Redesigned versions of the plant compound fisetin cleared aging cells better in lab tests and mice

As we age, some of our cells stop dividing but refuse to die. These senescent cells (often called “zombie cells”) pile up in tissues and release a mix of inflammatory signals, known as the SASP, that damages the cells around them. Drugs that kill senescent cells while sparing healthy ones are called senolytics, and in mice they have eased many problems of old age.

One of the best-known candidates is fisetin, a natural member of the flavonoids, a large family of plant compounds found in fruits and vegetables. It is being tested in several clinical trials, for conditions such as kidney disease and frailty. But it is a weak senolytic: it needs high doses, it dissolves poorly and the body absorbs little of it.

A team at the University of Minnesota set out to build a better version. Since it is unclear which protein fisetin acts on, they tested compounds directly on senescent cells and kept what worked, an approach called phenotypic drug discovery.

What they did

The researchers first screened 38 natural flavonoids on fibroblasts, the cells that make connective tissue, taken from mouse embryos. Some came from mice with a weakened DNA-repair gene called Ercc1, whose cells easily become senescent under stress; normal, dividing cells were the comparison. For each compound they measured the dose that killed half of the senescent cells (lower is stronger) and how much more it hit senescent than healthy cells, a “selectivity” score.

Fisetin did only moderately: it needed about 23 micromolar and was just 1.5 times more toxic to senescent cells than to healthy ones. The best natural compound, at least in these cells, was luteolin. By comparing which parts of the molecules helped or hurt, the team worked out design rules, then made and tested 35 new flavonoids. Two stood out: SR29384, the most selective, and SR31133, the most potent. Both were then tested on human lung fibroblasts and human blood-vessel cells, and in mice.

What they found

• In cells: SR29384 was 2.9 times and SR31133 27.5 times more potent than fisetin. SR31133 worked at about 0.8 micromolar, but was less selective than SR29384. Both reduced senescent cells in all the cell types tested.
• In old mice: mice aged 28 months (very old for a mouse) got fisetin or one of the new compounds through a feeding tube for five days, at 20 mg per kg of body weight a day. That is a fifth of the dose used in earlier fisetin studies, chosen on purpose as a tough test. Fisetin had only a small effect. The new compounds, especially SR29384, lowered the activity of genes that mark senescent cells (p16 and p21) and of several inflammatory SASP genes in organs including the kidney, brain, lung, spleen, fat and muscle.
• In fast-aging mice: the Ercc1 mice show signs of old age early, such as tremor, a hunched back and poor coordination. Those given SR29384 three times a week for five weeks had a lower overall symptom score, mainly in dystonia (abnormal muscle tension) and ataxia (unsteady movement). Senescence genes went down, or tended to, in the kidney, liver, lung and brain.
• How they may work: gene-activity data, computer predictions and enzyme tests pointed to two proteins. Both compounds blocked PARP1, an enzyme involved in DNA repair, more strongly than fisetin did, and blocked CDK2, an enzyme that drives cell division (SR29384 less strongly). A third suspect, BCL-xL, a protein that helps cells avoid self-destruction, was not blocked directly. Better antioxidant power did not explain the difference: all three compounds were similar there.

What to keep in mind

• Mice and dishes only. No person has received these compounds. They are experimental molecules, not approved drugs.
• Small mouse groups. There were 6 mice per group in the old-mouse test, and only 4 treated and 8 untreated fast-aging mice. In the fast-aging mice, the paper reports no significant difference in the overall symptom score at week 16.
• Mixed results across organs. The authors write that SR29384 beat fisetin on some measures but did not show “uniform superiority across all tissues”.
• Markers, not lifespan. The old mice were treated for five days and examined two days later; the team measured gene activity, not lifespan or physical fitness.
• How much reaches the body is unknown. The team did not measure how much of each compound got into the blood or organs, and call such studies important next steps. Safety was only estimated with computer predictions, which the authors call preliminary; long-term cancer risk, for example, still has to be tested.
• How they work is not settled. The authors say other pathways may also be involved.
• Conflicts of interest. Two of the researchers co-founded a company that develops drugs against senescent cells, and some team members and the university have filed a provisional patent on these compounds.

Why it matters

Many known senolytics are weak, poorly absorbed or toxic. This study shows that a natural compound with modest activity can be chemically reshaped into far more potent molecules, at least in the lab and in mice. The authors see the two compounds as candidates for further preclinical and, eventually, clinical testing. Whether they are safe and useful in people is a question for future studies, not this one.