Senolytics Explained: What Clearing ‘Zombie’ Cells Has and Hasn’t Shown

Some damaged cells don’t die. They stop dividing for good, linger in the tissue and release a steady stream of inflammatory signals. Biologists call them senescent cells; headlines call them zombie cells. In mice, clearing them delays several features of ageing, and since 2015 a class of compounds called senolytics has been developed to do exactly that. The human evidence so far comes from small trials with mixed results.

This guide explains what cellular senescence is, how senolytics were discovered, which compounds are studied, what the human trials have and haven’t shown, where the peptide FOXO4-DRI fits, and why no senolytic is approved for ageing anywhere.

What senescent cells are

Cellular senescence is a state in which a cell permanently stops dividing but does not die. It can be triggered by DNA damage, by telomeres wearing down after many rounds of division, by cancer-causing genes switching on, and by chemotherapy or radiation. In the short term this is protective: a damaged cell that cannot divide cannot grow into a tumour. Our guide to telomeres and telomerase covers one of the triggers in detail.

The trouble is that senescent cells linger. They resist apoptosis, the orderly self-destruction that removes most damaged cells, and they secrete inflammatory cytokines, chemokines and protein-cutting enzymes. In 2008, Coppé and colleagues named this output the senescence-associated secretory phenotype, or SASP. Senescent cells accumulate with age, and in 2013 “cellular senescence” was listed as one of nine hallmarks of ageing in an influential review in Cell; the 2023 update expanded the list to twelve. Their refusal to die, and the damage they do to the tissue around them, is what earned them the nickname “zombie cells”.

In the lab, senescent cells are usually identified by a combination of markers: the cell-cycle brakes p16Ink4a and p21, and a stain for senescence-associated β-galactosidase (SA-β-gal). No single marker is definitive, which matters when trials try to measure whether a drug cleared them.

From the Hayflick limit to senolytics

1. 1961Leonard Hayflick and Paul Moorhead report that normal human fetal cells stop dividing after about 40 to 60 divisions in culture. The term “Hayflick limit” came later, coined by Macfarlane Burnet in 1974.
2. 2008Coppé and colleagues name the senescence-associated secretory phenotype (SASP).
3. 2011A Mayo Clinic team shows that clearing p16Ink4a-positive cells in fast-ageing mice delays age-related disorders.
4. 2013Cellular senescence is named one of nine hallmarks of ageing.
5. 2015Mayo Clinic and Scripps Research scientists introduce the term “senolytics” and show that dasatinib plus quercetin reduces the burden of senescent cells in mice.
6. 2016Clearing p16Ink4a-positive cells extends median lifespan in normal mice, and the experimental cancer drug navitoclax is shown to kill senescent cells.
7. 2017Researchers at Erasmus MC in Rotterdam report the senolytic peptide FOXO4-DRI.
8. 2018Studies report that D+Q improves physical function in old mice and that fisetin extends mouse lifespan.
9. 2019The first small human pilot studies of D+Q are published.
10. 2020–2025Unity Biotechnology’s two lead candidates miss their main trial goals, and the company dissolves in 2025.
11. 2024–2026Randomized trials report a missed bone endpoint in 2024 and improved liver fibrosis, with more adverse events, in 2026.

Canadian labs are part of the field. In Montréal, groups led by Francis Rodier and Gerardo Ferbeyre at the Université de Montréal and its hospital research centre, the CRCHUM, study senescence and its links to cancer.

The mouse studies behind the idea

Genetic clearance. The case for senolytics rests on mouse genetics. A Mayo Clinic team built INK-ATTAC mice, in which cells that switch on the senescence marker p16Ink4a can be made to self-destruct on command. In 2011, clearing these cells in fast-ageing (BubR1 progeroid) mice “delayed onset” of age-related problems in fat, muscle and the eye. Survival was “not substantially extended”, because these mice die of heart failure, so the result was about healthspan rather than lifespan.

In 2016, the same approach in normal mice extended median lifespan by 27% in one genetic background and 24% in another. Maximum lifespan rose only in the mixed-background mice. Mayo’s announcement put the range at 17% to 35% across subgroups, which is where the often-quoted “35%” comes from. All of this used a genetic switch, not a drug.

The first drugs. In 2015, Yi Zhu, Tamara Tchkonia, James Kirkland and colleagues used gene-expression data to look for the survival pathways senescent cells depend on, the “Achilles’ heel” of the paper’s title, and for drugs that block them. They introduced “a new class of drugs termed senolytics”. Dasatinib eliminated senescent human fat-cell progenitors, quercetin worked better against senescent endothelial cells, and the combination reduced the burden of senescent cells in aged, irradiated and fast-ageing mice.

Function and survival. In 2018, Xu and colleagues showed that transplanting “relatively small numbers of senescent cells” into mice caused lasting physical dysfunction. In old mice, intermittent D+Q “increased post-treatment survival by 36% while reducing mortality hazard to 65%”. That 36% refers to the survival remaining from the start of treatment in very old mice, not to total lifespan.

The main senolytic candidates

“Senomorphics” is the term reviews use for SASP suppressors. A 2015 study found that rapamycin “selectively blunts the pro-inflammatory phenotype of senescent cells”, which makes it a commonly cited example.

What human trials have shown

Human senolytic research has mostly tested D+Q, given in short, intermittent courses. The studies below are the main published ones. Our guide to clinical trial phases explains what “open-label”, “phase 1” and “randomized” mean for how much weight each one can bear.

The 2026 liver trial, published on October 1, 2026 in Nature Metabolism, is the most encouraging controlled result so far: it was double-blind and placebo-controlled, and fibrosis is a meaningful outcome. It was also small, the authors describe it as “proof-of-principle”, and adverse events were nearly twice as common on D+Q. It needs to be repeated in a larger trial before it can be read as more than a promising lead.

The Unity Biotechnology story

Unity Biotechnology was one of the most prominent companies built on senolytics. Its first candidate, UBX0101, showed no significant benefit over placebo in knee osteoarthritis in August 2020. Its second, the BCL-xL inhibitor foselutoclax, missed the primary analysis of its ASPIRE eye-disease trial in May 2025. Unity laid off its staff that month, its shares were suspended from Nasdaq in July, and stockholders approved its dissolution in September 2025. The setbacks don’t disprove the biology, but they show how hard it has been to turn mouse results into medicines.

Why the results are hard to read

• Most studies are tiny and open-label. Trials of 5 to 14 people without a placebo group can’t separate a drug effect from chance or expectation.
• There is no agreed way to count senescent cells in people. Trials measure different markers in different tissues, such as p16 and p21, SA-β-gal staining or blood SASP factors, so “fewer senescent cells” can mean different things.
• Getting the drug to the target is not guaranteed. In the Alzheimer’s pilot, quercetin was not detected in cerebrospinal fluid.
• Mouse numbers are often stretched. The 2011 study showed healthier ageing, not longer life; the “35%” lifespan figure is the top of a subgroup range; and the 36% survival gain applies to time remaining in very old mice.
• Side effects are real. Dasatinib is a cancer drug, navitoclax lowers platelets, and adverse events were nearly twice as common on D+Q in the 2026 liver trial.
• Senescence has jobs. It is a barrier against cancer, so removing senescent cells repeatedly over years could carry costs that short trials cannot detect.

Where FOXO4-DRI fits

FOXO4-DRI is the best-known peptide senolytic. Reported in 2017 by Peter de Keizer’s group at Erasmus MC in Rotterdam, it is a D-retro-inverso peptide designed to break the bond between the proteins FOXO4 and p53 inside senescent cells. Released from FOXO4, p53 leaves the nucleus and the cell undergoes apoptosis. In mice, the peptide countered the toxicity of the chemotherapy drug doxorubicin and “restored fitness, fur density, and renal function in both fast aging XpdTTD/TTD and naturally aged mice”. No human trials have been published.

Unlike D+Q, which repurposes existing drugs, FOXO4-DRI was designed from the biology of senescence itself, and it has become one of the best-known senolytic research tools. Our full guide, FOXO4-DRI: the senolytic peptide explained, covers its design and the evidence in detail, and our anti-aging peptides guide maps it against other ageing-research peptides.

What’s established and what isn’t

Senolytic research reagents in Canada

Most senolytic candidates are small molecules, but FOXO4-DRI is a peptide, and it raises testing questions of its own. Before ordering, check that:

• The mass matches the expected sequence. For a peptide of this size, the report should give a deconvoluted molecular mass. See how to read a peptide mass spectrum.
• You know what was not tested. FOXO4-DRI is built from D-amino acids, which have the same mass as their L-forms, and routine purity methods generally cannot tell them apart. If chirality matters to your work, ask whether it was checked. Our guide to D-amino acids explains why.
• Net peptide content is taken into account, because counter-ions such as TFA or acetate add to the weight of the powder. See net peptide content.
• The batch number on the certificate matches your vial, and the peptide is stored cold, dry and dark. See our COA guide and peptide storage guide.

FOXO4-DRISenolytic D-retro-inverso peptideView product EpitalonTelomerase research tetrapeptideView product NAD+Coenzyme studied in ageing researchView product MOTS-cMitochondrial-derived peptideView product

Frequently asked questions