AI Longevity Daily

Today’s Summary

Blood protein screening identified 17 markers tied to mortality risk, some syncing up with younger biological age estimates.
An Alzheimer's blood test showed in a large study that the highest-risk group has about a 78% chance of cognitive impairment within a decade.
Both are observational associations — no go for individual diagnosis or definitive prediction.

📌 Today’s Focus Top 10

1. 17 Protein Markers Linked to Both Mortality Risk and Biological Age

Seventeen protein markers — pulled from a screen of 4,265 plasma proteins in a prospective cohort study of 1,025 older adults with an average 8-year follow-up — are significantly associated with all-cause mortality risk, and the findings replicated in an independent cohort of 1,124 people. Biological age, a composite metric in aging research that estimates “how old your body actually is” versus your chronological age, played a key role here. Higher ERBB1 and ANTXR2 levels lined up with lower mortality risk and younger biological age estimates, while higher GDF15, HE4, and IGFBP-2 pointed to higher mortality risk and older biological age. These are observational associations, though — causality isn’t confirmed, and biological age is itself an estimate that can’t directly stand in for individual health status. You can dig into the protein screening methods and cohort design over at Fight Aging! coverage .

2. Alzheimer’s Blood Test Could Flag Cognitive Decline About 10 Years Ahead

A blood test that can predict Alzheimer’s-related cognitive decline risk before symptoms even show up — that’s what a large international study uncovered, according to Science Daily coverage . The highest-risk group had about a 38% chance of cognitive impairment within five years and roughly 78% within ten. Heads up though: this conclusion comes from a short abstract, so the specific population tested, sample size, and exact biomarkers need verification from the full paper. And “predicted risk” is a population-level statistical probability — not a guarantee any individual will develop the disease, and it doesn’t mean we have interventions to stop it yet.

3. MTR4: The Vascular Endothelium’s “Guardian Protein” Declines With Age — Worm Experiments Show Lifespan Extension

MTR4, a protein that drops noticeably with aging in both human vascular cells and old mouse blood vessels, is shaping up to be a key guardian of vascular endothelial health. Vascular endothelial cell senescence — the aging of cells lining blood vessel walls — is considered an early driver of cardiovascular disease like atherosclerosis. According to Fight Aging! coverage , when researchers specifically overexpressed MTR4 in mouse vascular endothelium, it mitigated vascular functional decline. In short-lived nematodes, overexpressing mtr-4 “significantly extended lifespan, comparable to classic longevity models.” But let’s keep it real — worm results don’t automatically translate to humans, and the vascular endothelium covers a massive surface area, making gene delivery a serious engineering challenge.

4. Mitochondrial Fusion-Promoting Compounds Extend Lifespan in Nematode Worms

Mitochondrial fusion-promoting compounds — specifically Chimera and 8015-P2 — extended nematode lifespan in a dose-dependent manner, according to Fight Aging! coverage . Mitochondrial fusion, where multiple mitochondria merge into one within a cell, helps maintain overall mitochondrial function. 8015-P2 was the standout: it not only extended lifespan more effectively but also improved the worms’ swimming activity and stress resistance. Interestingly, 8015-P2 is already being developed as a candidate drug for human neurodegenerative diseases. But whether nematode results can replicate in mammals is still an open question, and the literature actually reports benefits from both fusion-promoting and fission-promoting experiments — the mechanism is far from settled.

5. H3K9me3 Epigenetic Mark: Inhibiting It Reduces Alzheimer’s Pathology in Mouse Models

H3K9me3, a histone mark that suppresses gene expression and rises with disease severity in post-mortem brain tissue and retinas of Alzheimer’s patients, became the target of a promising intervention. Epigenetic modifications are chemical marks that don’t change the DNA sequence but flip gene “on/off” switches. According to Fight Aging! coverage , researchers used the inhibitor ETP69 in two transgenic Alzheimer’s mouse models and saw reduced brain and retinal pathology, preserved synapses, and improved cognitive test performance. Still, this is animal model territory only — whether it works in human patients is the next big question.

6. Circular RNA Delivering Yamanaka Factors Achieves “Partial Reprogramming” Without Triggering Dedifferentiation in Cell Experiments

Circular RNA (circ-OSK) delivered via lipid nanoparticles is the star of this one — it expresses three Yamanaka factors (OCT4, SOX2, KLF4) in mesenchymal stem cells and pulls off “partial reprogramming” without pushing cells all the way back to a stem cell state. According to Fight Aging! coverage , digital PCR confirmed the treated cells did not convert into induced pluripotent stem cells. The core challenge of reprogramming therapy is making aging cells “rejuvenate” (the good part) without reverting them to stem cell status (potentially carcinogenic). After treatment, aging cells bounced back to near-early-passage levels across senescence markers, proliferative capacity, and differentiation ability. This is cell experiment stage stuff, though — in vivo efficacy is still waiting to be verified.

7. Senescent Cells Spread Aging Signals to Neighbors via “Small Extracellular Vesicles”

Senescent cells aren’t just secreting the usual inflammatory suspects (SASP) — they’re also pumping out “small extracellular vesicles” that can push healthy neighboring cells into senescence, according to Fight Aging! coverage . These tiny cell-secreted particles come from fibroblasts with high p16INK4A expression (a cell cycle arrest marker protein). The kicker: this propagation happens independently of classic inflammatory pathways like IL-6 and IL-8, pointing to a distinct mechanism. Cell experiment results showed vesicles from old donor cells had similar effects on young donor cells. But the in vivo environment is way more complex — targeting vesicles doesn’t automatically equal reversing aging.

8. AI-Assisted Radiology Reports: Currently FDA-Licensed Tools Cover Less Than a Third of Clinical Findings

FDA-licensed radiology AI tools can only recover about a third of “clinically actionable findings” — that’s the sobering takeaway from a study that built a radiology AI agent framework to simulate calling real FDA-licensed devices, according to a paper index/abstract . When researchers switched to an unrestricted tool catalog, coverage shot up dramatically. The most common reason for missing findings? No licensed detection tool exists for that imaging type, period. The researchers flagged that tool availability and effective task scheduling are two complementary bottlenecks — you can’t just throw a better orchestration model at the problem and call it solved. This is a preprint based on simulated tool outputs, not results from running actual devices.

9. One Extra Year of Education Linked to ~0.88 Percentage Point Drop in Dementia Cumulative Incidence — A Local Estimate From a UK Natural Experiment

One extra year of early-life education — that’s the lever this study zeroed in on, thanks to a UK policy change that raised the minimum school-leaving age from 15 to 16 in 1972. According to Fight Aging! coverage , researchers used birth dates as a natural grouping mechanism for “who got that extra year,” combined with administrative health records from UK Biobank participants. The original paper ’s preferred model, for those who left school at 18 or earlier, locally estimated that one additional year of education knocked down Alzheimer’s and related dementia cumulative incidence by about 0.88 percentage points. The effect held even in people with genetic risk, hinting that environmental intervention can partially offset genetic predisposition. That said, this estimate is fairly beefy relative to the 0.61% baseline incidence, with a 95% confidence interval of −1.763 to −0.154 percentage points. The study supports a protective direction in this policy context but can’t pin down the exact effect size — and it’s not meant for individual risk prediction.

10. AI Chatbots Miss One in Five Mental Health Emergencies in Multi-Turn Conversations

AI chatbots — 13 of the mainstream ones — missed about 20.3% of psychological emergencies when info came out gradually over 10 conversation rounds, compared to just 5.6% when everything was presented at once. According to a paper index/abstract , researchers drove these conversations with simulated users based on 112 structured mental health cases. They dubbed the gap “path dependence”: early judgments anchor subsequent reasoning, making it tough to correct course even when critical info surfaces later. Chatbots also had more rigid questioning trajectories than human clinicians and tended to loop on the same types of questions. Keep in mind this is all simulated conversations — the results don’t directly translate to real clinical settings.

🔬 Frontier Research

Hormone Replacement Therapy’s Relationship With Aging Clocks Is More Complex Than Expected

Hormone replacement therapy (HRT) — it’s associated with reduced all-cause mortality (HR 0.65) and cardiovascular mortality (HR 0.52), but the story gets messy from there. According to Fight Aging! coverage , a study using US NHANES data and Mendelian randomization found that in the subgroup with abnormal blood sugar, HRT’s protective effect was significantly weakened as epigenetic aging accelerated — an antagonistic interaction. Mendelian randomization uses genetic variants to infer causality, but the mortality analysis here is still observational. The takeaway: “who benefits from HRT” might split depending on blood sugar status and aging clock readings, and calibrating aging clocks against specific therapies is still wide open.

Cellular Senescence in Fatty Liver Disease: Mechanistically Plausible, But Pathogenic Importance Still Uncertain

Hepatocyte senescence in metabolic dysfunction-associated steatotic liver disease (MASLD, mostly driven by obesity) is a “plausible contributing factor” to fibrosis — but how big a player it really is remains up in the air. According to Fight Aging! coverage , a review synthesizing human, animal, and cell evidence mapped how senescent hepatocytes promote fibrosis through SASP inflammatory signals that affect immune cells and hepatic stellate cells. Its pathological importance depends on cell type, immune status, and how long senescence persists. Direct evidence that clearing senescent hepatocytes ameliorates damage in humans is still thin on the ground, and the lack of universal senescence markers isn’t helping research move forward.

💼 Industry Moves & Personal Impact

Sanofi Terminates Phase I Trial of STAT6 Degrader, Nurix Stock Plunges 43%

Sanofi pulled the plug on SAR448272 — an oral STAT6-targeting protein degrader co-developed with Nurix — halting it early in a Phase I human safety trial over “new preclinical findings.” According to GEN coverage , protein degraders are an emerging drug class that tags and destroys specific proteins rather than just blocking their function. The news hit Nurix hard: shares dropped about 43% over five days. But Nurix isn’t on the ropes — they’ve got enough cash to keep running through the second half of 2029, and their BTK degrader bexobrutideg, partnered with Roche, is currently in Phase III. The big picture: preclinical safety signals remain a wildcard in the protein degrader space, but a Phase I safety termination doesn’t doom the whole class.

📦 Open Source Top Project

Shape-Lab/SPHARM-Mamba: Analyzing Cortical Morphology and Predicting Brain Age Using Spherical Harmonics and Mamba Architecture

SPHARM-Mamba — that’s the tool that combines spherical harmonic representation with the Mamba sequence model for brain cortical surface morphological analysis and brain age prediction. According to the GitHub repository , the spherical harmonic side decomposes surface features into multi-scale frequency components, where spherical harmonic convolution preserves rotational equivariance and power aggregation within each harmonic order yields rotation-invariant descriptors. Brain age, the metric this tool predicts, estimates the degree of brain aging using imaging data. It’s a fit for computational neuroscience researchers studying how brain structure morphology relates to age. The repo currently sits at 1 star, with activity as of 2026-11-09. The mechanisms above have been verified against the README; license and runtime environment aren’t specified in the description.

🤔 Related Question

Can Blood Protein Levels Be Used to Predict an Individual’s Lifespan?

Short answer: nope, not for individual prediction. The study reported today found 17 proteins associated with all-cause mortality in a 1,025-person cohort, validated in an independent cohort — but that’s a population-level statistical association, and point-to-point prediction error is wide. “Biological age” is itself an estimated metric, so a single measurement can’t tell you whether an individual is aging faster than they should be. Check out the Fight Aging! original coverage for the study design and sources of error to get a better sense of where the boundaries lie.