Running on Empty: Why Refilling a Cell's NAD+ Rescued Old Male Mouse Hearts, and Did Nothing for Female Ones

RapAdmin · September 29, 2026, 8:01pm

A Weill Cornell team reports that aged male mouse hearts lose about 38 percent of their NAD+, a molecule that powers hundreds of metabolic reactions and fuels the sirtuin enzyme family. That loss silences two sirtuins, SIRT1 and SIRT6, allowing heart muscle cells to slip into a senescent, inflammatory state that stiffens the ventricle and impairs relaxation. Eight weeks of injected NAD+ precursors, either nicotinamide riboside or the more potent dihydronicotinamide riboside, reduced senescence markers, shrank enlarged heart cells, cut fibrosis, and improved a standard measure of diastolic function. The more potent compound worked better, and the authors argue this is because only a large NAD+ increase engages SIRT6 in addition to SIRT1. Aged female mice showed the same diastolic dysfunction but no NAD+ deficit and no response to either drug.

The heart does not fail in old age only because it pumps weakly. In most people over 65 who develop heart failure, the pumping stroke looks normal on an ultrasound. What has gone wrong is the refill. The ventricle has become stiff and cannot relax and fill properly between beats. This condition, heart failure with preserved ejection fraction, has almost no effective drug therapy, and the reason is that nobody has pinned down a single reversible mechanism driving it.

A group at Weill Cornell Medicine now proposes one. Their argument starts with NAD+, a molecule every cell uses as a metabolic go-between and which a family of enzymes called sirtuins consume as fuel. Measuring the hearts of 24-month-old male mice, roughly equivalent to humans in their seventies, the team found NAD+ had fallen by about 38 percent. The shortfall came from three directions at once. The machinery that builds NAD+ had been downregulated. The transporters that move it in and out of mitochondrial storage had declined. And the enzymes that destroy it, including CD38 and PARP1, had become more active.

The consequence was that sirtuins, which have a weak grip on NAD+ compared with those destructive enzymes, effectively ran out of fuel. Heart muscle cells then accumulated the classic hallmarks of cellular senescence: elevated p16 and p21, persistent DNA damage signals, and the secretion of inflammatory factors that make neighbouring cells swell and surrounding tissue scar. Those hearts were heavier, more fibrotic, and measurably worse at relaxing.

The team then tested whether this could be run in reverse. They gave aged male mice one of two NAD+ precursors by injection, three times a week for two months. Nicotinamide riboside, already sold as a supplement, produced a moderate NAD+ rise. Dihydronicotinamide riboside, a more potent compound discovered in the same laboratory, produced a larger one. Both reduced senescence markers, cell size and fibrosis. Both improved relaxation, though only the stronger compound reached statistical significance on the primary functional measure.

The most interesting claim is about why. The authors argue the two compounds are not simply different doses of the same thing. A moderate NAD+ rise wakes up SIRT1, which strips an activating chemical tag off the protein p53 and dials down the cell cycle brakes. A large rise additionally wakes up SIRT6, which works on chromatin and DNA repair. Crucially, SIRT6 needs SIRT1 to remove a tag from SIRT6 itself before it can engage. That makes sirtuin activation a threshold phenomenon rather than a smooth dial, which would explain a great deal of inconsistency in the NAD+ literature.

Then comes the complication. Aged female mice had the same stiff hearts, but their cardiac NAD+ was barely reduced, and neither compound helped them at all.

Actionable Insights

The magnitude first. In the treated old males, the relaxation measure improved by roughly ten units against a spread of about seven units between individual animals. In plain terms, the average treated mouse ended up better off than about nine in ten untreated mice. That sounds impressive, and on its face it is. But with only six to ten animals per group, the honest range around that estimate stretches from “barely detectable” to “enormous.” One or two animals moving would change the picture substantially.

Three practical points. The compound that actually worked, dihydronicotinamide riboside, is not sold, is not food-grade, has never been given to a human, and has shown cell-killing effects in at least one human liver cell line. Nicotinamide riboside, which is sold, missed significance on the main functional endpoint here, and the NIA Interventions Testing Program, cited in this paper’s own reference list, found it does not extend mouse lifespan in either sex.

Second, the drugs were injected, not swallowed, and the authors themselves say oral absorption is far lower.

Third, half the animals, the females, got no benefit whatsoever.

Context and Source

Open Access Paper: NAD+ Replenishment Reduces Cardiomyocyte Senescence and Improves Diastolic Function in the Aged Male Heart
Lead institution: Department of Pharmacology, Weill Cornell Medicine, New York, USA. Collaborating institutions: Changchun University of Chinese Medicine, China; North Carolina A&T State University, USA; Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, China.
Country: United States, with Chinese collaborating sites.
Journal and Impact Evaluation: bioRxiv is a preprint server, not a journal,

RapAdmin · September 29, 2026, 9:34pm

Dihydronicotinamide Riboside (NRH): Complete Evidence Review

Bottom Line Up Front

NRH is the most potent NAD+ precursor yet characterized, raising tissue NAD+ several-fold where nicotinamide riboside manages 1.5-fold. It is also the only NAD+ precursor with a documented, reproducible toxicity ceiling in mice, and the gap between the effective dose and the toxic dose is roughly three to four fold. There are zero human data of any kind: no pharmacokinetics, no safety study, no efficacy trial, no registered clinical trial, and no FDA new dietary ingredient notification. The entire evidence base is seven years old, comprises roughly a dozen primary papers, and rests on two laboratories plus a handful of collaborators.

Pharmacokinetics

This is the least developed part of the evidence base, resting essentially on two studies.

Absorption and circulating profile

Single 250 mg/kg intraperitoneal dose in young male C57BL/6J mice: blood levels of NRH peaked at 5 minutes with a Cmax of 65.7 micromolar, disappeared by 60 minutes, and showed a half-life of 8.2 minutes. Pre-treatment with ABT702, an adenosine kinase inhibitor, prolonged NRH serum presence with a Cmax of 111.6 micromolar. ScienceDirect

An independent report using the same dose found peak circulating concentrations at 15 minutes, declining steeply to near-undetectable by 4 hours. [Confidence: High that the compound clears within an hour, Medium on the precise half-life given single-study status]

An 8.2 minute half-life is extremely short. For comparison, most oral supplements aim for hours. The implication is that NRH does not produce a sustained plasma exposure; it produces a spike.

Oral bioavailability and biodistribution

This is the strongest PK dataset available, using a dual-labeled tracer.

Two hours after an oral dose of 250 mg/kg of isotopically labelled M+7 NRH (labelled on both the nicotinamide moiety with 18O and the ribose with 13C5), NRH was distributed in all tissues and body fluids analysed, including plasma, urine, liver, lung, pancreas, kidney, heart, muscle and brain. High relative values of NRH were found in urine, plasma and kidney; brain was the tissue with the poorest distribution, as expected due to the blood-brain barrier. Nature

Crucially, labelling showed two parallel routes. NRH led to labelling of liver NAD+ both through direct label incorporation (M+7) and via nicotinamide (M+2). So even orally, a portion of NRH is degraded to nicotinamide and recycled conventionally, exactly like NR. Intact delivery is real but partial. Nature

The comparison to NR is the point of the experiment. The detection of NRH in plasma and tissues contrasts with other NAD+ boosters such as NR and NMN, which can hardly be detected after oral administration, and Liu et al. failed to detect isotopically labelled NR in circulation up to 135 minutes after oral administration. Nature Nature

NRH is orally bioavailable and not degraded in plasma, which distinguishes it from NR, whose circulating levels are quickly curtailed by degradation to nicotinamide. ScienceDirect

What has never been measured: an actual bioavailability fraction. No study reports F, no study has done matched intravenous versus oral dosing, and no AUC comparison exists. “Orally bioavailable” in this literature means “detectable after oral dosing,” not a quantified percentage. [Confidence: High that F has not been determined]

Elimination

High relative values of NRH were found in urine, indicating substantial renal excretion of intact compound. No formal mass balance, clearance, or volume of distribution has been published. No metabolite profiling of excreted species exists beyond the pyridone chemistry described above. Nature

PK summary table

Parameter	Value	Species, route	Status
Cmax	65.7 uM	mouse, 250 mg/kg IP	single study
Tmax	5 to 15 min	mouse, IP	two reports
Half-life	8.2 min	mouse, IP	single study
Detectable duration	under 60 min	mouse, IP	consistent
Oral absorption intact	yes, all tissues at 2 h	mouse, 250 mg/kg gavage	single study, tracer
Oral bioavailability (F)	never measured	-	absent
Plasma stability	stable, unlike NR	mouse	replicated
BBB penetration	yes, lowest of all tissues	mouse	single study
Primary excretion	urinary, intact	mouse	single study
Clearance, Vd, AUC	never reported	-	absent
Human PK	none	-	absent

6. Pharmacodynamics

The PD profile contains the field’s most important and least discussed finding: acute and chronic dosing produce opposite effects on NAD+ redox state, and chronic dosing does not raise steady-state NAD+ at all.

Disease Model Evidence

Every entry below is rodent or cell culture. No human data exist for any indication.

Cisplatin acute kidney injury. NRH prevents cisplatin-induced acute kidney injury in mice, with measurement of kidney NAD+ metabolites and poly-ADP-ribosylation and PARP1 levels, consistent with cisplatin-triggered PARP activity depleting NAD+ and NRH restoring it. [Confidence: Medium, single study] Amsterdam UMC ResearchGate

Diet-induced obesity, prevention. Mice supplemented with NRH at 100 mg/(kg*day) in drinking water showed robust protection against diet-induced body weight gain, attributable to decreased fat depot weight, along with protection against high-fat-diet-induced increases in fasting glycemia and insulinemia, better glucose clearance, and prevention of increases in circulating ALT, creatinine and LDL-cholesterol. Across liver, brown adipose, and both white adipose depots, NRH reduced the number of high-fat-diet-induced differentially expressed genes by 90% or more. [Confidence: Medium to High, well powered, open access, independent of the discovering lab] Nature Nature

Toxicity and the Therapeutic Window

This is where NRH separates from every other NAD+ precursor, and it deserves the most attention.

Cell culture signals, 2020 onward

Hepatocellular carcinoma HepG3 cells show dose-dependent cytotoxicity when supplemented with 100 to 1000 micromolar NRH. At 100 micromolar, a significant increase in ROS was observed concurrent with changes in the NAD(P)H and GSH/GSSG pools. NRH altered mitochondrial membrane potential, increased mitochondrial superoxide formation, and induced mitochondrial DNA damage in those cells, and also caused metabolic dysregulation, altering mitochondrial respiration. The mechanism was PUMA and BAX-mediated apoptosis. HEK293T cells were unaffected at 100 micromolar while HepG3 showed significant loss of viability, establishing that the toxicity is cell-type specific rather than universal. Dihydronicotinamide riboside promotes cell-specific cytotoxicity by tipping the balance between metabolic regulation and oxidative stress | PLOS One +2

The definitive in vivo dose-response, 2026

The Nature Communications study ran 16 weeks of drinking-water dosing plus a two-week dose-ranging arm. The results:

At 100 mg/kg/day for 16 weeks: no alterations in basic blood biochemistry markers, including markers of liver damage (ALT, AST, alkaline phosphatase), bone, kidney (creatinine), or muscle (creatine kinase). Triglyceride and cholesterol levels were also comparable. Body weight, composition, food and water intake, activity, treadmill performance, energy expenditure and glucose tolerance were all unchanged. Nature

At 400 mg/kg/day for 16 weeks: male mice showed higher daily spontaneous activity during the light phase, driven by spikes of activity occurring every 3 to 4 hours, which were not observed at 100 mg/(kg*day). These measures were carried out in two independent batches of mice, all showing this abnormal behavioural pattern. The altered behaviour in males was also observed in the treadmill test, revealing lower performance that did not necessarily stem from fatigue per se, but from a failure to properly engage into the running test. High NRH doses led to increased circulating ALT and creatinine levels both in males and females, which may be reflective of hepatic and renal damage, as well as decreased levels of alkaline phosphatase. Nature

Mechanism of the toxicity: immunofluorescence analyses revealed a significant increase in the number of gamma-H2AX positive nuclei, but only in male mice treated with 400 mg/(kg*day), and poly-ADP-ribosylation levels were notably higher in liver homogenates from those mice. Nature

Dose-ranging at two weeks (0, 100, 200, 300, 400): a dose-dependent decrease in circulating alkaline phosphatase was already evident, reaching significance at 300 mg/(kgday). Transcriptomic analyses revealed significant gene expression changes across all tested doses, including at 100 mg/(kgday), a dose that showed no effect in the chronic supplementation study. Circadian rhythms emerged as the most significantly altered category at the toxic dose, with other enriched terms including cell cycle regulation, radiation response and glucose homeostasis. Nature

Causal attribution in cells: In AML12 hepatocytes, DNA damage occurred only at the higher dose of 0.5 mM, and 24-hour exposure to NRH 0.5 mM impaired both respiration and glycolytic rates, suggesting treated cells suffered a metabolic breakdown. These effects were reversed by inhibiting adenosine kinase with 5-iodotubercidin, indicating that toxicity arises from excessive NAD(H) synthesis rather than NRH itself. Nature

The authors’ own verdict: signs of toxicity and altered behaviour were detected when mice were dosed with NRH concentrations above 300 mg/(kgday). This suggests that the tolerance window of NRH is more limited than that of other NAD+ precursors, such as the closely related NR, where no adverse effects are observed at doses below 1 g/(kgday). Nature

Therapeutic window, expressed in human-equivalent terms

Using standard body surface area allometric scaling (mouse Km 3, human Km 37, conversion factor 0.081), for a 70 kg adult:

Mouse dose	Finding in mice	Human equivalent
100 mg/kg/day	clean at 16 weeks	about 570 mg/day
200 mg/kg/day	transcriptomic change only	about 1.13 g/day
300 mg/kg/day	ALP falls significantly	about 1.70 g/day
400 mg/kg/day	ALT and creatinine up, liver DNA damage, behavioural disruption	about 2.27 g/day
1000 mg/kg single	tolerated acutely	about 5.7 g single dose

This is an allometric estimate, not a pharmacokinetic bridge, and route differences are not captured. But the shape of the result is what matters: the margin between the clean dose and the first biochemical toxicity signal is about three-fold. People routinely take 1 to 2 g/day of NR, a dose range that in the NRH scaling falls between the first toxicity signal and frank hepatic and renal injury. [Confidence: Medium on the absolute numbers, High on the conclusion that the window is narrow]

Clinical Data

There are none.

No registered interventional trial of NRH or dihydronicotinamide riboside on ClinicalTrials.gov. Every NAD-precursor trial returned in searching is nicotinamide riboside, nicotinamide mononucleotide, or nicotinamide.
No GRAS notification, no New Dietary Ingredient notification, no FDA correspondence for dihydronicotinamide riboside.
No human pharmacokinetic study, no single ascending dose study, no repeat-dose tolerability study.
No genotoxicity battery, no carcinogenicity study, no reproductive toxicology, no second-species repeat-dose toxicology.

A labelling warning worth flagging for your audience. An FDA New Dietary Ingredient notification exists for something abbreviated “NRHM,” but NRHM is nicotinamide riboside combined with L-malic acid, where L-malic acid is affirmed as GRAS as a direct food substance. That is nicotinamide riboside hydrogen malate, an entirely different compound from dihydronicotinamide riboside. Anything marketed with “NRH” or “NR-H” on the label is far more likely to be the malate salt of ordinary NR than the reduced dihydro form. [Confidence: High on the identity of NRHM, Medium on the prevalence of the labelling confusion in the market] Regulations