damage and reduces brain inflammation.”
reproductive toxicology data were not complete in female models
“Discover” is generous. Hill’s paper, written with colleagues at King’s College London, appeared in FEBS Open Bio in July, when King’s and the journal’s publisher, Wiley, issued releases on July 8 and ScienceDaily ran a nearly identical headline, “New Alzheimer’s drug repairs DNA damage and reduces brain inflammation,” on July 17. The November version added one verb and four months of distance from anyone who had read the methods section.
A spinal cord drug, given by needle
KCL-286 was built for something else. Jonathan Corcoran’s group at King’s designed it to treat spinal cord injury, as a small molecule that switches on retinoic acid receptor-beta (RARβ), one of the receptors through which retinoic acid, a vitamin A metabolite, steers nerve growth. Earlier work suggested it helped neurons mend DNA double-strand breaks, which Corcoran, in the King’s release, likens to “a rope snapping completely in two, rather than just fraying at the edges.” Those breaks accumulate in neurons early in Alzheimer’s, so the team tried the spinal cord compound in a dementia model.
The releases sell a shortcut. Corcoran calls KCL-286 “a first-in-class, orally bioavailable small molecule” that “has already successfully cleared Phase 1 human safety and tolerability trials.” Maria Goncalves, who managed the drug’s development, says the results highlight its potential as “a disease-modifying therapy rather than simply addressing symptoms.” In Alzheimer’s, disease modification means changing the course of the illness, and the mouse study measured neither a course nor a symptom. It measured tissue, once, at the end.
The animals were male Tg2576 mice, engineered to overproduce human amyloid, plus normal littermates. From 15 to 18 months of age, the treated group received 1 mg/kg of KCL-286 three times a week by intraperitoneal injection, dissolved in a vehicle of 80 percent DMSO. The orally bioavailable molecule reached these mice through a needle.
The headline marker of broken DNA, neuronal γH2AX fluorescence, averaged 95.11 arbitrary units in untreated model mice, 73.74 in treated ones and 50.71 in normal mice, a drop of roughly 22 percent that stopped well short of healthy. The pooled result across all sampled regions carries a P value below 0.001, and the test behind it was run on 639 measurements, every one of which came from nine animals.
| Group | Value (arbitrary units) |
|---|---|
| Untreated model mice | 95.11 |
| Treated model mice | 73.74 |
| Normal mice | 50.71 |
The effect held in the frontal cortex and in one hippocampal subfield, CA2/CA3. Elsewhere in the hippocampus, the structure most closely tied to memory, the authors describe the reductions as “more modest” and say they “did not uniformly reach significance.” The repair protein BRCA1 climbed to 89 puncta per neuron in treated mice, against 58 in untreated mice and 42 in normal ones, overshooting the healthy level by a wide margin. Microglia, the brain’s resident immune cells, covered 5.32 percent of tissue in treated mice versus 8.56 percent in untreated ones, though regionally that reached significance only in CA1, and the astrocyte marker GFAP lost significance after statistical correction.
Then there are the measurements a reader would assume had been made. The paper reports no behavioral or memory testing, no quantification of amyloid plaques, and no drug levels in blood or brain from this experiment. The model also, as the authors concede, “lacks overt tau pathology,” which means the tau-driven loss of BRCA1 they consider most relevant to human Alzheimer’s was never in the room. The researchers measured the hippocampus. They did not measure what it does.
Secondhand coverage wandered further. BioNews reported on July 20 that the drug increased “the BRCA2 protein”; the paper says BRCA1. In Medical News Today, consultant neurologist Steven Allder of Re:Cognition Health said the treatment targets the PTPσ receptor, which is not the receptor in the paper’s title. The same piece was one of the few to tell readers the mice were all male, three to a group, and it quoted physician Dung Trinh asking for “a clear dose-response relationship, adequate penetration into the brain and durable effects after treatment.” The study tested one dose and measured no brain levels at all.
Seven days, all men
The human data come from a Phase 1 trial published in the British Journal of Clinical Pharmacology in 2023, funded by the UK Medical Research Council and registered as ISRCTN12424734. It enrolled 109 healthy men at two UK sites, with inclusion limited to ages 19 to 45. Of those given the drug, 42 received a single dose between 1 and 100 mg, and 30 took it once daily for up to a week. There were no serious adverse events. The side effects were the ones you would predict from a drug working the vitamin A pathway: dry skin, peeling, rash, eye complaints, and one mild, reversible rise in liver enzymes after a single 12 mg dose. The authors report that the doses exceeded the plasma exposures their preclinical models suggested might matter clinically.
The trial included no women, because “reproductive toxicology data were not complete in female models,” a standard precaution for a retinoid-family drug and an awkward one for a disease whose patients are mostly older women. And the longest anyone took KCL-286 was a week, in men whose average age sat in the late twenties, for a condition that would be treated for years. “Cleared Phase 1” is accurate. What it cleared was a week of pills in young men.
Who pays, and who holds the patent
KCL-286 is an academic molecule, developed on Medical Research Council and Wellcome Trust grants, and Goncalves and Corcoran hold a composition-of-matter patent on it, which the paper discloses and which belongs next to every use of the phrase “disease-modifying.” When the Phase 1 results came out in 2023, King’s said the team was seeking funding for a Phase 2a trial in spinal cord injury. The July Wiley release says that “with appropriate funding,” investigators can next test whether the approach helps patients, and quotes Corcoran: “The opportunity is immediate, and the science is ready to advance.”
Any Alzheimer’s drug that gets that far will be measured against the amyloid antibodies. Eisai and Biogen’s lecanemab was tested in a 1,795-person trial in which, after 18 months, patients on the drug had declined 1.21 points on the CDR-SB scale against 1.66 on placebo, a difference of 0.45 on an 18-point scale. Brain scans showed swelling or effusions in 12.6 percent of treated patients and micro- or macrohemorrhages in 17.3 percent. That is what the commercial standard in this disease looks like, a sliver of slowed decline bought with a measurable rate of bleeding and swelling in the brain.
The King’s group has noticed. In its discussion section, the paper proposes using KCL-286 alongside amyloid-directed immunotherapies, arguing the pairing “may therefore amplify therapeutic benefit.” A compound that has never been shown to improve a single mouse’s memory already has a suggested place beside the best-selling drugs in the field.
Corcoran told Wiley’s publicists that the team thinks of KCL-286 “as repairing potholes in a road.” His own paper counted fewer potholes in the brains of three treated mice, and never checked whether any of them drove better afterward.
Sources
- Hill et al., FEBS Open Bio (2026), full text
- Goncalves et al., British Journal of Clinical Pharmacology (2023), Phase 1 trial
- King’s College London via EurekAlert!, “Human-safe drug repairs DNA in a mouse model of Alzheimer’s” (July 8, 2026)
- Wiley via EurekAlert!, “Could an investigational drug that targets DNA damage help treat Alzheimer’s disease?” (July 8, 2026)
- ScienceDaily, “Scientists discover new Alzheimer’s drug that repairs DNA damage and reduces brain inflammation” (November 2, 2026)
- ScienceDaily, “New Alzheimer’s drug repairs DNA damage and reduces brain inflammation” (July 17, 2026)
- Medical News Today, “DNA repair drug fixes early Alzheimer’s brain damage in mouse model”
- King’s College London via EurekAlert!, Phase 1 spinal cord injury results (August 10, 2023)
- Healio, lecanemab CLARITY AD results (2022)
- BioNews, “Drug candidate targets DNA damage in Alzheimer’s disease model” (July 20, 2026)