Plasminogen Activator Inhibitor-1 (PAI-1), An Important Factor in Aging

#1

I. Executive Summary

The core thesis presented by Douglas Vaughan, PhD, establishes Plasminogen Activator Inhibitor-1 (PAI-1), encoded by the SERPINE1 gene, as a direct upstream mediator of human organismal aging rather than a passive downstream biomarker of cellular senescence. Historically characterized as a regulator of fibrinolysis via tissue plasminogen activator (tPA) and urokinase plasminogen activator (uPA) inhibition, modern geroscience identifies PAI-1 as a critical hub linking cellular senescence, chronic inflammation, metabolic dysfunction, and immune evasion. Circulating PAI-1 levels rise with chronological age, driven by visceral adiposity, hyperinsulinemia, and pro-inflammatory cytokines like Tumor Necrosis Factor-alpha (TNF-alpha), establishing a feed-forward loop that accelerates multi-organ degeneration.

This causal paradigm is validated by human loss-of-function genetics. Longitudinal evaluation of an Old Order Amish kindred in Adams County, Indiana, carrying a rare SERPINE1 null mutation (c.699_700dupTA) demonstrates that heterozygous carriage results in a ~50% reduction in circulating PAI-1 antigen. This lifelong partial deficiency confers profound systemic healthspan advantages, including a median lifespan extension of approximately 10 years, 10% longer leukocyte telomeres, lower fasting insulin, lower hemoglobin A1c, and near-absolute protection against type 2 diabetes. These individuals exhibit preserved cardiovascular compliance, characterized by reduced aortic pulse wave velocity and delayed vascular stiffness. Conversely, absolute homozygous deficiency induces clinical bleeding disorders and spontaneous cardiac fibrosis, defining a narrow therapeutic window where partial, rather than complete, inhibition is mandatory to optimize organismal fitness.

At the molecular level, PAI-1 exerts pleiotropic control over biological pathways by blocking the proteolytic processing of essential signaling peptides. It halts the activation of brain-derived neurotrophic factor (BDNF) and hepatocyte growth factor (HGF), impairing neuroplasticity and tissue regeneration. Furthermore, PAI-1 reinforces cellular senescence via intracellular p53 stabilization and upregulates programmed death-ligand 1 (PD-L1) to shield senescent cells from immune clearance. Epigenetically, these mechanisms map directly onto the Horvath GrimAge clock, where the DNAm PAI-1 estimator serves as the strongest surrogate predictor of all-cause mortality and multi-system comorbidities. Translationally, this validates the deployment of orally bioavailable small-molecule PAI-1 inhibitors like TM5614, which have shown safety in Phase I/II trials and are entering human longevity testing.

II. Insight Bullets

  • PAI-1 as a Functional Senescence Driver: PAI-1 is an active, causal mediator of cellular senescence and part of the Senescence-Associated Secretory Phenotype (SASP), rather than a passive downstream marker of cellular aging Vaughan et al., 2024.
  • Geroscience Paradigm Target: The core hypothesis mandates that targeting upstream molecular hubs like PAI-1 can delay or compress multi-system morbidities simultaneously, extending healthy human healthspan.
  • Biological vs. Chronological Disconnect: Chronological age is cosmically fixed and unyielding, whereas biological age is a mutable, non-linear metric counting multi-system homeostatic changes over time.
  • Adiposity as Primary PAI-1 Regulator: Plasma PAI-1 antigen levels correlate directly with body mass index (BMI); visceral adiposity and hepatic lipid accumulation drive the modern epidemic of elevated PAI-1 ResearchGate Source.
  • Inflammatory Acute-Phase Induction: PAI-1 operates as an acute-phase reactant; systemic inflammatory cytokines, primarily Tumor Necrosis Factor-alpha (TNF-alpha) and Interleukin-6 (IL-6), strongly induce its expression.
  • The SERPINE1 Null Mutation: A rare human loss-of-function founder mutation (c.699_700dupTA) exists within an Old Order Amish community in Indiana, providing a unique population genetics model for geroscience Vaughan et al., 2017.
  • Heterozygous Longevity Phenotype: Heterozygous carriers of the SERPINE1 null mutation have a 50% reduction in circulating PAI-1, which extends median lifespan by approximately 10 years Vaughan et al., 2017.
  • Leukocyte Telomere Preservation: Lifelong 50% reduction of PAI-1 is associated with approximately 10% longer leukocyte telomeres, demonstrating protected replicative capacity in humans Vaughan et al., 2017.
  • Near-Absolute Diabetes Protection: Heterozygous SERPINE1 mutation carriers exhibit lower fasting insulin and lower HbA1c, rendering them highly resilient against type 2 diabetes development compared to unaffected cohort matches.
  • Homozygous Cardiac Pathology: Complete absence of PAI-1 (homozygous mutation) results in localized subendocardial bleeding and spontaneous cardiac fibrosis/scarring, proving that complete PAI-1 knockout is highly deleterious.
  • Smooth Muscle Origin in Vascular Aging: Single-cell RNA sequencing reveals that the pathologically elevated PAI-1 seen in advancing vascular wall aging originates from vascular smooth muscle cells, not healthy endothelial cells Vaughan et al., 2023.
  • Hepatic PAI-1 Knockdown and FGF-21: Targeted hepatic reduction of PAI-1 using siRNA significantly upregulates Fibroblast Growth Factor 21 (FGF-21), demonstrating an intimate connection between fribrinolytic signaling and systemic metabolic adaptation.
  • PAI-1 and PCSK9 Co-Regulation: Lowering PAI-1 by half significantly downregulates Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) in the liver, identifying PAI-1 as a critical post-translational regulator of LDL receptor clearance.
  • GrimAge Epigenetic Hierarchy: In Steve Horvath’s GrimAge clock, the DNAm PAI-1 surrogate is the single most powerful predictor of all-cause mortality, coronary artery disease, and comorbidity accumulation Lu et al., 2019.
  • Surrogate Superiority over Direct Plasma Tests: Epigenetic estimators (DNAm PAI-1) often outperform single-timepoint blood draws in predictive power due to the short half-life and massive acute-phase volatility of raw plasma proteins Lu et al., 2019.
  • Circadian Volatility Limitations: Plasma PAI-1 concentrations follow a powerful circadian curve, peaking sharply in the early morning and dropping in the late afternoon, requiring rigid standardization for clinical sample processing.
  • Serum Sample Invalidation: PAI-1 must be analyzed strictly in anticoagulated plasma; serum tests are completely invalidated by the massive dump of PAI-1 from platelet intracellular granules during active coagulation.
  • Neuroplasticity Blockade via BDNF Inhibition: Elevated PAI-1 prevents tPA from cleaving pro-BDNF into its active, mature form, directly arresting neuroplasticity, memory consolidation, and downstream cognitive resilience.
  • Immune Evasion via PD-L1 Upregulation: Local tissue PAI-1 accumulation overexpresses Programmed Death-Ligand 1 (PD-L1), constructing an immune checkpoint shield that blocks immune-mediated clearance of senescent cells.
  • Regenerative Suppression via HGF Inhibition: PAI-1 actively suppresses the proteolytic activation of Hepatocyte Growth Factor (HGF), thereby halting cell proliferation and downstream organ regenerative kinetics.
  • Vascular Elasticity Preservation: Heterozygous SERPINE1 mutation carriers exhibit low aortic pulse wave velocity, indicating lifelong resistance to age-related mechanical vascular stiffening Vaughan et al., 2023.
  • The Small-Molecule Inhibitor TM5614: TM5614 (RS5614) is an orally bioavailable, structure-guided small molecule that targets human PAI-1, currently showing safety in phase II trials for hematological and solid tumors Vaughan et al., 2024.
  • Transition to Longevity IND Trials: Researchers have applied for an FDA Investigational New Drug (IND) clearance to test small-molecule PAI-1 inhibitors in non-Amish human cohorts to evaluate direct biological clock reversal.
  • Inadequacy of Isolated Epigenetic Clocks: Epigenetic age rollbacks are clinically meaningless if they occur in isolation from physical, proteomic, glycomic, and multi-system functional physiological performance assessments.
  • Multi-Dimensional Diagnostics Paradigm: The Northwestern Human Longevity Lab integrates deep molecular profiling (OLink proteomics, GlycanAge) with rapid, AI-driven functional stations (ECG machine learning, retinal imaging, and video gait kinetics) to establish valid organ-specific age profiling.

IV. Actionable Protocol

High Confidence Tier (Level A/B Evidence)

  • Targeted Metabolic Regulation for PAI-1 Reduction: Visceral fat and hepatic steatosis are the primary drivers of pathologically elevated PAI-1. Comprehensive meta-analyses establish that targeted weight reduction, visceral fat loss, and resolution of hyperinsulinemia drastically lower circulating PAI-1 levels while downregulating systemic inflammatory checkpoints ResearchGate Source.
  • Rigidly Standardized Biomarker Surveillance: Due to high biological variation, PAI-1 must never be tested in serum (confounded by platelet dumping). Clinical evaluation must utilize plasma specimens drawn strictly at an identical early morning timepoint to control for its rapid circadian fluctuations. Epigenetic tracking via second-generation mortality predictors (e.g., GrimAge DNAm PAI-1) provides a more stable, long-term reflection of tissue PAI-1 burden than singular blood draws Lu et al., 2019.

Experimental Tier (Level C/D Evidence with High Safety Margins)

  • Multi-System Functional and Proteomic Mapping: Clinicians should cross-reference epigenetic age scores with real-world physiological capacity testing. This includes monitoring arterial compliance via aortic pulse wave velocity (a validated indicator of vascular preservation in low PAI-1 phenotypes), alongside deep proteomic panels (e.g., OLink platform), retinal microvascular photography, and AI-guided gait velocity assessments to gauge functional tissue reserves.
  • Structured Neuroendocrine Stress Mitigation: Chronically elevated cortisol and sympathetic overdrive trigger acute-phase reactants, including hepatic PAI-1 synthesis. Deploying validated, digital positive-emotion and cognitive stress-reduction architectures can attenuate the systemic inflammatory drive, though direct, long-term clinical trial readouts on biological clocks are still actively accumulating.
  • Empirical Precision Nutrition Interventions: Incorporating high-fiber dietary protocols (~70g/day) and evaluating functional food targets such as specific mushrooms (which contain natural trace HMG-CoA reductase inhibitors, ergothioneine, and spermine) to optimize cardiometabolic risk parameters. Longevity-specific dosing ranges for absolute lifespan extension in humans remain speculative and unverified in large-scale randomized controlled trials (RCTs).

Red Flag Zone (Safety Data Absent)

  • Unregulated Off-Label Deployment of Small-Molecule PAI-1 Inhibitors: Utilizing investigational PAI-1 small-molecule inhibitors (e.g., TM5614 or RS5614) outside of tightly controlled clinical trial infrastructure is highly dangerous and classified as “Safety Data Absent” for healthy aging cohorts. Complete or excessive titration of PAI-1 introduces severe toxicological risks, including uncompensated clinical bleeding disorders, compromised innate immune surveillance, and catastrophic spontaneous subendocardial cardiac fibrosis. Partial, targeted modulation is mandatory; complete knockdown mimics a deleterious disease state.
Mike Lustgarten Video Series
#2

Thought it sounded familiar.

A methylation estimate of PAI-1 is included as one of 9 factors in the GrimAge clock.

2 Likes
#3

Yes, we’ve covered it in discussions here before:

2 Likes
#4

PAI-1 is a Marker and a Mediator of Senescence

Vaughan DE, Rai R, Khan SS, Eren M, Ghosh AK. “PAI-1 is a Marker and a Mediator of Senescence.” Arteriosclerosis, Thrombosis, and Vascular Biology (ATVB), 2017;37(8):1446–1452.

A blood-clotting brake called PAI-1 is not just a passive flag of aged, “senescent” cells — it appears to actively causethem to age, and blocking it (with drugs or lifestyle) delays multiple features of aging in mice. This is a review arguing PAI-1 is a druggable master-switch for senescence.

For decades, plasminogen activator inhibitor-1 (PAI-1) was filed under “clotting.” It is the body’s main off-switch for the enzymes that dissolve blood clots, and too much of it leaves people prone to dangerous thrombosis. But a team at Northwestern University in Chicago argues that PAI-1’s real story is far bigger: it may be a central driver of biological aging itself.

The big idea is a shift from marker to mediator. Scientists have long noticed that PAI-1 levels climb as cells and tissues age. Old fibroblasts, aged blood-vessel linings, and patients with premature-aging syndromes such as Werner syndrome all show elevated PAI-1. That made it a convenient biomarker of “cellular senescence” — the state in which cells stop dividing but linger, leaking inflammatory and tissue-damaging signals (the so-called senescence-associated secretory phenotype, or SASP). PAI-1 is one of the most prominent molecules in that toxic secretion.

The provocative claim the authors marshal evidence for is that PAI-1 doesn’t just accompany senescence — it helps switch it on. In landmark cell experiments, forcing PAI-1 expression was enough to push cells into senescence even when the usual master regulator, the tumor-suppressor p53, was absent. Conversely, removing PAI-1 let cells keep dividing.

The most striking support comes from “Klotho” mice, a strain engineered to age rapidly and die young. These animals carry PAI-1 levels roughly 45 times higher than normal. When the researchers genetically deleted PAI-1 — or blocked it with an experimental drug called TM5441 — the mice were protected against multi-organ aging, and in the underlying study their median lifespan increased about fourfold.

The authors then make the leap to humans. They note that the very interventions associated with healthier aging — eating more fruit and vegetables, modest caloric restriction, better sleep, and regular exercise — all lower circulating PAI-1. The implication is seductive: PAI-1 might be a shared chokepoint where lifestyle and future drugs converge to slow aging.

It is a compelling, coherent thesis. But it is a thesis. This is a perspective piece assembling correlative and model-organism evidence into a hypothesis — not proof that lowering PAI-1 extends healthy human lifespan. The headline mouse result comes from an extreme premature-aging model, not normal animals. [Confidence: the biology of PAI-1 in senescence is Medium-High; the human anti-aging payoff is Low and unproven.]

Actionable Insights (take-home, with effect sizes)

The practical message is narrow but real: the lifestyle levers that lower PAI-1 are the same boring, evidence-based ones — and PAI-1 gives a plausible mechanism for why they help vascular aging.

What the review actually supports, and the magnitude where stated:

  • Fruit and vegetable intake is inversely correlated with circulating PAI-1 (Nilsson 1990, cited): highest consumers showed the lowest PAI-1, with plasminogen-activator levels unchanged — i.e., the benefit is specifically a fall in the inhibitor. The review reports direction, not a clean effect size; no standardized magnitude is recoverable from the text.
  • Caloric restriction lowers PAI-1 while raising t-PA. CR is the most reproducible longevity intervention across species, so this is mechanistically consistent, but the review provides no human PAI-1 effect size.
  • Sleep quality is inversely correlated with PAI-1 (Chicago sleep study, cited).
  • Exercise / whole-body vibration “significantly reduces” PAI-1 (cited), no magnitude given.

The one large effect size in the whole piece is from mice, not people: in the cited Klotho model, PAI-1 was ~45-fold elevated, and PAI-1 blockade produced a ~4-fold (≈300% relative) median-lifespan increase, plus a 98% reduction in the aging hormone FGF23. That number is dramatic but comes from a sick, short-lived engineered strain (see Part 2) and should not be read as a human longevity multiplier.

Bottom line for the reader: there is no validated PAI-1-lowering supplement or drug for healthy people today (TM5441 is experimental). The honest take-home is that diet, CR, sleep and exercise plausibly act partly through PAI-1 — a reason to keep doing them, not a new intervention.

Source:

  • Institution: Feinberg Cardiovascular Research Institute, Feinberg School of Medicine, Northwestern University.
  • Country: United States (Chicago, Illinois).
  • Journal: Arteriosclerosis, Thrombosis, and Vascular Biology (ATVB) — an American Heart Association journal.
  • Impact Evaluation: The impact score of this journal is 7.4 (2024 Journal Impact Factor, Clarivate); CiteScore 15.5 (Scopus), evaluated against a typical high-end range of 0–60+ for top general-science and elite specialty journals, therefore this is a Medium-to-High impact journal
1 Like
#5

More in-depth look at actionable items…

Translational Longevity Interventions Derived from PAI-1 / Senescence Biology

Source data: Vaughan et al., PAI-1 is a Marker and a Mediator of Senescence, ATVB 2017 — a narrative review. The interventions below are derived from the review’s thesis (lower or normalize PAI-1; clear the senescent cells that secrete it) and then independently validated against external primary literature. Because the source is a review, none of these doses come from a controlled longevity trial in normal humans; translational uncertainty is high throughout and is flagged per intervention.

HED methodology (used consistently below): FDA Body Surface Area normalization. HED (mg/kg) = Animal dose (mg/kg) × (Animal Km ÷ Human Km). Standard Km values: mouse = 3, rat = 6, human = 37. For mouse → human the multiplier is 3 ÷ 37 = 0.081. Human absolute dose assumes a 70 kg adult unless noted. These are scaling estimates, not validated clinical doses.

PART 1 — ACTIONABLE INTERVENTIONS & EVIDENCE VALIDATION

Intervention 1 — Direct PAI-1 small-molecule inhibition (TM5614; research analog TM5441)

The Core Strategy. This is the most literal translation of the paper’s thesis: pharmacologically inhibit PAI-1 (SERPINE1) itself. PAI-1 is both a SASP component and, per the review, a causal mediator of senescence acting via the p53→PAI-1→PI3K-AKT-GSK3-cyclin D1 axis and via stabilization of IGFBP-3 (blocking t-PA-mediated IGFBP-3 proteolysis). Mechanistic target: direct antagonism of PAI-1, downstream suppression of p16^INK4a/p21/p53. Intended outcome: reduced vascular/pulmonary/renal senescence, improved fibrinolytic balance, delayed multi-organ aging.

Translational Dosing Protocol.

  • Mouse efficacy dose (vascular senescence, NAFLD, Klotho models): 20 mg/kg/day oral.
  • HED = 20 mg/kg × (3 ÷ 37) = 1.62 mg/kg → × 70 kg = ≈113 mg/day.
  • Validation of the math: human TM5614 trials independently used 120–180 mg/day orally — within ~6–60% of the BSA-predicted figure, an unusually clean cross-species convergence.
  • Pharmacokinetics: orally bioavailable small molecule; designed from the human PAI-1 crystal structure. Precise human half-life is not well published in the open literature (dosed once daily in trials, implying a half-life supporting QD dosing). PK detail partially absent.

Literature Validation & Source Verification.

Safety, Toxicity & Interaction Profile. Human trials (CML 180 mg/day × 48 weeks; COVID-19 120–180 mg/day) reported no serious drug-attributable adverse events and, notably, no prolongation of PT/aPTT or bleeding time — distinguishing PAI-1 inhibition from conventional anticoagulants. A formal NOAEL in mg/kg is not published in the accessible literature — Safety Data partially Absent for chronic healthy-population use. Long-term oncology context raises an unresolved theoretical concern: senescence is also tumor-suppressive, so chronically disabling a senescence mediator warrants caution. CYP450 interaction profile is not clearly published — Interaction Data Absent.

  • Longevity-stack compatibility: No published interaction data with rapamycin, metformin, acarbose, 17α-estradiol, SGLT2 inhibitors, or PDE5 inhibitors. Mechanistically, the lack of effect on coagulation times suggests low additive bleeding risk; but all combinations are unstudied — treat as unknown. [Confidence: Low for stack safety.]

Intervention 2 — Senolytic clearance with Fisetin

The Core Strategy. PAI-1 is a SASP factor; senescent cells are the source. Rather than inhibit PAI-1, remove the cells producing it. Fisetin is a flavonoid senolytic that preferentially triggers apoptosis in senescent cells (acting on senescent-cell anti-apoptotic pathways / SCAPs). Mechanistic target: senescent-cell apoptosis → SASP reduction (incl. SERPINE1/PAI-1, IL-6, p16, p21). Intended outcome: lower systemic senescent burden and inflammatory SASP, improved tissue homeostasis.

Translational Dosing Protocol.

  • Mouse senolytic dose: 100 mg/kg/day (acute pulse, ~5 days) or 500 ppm in chow.
  • HED = 100 mg/kg × (3 ÷ 37) = 8.1 mg/kg → × 70 kg = ≈567 mg/day during a pulse.
  • Real-world human senolytic protocols (and the Mayo “hit-and-run” trials) typically use 1,000–1,500 mg/day × 2 consecutive days, intermittently — i.e., 2–3× the simple BSA estimate, reflecting fisetin’s poor bioavailability rather than a higher target tissue exposure.
  • Pharmacokinetics: low oral bioavailability, extensive first-pass metabolism, short plasma half-life (~hours) — the pharmacological rationale for intermittent high-dose pulsing rather than daily low dosing.

Literature Validation & Source Verification.

Safety, Toxicity & Interaction Profile. Fisetin has a strong preclinical safety margin (food-derived flavonoid, found in strawberries). A formal human NOAEL is not established; published human efficacy data remain limited (trials ongoing). No major liver/kidney toxicity signal at studied doses, but **chronic high-dose human safety data are Absent.**Flavonoids including fisetin can inhibit CYP3A4 and CYP2C and modulate P-glycoprotein in vitro — a plausible interaction route at supratherapeutic doses. [Confidence: Medium on senolytic efficacy in mice; Low on human longevity benefit.]

  • Longevity-stack compatibility: CYP3A4 inhibition is the key flag — rapamycin is a CYP3A4/P-gp substrate, so concurrent high-dose fisetin could theoretically raise rapamycin exposure. Intermittent (2-day) fisetin pulsing limits overlap. No specific contraindication data with metformin, acarbose, SGLT2i, 17α-estradiol, or PDE5i. [Confidence: Low — interactions inferred, not measured.]

Intervention 3 — Caloric Restriction / dietary PAI-1 lowering

The Core Strategy. The review explicitly cites caloric restriction (CR) and high fruit/vegetable intake as lowering PAI-1 while raising t-PA, shifting fibrinolytic balance favorably. CR is the most reproducible longevity intervention across species. Mechanistic targets: AMPK activation, mTOR/IGF-1 suppression, reduced adipose-derived PAI-1 (adipose tissue is a major PAI-1 source). Intended outcome: lower circulating PAI-1, improved metabolic and vascular aging markers.

Translational Dosing Protocol.

  • No mg “dose.” Human CALERIE protocol targeted 25% restriction (achieved ~12%). Practical longevity target: ~10–20% sustained calorie reduction, or high produce intake. No BSA conversion applies.
  • “Pharmacokinetics” not applicable; effect is chronic and behavior-dependent.

Literature Validation & Source Verification (with effect sizes).

Effect-size note: the PAI-1 reductions above (−26% to −74%) are large and human-verified, but are largely a function of fat-mass loss in overweight subjects; the absolute reduction in lean, normal-weight individuals is likely smaller. One study found subcutaneous adipose PAI-1 mRNA paradoxically rose during VLCD even as plasma PAI-1 fell — a caution against assuming uniform tissue-level effects.

Safety, Toxicity & Interaction Profile. Excellent safety at moderate restriction; risks are lean-mass loss, bone density loss, and disordered-eating potential at aggressive restriction. No CYP issues. NOAEL framing not applicable.

  • Longevity-stack compatibility: Synergistic in mechanism with metformin, acarbose, SGLT2i, and rapamycin (all converge on nutrient-sensing). Practical caution: CR + SGLT2 inhibitors or aggressive CR + glucose-lowering agents raises hypoglycemia and euglycemic ketoacidosis risk. [Confidence: High for PAI-1 lowering in overweight; Medium for added longevity benefit beyond weight normalization.]

Intervention 4 — SIRT-1 activation (Resveratrol)

The Core Strategy. The review notes SIRT-1 suppresses PAI-1 and raises eNOS, while SIRT-1 inhibition induces senescence and raises PAI-1. Resveratrol is the prototypical SIRT-1 activator. Mechanistic target: SIRT-1/NAD+ axis → PAI-1 suppression, endothelial protection. Intended outcome: reduced endothelial senescence and PAI-1.

Translational Dosing Protocol.

  • Mouse metabolic/longevity doses span ~22 mg/kg (Baur high-calorie study) to higher. Using 22 mg/kg: HED = 22 × 0.081 = 1.78 mg/kg → × 70 kg = ≈125 mg/day; higher murine SIRT studies (100+ mg/kg) scale to ~570+ mg/day.
  • Human trials commonly use 150 mg to 1,000 mg/day.
  • Pharmacokinetics: notoriously poor — extensive first-pass glucuronidation/sulfation, very low systemic bioavailability, rapid urinary excretion (largely within ~4 h), enterohepatic recirculation. This is the central weakness: plasma exposure of free resveratrol is low, and human SIRT-1 activation is inconsistent.

Literature Validation & Source Verification.

Effect-size / dissent note: human resveratrol data are mixed and frequently null; the high-profile “SIRT1 direct activator” interpretation has been contested (assay-artifact debate). Direct human PAI-1 reduction by resveratrol is not robustly established. This is the weakest-evidence intervention in the report.

Safety, Toxicity & Interaction Profile. Well tolerated to ~1 g/day; **GI distress and diarrhea at higher doses (2.5–5 g).**No major organ-toxicity signal at supplement doses. Resveratrol inhibits CYP3A4, CYP2C9, and CYP1A2 — a meaningful interaction route. NOAEL effectively high but loosely defined.

  • Longevity-stack compatibility: CYP3A4 inhibition → can raise rapamycin and dasatinib levels; CYP2C9 inhibition → relevant to many drugs. Mild antiplatelet effect → caution with PDE5 inhibitors and other antiplatelet agents. No specific contraindication with metformin/acarbose/SGLT2i/17α-estradiol. [Confidence: Low — both efficacy and the SIRT-1→PAI-1 human link are unproven.]

PART 2 — STRATEGIC FEASIBILITY & TARGET ENGAGEMENT

Biomarker Verification (proving target engagement in humans)

Intervention Primary target-engagement biomarker Supporting biomarkers
PAI-1 inhibitor (TM5614) Plasma PAI-1 antigen + PAI-1 activity (direct target); fibrinolytic shift (↑ t-PA activity, ↓ D-dimer normalization) p16^INK4a in tissue/T-cells, IGFBP-3
Fisetin p16^INK4a (CDKN2A) in CD3+ T-cells / adipose; SA-β-Gal in biopsy Circulating SASP panel: IL-6, PAI-1/SERPINE1, MMP-9
Caloric restriction Plasma PAI-1 (human-validated to fall 26–74% with fat loss); fasting insulin, IGF-1 Epigenetic clock pace (DunedinPACE), adiponectin
Resveratrol Plasma PAI-1 + SIRT-1 activity / target-gene readout eNOS-dependent flow-mediated dilation; free vs conjugated resveratrol PK

The common, cheap, directly-relevant readout across all five is a plasma PAI-1 antigen + activity assay — the exact marker this paper validates. p16^INK4a in peripheral T-cells is the best general senescence-burden surrogate but is largely research-grade, not routine clinical.

Sourcing & Financial ROI

Intervention Procurement status Approx. monthly cost (HED) Cost-to-benefit assessment
TM5614 / TM5441 Research chemical / investigational Rx (TM5441 sold for research use only; TM5614 in trials, not approved for longevity) Effectively N/A for legitimate consumer use; research-grade material highly variable Poor/unavailable — strongest mechanistic fit to the paper, but not legitimately obtainable or quality-assured for human use. Highest translational uncertainty.
Fisetin Supplement (OTC) ~$10–25/month (intermittent pulsing uses little material) Best compound ROI — cheap, oral, intermittent, decent mouse data. Capped by absent human efficacy data and bioavailability.
Caloric restriction Lifestyle (free) ~$0 (potentially net savings) Highest practical ROI — free, human-validated PAI-1 reduction, broad benefits. Cost is adherence/discipline, not money.
Resveratrol Supplement (OTC) ~$15–40/month Weak ROI — cheap but efficacy/bioavailability poor; the SIRT-1→PAI-1 human link is the least substantiated.

Strategic Synthesis

Ranked by evidence-weighted, risk-adjusted practicality for the stated longevity goal (lowering PAI-1 / senescent burden):

  1. Caloric restriction / fat-mass normalization — only intervention with direct, large, human-verified PAI-1 reduction; free; lowest risk. The default high-yield action.
  2. Fisetin (intermittent pulses) — best benefit-to-risk-to-cost among compounds; strong mouse senolytic + lifespan data; main gap is human efficacy.
  3. TM5614 / TM5441 — the purest mechanistic match and the cleanest HED convergence (~113 mg predicted vs 120–180 mg used), but not legitimately available and unproven for longevity.
  4. Resveratrol — weakest: poor bioavailability, contested SIRT-1 mechanism, no robust human PAI-1 effect.

Overarching Limitations & Dissent

  • The source is a review by the lab commercializing PAI-1 inhibitors — directional bias toward the PAI-1-as-driver thesis.
  • No intervention here has demonstrated extended lifespan in normal-aged humans. The flagship mouse lifespan result (TM5441/Eren) is in a progeroid Klotho model with extremely short-lived controls — the weakest evidence class for longevity translation.
  • Senescence is tumor-suppressive; chronically inhibiting senescence mediators or clearing senescent cells carries a theoretical oncogenic/wound-healing trade-off that long-term human data have not resolved.
  • CYP3A4 is the recurring interaction hub: fisetin and resveratrol (inhibitors) and rapamycin (substrate) converge there — the single most important practical caution for anyone stacking these compounds.

Consolidated Sources

Disclaimer: This is a research analysis, not medical advice. Several interventions involve research chemicals and off-label use with real toxicity and interaction risks. Consult a qualified physician before acting; do not self-administer research chemicals.

#6

For me the key questions are:

  1. What is the median and standard deviation of PAI-1 levels at different ages.

  2. How is it measured and how reliable is the measurement?

Probably the paper on the amish population will be most useful.

Then target 50% of the median PAI-1 levels (heterozygous SERPINE mutation).

PAI-1 levels that are too low are also harmful (as mentioned by the video)

1 Like
#7

The paper:

A Natural “Off Switch” for Aging? The Amish Who Inherited Longer Telomeres and a Decade of Extra Life

An analysis of Khan et al., “A null mutation in SERPINE1 protects against biological aging in humans,” Science Advances, 2017.

A rare, naturally occurring loss-of-function mutation in the SERPINE1 gene, found in a small Amish community, halves blood levels of the protein PAI-1 and is linked to longer telomeres, better metabolic health, and roughly a decade of extra lifespan. This is the first human evidence that a single anti-aging molecular target validated in mice may also operate in people.

For decades, anti-aging research has lived mostly in petri dishes and mouse cages. A 2017 study from Northwestern University finally brought one of its most promising targets into humans — and it did so by studying an Amish community in Berne, Indiana, that carries a genetic quirk found almost nowhere else on Earth.

The protein at the center is PAI-1 (plasminogen activator inhibitor-1), encoded by the gene SERPINE1. PAI-1 is best known for regulating blood clotting, but it has a second life: it is a core component of the “senescence-associated secretory phenotype” — the toxic cocktail of signals that aging, “zombie” cells pump out into surrounding tissue. In mice engineered to age rapidly, switching off PAI-1 protects organs, preserves telomeres, and dramatically extends life. The open question was always whether the same is true for us.

The Berne Amish offered a rare natural experiment. Because of their geographic and genetic isolation, a single frameshift mutation in SERPINE1 — traced back to one married couple roughly six generations ago — has spread through the kindred. Carriers make far less PAI-1 for their entire lives, essentially mimicking a lifelong, gentle drug treatment.

The researchers studied 177 community members, 43 of whom carried one copy of the null mutation. The carriers stood out. Their telomeres — the protective caps on chromosomes that fray with age — were about 10 percent longer on average. They had lower fasting insulin and, strikingly, not a single case of diabetes among the 43 carriers, compared with 7 percent of non-carriers. And when the team reconstructed the family tree using birth and death records stretching back generations, carriers lived a median of about 10 years longer.

The “big idea” is that aging may have manipulable molecular levers, and PAI-1 looks like one of them. Crucially, orally active PAI-1 inhibitor drugs already exist and were in early human trials in Japan. Decades of observing these mutation carriers — who show no abnormal bleeding despite low PAI-1 — hint that partially blocking the protein could be safe.

This is an observational study of a tiny, unusual population, so it is a beginning rather than a verdict. But it is a rare moment when a longevity target jumps convincingly from mouse to human.

Actionable Insights

The honest take-home is that the headline benefit here is genetic, not a lifestyle hack — you cannot acquire this mutation. But the paper points to actionable directions and quantifies what lowering PAI-1 is worth.

The effect sizes are substantial. Carriers averaged a roughly 10-year longer median lifespan (85 vs 75 years; about 13 percent) and a 7-year longer mean age at death (82 vs 75; Cohen’s d about 0.63, a medium-to-large effect). Their telomeres were about 10 percent longer, which — given the cohort lost roughly 9 percent of telomere length per decade — translates to looking biologically about 11 years younger at the chromosomal level. Fasting insulin was 28 percent lower (model-adjusted), and diabetes prevalence was 0 percent versus 7 percent.

What is actionable for the rest of us: PAI-1 is not fixed by genetics alone. It is raised by insulin, glucose, free fatty acids, obesity, and visceral fat, and it is lowered by interventions that already have evidence behind them — caloric restriction, weight loss, exercise, and metabolic drugs such as metformin. The paper explicitly notes these all converge on reducing PAI-1 expression. So the practical message is that strategies that improve insulin sensitivity and reduce adiposity may be partly working through the same pathway this fortunate Amish family inherited. Pharmacological PAI-1 inhibitors are an emerging “watch this space” intervention, not yet available for longevity use.

Context

  • Institution: Northwestern University Feinberg School of Medicine (with collaborators at New Jersey Medical School, University of British Columbia, the Indiana Hemophilia and Thrombosis Center, and Tohoku University, Japan).
  • Country: United States.
  • Journal: Science Advances (American Association for the Advancement of Science, AAAS).

Impact Evaluation

Science Advances has a 2024 Journal Impact Factor of approximately 12.5 (CiteScore approximately 19.6).

The impact score of this journal is 12.5, evaluated against a typical high-end range of 0–60+ for top general-science journals, therefore this is a High impact journal. (It sits clearly in Q1 multidisciplinary science but below the very top tier occupied by Nature, Science, and Cell, which carry impact factors in the 40–60+ range — hence “High” rather than “Elite.”)

PART 2 — THE BIOHACKER ANALYSIS (Technical)

Study Design Specifications

  • Type: Human observational study — cross-sectional for the biomarker and metabolic endpoints, plus a retrospective genealogical (pedigree) analysis for lifespan. This is not an interventional clinical trial, and not an in vivo animal or in vitro study. [Confidence: High]
  • Subjects: Old Order Amish (Berne, Indiana founder population), age 18+, both sexes.
    • Total enrolled: 177 (of 450 invited).
    • Non-carriers (SERPINE1 +/+, control group): n = 127.
    • Heterozygous carriers (+/-): n = 43.
    • Homozygous null (-/-): n = 7 (excluded from primary analyses due to small number and young age, 18–34 years).
    • Lifespan analysis: 221 deceased extended-kindred individuals with known birth/death dates; genotype ascertained (direct or obligate) in 56.
    • External validation cohort: CARDIA, a U.S. population-based study (n = 2793 at the relevant exam).

Lifespan

  • Median survival: carriers 85 years (IQR 73–88) vs non-carriers 75 (IQR 70–83); P = 0.037.
    • Absolute median extension: +10 years; relative: +13.3%.
  • Mean age at death: carriers 82 ± 10 vs non-carriers 75 ± 12; P = 0.037 (Wilcoxon).
    • Absolute: +7 years; relative +9.3%.
    • Cohen’s d ≈ 0.63 (calculated from pooled SD ≈ 11.05) — a medium-to-large effect. [Confidence: Medium]
  • Maximum lifespan: not formally reported as a distinct statistic; the upper IQR bound (88 vs 83) is not a true maximum-lifespan measure. Treat maximum-lifespan extension as not established by this paper. [Confidence: High]
  • Note: lifespan analysis was restricted to individuals who died at ≥45 years to exclude premature death from accident, infection, and childbirth.

PAI-1 (the mechanistic readout)

  • Heterozygotes: 5.9 ± 6.6 ng/ml vs non-carriers 12.7 ± 9.8 ng/ml; P < 0.0001 (about 50% lower).
    • Cohen’s d ≈ 0.81 (large). Caveat: distributions are highly skewed (SD approaches or exceeds the mean), so d is an approximation. [Confidence: Medium]
    • Homozygotes: no detectable PAI-1 antigen (consistent with true loss of function).

Telomere length (primary endpoint)

  • Carriers had ~10% longer mean leukocyte telomere length (LTL) after adjustment for age, sex, and family structure; P = 0.007 (qPCR) and P = 0.039 (Southern blot/TRF).
  • Translation: at ~9% LTL loss per decade (0.9%/year) in this cohort, a 10% longer LTL corresponds to roughly 11 years younger biological telomere age. [Confidence: Medium]
  • A formal standardized effect size (Cohen’s d) cannot be computed precisely because group-level SDs for LTL are not reported; only the adjusted polygenic-model estimate and p-value are given. Heritability of LTL h² = 0.55. [Confidence: High that d is not computable from the published numbers]

Metabolic endpoints

  • Fasting insulin: 28% lower in heterozygotes (model-adjusted, P = 0.035). Raw medians: 4.0 (2.9–5.1) vs 4.9 (3.3–6.7) uIU/ml, an ~18% raw reduction. [Confidence: High]
  • Type 2 diabetes prevalence: 0% (0/43) in carriers vs 7% (8/127) in non-carriers; P = 0.001.
    • Relative risk / odds ratio are undefined (division by zero events) in carriers; report as absolute risk reduction of ~7 percentage points. This is a real but statistically fragile signal given zero events in a small group. [Confidence: Medium]
  • PAI-1 correlated positively with fasting insulin in both Amish (R = 0.55) and CARDIA (R = 0.48). Moderate correlation. [Confidence: High]

Cardiovascular composite scores (e′ velocity, brachial pulse pressure, carotid IMT): all trended favorably in carriers but the individual cardiovascular composite did not reach significance (P = 0.09); only the comprehensive score including LTL was significant (0.53 units lower, P = 0.005). [Confidence: Medium]

Mechanistic Deep Dive

The unifying node is cellular senescence and the SASP. PAI-1 is not just a marker of senescence; per the cited mechanistic work it is “necessary and sufficient” for replicative senescence in vitro and is a key downstream effector of p53. Lowering PAI-1 plausibly reduces senescent-cell burden and the inflammatory secretome that drives tissue aging. [Confidence: Medium]

Mapping to canonical longevity pathways:

  • Deregulated nutrient sensing (insulin/IGF-1): This is the best-supported axis here. PAI-1 is induced by insulin/glucose/free fatty acids and, conversely, impairs degradation of IGFBP-3 and IGF-1, both of which can trigger senescence. Lower PAI-1 tracks with lower fasting insulin and no diabetes — consistent with improved insulin sensitivity. This intersects the insulin/IGF-1/FOXO longevity pathway, the same one targeted by caloric restriction and metformin. [Confidence: Medium-High]
  • mTOR / AMPK: Not directly measured. The paper notes metformin, resveratrol, and caloric restriction (AMPK activators / mTOR modulators) all reduce PAI-1 expression, implying PAI-1 sits downstream of or parallel to these pathways, but this study provides no direct mTOR/AMPK data. [Confidence: Low]
  • Telomere attrition: Strongly featured. Carriers preserve LTL, consistent with mouse data showing PAI-1 inhibition prevents telomere shortening. Whether this is causal (PAI-1 protecting telomeres) or a downstream marker of less senescence/metabolic stress is unresolved. [Confidence: Medium]
  • Autophagy, cGAS-STING, mitochondrial dynamics: Not assessed. No data. Any link is speculative. [Confidence: Low]
  • Organ-specific aging priorities: The signal points to cardiometabolic/vascular tissue (insulin sensitivity, diabetes, vascular stiffness, myocardial relaxation) as the primary beneficiaries, consistent with PAI-1 being liver- and adipose-derived. [Confidence: Medium]

Novelty — What This Adds

This is the first human demonstration that a private loss-of-function mutation lowering PAI-1 is associated with longer telomeres, healthier metabolism, lower diabetes, and longer lifespan. Prior PAI-1/longevity causal data came almost entirely from mouse models of accelerated aging (Klotho, BubR1). This study translates that target into a human population whose effect on circulating PAI-1 is large (a genetic “knockdown” rather than a small common-variant nudge), and links it to a near-term druggable target — oral PAI-1 inhibitors already in early human testing. It also introduces composite “biological aging” scores validated against 5-year cardiovascular outcomes in CARDIA. [Confidence: High]

Critical Limitations (told straight)

  • Observational, not interventional. No randomization, no causal proof. “Causal effect” in the abstract is an overstatement relative to the design; the data are associative. [Confidence: High]
  • Tiny, genetically confounded sample. 43 heterozygotes and only 7 homozygotes. All carriers descend from a single ancestral couple ~6 generations ago, so the mutation could be co-inherited with other longevity-favoring genetic or epigenetic factors. Adjusting for relatedness in SOLAR mitigates but does not eliminate this. [Confidence: High]
  • Lifespan data are retrospective genealogy. Genotype was directly or inferentially ascertained in only 56 of 221 deceased; obligate ascertainment introduces classification uncertainty. The ≥45-year survival cutoff is reasonable but the wide IQRs (carriers 73–88) signal substantial uncertainty around the ~10-year estimate. [Confidence: Medium]
  • Wide CIs / fragile p-values. Several key results sit just under 0.05 (lifespan P = 0.037, insulin P = 0.035). The diabetes finding rests on zero events in carriers — impressive but unstable; one or two cases would shift it materially. [Confidence: High]
  • Cardiovascular endpoints largely non-significant on their own (P = 0.09); the “comprehensive” score’s significance is partly driven by LTL it already contains (some circularity). [Confidence: Medium]
  • Limited generalizability. A homogeneous-diet, lifestyle-isolated, relatively healthy founder population — even non-carrier Amish had PAI-1 ~50% below U.S. norms. Effects may not transfer to a typical Western population with high baseline PAI-1, obesity, and varied genetics. [Confidence: High]
  • Missing data: no direct mechanistic readouts (mTOR, AMPK, autophagy, mitochondrial, senescence-cell quantification); no longitudinal within-person telomere tracking; no incident-disease follow-up; effect sizes for LTL and insulin not reported with the SDs needed for clean standardized effect-size computation. [Confidence: High]

Bottom Line

This paper provides the strongest human signal to date that PAI-1 is a real, possibly causal, and druggable aging lever — but the evidence is preliminary, drawn from a small, genetically peculiar, healthier-than-average population, with borderline statistics and an observational design. It strongly motivates trials of PAI-1 inhibitors and supports the broader idea that targeting senescence and the insulin/IGF-1 axis matters, without proving that lowering PAI-1 will extend life in the general population. [Overall Confidence: Medium]

Sources for journal metrics:

Primary source: Khan SS, Shah SJ, Klyachko E, et al. A null mutation in SERPINE1 protects against biological aging in humans. Sci Adv. 2017;3(11):eaao1617.

#8

My follow-on Prompt:

From this, and a broader research review, identify

  1. What is the median and standard deviation of PAI-1 levels at different ages.
  2. How is it measured and how reliable is the measurement?
  3. What was the median and standard deviation of PAI-1 levels in the Amish group with the mutation in the SPERPINE1 gene?
  4. What would be the levels to target if the target is 50% of the median PAI-1 levels (heterozygous SERPINE mutation).

Claude Opus 4.8 Response:

A key caveat before the numbers: PAI-1 is one of the least standardized common plasma analytes. Values depend on whether you measure antigen (ng/mL) or activity (IU/mL or AU/mL), the assay vendor, time of day, fasting state, and the 4G/5G genotype. So cross-study “medians by age” should be treated as ballpark, not canonical.

1. PAI-1 levels by age (general population)

There is no single clean median±SD-by-decade table in the literature; reported reference ranges overlap heavily. The consensus picture:

Group Typical PAI-1 antigen Notes
Healthy adults (broad) ~5–40 ng/mL; mean ~18 ± 10 Wide fasting reference range cited as 3–72 ng/mL
Young adults lower end Rises after middle age
Middle-aged (esp. males) 50–60 ng/mL “not uncommon” Driven largely by insulin resistance/adiposity
Elderly Higher still Steep rise after midlife, paralleling senescent-cell load

The honest summary: PAI-1 rises progressively with age, with the steepest climb after middle age. But the dominant driver is not chronological age per se — it is insulin resistance, BMI, triglycerides, and the 4G/5G polymorphism. One large analysis found t-PA/PAI-1 complex rises clearly with age while free PAI-1 age-trends were weaker, underscoring how method-dependent the “age effect” is. I would not quote a precise median±SD for, say, a 40-year-old, because the literature doesn’t support that level of precision. [Confidence: Medium on the trend; Low on exact per-decade values]

2. How it’s measured and reliability

Two assay families:

  • Antigen (ELISA, ng/mL) — measures total PAI-1 protein (active + latent + complexed). This is what the Amish paper used (Molecular Innovations ELISA). Good analytical precision: intra-assay CV ~5–6%, inter-assay CV ~3–8% across common kits (TintElize, Imubind). The paper itself reported intra-assay CV 6.15%, inter-assay 5.98% — solidly reliable analytically.
  • Activity assay (IU/mL) — measures only functional PAI-1. Higher CVs (activity assays ~5% intra, up to ~16% inter in some validations) and poor sensitivity at very low levels, so it can’t reliably distinguish “low-normal” from true deficiency.

The reliability problem is biological/pre-analytical, not analytical:

  • Diurnal variation is large — PAI-1 peaks in the early morning and falls through the afternoon. This rhythm is strong in 4G/4G and 4G/5G genotypes and largely absent in 5G/5G. This swamps the ~6% assay CV.
  • Sensitive to fasting state, recent exercise, acute illness/inflammation (it’s an acute-phase reactant), and sample handling (platelet contamination raises it).

Bottom line: the assay is precise, but a single PAI-1 draw is a noisy estimate of a person’s “true” level. Standardize timing (early-morning fasting), and ideally repeat-measure, before acting on a number. [Confidence: High]

3. The Amish mutation group, and your 50% target

One clarification: the paper reported PAI-1 as mean ± SD, not median:

  • Heterozygous carriers (+/-, the mutation group): 5.9 ± 6.6 ng/mL
  • Non-carriers (+/+): 12.7 ± 9.8 ng/mL
  • Homozygous (-/-): no detectable antigen

Note the SD (6.6) exceeds half the mean — the distribution is heavily right-skewed, so the median is almost certainly lower than 5.9 (likely ~4–5 ng/mL), though the paper doesn’t publish it. Treat 5.9 ± 6.6 as the best available figure.

Target calculation (50% of the heterozygous level):

  • 0.5 × 5.9 ng/mL ≈ 2.95, i.e. roughly 3 ng/mL.

Two ways to read your goal, worth separating:

  • If you literally want half of the carriers’ level: target ~3 ng/mL. That is approaching the homozygous (undetectable) state — biologically aggressive and, in drug terms, near-complete inhibition.
  • If your goal is to mimic the protected carriers (which is what the longevity signal actually supports): the carriers themselves already sit at ~50% of non-carriers (5.9 vs 12.7) and ~50% of typical U.S. population levels. So the evidence-backed target is ~6 ng/mL, i.e. roughly halving a typical baseline, not halving the carriers’ already-low level.

The paper’s own safety argument supports the second framing: heterozygotes (partial reduction, ~6 ng/mL) had 20+ years with no bleeding, whereas homozygotes (near-zero) have a bleeding disorder. Driving toward ~3 ng/mL moves you toward that homozygous deficiency zone. [Confidence: Medium — the “optimal” target for longevity is not established by any trial.]

A practical note: because of the diurnal and biological variability above, hitting “3” vs “6” ng/mL on a single morning draw is within the noise band of the measurement itself, so any targeting should be based on repeated, time-standardized sampling.

Sources:

#9

Follow-on Prompt:

Identify and provide details on all the orally active PAI-1 inhibitor drugs that already exist and have at least phase 1 clinical trials completed.

Claude Opus 4.8 Response:

The honest bottom line: despite decades of interest, only two orally active PAI-1 inhibitors have ever reached human clinical trials. One is essentially the entire clinical-stage field today; the other failed years ago. Everything else (TM5275, TM5441, TM5484, TM5509, S35225, MDI-2268, etc.) is preclinical or research-tool only.

1. TM5614 (also called RS5614)

This is the drug the Amish paper was referring to (“orally active PAI-1 inhibitors… early-phase clinical testing in humans in Japan”).

  • Developer: Tohoku University (Toshio Miyata’s group — a co-author on the Amish paper) and Renascience Inc., Japan. Developed almost entirely in academia over 20+ years.
  • Chemistry/mechanism: Small-molecule N-acylanthranilic acid derivative, selected from ~1,400 synthesized analogues of the hit compound TM5275 via in-silico screening against the human PAI-1 crystal structure. It inserts into the vitronectin-binding site, blocking PAI-1 and promoting its degradation. IC50 ≈ 3.6 µM. Oral, once-daily.
  • Clinical status: Developed through the full Phase I → II → III pathway. Phase 1 safety was established (the program describes first-in-human dosing confirming safety before efficacy trials), and it has since gone well beyond Phase 1.
  • Typical dosing in trials: ~120–180 mg/day.
  • Completed trials (proof-of-concept obtained):
    • Chronic myeloid leukemia (Phase II, n=33, 48 weeks + TKI): deep molecular response rate 33% vs historical 8–12%; no serious drug-related adverse events. Phase III ongoing (ends ~2026).
    • Refractory malignant melanoma (Phase II, n=29, + nivolumab): response rate ~24%, disease control ~62%; granted orphan-drug designation; Phase III started Feb 2025.
    • COVID-19 pneumonia (Phase IIa open-label n=26, then placebo-controlled Phase IIb n=75): safe; primary efficacy endpoint not significant but signals of benefit in moderate disease.
  • Ongoing Phase II: non-small-cell lung cancer, angiosarcoma, systemic sclerosis–associated interstitial lung disease; planned for cancer-associated thrombosis and pancreatic cancer.
  • Safety signal relevant to your longevity interest: No notable bleeding across these trials — consistent with the Amish heterozygotes’ 20+ years without bleeding. The main flagged adverse event was occasional liver dysfunction (~6% in the melanoma combo arm).
  • Important caveat: Every completed human trial is for cancer, COVID, or fibrosis — none for aging/longevity per se. The anti-aging rationale remains preclinical (mouse). So “exists and clinically tested” is true; “validated for longevity in humans” is not.

2. Tiplaxtinin (PAI-039; INN: tiplasinin)

  • Developer: Wyeth (now Pfizer), USA.
  • Chemistry/mechanism: Indole-oxoacetic acid small molecule, orally bioavailable, inactivates active PAI-1 (Kd ≈ 480 nM; IC50 ≈ 2.7 µM). Strong preclinical anti-thrombotic data in rat, dog, and primate models.
  • Clinical status: Advanced to Phase 1 for thrombosis, then discontinued. Per Wikipedia and NCATS, it was unsuccessful in human testing due to an unfavorable risk/benefit ratio — it required tight dose control to avoid provoking bleeding. (Note: sources are slightly inconsistent; some describe it as never progressing past preclinical, but the NCATS drug registry records a discontinued Phase 1.) It is now used only as a research reagent.

Summary

Drug Developer Oral Highest stage Status
TM5614 / RS5614 Tohoku Univ. / Renascience (Japan) Yes Phase III (ongoing) Active; multiple completed Phase I/II trials, POC obtained
Tiplaxtinin (PAI-039) Wyeth/Pfizer (USA) Yes Phase I Discontinued (bleeding risk / risk-benefit)

Sources:

#10

Ashitaba exudate is an herb that will work. A standardized dose is from a company called Chalcurb. You don’t want the leaves. I actually can’t find this product. Does anybody know somebody in Japan?

#11

https://www.amazon.com/Swanson-Chalcurb-Ashitaba-Milligrams-Capsules/dp/B07C4XWHLL

This is what is needed but it’s out of stock

1 Like
#12

Maybe…

Short answer: yes, plausibly — but only as an indirect PAI-1 suppressor, and the evidence is entirely preclinical (cell + mouse). No human study has ever directly measured a PAI-1 reduction from Ashitaba. It’s a fundamentally different mechanism from the direct inhibitors (TM5614, tiplaxtinin) discussed earlier.

What Ashitaba is and the active compounds

Ashitaba (Angelica keiskei Koidzumi) is a Japanese plant whose yellow stem exudate (sap) is rich in two prenylated chalcones — xanthoangelol (XA) and 4-hydroxyderricin (4-HD) — which together make up ≥90% of the chalcone content. Critically for PAI-1, the activity is structure-dependent and not shared by both compounds.

The PAI-1 evidence (all preclinical)

In vitro (human endothelial cells):

  • Ohkura et al., 2011 (BioFactors): Xanthoangelol and its analogs (XA-B, XA-D) suppressed TNFα-induced PAI-1 production in human umbilical vein endothelial cells.
  • Ohkura et al., 2015 (Biosci Biotechnol Biochem): PAI-1 release inhibition from human endothelial cells is structure-dependent — a long hydrocarbon chain with moderate hydrophobicity on the chalcone A-ring is essential. XA suppressed PAI-1, but 4-HD did not. The effect occurs at the mRNA level (it blocks PAI-1 expression/transcription), not by binding the PAI-1 protein.

In mice:

  • Low-grade inflammation (LPS) model: Exudate (oral and intraperitoneal) blocked plasma PAI-1 elevation and suppressed PAI-1 in adipose, liver, and heart tissue.
  • Ohta et al., 2018 (J Diet Suppl), Tsumura Suzuki Obese Diabetic mice: Ashitaba exudate prevented the obesity-induced rise in plasma PAI-1, bringing it nearly to lean-control levels, and also lowered glucose, insulin, TNFα, and body-weight gain. Notably it had little/no effect in lean mice — the PAI-1 benefit appeared only where PAI-1 was pathologically elevated.
  • A long-term 52-week study reported that exudate reduced the age-associated increase in plasma PAI-1 in mice.

In humans (no direct PAI-1 data):

  • Small trials exist for metabolic syndrome / obesity — e.g. a pilot study (green juice ~6.2 g/day, 8 weeks, n=9) and randomized double-blind chalcone-powder studies (ChalCurb, ~200 mg/day, 8 weeks) — showing reductions in visceral fat, body weight, glucose, and lipids. But these measured metabolic endpoints; none published a direct PAI-1 measurement. Any human PAI-1 benefit is inferred, not demonstrated.

Mechanism — how this differs from the drugs you asked about earlier

Ashitaba (XA) TM5614 / Tiplaxtinin
Action on PAI-1 Lowers production (↓ mRNA expression, esp. inflammation-driven) Directly inhibits the PAI-1 protein’s activity
Driver targeted Obesity + chronic inflammation (TNFα), via AMPK activation, anti-adipogenesis The molecule itself, regardless of source
Evidence level Cell + mouse only Human Phase I–III

XA also activates AMPK (the same nutrient-sensing longevity pathway flagged in the original paper) and suppresses adipocyte differentiation — so it plausibly reduces the upstream metabolic-inflammatory drivers that raise PAI-1, rather than neutralizing PAI-1 directly.

Honest assessment of whether it would help [Confidence: Low–Medium]

  • The PAI-1-lowering signal is biologically coherent and consistent across cell and mouse models, and it links cleanly to Ashitaba’s better-established metabolic effects. [Confidence: Medium for the preclinical effect]
  • But three real caveats: (1) No human PAI-1 data exist — the leap from mouse plasma to human plasma is unproven. (2) The active compound is XA specifically; 4-HD is inactive for PAI-1, so a supplement’s effect depends heavily on its XA content/standardization. (3) Prenylated chalcones have low oral bioavailability, so whether oral Ashitaba reaches the µM tissue concentrations used in vitro is unclear. (4) The mouse data suggest the benefit appears mainly when PAI-1 is already elevated by obesity/inflammation — not necessarily in a healthy, lean person. [Confidence: High on these limitations]

So: a reasonable, low-risk dietary candidate for modestly dampening elevated PAI-1 in a metabolic-syndrome context, but it is nowhere near the evidence base — or the direct, potent mechanism — of the clinical PAI-1 inhibitors. It complements rather than replaces the insulin-sensitizing/weight-loss strategies that lower PAI-1.

Sources:

#13

I’m convinced haha. This is going on my list of low risk things to try.

#14

I am taking TM5614 for about 2 years now. Cycle and limit dosing with the aim to limit bleeding risk. Never had any issues but have no objective way to judge if it is making a positive difference. My hair quality is very good for my age (57). Virtually no grey hair or hair loss, but lifespan is more important to me.

1 Like
#15

Vendors related to this:

One important distinction up front: **most retail “Ashitaba” products are leaf/stem powder, which is not the exudate.**The exudate is the thick yellow sap squeezed from cut stems; when freeze-dried it concentrates to roughly 8% chalcone, versus a fraction of that in leaf powder. Since the PAI-1 research used the exudate/sap, the standardized-chalcone products below are the closest match. I can’t verify live stock, pricing, or that every label’s “chalcone” truly comes from sap vs. leaf — supplement marketing routinely blurs this, so confirm chalcone standardization before buying.

Tier 1 — Sap/exudate-derived or standardized-chalcone products (closest to the research material)

  • ChalCurb® — A standardized powder derived from the sap of Angelica keiskei, ~8% chalcone. This is the exact material used in the published human pilot trials. It’s primarily a branded ingredient sold to formulators, so you’ll usually find it inside finished products rather than sold direct.
  • AshiGuard (Dyna Nutrition) — 500 mg capsules, marketed as standardized Ashitaba chalcone.
  • Ace Ashitaba (Singapore; Hachijō Island, Japan source) — Sells a liquid “chalcone extract” (100 ml) and capsules, explicitly describing the yellow-sap chalcones xanthoangelol and 4-HD. Sold via aceashitaba.com and Amazon SG.
  • Toki Botanicals — Ashitaba 100:1 organic extract standardized to 10% chalcone.

Practical guidance

If your goal is the PAI-1/chalcone effect specifically, look for a label that states sap- or exudate-derived and a standardized chalcone percentage (≈8–10%) — Tier 1. Tier 2 leaf powders are fine as a general superfood but deliver far less chalcone per gram and aren’t what the studies tested. Also note the EU classifies **ashitaba sap as a “Novel Food”**with assessed safety limits, which is a useful quality/dose reference point.

A standard caution: dietary supplements aren’t tightly regulated for content accuracy, and ashitaba contains furanocoumarins (potential drug-interaction/photosensitivity concerns), so verify third-party testing and check interactions if you’re on medication. This is information, not a purchase recommendation.

Sources:

#16

Interesting… how did you get it? What dose are you taking?

1 Like
#17

I source it from a supplier in China without independent testing. I appreciate that some people will find that too risky. We will all have to make our own trade offs in terms of risks and benefits. I take about 120mg daily in my coffee. Happy to share contact details of supplier to anyone interested. It helps to pool demand to get a lower price.

1 Like
#18

Have you measured PAI-1 and seen it has gone down? If so what assay?

#19

Unfortunately I have no measurements of PAI-1

#20

Where do you buy it?