# NAD⁺ in aging: compartments, CD38, PARP1, SARM1 and intervention limits

- Canonical: https://nutrifit.health/knowledge/nad-depletion-cd38-parp1-dna-repair
- Language: en
- Published: 2026-10-08T05:14:10.892Z
- Updated: 2026-10-08T05:14:10.892Z
- Author: NutriFit Editorial

> NAD⁺ synthesis and consumption across cellular compartments, the structural and kinetic logic of CD38/PARP1/SARM1, and critical reading of human NR and NMN studies.

NAD⁺ concentration reflects two simultaneous fluxes: cofactor formation from vitamin B3 and tryptophan-processing products, and consumption in signaling reactions. A biopsy concentration is a snapshot of that balance, not a direct measure of the “rate of aging.” CD38, PARP1 and SARM1 consume NAD⁺ in different compartments for different reasons; treating them as one harmful “drain” ignores the necessity of PARP-dependent repair and calcium signaling. Their catalytic reactions, topology and limits of animal-to-human translation are examined below.

## 1. Pool chemistry: redox cofactor and consumed substrate

### 1.1. Electron turnover is not molecular loss

The oxidized pyridine ring accepts a hydride in dehydrogenase reactions: RH₂ + NAD⁺ ⇄ R + NADH + H⁺. NADH reoxidation returns NAD⁺, so this cycle need not reduce the total dinucleotide pool. Cytosolic glyceraldehyde-3-phosphate dehydrogenase, mitochondrial pyruvate and α-ketoglutarate dehydrogenase complexes, IDH3 and MDH2 connect different substrates to the same NAD⁺/NADH pair. Complex I accepts NADH electrons and transfers them through FMN and Fe–S centers to ubiquinone; coupled conformational changes generate a proton gradient. NADH does not directly reduce only NDUFV1, and one turnover does not invariably take milliseconds under all conditions.

Other reactions cleave the NAD⁺ N-glycosidic bond and release nicotinamide (NAM). Hydrolytic CD38 produces ADP-ribose; PARP transfers ADP-ribose to an acceptor and can build chains; sirtuins use NAD⁺ during protein deacylation and generate O-acyl-ADPR. Writing them all as NAD⁺ + H₂O → NAM + ADPR is chemically inaccurate. Consumption requires resynthesis, but consumption during DNA repair is not itself disease. NAM can inhibit individual NAD⁺-dependent enzymes; its intracellular concentration reflects formation, reutilization and elimination.

### 1.2. Free and bound fractions in cytosol and matrix

Total extracted NAD(H), the free NAD⁺/NADH ratio and a method's spatial/temporal resolution must be distinguished. Free cytosolic ratios are usually much higher than matrix ratios: lactate dehydrogenase and other equilibria maintain an oxidized cytosol, while substrate oxidation supplies matrix reducing equivalents. Ratios such as 700:1 and 8:1 illustrate specific systems, not norms for every cell and age. The Nernst potential depends on activity ratios, pH and temperature; standard E°′ is not actual matrix potential. Inadequate oxidized cofactor can slow GAPDH, but glycolytic direction also depends on glyceraldehyde-3-phosphate, 1,3-bisphosphoglycerate, phosphate and coupled reactions. One total NAD⁺ measurement does not establish methylglyoxal accumulation.

The inner mitochondrial membrane restricts free NAD(H) diffusion, but SLC25A51 transports NAD⁺ into the matrix, with contributions depending on tissue and conditions. Cytosolic NADH transfers reducing equivalents through the malate–aspartate shuttle without itself crossing. MDH1 reduces oxaloacetate to malate using NADH; SLC25A11 exchanges malate for α-ketoglutarate; matrix MDH2 oxidizes malate and generates NADH; GOT2, cytosolic GOT1 and aspartate/glutamate carriers close the carbon–nitrogen cycle. The glycerol-3-phosphate shuttle is an alternative electron-transfer route. Nuclear and cytosolic pools exchange small metabolites through pores, yet local NMNAT1 and consumers create functional heterogeneity. [Melatonin and mitochondrial pathways](/knowledge/melatonin-mitochondria-circadian-receptors) illustrate why matrix and membranes must be distinguished rather than summarized as “mitochondrial protection.”

### 1.3. De novo, Preiss–Handler and salvage: three entries with shared final chemistry

The *de novo* route takes L-tryptophan through N-formylkynurenine, kynurenine, 3-hydroxykynurenine, 3-hydroxyanthranilate and 2-amino-3-carboxymuconate semialdehyde (ACMS) to quinolinate, involving TDO/IDO, AFMID, KMO, KYNU and HAAO. ACMSD diverts ACMS from spontaneous quinolinate cyclization; QPRT converts quinolinate + PRPP → NaMN + PPᵢ. NMNAT adenylates NaMN + ATP → NaAD + PPᵢ; glutamine-dependent NADSYN1 converts NaAD + glutamine + ATP + H₂O → NAD⁺ + glutamate + AMP + PPᵢ. Inflammation can induce IDO1 and alter tryptophan/kynurenine distribution, but does not imply universal “QPRT blockade” and neurotoxic quinolinate accumulation. “60 mg tryptophan per 1 mg niacin equivalent” is a conditional dietary mass factor, not 60-molecule-to-one-NAD⁺ stoichiometry.

Preiss–Handler begins with nicotinic acid: NA + PRPP —NAPRT→ NaMN + PPᵢ; subsequent NMNAT/NADSYN1 steps are shared with de novo. Salvage uses NAM + PRPP —NAMPT→ NMN + PPᵢ and NMN + ATP —NMNAT→ NAD⁺ + PPᵢ. NAMPT depends on PRPP, ATP and transcriptional regulation; “rate-limiting” depends on tissue and regime. NR enters through NRK1/2: NR + ATP → NMN + ADP. Much dietary precursor undergoes intestinal/blood transformation before intracellular NAD⁺, so a capsule formula does not prove direct NMN delivery to every cell. NMNAT1 is mainly nuclear, NMNAT2 cytosolic/vesicular and essential to axons; NMNAT3 distribution and matrix roles differ by species/tissue. SLC25A51 also imports NAD⁺ into matrix. No salvage percentage describes universal daily turnover across tissues. Experimental CLOCK:BMAL1/SIRT1–NAMPT relationships do not determine the best human dosing hour. Vitamin B3 forms and dose effects are discussed in [niacin and lipid metabolism](/knowledge/vitamin-b3-niacin-lipids-safety).

### 1.4. Tissue aging measurements: method matters more than one summary number

Age-related declines in liver, muscle, fat, brain and kidney are heterogeneous; no unified human cross-organ series has fixed percentages. Cell composition, collection time, freezing, extraction and whether NAD⁺, NADH or their sum is measured affect results; cells within a tissue have different pools. Human brain ³¹P-MRS measures tissue signals *in vivo* but has limited spatial resolution and spectral separation; it does not distinguish neuronal cytosol from microglia. Skin biopsies also contain variable cellular/extracellular fractions. An age association does not establish that NAD⁺ decline primarily causes sarcopenia, steatosis or cognitive decline: disease, inflammation and composition may cause the measured change. Tissue comparisons require comparable methods and cell-composition control.

### 1.5. Nuclear–mitochondrial communication and “pseudohypoxia”

Gomes and colleagues connected mouse NAD⁺ decline with weaker SIRT1-dependent HIF-1α control, altered c-MYC/TFAM signaling and imbalanced nuclear/mitochondrial respiratory-protein expression. Normoxic HIF-1α activity was termed “pseudohypoxia,” but obligatory respiratory collapse in every aging human was not demonstrated. HIF-1α is not regulated only by deacetylation: prolyl hydroxylases use O₂, Fe²⁺ and 2-oxoglutarate; pVHL recognizes hydroxylated protein and targets degradation. Deacetylation can change transcriptional activity, but protein accumulation and promoter action are different levels. TFAM packages/transcribes mtDNA; the genome encodes 13 respiratory polypeptides, yet “all 13 cease transcription” is excessive. Entirely nuclear-encoded complex II allows imbalance without loss of all OXPHOS. Clinical inference requires compartment pools, HIF-1α, transcripts and respiration measured in one human tissue, not a mouse mechanistic arrow.
## 2. CD38 and CD157: catalytic topology matters more than a slogan about “destroying NAD⁺”

### 2.1. CD38 architecture and its active site

The usual form of CD38 is a type II membrane protein: its short N-terminus faces the cytosol, and a single transmembrane helix holds a large glycosylated catalytic domain outside the cell. Crystal structures of the ectodomain reveal the ADP-ribosyl cyclase family fold, disulfide bridges and a pocket positioning NAD⁺ for cleavage of the nicotinamide–ribose bond. Glu226 is critical for catalysis; nearby Glu146, Trp189 and other residues affect binding and product selection. An active-site mutation and an enzyme structure support discussion of the mechanism, but a detailed claim about a stable covalent “Glu226–ribose” intermediate, or an obligatory role for every listed residue, requires direct structural and kinetic evidence. Water access, pH and the position of the nicotinamide moiety alter the branching between hydrolysis and cyclization. The full protein also acts as a cell-surface marker and participates in cellular contacts, so a CD38-knockout phenotype need not be explained solely by the NADase reaction.

### 2.2. Hydrolysis, cyclization and base exchange

In the hydrolase reaction, NAD⁺ + H₂O —CD38→ NAM + ADPR; an alternative intramolecular attack yields NAD⁺ —CD38→ NAM + cyclic ADPR (cADPR). Hydrolysis strongly predominates for human CD38 under several *in vitro* conditions, but the product ratio varies with substrate, pH and measurement system. “Exactly 99% in every tissue” or “one cADPR always costs one hundred NAD⁺ molecules” is therefore inaccurate. CD38 also catalyzes base exchange: NADP⁺ + nicotinic acid → NAADP + NAM, particularly in an acidic environment. cADPR is associated with the regulation of Ca²⁺ release from intracellular stores, whereas NAADP triggers calcium responses through lysosomal and endosomal systems involving two-pore channels, TPCs, and accessory proteins. Calling NAADP a direct “TPC agonist” without acknowledging these intermediaries oversimplifies the pathway. CD38 products are not waste: at suitable concentrations they act as signals, and complete suppression of the enzyme can alter immune function.

### 2.3. CD157/BST1: a paralog with a different anchor

CD157, or BST1, belongs to the same family but is held on the outer surface of the plasma membrane by a glycosylphosphatidylinositol anchor rather than a type II transmembrane helix. Its catalytic properties and cellular interactions differ from those of CD38; measured BST1 expression therefore does not equal a proportional loss of intracellular NAD⁺. CD157 occurs in some immune and stromal populations and may participate in contacts, migration and local calcium signaling. A single kcat across different conditions, or a causal chain from CD157 in a Paneth cell to restoration of an Lgr5⁺ crypt, should not be accepted as an established explanation for age-related NAD⁺ depletion without a direct experiment. Functional experiments must distinguish the protein’s catalytic and adhesive roles.

### 2.4. An ectoenzyme and an intracellular pool

The catalytic domain of most CD38 molecules faces the extracellular space, while most measured NAD⁺ is inside cells. A direct scheme in which “surface CD38 consumed all nuclear NAD⁺” violates this topology. Possible bridges across the boundary include CD38 processing extracellular NAD⁺ and NMN to ADPR or ribose-5-phosphate and NAM; cells releasing NAD⁺ through regulated channels or following injury; reverse CD38 orientation, with the catalytic domain facing inward, described in particular experimental systems; and CD38 on intracellular membranes with different substrate access. Each mechanism requires evidence for protein localization, membrane integrity and substrate flux. A fixed proportion of reverse-oriented CD38, continuous NAD⁺ release through Cx43/Panx1, and delivery of internalized CD38 into the nucleoplasm all require separate measurements. NR cannot be equated with NMN as a precursor identically cleaved by CD38: substrate specificity matters. Extracellular NAD⁺ and NMN concentrations change locally and rapidly, so comparison with one cytosolic number does not calculate tissue flux.

[![Scientific diagram separates extracellular NAD⁺ and CD38, cytosolic NAMPT/NMNAT synthesis, nuclear PARP1, mitochondrial SLC25A51/SIRT3 and local axonal NMNAT2/SARM1; flux rates are unspecified.](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-nad-compartment-flux.en.png)](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-nad-compartment-flux.en.png)

### 2.5. Macrophages, senescent cells and causality

In the mouse experiments of Camacho-Pereira and colleagues, genetic absence of Cd38 was accompanied by preservation of NAD⁺ and changes in mitochondrial measures during aging. Later work by Chini and Covarrubias linked age-associated increases in CD38 activity to myeloid cells and the influence of the secretory environment of senescent cells. In some tissues, CD38⁺ macrophages become substantial NAD⁺ consumers; this does not mean that all CD38 is confined to macrophages or that each such cell is “M1.” The senescence-associated secretory phenotype, SASP, is heterogeneous and includes IL-6, IL-1β and other molecules depending on stimulus and tissue; IFN-γ should not automatically be attributed to every senescent cell. CD38 induction can involve inflammatory transcriptional networks, but obligatory CpG demethylation as a universal cause of a several-fold CD38 increase has not been demonstrated. A mouse knockout also alters immune-system development, so the contribution of catalytic consumption to the phenotype must be isolated. The published [analysis of apigenin and CD38](/knowledge/apigenin-gaba-cd38-nad-sleep) discusses the distinction between biochemical inhibition and an effect in humans.
## 3. PARP1: the cost of preserving the genome

### 3.1. Damage-sensor domains

Nuclear PARP1 combines three N-terminal zinc-binding modules, nuclear-localization and automodification regions with a BRCT domain, WGR and a C-terminal catalytic domain. Zn1/Zn2 recognize break structure; WGR connects DNA binding to catalysis, while the helical HD subdomain limits substrate access before activation. Catalysis transfers ADP-ribose from NAD⁺ and releases NAM. This explains sharply increased local nuclear consumption after DNA damage. PARP2 has different recognition architecture and participates in repair; PARP1's activity fraction depends on the stress type and cellular assay, rather than always being 90%.

### 3.2. HPF1 changes the amino-acid acceptor

DNA-break binding brings Zn modules, WGR and HD together, relieving active-site inhibition. HPF1 completes PARP1/2 catalysis and promotes serine ADP-ribosylation of histones and other proteins; HPF1 Glu284 is important. Initiation: protein–Ser–OH + NAD⁺ → protein–Ser–ADPR + NAM; elongation: protein–ADPRₙ + NAD⁺ → protein–ADPRₙ₊₁ + NAM. Some chains branch, with sizes/fractions depending on the enzyme and reaction conditions. PARG cleaves polymer ribose–ribose links, while ARH3 removes the final serine-linked ADP-ribose; “PARG erases every trace” is inaccurate. PAR recruits damage-response proteins and changes chromatin interactions, not merely electrostatic DNA repulsion. XRCC1, POLβ and LIG3 participate in single-strand/base-excision repair in appropriate contexts.

### 3.3. Damage, aging and excessive consumption

8-oxo-dG, abasic sites, single-strand breaks and dysfunctional telomeres activate different response nodes. An oxidized base does not imply direct PARP1 sensing: base-excision repair first processes it and creates intermediate breaks. Telomeric foci likewise are not identical to free-DNA-end flux. Intense genotoxic stress can rapidly consume NAD⁺ through PARP1; PARG returns polymer to monomer products but does not automatically regenerate NAD⁺. Depletion depends on damage, NAMPT/NMNAT, available ATP and time. Universal “80–90% within hours during normal aging” extrapolates acute models. Inhibiting PARP1 in ordinary aging without demonstrated overactivity is risky because it preserves the genome.

### 3.4. Parthanatos: a distinct severe-stress outcome

Excessive PARP1 activation can combine free PAR chains and marked NAD⁺ consumption with energy failure. In the described experimental branch PAR promotes mitochondrial AIF release; AIF interacts with MIF, whose nuclease activity contributes to large-scale chromatin fragmentation. This differs from classical caspase apoptosis and is termed parthanatos. Before release, AIF is functionally associated with the inner mitochondrial membrane, not solely the outer membrane. The contributions of calpains and MIF and fragment sizes depend on model; severe damage may activate other death routes simultaneously. Low NAD⁺ alone does not diagnose parthanatos: PAR dependence, AIF translocation, MIF and exclusion of alternatives are required. NMNAT uses ATP to resynthesize NAD⁺ from NMN, but every cycle does not automatically irreversibly burn all cellular ATP. Cancer PARP inhibitors involve DNA trapping and synthetic lethality in homologous-repair defects, not simply NAD⁺ conservation.

## 4. SARM1 and the local fate of an injured axon

### 4.1. ARM, SAM and TIR in one octamer

SARM1 contains an N-terminal regulatory ARM, paired oligomerization SAM domains and a C-terminal NADase-active TIR. Structures show a ring-shaped octamer; basal interdomain contacts restrain productive TIR interaction. One closed crystal structure does not capture every living-axon state: nucleotides, partners and damage change conformation. SARM1 catalysis is important to the *local* post-axotomy response, not every aging-associated NAD⁺ decline in liver or blood.

### 4.2. NMNAT2 and the regulatory NMN/NAD⁺ ratio

Short-lived NMNAT2 is transported into axons and converts NMN + ATP → NAD⁺ + PPᵢ. Interrupted delivery and breakdown after injury change local NMN/NAD⁺. Competitive NMN/NAD⁺ binding to the ARM pocket affects autoinhibition; rising NMN and falling NAD⁺ favor activation. Multiple regulators affect NMNAT2 ubiquitination and transport, so one obligatory Phr1/Highwire cascade or fixed half-life does not describe every axon. The diagram needs time-resolved local ratios, not culture thresholds transferred to every human nerve.

### 4.3. NAD⁺ hydrolysis, calcium and cytoskeleton

Active SARM1 TIR consumes NAD⁺ to NAM and ADPR, with cyclic products also possible. Energy deficits and metabolites can alter calcium signaling; activated calpains cleave spectrin, neurofilaments and other cytoskeletal proteins. Blocking SARM1 often preserves the morphology of a severed axon's *distal segment*, but does not establish restored conduction, regeneration of the neuronal cell body, or a universal treatment for ALS, Alzheimer’s disease or glaucoma. Depletion speed after one axotomy is not a universal “20–40 minutes to zero” in neuropathy. Wallerian degeneration is a multistep local-injury outcome with an important but contextual SARM1 axis.
## 5. Consumer kinetics: why one Km table does not establish a hierarchy

### 5.1. What Km tells us, and what it does not

For an isolated enzyme and a simplified reaction, v = Vmax·[S]/(Km + [S]) describes the dependence of rate on available substrate. Km is the concentration at half Vmax under specified conditions; it is not identical to equilibrium binding affinity, Kd, especially with a complex mechanism and coupled reactions. Comparing the Km of CD38 on the outer surface, PARP1 in the nucleus, SIRT3 in the matrix and SARM1 in the axon as if they occupied one vessel is inappropriate. Even within one compartment, rate depends on active-enzyme concentration, protein substrate, the inhibitor NAM, pH, post-translational modifications and access to free NAD⁺. For SIRT6, the nature of the acyl group and activators is particularly important; one “SIRT6 Km value” does not characterize every reaction. Precise Km ranges obtained for different enzymes under incomparable conditions cannot establish a general ranking of physiological NAD⁺ consumption.

### 5.2. Mass balance instead of invented competition

Within a local compartment, NAD⁺ dynamics can be written conceptually as d[NAD⁺]/dt = J_synthesis + J_in − J_CD38 − J_PARP − J_SIRT − J_SARM1 − J_other, with NAD⁺⇄NADH redox conversion and exchange between regions added to the description. Each flux must be measured or justified; its magnitude cannot be inferred from a Km ratio alone. A mathematical example using 400 and 120 µM free NAD⁺ and arbitrarily chosen Km values of 20 or 250 µM illustrates the curve’s shape, but is not a measured transition from a young cell to an old one. As substrate falls, the rate of any enzyme with unchanged Vmax decreases; CD38 or PARP1 can nevertheless increase *total* flux only if their active abundance or stimulation changes. They do not automatically “switch off” sirtuins: consumption must be shown to exceed salvage compensation within the same accessible pool. Nuclear PARP1 and matrix SIRT3 are connected through transport and metabolism, making their competition indirect rather than a direct contest for one molecule.

[![Local NAD⁺ balance: synthesis and transport replenish the accessible pool, while CD38, PARP, sirtuins and axonal SARM1 use NAD⁺ in their respective compartments; the Michaelis–Menten equation does not rank overall consumption.](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-nad-balance-kinetics.en.png)](https://api.nutrifit.health/api/v2/public/knowledge/files/images/nutrifit-scientific-20261002-nad-balance-kinetics.en.png)

### 5.3. Sirtuin substrates and ambiguity in the causal arrow

SIRT1-mediated deacetylation of PGC-1α, p53 and RelA/p65, SIRT3-dependent changes in SOD2/IDH2, and SIRT6-dependent maintenance of chromatin have been studied in different models. PGC-1α coordinates mitochondrial biogenesis through several factors, but its acetylation also depends on GCN5, nutritional signaling and tissue activity; loss of deacetylation does not mean a “complete halt to renewal” of organelles. For SOD2, specific residues and enzyme activity matter; for IDH2, NADPH production matters. A universal loss of 80% of activity should not be assumed. Deacetylation of p53 changes the stress response, so arbitrary enhancement of SIRT1 need not be beneficial during DNA damage either. The effect of RelA Lys310 acetylation on transcription does not prove that every cytokine is caused by NAD⁺ shortage. SIRT6 is associated with H3K9/H3K56, LINE-1 repression and genome stability; subsequent cGAS–STING involvement requires separate evidence in the particular cell. These pathways can interact, but “less NAD⁺ → all sirtuins switched off → every feature of aging” does not follow from an individual biomarker.
## 6. Interventions: from an enzyme target to an outcome in humans

### 6.1. CD38: flavonoids, 78c and the limits of specificity

Apigenin, quercetin and luteolin inhibit CD38 NADase activity in biochemical systems. For apigenin, Escande and colleagues demonstrated CD38 inhibition and changes in NAD⁺-dependent metabolism in cells and mice; the IC₅₀ depends on substrate, enzyme construct and assay conditions. Micromolar activity of a purified compound therefore cannot be equated with an achievable free tissue concentration after oral intake. Glucuronidation, sulfation, protein binding and metabolite distribution complicate that translation. A claim that each flavonoid directly contacts Glu226 requires structural or mutational confirmation for that particular molecule. Quercetin and luteolin also have other targets, and quercetin’s proposed senolytic activity depends on the cellular context and combination with other agents. The [apigenin review](/knowledge/apigenin-gaba-cd38-nad-sleep) distinguishes its biochemical mechanism, GABA signaling and human evidence in more detail.

78c is an investigational small-molecule CD38 inhibitor. In the study by Tarragó and colleagues, administration to old mice increased tissue NAD⁺ and changed some metabolic and functional measures. Nanomolar IC₅₀ values characterize an enzyme-assay system, not a dose, efficacy or safety in humans. Selectivity should be tested against CD157, other NAD⁺ consumers and tissue concentrations within the same experimental setting; an absolute claim that it “affects no other target” is not justified. No clinical trial of 78c establishing prevention of sarcopenia or restoration of human mitochondrial function is provided here. An anti-CD38 antibody used in oncology has immune and cellular consequences beyond catalytic inhibition; it cannot be considered equivalent to a dietary flavonoid.

### 6.2. PARP: genome protection and the cost of inhibition

Olaparib and other PARP inhibitors are used for specific oncological indications. Their effect in homologous-recombination deficiency includes catalytic inhibition and, to a degree varying between molecules, trapping PARP on damaged DNA. That benefit–harm balance is fundamentally different from a hypothetical reduction in NAD⁺ consumption in a healthy person. Myelosuppression and other toxic effects, drug interactions and the tumor-genome context make transfer of cancer therapy to aging prevention unjustified. Yet the reverse absolute assertion that “any brief PARP reduction immediately increases mutation numbers” also exceeds the evidence: the outcome depends on damage, PARP1/2, dose and other repair pathways.

Nicotinamide is a product of the PARP reaction and can inhibit NAD⁺-consuming enzymes *in vitro*, while also serving as a NAMPT substrate for salvage. A NAM dose therefore cannot be derived from a single IC₅₀ as a means of “gently adjusting PARP”: both opposing fluxes change at different tissue concentrations. Curcumin, resveratrol and other polyphenols have been studied in oxidative-stress models; a reduction in the marker 8-oxo-dG does not prove that pathological PARP hyperactivation is suppressed in humans without impairing repair. Preventive use would require direct measures of DNA damage and repair, pharmacodynamics and clinical outcomes. Dietary vitamin B₃ forms and high-dose nicotinamide pharmacology are distinguished in the [article on niacin](/knowledge/vitamin-b3-niacin-lipids-safety).

### 6.3. NR and NMN in humans: a biomarker is not geroprotection

The table separates study design, the measured pool and the limit of inference. An increase in blood NAD⁺ cannot be translated into free NAD⁺ concentration in the nucleus, muscle matrix or axon without additional measurement.

| Primary study | Population and intervention | What was found | What was not established |
| :--- | :--- | :--- | :--- |
| Trammell and colleagues, 2016 | Small pharmacokinetic series in healthy people, single NR doses | Appearance of NR-dependent metabolites and an increase in the measured blood NAD⁺ metabolome; the NAM/NAAD pathway and sampling time matter | Long-term benefit, tissue restoration or reduced disease risk |
| Martens and colleagues, 2018 | Randomized crossover NR trial in middle-aged and older adults, six weeks per phase | Increased blood NAD⁺ and tolerability over the studied period | The subgroup blood-pressure signal was exploratory and does not establish hypertension treatment; clinical rejuvenation or reduced vascular events was not demonstrated |
| Yoshino and colleagues, 2021 | 25 postmenopausal women with overweight/obesity and prediabetes, NMN 250 mg/day for ten weeks | Improved measured skeletal-muscle insulin sensitivity using a hyperinsulinemic–euglycemic clamp, and changes in muscle signaling | An effect in men or other populations, diabetes prevention or long-term safety |
| Pencina and colleagues, 2023 | 32 people aged 55–80 with overweight/obesity, MIB-626 1000 mg once or twice daily or placebo for 14 days | Increased blood NMN/NAD metabolites and short-term tolerability | Efficacy for physical, cognitive or clinical outcomes, or long-term safety |

Methylated NAM metabolites inform us about NNMT flux, but their appearance alone does not prove pathological SAM depletion or hyperhomocysteinemia. The blood matrix, extraction procedure, LC–MS standards and time after dosing change the measured effect size. A participant with baseline vitamin deficiency, obesity or tissue injury represents a different biological situation in each case. Alongside Yoshino’s positive muscle outcome, clinical studies have found no convincing change in several metabolic endpoints; every increase in blood NAD⁺ cannot be declared a functional improvement. The table’s doses describe research protocols, not a prescription for the reader.
### 6.4. CD38 inhibition with a precursor: a logical hypothesis, not a validated combination

If CD38 reduces the availability of a precursor in a particular tissue, inhibition together with a precursor could theoretically improve replenishment of the local NAD⁺ pool. That hypothesis requires measurements of tissue flux, inhibitor availability and the combination’s benefit compared with either intervention alone. An ectoenzyme does not control every intracellular entry route, and NR and NMN differ in their transport and metabolism. Neither “70–85% of all ingested NR/NMN is destroyed by CD38” nor a universal concentration above which all human sirtuins work optimally has been established. These figures cannot serve as a clinical rationale for a combination.

NNMT catalyzes NAM + SAM → 1-MNA + SAH; SAH metabolism is linked to homocysteine. This provides a possible connection between NAM flux and methyl-group balance, but does not establish inevitable SAM depletion after every NR/NMN dose. Tissue expression, diet, methionine metabolism and excretion of methylated nicotinamide all matter. Betaine–homocysteine methyltransferase uses betaine to remethylate homocysteine predominantly in the liver and kidney; adding TMG to any precursor has not become an evidence-based standard and does not “insure” all tissues against an uncertain risk. No studies confirming the triple combination NR/NMN + apigenin + TMG for functional outcomes and safety are provided here. Sleep and circadian state may alter NAD⁺ metabolism, but [melatonin mechanisms](/knowledge/melatonin-mitochondria-circadian-receptors) likewise do not establish a universal schedule for this combination.

### 6.5. The senescent environment: removing cells and changing secretion

In model systems, senescent cells can create a SASP, and their signals can increase CD38 in neighboring myeloid cells. Reducing the number of particular senescent cells or altering their secretory phenotype could therefore change local NAD⁺ consumption. These interventions must be distinguished: a senolytic effect means preferential death of selected senescent cells, while a senomorphic effect means changing SASP without necessarily removing the cells. Neither guarantees systemic disappearance of CD38⁺ macrophages or a binary conversion from “M1 to M2” throughout every tissue.

Fisetin has shown context-dependent senolytic and senomorphic activity in preclinical models, but a molecular target and dose obtained in a particular culture or mouse do not demonstrate safe “pulse therapy” in humans. Dasatinib with quercetin has been investigated in small open-label clinical studies with selected biomarkers and tolerability measures; those studies have not established prevention of age-related NAD⁺ depletion, improved survival or disease prevention. Dasatinib is a prescription anticancer drug with substantial risks. The causal chain “fewer senescent cells → lower tissue CD38 → higher NAD⁺ → clinical benefit” needs to be tested in one human population with appropriate tissue measurements, rather than assembled from separate experiments.

### 6.6. Circadian timing, combinations and interpretation of laboratory data

NAMPT and NAD⁺-dependent processes are linked to circadian clocks in experimental tissues. This does not imply a fixed morning NAMPT peak in every organ of every person, or an established advantage of taking NR/NMN fasting at specified hours. Neither a universal “night-time CD38 blockade” by apigenin nor compatibility of the proposed doses with demonstrated sleep improvement has been established. Daily rhythms, food intake, shift work, age and tissue type alter the phase; a clinical comparison of administration times using functional endpoints remains a separate task.

A combination of NAD⁺ precursors, polyphenols, TMG, senolytics and sirtuin activators represents several interacting hypotheses rather than one validated protocol. Experimental animal doses cannot automatically be transferred to humans, and resveratrol cannot be declared a universal direct activator for all SIRT1 substrates: results depend on the particular assay, including its fluorophore-bearing substrate. Safety measures for each substance and combination must be assessed in their own regulatory and clinical context.

Investigating the mechanism requires a predefined matrix and NAD⁺/NADH measurement method, fractionation with purity controls, sampling time, NR/NMN/NAM/NAAD/1-MNA, and, for a methyl-depletion hypothesis, SAM/SAH/homocysteine with diet and renal function taken into account. General markers of inflammation, such as hs-CRP and IL-6, liver and kidney function, and clinical outcomes answer different questions; they are not specific tests of CD38 or PARP. The percentage NAD⁺ increase in mononuclear cells, an “ideal” homocysteine level or a single hs-CRP threshold cannot be set as a therapeutic target for healthy people without standardized methods and an outcome-focused study. Interpretation of laboratory abnormalities must rest on the clinical situation rather than a homemade algorithm for increasing TMG.

## 7. Conclusion: measure flux and function where the cause is proposed

NAD⁺ connects redox chemistry to ADP-ribose transfer and deacetylation reactions. The contribution of each source and consumer varies by tissue, cellular state and compartment. CD38 has a particularly important experimental role in aging-related immune contexts; PARP1 protects the genome but can consume a large nuclear pool during severe damage; SARM1 initiates a local program in an injured axon. These mechanisms do not add up to one obligatory cause of aging in every organ. Enzyme Km ratios do not establish a general priority for NAD⁺ until the available substrate and opposing fluxes have been measured.

Human NR/NMN studies most convincingly demonstrate changes in accessible blood metabolites at the doses and durations studied; particular functional signals remain limited by population and design. Neither higher blood NAD⁺ nor a successful intervention in a mouse CD38 or SARM1 axis demonstrates slower human aging. The next scientific step is to connect an anatomically defined mechanism, tissue NAD⁺ flux, a prespecified functional outcome and adequate safety assessment within one study. Until that connection exists, combination regimens remain hypotheses rather than clinical recommendations.

## Method and limits of selection

The editorial basis is the supplied manuscript with 37 bibliographic entries. Mechanistic nodes draw on specifically selected primary studies of CD38, PARP1–HPF1 and SARM1, and clinical NR/NMN protocols; review sources help orient the compartment discussion but do not replace primary evidence. This is a critical narrative review, not a systematic search, GRADE assessment or external peer review. An exhaustive comparison of all doses and registered trials was not performed.

The saved targeted check of 4 October 2026 identified a substantial bibliographic error for Pencina: the described metabolome study is the 14-day MIB-626 experiment, published in *The Journals of Gerontology: Series A* in 2023, using 1000 mg once or twice daily. A separate physiological publication was not substituted for that study. Yoshino used a hyperinsulinemic–euglycemic clamp, not a “hyperglycemic–euglycemic” clamp. Mouse-model findings and short-term human biomarkers do not validate a combined geroprotection protocol. This continuation translated and preserved the saved scientific treatment; it did not conduct a new systematic literature search or external peer review.

## FAQ

### Does increased blood NAD⁺ mean that its levels are restored in the brain and muscle?

No. Blood cells, the nucleus, mitochondrial matrix and axon have different accessible pools; a direct tissue method and a functional outcome are needed.

### Can NMN/NR, apigenin and TMG be prescribed together to prevent aging?

The clinical effect and safety of that combination have not been established. Interaction between the CD38 and NNMT pathways remains a mechanistic hypothesis.

### Why is an oncology PARP inhibitor unsuitable for healthy aging?

PARP participates in DNA repair; oncology inhibitors have a specific mechanism and risks, and preventive benefit in a healthy population has not been demonstrated.

## Sources

- [Camacho-Pereira et al. (2016). CD38 dictates age-related NAD decline and mitochondrial dysfunction through an SIRT3-dependent mechanism. Cell Metabolism.](https://doi.org/10.1016/j.cmet.2016.05.006)
- [Chini et al. (2020). CD38 ecto-enzyme in immune cells is induced during aging and regulates NAD+ and NMN levels. Nature Metabolism.](https://doi.org/10.1038/s42255-020-00298-z)
- [Covarrubias et al. (2020). Senescent cells promote tissue NAD+ decline during ageing via the activation of CD38+ macrophages. Nature Metabolism.](https://doi.org/10.1038/s42255-020-00305-3)
- [Covarrubias et al. (2021). NAD+ metabolism and its roles in cellular processes during ageing. Nature Reviews Molecular Cell Biology.](https://doi.org/10.1038/s41580-020-00313-x)
- [Gomes et al. (2013). Declining NAD+ induces a pseudohypoxic state disrupting nuclear-mitochondrial communication during aging. Cell.](https://doi.org/10.1016/j.cell.2013.11.037)
- [Escande et al. (2013). Flavonoid apigenin is an inhibitor of the NAD+ ase CD38: implications for cellular NAD+ metabolism, protein acetylation, and treatment of metabolic syndrome. Diabetes.](https://doi.org/10.2337/db12-1139)
- [Tarragó et al. (2018). A potent and specific CD38 inhibitor ameliorates age-related metabolic dysfunction by reversing tissue NAD+ decline. Cell Metabolism.](https://doi.org/10.1016/j.cmet.2018.03.016)
- [Wang et al. (2016). A nuclease that mediates cell death induced by DNA damage and poly(ADP-ribose) polymerase-1. Science.](https://doi.org/10.1126/science.aad6872)
- [Essuman et al. (2017). The SARM1 Toll/interleukin-1 receptor domain possesses intrinsic NAD+ cleavage activity that promotes pathological axonal degeneration. Neuron.](https://doi.org/10.1016/j.neuron.2017.02.022)
- [Figley et al. (2021). SARM1 is a metabolic sensor activated by an increased NMN/NAD+ ratio to trigger axon degeneration. Neuron.](https://doi.org/10.1016/j.neuron.2021.02.009)
- [Bai et al. (2011). PARP-1 inhibition increases mitochondrial metabolism through SIRT1 activation. Cell Metabolism.](https://doi.org/10.1016/j.cmet.2011.03.004)
- [Fang et al. (2016). NAD+ replenishment improves lifespan and healthspan in Ataxia telangiectasia models via mitophagy and DNA repair. Cell Metabolism.](https://doi.org/10.1016/j.cmet.2016.09.004)
- [Cantó et al. (2015). NAD+ metabolism and the control of energy homeostasis: a balancing act between mitochondria and the nucleus. Cell Metabolism.](https://doi.org/10.1016/j.cmet.2015.05.023)
- [Zhu et al. (2015). In vivo NAD assay shows the intracellular NAD contents and redox state in healthy human brain and their changes with age. Proceedings of the National Academy of Sciences.](https://doi.org/10.1073/pnas.1417921112)
- [Massudi et al. (2012). Age-associated changes in oxidative stress and NAD+ metabolism in human tissue. PLoS ONE.](https://doi.org/10.1371/journal.pone.0042357)
- [Liu et al. (2005). Crystal structure of human CD38 in complex with substrate analog. Journal of Biological Chemistry.](https://doi.org/10.1074/jbc.M509208200)
- [Dawicki-McKenna et al. (2015). PARP-1 activation requires local unfolding of an autoinhibitory domain. Molecular Cell.](https://doi.org/10.1016/j.molcel.2015.10.013)
- [Suskiewicz et al. (2020). HPF1 completes the PARP active site for DNA damage-induced ADP-ribosylation. Nature.](https://doi.org/10.1038/s41586-020-2125-8)
- [Gerdts et al. (2015). SARM1 activation triggers axon degeneration locally via NAD+ destruction. Science.](https://doi.org/10.1126/science.1258366)
- [Bratkowski et al. (2020). Structural and mechanistic regulation of SARM1 auto-inhibition and activation. Structure.](https://doi.org/10.1016/j.str.2020.08.010)
- [Xu et al. (2018). Senolytics improve physical function and increase lifespan in old age. Nature Medicine.](https://doi.org/10.1038/s41591-018-0092-9)
- [Yousefzadeh et al. (2018). Fisetin is a senolytic that extends health and lifespan. EBioMedicine.](https://doi.org/10.1016/j.ebiom.2018.09.015)
- [Trammell et al. (2016). Nicotinamide riboside is uniquely and orally bioavailable in mice and humans. Nature Communications.](https://doi.org/10.1038/ncomms12948)
- [Martens et al. (2018). Chronic nicotinamide riboside supplementation is well-tolerated and elevates NAD+ in healthy middle-aged and older adults. Nature Communications.](https://doi.org/10.1038/s41467-018-03421-7)
- [Yoshino et al. (2021). Nicotinamide mononucleotide increases muscle insulin sensitivity in prediabetic women. Science.](https://doi.org/10.1126/science.abe9985)
- [Pencina KM et al. (2023). MIB-626, an Oral Formulation of a Microcrystalline Unique Polymorph of β-Nicotinamide Mononucleotide, Increases Circulating Nicotinamide Adenine Dinucleotide and its Metabolome in Middle-Aged and Older Adults. The Journals of Gerontology: Series A, 78(1), 90–96.](https://doi.org/10.1093/gerona/glac049)