🧬 Aging Wiki

senolytics

Properties1
aliasessenolytic drugs, senolytic agents, SCAP-targeting drugs, senotherapeutics

23 min read

Senolytics

A class of pharmacological agents that selectively kill senescent cells by disrupting the Senescent-Cell Anti-Apoptotic Pathways (SCAPs) that protect senescent cells from their own pro-apoptotic SASP environment. The concept and naming were established by Zhu, Tchkonia, Kirkland and colleagues at Mayo Clinic in 2015 1. Senolytics have moved from preclinical proof-of-concept to multiple Phase 2 human trials over a single decade — the most clinically-advanced anti-aging intervention class as of 2026.

This page covers the drug class as a whole. Specific compounds live in molecules/compounds/:

  • fisetin (verified) — flavonol; cell-type-specific clearance; Phase 2 trials ongoing
  • quercetin (verified) — flavonol; D+Q “Q”; HUVEC-active via BCL-xL / EFNB1 (not PI3Kδ, which is preadipocyte-selective per Zhu 2015)
  • dasatinib (verified) — TKI; D+Q “D”; preadipocyte-active via ephrin/EPH receptors
  • navitoclax (planned) — Bcl-2/Bcl-xL inhibitor; thrombocytopenia-limited
  • a1331852 (planned) — selective Bcl-xL inhibitor
  • procyanidin-c1 (verified) — grape-seed B-type procyanidin trimer; dual senomorphic (low-dose) / senolytic (≥50 µM) via NOXA/PUMA→ROS→mitochondrial apoptosis; aged-mouse lifespan extension (Xu 2021); preclinical, no human senolytic trial
  • piperlongumine (verified) — long-pepper alkaloid; OXR1-targeting → chronic ROS → senolysis; ~2.5–3.3× selectivity; navitoclax-synergistic; preclinical

Definitional clarifications

The literature uses several overlapping terms; the distinctions matter:

TermMeaning
SenolyticSelectively kills senescent cells (induces apoptosis or other death)
SenomorphicSuppresses SASP without killing senescent cells (e.g., rapamycin, JAK inhibitors, metformin, NF-κB inhibitors)
SenotherapeuticUmbrella term covering both senolytics and senomorphics
Hit-and-runIntermittent dosing strategy: senescent cells take days–weeks to repopulate after clearance, so continuous exposure isn’t required. Sublethal exposure carries a theoretical senescence-escape risk; the endogenous alternative is senescence-immune-surveillance

The SCAP framework

Zhu et al. 2015 (Aging Cell) 1 is the foundational paper that established the SCAP concept and identified the first senolytic combination (dasatinib + quercetin). The key insight: senescent cells that resist their own pro-apoptotic SASP environment do so by upregulating specific anti-apoptotic networks, and these networks are distinct across cell types.

Cell-type-specific SCAPs identified (Zhu 2015 Fig 1D for preadipocytes; refined by Zhu 2016 + Yosef 2016 + Zhu 2017 across HUVECs and IMR90):

Cell typeSCAP nodesEffective senolytics
Senescent HUVECsBCL-xL, EFNB1quercetin (Q); navitoclax; A1331852; fisetin (per Zhu 2017 — HUVECs only in that assay)
Senescent IMR90 fibroblastsMulti-target Bcl-2 family — no single member sufficient (see note below)navitoclax (hits BCL-2/BCL-xL/BCL-w); A1331852 (per Zhu 2017)
Senescent human preadipocytesEphrins (EFNB1/3); EPH receptors; PI3KCD; serpins (PAI-1/PAI-2); p21dasatinib (D); D+Q combination

Note on IMR90 SCAP (corrected from earlier “BCL-2 + BCL-w dominant” framing per bcl-2 and bcl-xl verifier rounds):

  • Zhu 2016 Fig 4A-B: BCL-xL siRNA alone or BCL-2 + BCL-xL together were insufficient for IMR90 senolysis; the full triple BCL-2 + BCL-xL + BCL-w combination was required.
  • Yosef 2016: BCL-W + BCL-xL dual knockdown produced ~53% IMR90 viability reduction; each alone had only minor effect.
  • Practical implication: IMR90 senescent cells require multi-target Bcl-2 family inhibition. Navitoclax (BCL-2/BCL-xL/BCL-w pan-inhibitor) is well-suited; single-target agents (e.g., venetoclax, BCL-2 only) are not.

Critical implication: no single senolytic clears all senescent cell types. D+Q was designed as a polypharmacological combination specifically to cover complementary cell-type populations. The clinical translation challenge is largely a cell-type-matching problem rather than a potency problem.

Major senolytic drug families

BH3-mimetics (Bcl-2 family inhibitors)

The mechanistically cleanest senolytic class — directly inhibit anti-apoptotic Bcl-2/Bcl-xL/Bcl-w proteins, releasing the BAX/BAK-mediated mitochondrial apoptosis cascade in cells that are already primed for apoptosis (as senescent cells are, due to chronic SASP-driven stress). See apoptosis-pathway § BH3-mimetics for mechanism.

AgentTargetStatusKey limitation
Navitoclax (ABT-263)BCL-2 / BCL-xL / BCL-wOncology Phase 2; senolytic preclinicalThrombocytopenia — BCL-xL is required for platelet survival; on-mechanism dose-limiting toxicity
A1331852BCL-xL selectivePreclinical senolyticSame platelet liability as navitoclax
ABT-737BCL-2 / BCL-xL / BCL-w (tool compound)Preclinical onlyPoor oral bioavailability
UBX0101 (Unity Biotechnology)MDM2 → p53/p21 axis (claimed)Phase 2 failed for osteoarthritis (UNITY-OA-101, 2020)Mechanism debated; may not act primarily as a clean senolytic
UBX1325 / foselutoclax (Unity Biotechnology)BCL-xL-selective small moleculePhase 2 positive trend (BEHOLD, NCT04857996) in diabetic macular edema 2; intravitreal routeLocal-delivery design sidesteps systemic thrombocytopenia. First BCL-xL-axis senolytic to show positive Phase 2 RCT signal in humans; class proof-of-concept for local-delivery BH3-mimetics.

Navitoclax in aged-mouse studies improved hematopoietic stem cell function (Chang et al. 2016 Nat Med 3 — the primary finding per the paper’s title; lung/liver/bone marrow senescent-cell clearance data attributed to this paper cannot be confirmed from the PDF, which is unavailable in the archive gap/no-fulltext-access — note: this is the correct attribution for the aged-mouse navitoclax data; it is NOT in Zhu 2016, which is in vitro only).

Tyrosine kinase inhibitors

AgentSenolytic mechanismStatus
dasatinibDisrupts ephrin/EPH SCAP arm in preadipocytes; complementary to BH3-mimeticsPhase 2 in D+Q regimen

Flavonoid / natural-product senolytics

AgentMechanismStatus
fisetinPI3K/AKT/mTOR pathway disruption (proposed); cell-type-specific clearance; mechanism incompletely characterizedPhase 2 trials ongoing; but NIA ITP-null for lifespan + p16 qPCR clearance 4 and first efficacy RCT (knee OA, n=74) null 5
quercetinBCL-xL / EFNB1 in HUVECs (PI3Kδ is preadipocyte-selective per Zhu 2015, not HUVEC); D+Q “Q” componentPhase 1/2 in D+Q regimen
procyanidin-c1 (grape seed)Dual dose-dependent: senomorphic (SASP-suppressing) at low concentration, senolytic at ≥50 µM; kills via NOXA/PUMA upregulation → ROS → mitochondrial depolarization → caspase 3/7 (NOT BCL-2/BCL-xL). Aged-mouse lifespan +9.4% whole-life / +64% post-treatment (Xu 2021 Nat Metab)Preclinical (no human senolytic trial)
piperlongumineLong-pepper (Piper longum) alkaloid; targets OXR1 → chronic antioxidant-defense erosion → ROS-dependent senolysis; ~2.5–3.3× senescent-vs-normal selectivity; synergistic with navitoclax (Wang 2016)Preclinical (no human trial)

Other / investigational

  • Cardiac glycosides (digoxin, ouabain) — preclinical senolytic evidence (Triantafyllou 2019); narrow therapeutic index complicates translation
  • HSP90 inhibitors (17-AAG, 17-DMAG) — preclinical only
  • FOXO4-DRI peptide — disrupts FOXO4–p53 interaction in senescent cells; Baar 2017 preclinical evidence
  • Galactose-conjugated prodrugs (lysosomal-targeting) — exploit elevated SA-β-galactosidase in senescent cells; preclinical / first-in-human trials beginning

Clinical evidence status (as of 2026)

TrialAgent(s)PopulationNCTStatus / key results
Justice 2019 6D + QIdiopathic pulmonary fibrosis (n=14)NCT02874989Phase 1 pilot — primary endpoints feasibility (achieved); secondary 6MWT +21.5 m (p=0.012), gait speed +0.12 m/s (p=0.024), chair-stands −2.2 s (p=0.013); pulmonary function unchanged
Hickson 2019 7D + QDiabetic kidney disease (n=9)NCT02848131Phase 1 — adipose biopsy: p16 −35% (p=0.001), p21 −17% (p=0.009), SA-βgal+ −62% (p=0.005); macrophages −28% (p=0.0001); circulating SASP factors (IL-1α, IL-6, MMPs-9/-12) reduced
Mayo Frailty FisetinFisetinOlder women with frailtyNCT03675724Phase 2 — ongoing as of 2026
DKD FisetinFisetinPatients with diabetic kidney diseaseNCT03325322Phase 2 — suspended (per ClinicalTrials.gov 2026-05-04; reason not posted)
UNITY-OA-101UBX0101Knee osteoarthritisNCT04129944Phase 2 failed — no improvement vs placebo; UBX0101 program subsequently abandoned. The first major senolytic clinical setback.
BEHOLD 2UBX1325 (foselutoclax)Diabetic macular edema (suboptimal anti-VEGF response)NCT04857996Phase 2 RCT — first positive BCL-xL-axis senolytic human trial. Single 10 µg intravitreal injection vs sham; n=65 (UBX1325 n=32, sham n=33); 48-wk follow-up. BCVA difference UBX1325 − sham at week 48 = +5.6 ETDRS letters (95% CI −1.5 to 12.7); no systemic safety signals (no platelet effects via local intravitreal route); 5 vs 4 Grade ≥3 TEAEs (UBX1325 vs sham; CI overlaps zero). UNITY pivoted to UBX1325/foselutoclax (a BCL-xL-selective small molecule) after UBX0101 failure; intravitreal local delivery sidesteps the systemic thrombocytopenia liability that has blocked navitoclax.
TROFFi 8FisetinPostmenopausal breast cancer survivors with 6MWD <400 mNCT05595499Phase 2 RCT — protocol published 2026; 88 women; fisetin 20 mg/kg/day d1–3 of 14-day cycles × 4 cycles vs placebo; primary endpoint Δ6MWD baseline-to-end. Multicenter, double-blind, placebo-controlled. First placebo-controlled fisetin senolytic RCT with hard physical-function endpoint. Recruiting as of 2026-05.
Tashman 2025 5FisetinKnee osteoarthritis (n=74)NCT04210986First completed fisetin senolytic RCT to report efficacy — NEGATIVE on all endpoints. Phase 1/2 randomized double-blind placebo-controlled (fisetin 34 / placebo 40); ~20 mg/kg/day × 2 consecutive days, repeated; no benefit vs placebo on pain (NRS), function (WOMAC), cartilage (MRI T2), or biomarkers (CRP, COMP); safety comparable. The second senolytic OA failure after UNITY-OA-101 (different agent/mechanism). Caveats: safety-powered Phase 1/2, single-site, ~4 days’ total exposure, no pharmacodynamic senescent-cell readout. Abstract-level (OARSI 2025); full paper pending. See tashman-2025-fisetin-knee-oa-rct.
Harrison 2024 (NIA ITP) 4Fisetin (mouse)UM-HET3 mice, 3 sites—Highest-rigor preclinical test — NULL. No lifespan extension (either sex, continuous or cyclic dosing) at 600 ppm from 20 mo; antibody-independent Cdkn2a qPCR sub-cohort showed no senescent-cell clearance in kidney/brain/liver. Directly stress-tests the foundational fisetin lifespan/clearance claim. (Mouse, not human — included here because it anchors the fisetin-efficacy reassessment.) See harrison-2024-itp-astaxanthin-meclizine.
Farr 2024 9D + QPostmenopausal women (osteoporosis risk; n=60)NCT04313634First placebo-controlled D+Q RCT to report results; primary endpoint NEGATIVE. 20-week open-label RCT: D 100 mg × 2 days + Q 1,000 mg × 3 days, every 28 days × 5 cycles vs placebo; primary endpoint CTx (bone resorption): p=0.611 (NS). P1NP (formation) significantly elevated at 2 wk (+16%, p=0.020) and 4 wk (+16%, p=0.024), returned to NS at 20 wk (−9%, p=0.149). 36 circulating SASP factors measured at baseline and 2 wk — no significant D+Q vs control difference. Exploratory high-senescent-burden tertile (highest T-cell p16 mRNA [variant 5], n≈10/arm): P1NP +34% (p=0.035) and CTx −11% (p=0.049) at 2 weeks post-dosing; radius BMD +2.7% (p=0.004) at 20 weeks. Materially qualifies the senolytic-for-bone framing: benefit appears NOT universal but stratifiable by senescent-cell burden. The biomarker-stratified follow-up is in progress (Farr 2025 Aging Cell). See bone for the bone-specific verification context.

gap/needs-human-replication — All non-UBX1325 positive senolytic results in humans are from small open-label or pilot trials. No large randomized blinded outcome trials have completed. The UNITY-OA-101 failure is a cautionary signal that not all SCAP-targeting strategies translate; the UBX1325/BEHOLD result (Klier 2025) is the first BCL-xL-axis senolytic Phase 2 RCT to show a positive efficacy trend with a clean safety profile, vindicating the BCL-xL mechanism in a localized-delivery context.

Active senolytic trials count (ClinicalTrials.gov, 2026-05-08): 14 RECRUITING + ACTIVE_NOT_RECRUITING + ENROLLING_BY_INVITATION (excludes oncology-only navitoclax/venetoclax monotherapy trials). Fisetin dominates the senolytic-aging trial roster (17 active/enrolling fisetin trials, several spanning aging-relevant indications: PAD, mild Alzheimer’s [NCT07279714], healthy aging [NCT07195318], breast-cancer-survivor frailty [NCT06113016, NCT05595499], multimorbidity [NCT06431932], cancer-survivor fatigue [NCT06819254]).

Evidence caveat: the p16 antibody confound (2026)

Much of the correlative case for senolytics — “senescent (p16+) cells accumulate with age and in disease X, and our senolytic lowers their count” — rests on tissue p16INK4a immunohistochemistry. A 2026 research-integrity analysis (S. David) indicates that across the senescence literature a large majority of papers using certain catalog antibodies were in fact detecting the unrelated actin-cytoskeleton protein p16-ARC (ARPC5), not p16INK4a — see p16-immunodetection. Implications for this page:

  • The causal core is intact. The strongest evidence that clearing p16-high cells extends healthspan/lifespan is genetic (INK-ATTAC, p16-3MR) and antibody-independent 10 — it does not depend on the implicated reagents.
  • Antibody-based clearance readouts should be weighted cautiously. Where a senolytic’s target-engagement claim rests solely on p16 IHC with an unvalidated clone, treat it as provisional. The human D+Q proof-of-target-engagement data (Hickson 2019: adipose/skin p16+ cell counts) and similar should be re-checked for clone/catalog validation.
  • Antibody-independent readouts carry more weight — and for fisetin they were null (ITP Cdkn2a qPCR 4). This convergence (failed antibody-free tests + a null human RCT) is why fisetin’s senolytic case has weakened on close inspection, independent of the antibody scandal.

Cell-type specificity and clinical translation challenges

The cell-type-specific SCAP framework has direct clinical-design implications:

  • D+Q clears preadipocytes, mesenchymal stem cells, and some stromal populations — well-suited for fibrotic diseases (where mesenchymal senescent cells drive pathology) and metabolic syndromes (preadipocyte senescence drives adipose dysfunction).
  • Navitoclax / A1331852 are most effective on HUVECs and other Bcl-xL-dependent senescent cells — well-suited for vascular and hematologic contexts, but the dose-limiting thrombocytopenia restricts use to cancer indications where the risk/benefit profile permits.
  • Fisetin shows broader cell-type activity than pure Bcl-2-family targeting would predict — its mechanism remains incompletely characterized, which is both its appeal (broader spectrum) and its limitation (harder to design rational combinations).

Trial design implication: the indication should be chosen for the senescent-cell type most relevant to the disease, not just for the most easily-recruited patient population. This is a partial explanation for the UNITY-OA-101 failure (osteoarthritis senescent cells may not have been the right SCAP profile for UBX0101’s mechanism).

Senolytic resistance and heterogeneity within cell populations

Beyond the between-cell-type SCAP differences above, evidence is emerging that even within a single senescent cell population, a molecularly defined subset is resistant to BH3-mimetic senolysis. A 2025 Weizmann Institute (Krizhanovsky lab) preprint identifies cell-surface ATP6V1B2 (csV1B2; atp6v1b2) as a marker of such a subset 11: in response to DNA damage, a subset of senescent cells upregulates the v-ATPase subunit ATP6V1B2 on the cell surface, and csV1B2 expression correlates with resistance to ABT-737-induced apoptosis in culture (ABT-737 being the in-vitro analog of navitoclax/ABT-263, a canonical BCL-2/BCL-xL/BCL-w senolytic). This subset carries a transcriptional signature associated with DNA repair and apoptosis resistance, and is present in aging and fibrotic lungs in vivo. The csV1B2 surface marker is therefore a candidate biomarker for BH3-mimetic treatment failure and a mechanistic explanation for residual senescent-cell persistence after navitoclax-class drug exposure. gap/preprint-not-peer-reviewed gap/needs-replication — preprint only; ABT-737 to navitoclax/foselutoclax extrapolation requires full-text dose-response data.

Interaction with skeletal-muscle hypertrophy and regeneration

A distinct and somewhat counterintuitive axis from the static “senescent cells accumulate in aged muscle” framing: the hypertrophic / regenerative stimulus itself generates senescent cells, and in aged muscle their impaired clearance blunts the adaptive response. This is the most direct evidence that senolytics interact with muscle growth (not just resting sarcopenic muscle).

  • A growth stimulus induces senescent cells. They are nearly absent in resting muscle but emerge ~2 weeks after a single bout of resistance exercise in humans (SA-β-Gal + p21 IHC), and during mechanical overload in mice 12. Aged mice accumulate significantly more post-overload senescent cells than young mice (days 7 and 14).
  • In aged muscle this blunts hypertrophy, and senolytics rescue it. Old mice have attenuated overload hypertrophy (smaller muscles, smaller glycolytic type 2x/2b fibers); a “hit-and-run” senolytic course (D+Q, 5 mg/kg dasatinib + 50 mg/kg quercetin, days 7 and 10 of a 14-day overload) gave old mice larger muscles and fibers, fewer senescent cells, and shifted hypertrophy-associated genes (↑Igf1, ↓Ddit4, ↑Mmp14) — though the authors note D+Q may act partly via M2 macrophages (the main plantaris Mmp14 source) rather than purely senescent-cell killing, and SASP inflammatory genes did not change 12. See dungan-2022-senolytic-muscle-hypertrophy.
  • Mechanism — the inflamed niche suppresses satellite cells. Senescent cells are integral, regeneration-repressing components of the muscle niche at all ages; they build an aged-like inflamed (inflammageing-mirroring) niche that arrests satellite-cell proliferation. Reducing their burden — or neutralizing their CD36-dependent inflammatory secretome — accelerates regeneration in both young and old mice, while transplanting senescent cells delays it 13. See moiseeva-2023-senescence-atlas-muscle-regeneration.
  • Non-senolytic route to the same end: a p53–MDM2 inhibitor (senomorphic) likewise reduced senescent-cell abundance and improved the adaptive response of aged mouse muscle 14.

Age-specificity and a caution. The hypertrophy benefit is demonstrated in old / clearance-impaired muscle. Young animals make these cells too but clear them and hypertrophy normally — there is no evidence senolytics add to a young, healthy growth response, and transiently induced senescent-like (p21+) cells are plausibly part of normal remodeling, so indiscriminate senolysis during training in young muscle is untested and possibly counterproductive. Exercise is itself partly senomorphic: resistance training in middle-aged humans lowered muscle senescent-cell and denervated-fiber markers 15, so a trained person is already reducing the burden a senolytic would target. No human senolytic-plus-resistance-training trial exists gap/needs-human-replication. The strongest rationale is aged / sarcopenic muscle, not athletic hypertrophy.

DimensionStatus
Mechanism (load → senescent cells → blunted growth) conserved in humans?partial — senescent-cell appearance post-exercise shown in humans; senolytic rescue is mouse-only
Phenotype (senolytic restores hypertrophy) conserved in humans?unknown — no human senolytic + resistance-training trial
Generalizes to young/trained muscle?likely no / possibly harmful — benefit specific to aged, clearance-impaired muscle

SENS / hallmark mapping

  • Genetic senescent-cell clearance — INK-ATTAC and p16-3MR transgenic mouse models (Baker 2011 16; Baker 2016 10) — preclinical proof-of-principle that motivated senolytic drug development. Baker 2016 demonstrated 27% (mixed background) / 24% (C57BL/6) lifespan extension from p16+ cell clearance in naturally-aged mice (per p21 verification round; tissue-selective effect).
  • Senomorphics (suppress SASP without killing): rapamycin, JAK inhibitors (ruxolitinib), metformin (partial), NF-κB inhibitors, BET inhibitors — see sasp for mechanism details
  • CAR-T senolytic approaches (preclinical) — uPAR-targeted CAR-T cells that clear senescent cells (Amor 2020 Nature) — an early-stage cellular-immunotherapy translation of the senolytic concept
  • Transcript-activated cell ablation (preclinical; cancer-only) — transcript-activated-cell-ablation kills cells by an intracellular RNA signature via Cas12a2 collateral chromatin shredding (Doudna lab, Nature 2026; see zeng-2026-cas12a2-chromatin-shredding). A non-SCAP, non-apoptosis-priming kill switch — conceptually a candidate for senotype-targeted clearance (program the guide to a senescence-restricted transcript), but untested in senescence and bounded by the no-universal-marker problem (cf. freizus-2025-atp6v1b2-persistent-senescence apoptosis-resistant subset that a non-apoptotic killer could in principle reach).

Limitations and open questions

  • Cell-type specificity is the dominant unsolved problem. No single agent or combination clears all senescent cells; tissue-level efficacy depends on which senescent populations dominate.
  • Long-term safety in humans is unknown for all senolytics — acute trials (single or short courses) have favorable safety; chronic intermittent exposure data is accumulating but limited. gap/long-term-unknown
  • Optimal dosing schedule remains debated — “hit-and-run” intermittent (D+Q 3-day courses, weekly cycles per Justice 2019) vs continuous low-dose. Mechanistic basis for hit-and-run is solid but timing optimization is empirical. gap/dose-response-unclear
  • Trial enrollment biomarkers — no validated way to identify patients with high senescent-cell burden a priori. Trials enroll on disease state. gap/needs-replication for biomarker validation
  • Senolytic vs senomorphic comparative efficacy — head-to-head clinical comparisons are scarce.
  • Off-target toxicity at therapeutic doses — D+Q dosing is short and intermittent partly to minimize TKI cardiotoxicity (D) and bioavailability concerns (Q).
  • D+Q-specific CNS safety signal — In healthy young and aged mice, D 5 mg/kg + Q 50 mg/kg (six oral doses over four weeks) produced thinner myelin in the rostral corpus callosum without detected oligodendrocyte loss 17. The displayed TEM result used three mice/group and an axon-level Kolmogorov–Smirnov test, so pseudoreplication and the absence of an animal-level effect estimate limit precision. Individual D and Q each retracted processes in cultured rat oligodendrocytes, but neither was tested alone in vivo. Because brain senescent-cell burden and target engagement were not measured, this is an off-target D+Q-regimen signal—not evidence that senolysis caused the injury or that all senolytics share it. No human CNS endpoint or recovery experiment is available. gap/needs-replication gap/needs-human-replication gap/no-mechanism
  • Cardiac-glycoside senolytic activity — promising preclinical signal but interpretation complicated by their narrow therapeutic index and existing clinical use in heart failure.
  • UBX0101 failure interpretation — was the failure due to wrong cell-type targeting? Wrong indication? Wrong mechanism (UBX0101 may not act as a senolytic at all)? The post-mortem is incomplete and matters for the next generation of senolytic trials.

Recent class-level developments (2024–2026)

The 2024–2026 senolytic literature has shifted from generic “BH3-mimetic monotherapy” toward:

  1. Local-delivery BCL-xL inhibition — UBX1325/foselutoclax’s intravitreal route (Klier 2025 2) demonstrates that the systemic-thrombocytopenia barrier can be bypassed for tissue-localized senescent-cell pathology (retina, CNS, joint). This is the first BCL-xL-axis senolytic to clear Phase 2 with a positive efficacy trend.
  2. Tissue-specific PROTAC degraders (DT2216, PZ15227-class; recent reviews 18) — recruit organ-restricted E3 ligases (VHL, CRBN) to degrade BCL-xL selectively in nucleated senescent cells while sparing platelets. Reviewed as one of three “next-generation” senotherapy strategies in Zhang 2026.
  3. Immune-based senolysis — uPAR-targeted CAR-T (Amor 2020), NK cell senotherapy approaches (Front Immunol 2025), and immune-checkpoint-modulated SnC clearance.
  4. Microbiome-senolysis crosstalk — SCFAs (butyrate) modulate drug-transporter expression and SASP; emerging gut-liver axis modifiers of senolytic efficacy 18.
  5. Indication-specific cell-type matching — community-wide rethinking of trial design (Nature Aging 2025 perspective “Towards a personalized approach in senolytic trials”) in light of UBX0101-OA failure vs UBX1325-DME positive signal.

gap/long-term-unknown — long-term effects of repeated intravitreal BCL-xL inhibition (UBX1325) not yet characterized; BEHOLD followed for 48 wk.

Cross-references

Footnotes

Footnotes

  1. doi:10.1111/acel.12344 · in-vitro + in-vivo (mouse) · transcriptomic SCAP identification + first senolytic combination · model: human cells (HUVECs, IMR90, preadipocytes) + Ercc1-/Δ progeroid mice · Aging Cell — the foundational senolytic paper ↩ ↩2

  2. doi:10.1056/EVIDoa2400009 · Klier S et al. (UNITY Biotechnology) · NEJM Evidence 2025 May;4(5):EVIDoa2400009 · Phase 2 randomized sham-controlled · n=65 (UBX1325 32, sham 33) DME patients with prior suboptimal anti-VEGF response · single 10 µg intravitreal injection vs sham; 48-week follow-up · BCVA Δ at week 48: UBX1325 − sham +5.6 ETDRS letters (95% CI −1.5 to +12.7) · 5 vs 4 Grade ≥3 TEAEs UBX1325 vs sham; no apparent between-group differences in vital signs, ECG, or routine chemistries · NCT04857996 · UBX1325/foselutoclax is a BCL-xL-selective small-molecule senolytic; local intravitreal delivery sidesteps systemic platelet toxicity. Verifier note: abstract-level extraction; full-text not yet end-to-end verified. ↩ ↩2 ↩3

  3. doi:10.1038/nm.4010 · in-vivo (mouse) · navitoclax (ABT-263) in naturally-aged mice · Nature Medicine · confirmed title: “Clearance of senescent cells by ABT263 rejuvenates aged hematopoietic stem cells in mice” — HSC rejuvenation is the primary finding; lung/liver/bone marrow senescent-cell clearance claims gap/no-fulltext-access (DOI lookup failed after retry) ↩

  4. harrison-2024-itp-astaxanthin-meclizine · doi:10.1007/s11357-023-01011-0 · PMID 38041783 · PMC10828146 (gold OA) · Harrison DE, Strong R, … Miller RA · GeroScience 2024;46(1):795–816 · in-vivo, NIA Interventions Testing Program (UM-HET3, 3 sites, both sexes) · fisetin 600 ppm chow (~60–80 mg/kg/day) from 20 mo, continuous or cyclic → no lifespan extension either sex (♂ p=0.85 / 0.54; ♀ p=0.37 / 0.75); 22–24-mo sub-cohort: no p16^Ink4a^ mRNA reduction (qPCR) in kidney/brain/liver. (Astaxanthin +12% ♂ / meclizine +8% ♂ were the positives.) verified: false — seeded 2026-06-30, full-text verifier pass pending. ↩ ↩2 ↩3

  5. tashman-2025-fisetin-knee-oa-rct · doi:10.1016/j.joca.2025.02.667 · Tashman S, Philippon M, Dornan G, Huard J · Osteoarthritis and Cartilage 2025;33(Suppl):S456 · conference abstract · randomized double-blind placebo-controlled Phase 1/2 · n=74 (fisetin 34 / placebo 40), knee OA (KL II–IV) · fisetin 100 mg ~20 mg/kg/day × 2 consecutive days, 28 off, second course d31–32 · NCT04210986 · null on all efficacy endpoints (NRS pain, WOMAC, MRI T2 cartilage, CRP, COMP, performance); safety comparable (efficacy data from ClinicalTrials.gov posted results; abstract body Cloudflare-blocked) · full paper unpublished as of mid-2026 · verified: false gap/needs-full-paper-publication ↩ ↩2

  6. doi:10.1016/j.ebiom.2018.12.052 · open-label Phase 1 pilot · n=14 IPF · D 100 mg + Q 1250 mg/day, 3 consec days/week × 3 weeks (9 dosing days) · primary endpoints feasibility; secondary 6MWT +21.5 m (p=0.012), gait speed +0.12 m/s (p=0.024), chair-stands −2.2 s (p=0.013) · model: humans ↩

  7. doi:10.1016/j.ebiom.2019.08.069 · open-label Phase 1 · n=9 DKD · 3-day oral course: D 100 mg/day + Q 1000 mg twice daily · adipose biopsy: p16 −35% (p=0.001), p21 −17% (p=0.009), SA-βgal+ −62% (p=0.005); macrophages −28% (p=0.0001); circulating SASP factors reduced (IL-1α, IL-2, IL-6, IL-9, MMP-2, -9, -12 all decreased) · model: humans ↩

  8. doi:10.1177/17588359261424668 · Ji J, Crespi CM, Yee L, Zekster YA, Al-Saleem A, Petersen L, Lee C, Son N, Smith C, Evans T, Tchkonia T, Kirkland JL, Kuchel GA, Cohen HJ, Sedrak MS · Ther Adv Med Oncol 2026 Mar 11;18:17588359261424668 · TROFFi study rationale and trial design publication (not yet results) · Phase 2 multicenter randomized double-blind placebo-controlled trial · planned n=88 postmenopausal early-stage breast cancer survivors with 6MWD <400 m, completed neo/adjuvant chemotherapy within 12 mo · 1:1 randomization to fisetin 20 mg/kg/day d1–3 of 14-day cycle × 4 cycles vs placebo · primary endpoint Δ6MWD baseline-to-end-of-treatment · NCT05595499 · status: recruiting as of 2026-05. ↩

  9. doi:10.1038/s41591-024-03096-2 · Farr JN et al. · Nature Medicine 2024;30(9):2605-2612 · randomized open-label placebo-controlled · n=60 postmenopausal women age 60–90 y · 20-week intermittent D+Q: D 100 mg/day × 2 consecutive days + Q 1,000 mg/day (250 mg ×4) × 3 consecutive days, every 28 days × 5 cycles vs placebo · primary endpoint CTx % change at 20 wk: D+Q −4.1% (IQR −13.2, 2.6) vs control −7.7% (IQR −20.1, 14.3); p=0.611 (NS) · secondary P1NP +16% vs control at 2 wk (p=0.020) and 4 wk (p=0.024); −9% at 20 wk (p=0.149, NS) · 36 SASP factors measured at baseline and 2 wk; no significant D+Q vs control difference in SASP in any group · T-cell p16 mRNA measured at baseline only (not post-treatment) · exploratory high-T-cell-p16 tertile (T3; n≈10/arm): P1NP +34% (p=0.035) and CTx −11% (p=0.049) at 2 weeks post-dosing; radius BMD +2.7% (p=0.004) at 20 weeks · NCT04313634 · the first placebo-controlled D+Q clinical bone trial; established the need for senescent-burden stratification in trial design (added 2026-05-23 via bone verifier pass; timing-split corrected 2026-05-23 via osteoporosis verifier pass; schedule/design/SASP corrected 2026-06-12 via Farr 2024 PDF read) ↩

  10. doi:10.1038/nature16932 · in-vivo (mouse) · INK-ATTAC; naturally-aged mice · 27% (mixed bg) / 24% (C57BL/6) median lifespan extension from p16+ cell clearance (range 17–35% by sex); tissue-selective · per p21 verification round — values directly read from PDF ↩ ↩2

  11. freizus-2025-atp6v1b2-persistent-senescence · Freizus N, Majewska JM, Ovadya Y et al. (Krizhanovsky lab, Weizmann Institute) · preprint (bioRxiv) · doi:10.1101/2025.11.30.691415 · 2025-12-01 · in-vitro (human + mouse senescent cells) + in-vivo (aging + fibrotic lungs) · abstract-only; not peer-reviewed · csV1B2 (cell-surface ATP6V1B2) marks apoptosis-resistant senescent subset; ABT-737 resistance correlated with csV1B2; altered lysosomal activity and intracellular pH gap/preprint-not-peer-reviewed ↩

  12. dungan-2022-senolytic-muscle-hypertrophy · doi:10.1007/s11357-022-00542-2 · Dungan CM et al. · GeroScience 2022;44(4):1925-1940 · in-vivo (mouse) + human IHC · adult (5–6 mo) vs old (23–24 mo) male C57BL/6J, synergist-ablation mechanical overload; senolytic D+Q 5 mg/kg dasatinib + 50 mg/kg quercetin on days 7+10 of 14-day MOV · load-induced senescent cells (SA-β-Gal + p21) emerge ~2 wk after a single resistance bout in humans; old mice over-accumulate them and have blunted overload hypertrophy (smaller type 2x/2b fibers); D+Q gave old mice larger muscles/fibers, fewer senescent cells, ↑Igf1 + altered Ddit4/Mmp14 · gap/needs-human-replication ↩ ↩2

  13. moiseeva-2023-senescence-atlas-muscle-regeneration · doi:10.1038/s41586-022-05535-x · Moiseeva V et al. · Nature 2023;613(7942):169-178 (Author Correction 10.1038/s41586-023-05765-7) · scRNA-seq + senescent-cell enrichment sorting · young + old mice · senescent cells are integral regeneration-repressing niche components at all ages, building an inflamed (inflammageing-mirroring) niche that arrests satellite-cell proliferation; reducing burden or CD36 neutralization accelerates regeneration in young AND old; transplanting senescent cells delays it; senescent cells also accumulate in human muscle ↩

  14. doi:10.1007/s11357-023-00976-2 · Nolt GL et al. · GeroScience 2024 · in-vivo (aged mouse) · p53–MDM2 binding inhibitor (senomorphic/senostatic, not senolytic) reduced senescent-cell abundance and improved the adaptive response of skeletal muscle from aged mice · AI-extracted from abstract; PDF not yet verified · gap/needs-human-replication ↩

  15. doi:10.1096/fj.202302103rrr · Ruple BA et al. · FASEB J 2024 · human · resistance training in middle-aged adults reduced denervated myofibers and muscle senescent-cell / associated protein markers — evidence that exercise is partly senomorphic · AI-extracted from abstract; PDF not yet verified ↩

  16. doi:10.1038/nature10600 · in-vivo (mouse) · INK-ATTAC transgenic; p16+ cell genetic clearance · model: BubR1 progeroid mice · proof-of-principle for senescent-cell clearance benefit. Note: green OA but DOI lookup has failed; full PDF unverified — #gap/no-fulltext-access ↩

  17. lombardo-2026-senolytic-demyelination · displayed TEM n=3 mice/age/treatment group; 1,500–2,000 nested axons/group · in-vivo (young and aged C57BL/6J mouse) + in-vitro (primary rat oligodendrocyte lineage) · D 5 mg/kg + Q 50 mg/kg oral gavage, six doses over four weeks · regional myelin thinning/loss and nonlethal oligodendrocyte dysfunction; UPR mechanism associative · doi:10.1073/pnas.2524897123 ↩

  18. doi:10.1038/s41514-026-00355-z · Zhang W, Song S et al. · NPJ Aging 2026 Mar 10 · review · model: literature synthesis · catalogues three next-generation senotherapy strategies — (1) immune-based senolysis (CAR-T uPAR, GD3-blocking, ferroptosis sensitization); (2) tissue-precision PROTACs (VHL, CRBN E3-ligase recruiters degrading BCL-xL in nucleated cells but sparing platelets); (3) microbiome-epigenetic interplay (SCFAs/butyrate modulating drug transporters + SASP) · review-level claims; not a primary-source experiment. ↩ ↩2

Graph View

Backlinks