The biological premise behind the trial rests on one of the hallmarks of ageing: cellular senescence. Senescent cells are cells that have essentially permanently stopped dividing, usually in response to stressors such as DNA damage, oxidative stress or oncogenic signalling. In the short term, senescence is useful — it helps wounds heal and stops damaged cells from becoming cancerous. But when senescent cells linger, they become a problem. They resist apoptosis through activated anti-death pathways, evade clearance by the immune system, and pump out a cocktail of proinflammatory and profibrotic signalling molecules collectively known as the senescence-associated secretory phenotype. In the liver, this chronic inflammatory noise has been implicated in driving both the inflammation and the fibrogenesis that define MASH, and senescence signatures have also been documented in alcohol-associated hepatitis and primary sclerosing cholangitis.

Senolytic drugs exploit a specific vulnerability. Because senescent cells depend on particular prosurvival networks to avoid dying, drugs that interfere with those networks can selectively eliminate them while sparing healthy, proliferating cells. Dasatinib, an approved tyrosine kinase inhibitor, and quercetin, a naturally occurring flavonoid, target complementary senescence-associated survival pathways, and the pairing emerged from transcriptomic analyses showing which survival nodes senescent cells rely on most. Earlier phase-1 studies in idiopathic pulmonary fibrosis and diabetic kidney disease had suggested that intermittent dosing could reduce senescent cell burden in humans with an acceptable safety profile, but no one had yet shown that this approach could change liver histology in MASH.

The Amsterdam trial enrolled 31 participants with biopsy-proven fibrotic MASH, with a median age of 56 years; 76 percent were male and 58 percent had type 2 diabetes. Participants were randomized one-to-one to receive either dasatinib 100 milligrams per day plus quercetin 1,000 milligrams per day, or matching placebo. The dosing schedule was deliberately intermittent: three consecutive days per week for three weeks, followed by a four-week drug-free period, with the seven-week cycle repeated three times. The logic is that senescent cells take weeks to redevelop and do not divide, so brief pulses of treatment should be enough to clear them while limiting cumulative drug exposure. Twenty-seven participants completed the study, and all randomized participants were included in the intention-to-treat analysis, with those who discontinued counted as non-responders.

The primary endpoint was demanding: at least a one-stage improvement in fibrosis score on the Steatosis Activity Fibrosis system without any worsening of MASH, assessed on paired liver biopsies read centrally by three blinded pathologists. Forty-seven percent of participants in the treatment arm met this endpoint versus 7 percent on placebo, an absolute difference of 40 percentage points with a P value of 0.021. MASH resolution — improvement in the hepatocyte ballooning that defines the disease — occurred in 53 percent of treated patients versus 7 percent of controls, and the NAFLD Activity Score fell substantially more in the treated group. The SAF-activity score also improved significantly, while improvement on the EPoS fibrosis staging system showed a similar directional trend that narrowly missed statistical significance.

What elevates the trial beyond a simple histological result is the molecular layer. The team performed single-nucleus RNA sequencing on 53 frozen liver biopsy samples, profiling gene expression in individual liver cells before and after treatment. Because canonical senescence markers such as p21, p16 and p53 were expressed at low levels in the sequencing data, the researchers used curated gene signatures — the SenMayo set and a senescent hepatocyte gene signature — to capture senescence activity across the tissue. In the treated group, both senescence signatures were significantly depleted by the end of the study, while the placebo group showed no such change or even the opposite pattern.

The sequencing also revealed a striking remodelling of the cellular composition of the liver. Treated participants showed a significant reduction in the proportional abundance of hepatic stellate cells — the fibrogenic workhorses of the liver — along with decreases in myeloid cells, B cells and T cells, accompanied by a relative increase in hepatocytes. Mixed-effects logistic regression confirmed the reduction in immune cell and stellate cell presence. A fibrosis gene signature, including collagen genes COL1A1 and COL1A2, the profibrotic signalling molecule TGFB1 and the stellate cell activation marker ACTA2, was significantly downregulated after treatment, whereas the placebo group showed enrichment of the same fibrosis program. Notably, no cancer-related transcriptional pathways were enriched despite the reduced senescence signatures, an important check given that hepatocyte senescence can both suppress and, in some contexts, promote tumour development.

Safety findings were nuanced. Adverse events were significantly more frequent in the treatment group, affecting 82 percent of participants versus 43 percent on placebo, with headache the most commonly reported complaint in the treated arm. However, nearly all events were mild to moderate and self-limiting, and there were no cases of myelosuppression or QTc prolongation — the cardiac and bone marrow toxicities most feared with dasatinib. One serious adverse event, a planned knee operation, was deemed unrelated to treatment, and one severe event involving difficulty swallowing was traced to an oesophageal malignancy in a participant who had not disclosed an active oncological condition during screening. On quality-of-life questionnaires, treated participants reported significantly better general health on the SF-36, though most other domains showed no significant differences.

The authors are candid about the limitations. The number of responder events was small, producing a wide confidence interval around the relative risk of 6.59 whose lower bound approached the null. Four sensitivity analyses using different assumptions about missing biopsy data consistently preserved the direction of the effect, with relative risks ranging from 3.29 to 9.06, and statistical significance held in three of the four scenarios; only the most conservative worst-case analysis failed to reach significance. The team emphasizes that the histological response finding should be considered hypothesis-generating rather than confirmatory pending replication in a larger trial. The 21-week follow-up is also short for a fibrosis endpoint, and the modest sample size limited the ability to detect changes in non-invasive markers such as FIB-4, liver stiffness measurements and the enhanced liver fibrosis test, which did not differ between groups.

Even with those caveats, the implications are considerable. Metabolic dysfunction-associated steatotic liver disease is the most common liver disorder worldwide, with an estimated prevalence approaching 40 percent, and fibrosis stage is the strongest predictor of liver-related complications and mortality. Current approved therapies for MASH act primarily on metabolic drivers, whereas the single-cell data here suggest that senolytics may strike at the fibrogenic machinery itself, depleting stellate cells and inflammatory populations rather than merely resolving steatosis as a secondary effect. The antifibrotic reach of senolytics may extend beyond the liver, with prior studies reporting benefits in idiopathic pulmonary fibrosis and diabetic kidney disease. The Amsterdam team argues that future trials must now define optimal dosing, cycle number, drug-free intervals, patient selection and long-term safety — but the proof of principle, that ageing cells can be evicted from a scarred human liver and that the tissue responds by healing, has now been demonstrated under blinded, randomized, placebo-controlled conditions.

Subject of Research: Senolytic therapy with dasatinib and quercetin for fibrotic metabolic dysfunction-associated steatohepatitis

Article Title: Senolytics dasatinib and quercetin in metabolic dysfunction-associated steatohepatitis: a proof-of-principle randomized, controlled trial

Article References: Koning, M., Kovynev, A., Fondevila, M. F., Ruhe, E. J. M., Mommersteeg, M. C., Ramakers, C., Hutten, B. A., Verwer, B., Vlug, M., Kuiper, T., van den Berg, M., Vrieze, A., Bizino, M., van den Beld, A., Berk, L., Boerlage, T., Augustijn, Q. J. J., Brouwer, W. P., Chen, W. J., … Meijnikman, A. S. (2026). Senolytics dasatinib and quercetin in metabolic dysfunction-associated steatohepatitis: a proof-of-principle randomized, controlled trial. Nature Metabolism. https://doi.org/10.1038/s42255-026-01643-4

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01643-4

Keywords: senolytics, dasatinib, quercetin, MASH, liver fibrosis, cellular senescence, randomized controlled trial, hepatic stellate cells, single-nucleus RNA sequencing, Nature Metabolism, fatty liver disease, senescence-associated secretory phenotype