Research digest / mechanism lens · 01 sirtuins · 02 PARP1 · 03 CD38
Sirtuins and NAD+: the NAD-consuming enzymes in the research.
Three enzyme families draw on one NAD+ pool. This is what the mechanistic literature established about how they compete.
The short version
Sirtuins and NAD+ are inseparable: sirtuins (a family of cellular-maintenance enzymes that can't work without NAD+) burn a molecule of NAD+ every time they do their job. But sirtuins are not the only customers. PARP1, a DNA-repair enzyme, and CD38, a surface enzyme, also consume NAD+ — and all three pull from the same limited pool. That competition is the heart of the mechanism: when one enzyme (especially CD38) ramps up, the others have less NAD+ to spend. This page walks the three stations — sirtuins, PARP1, CD38 — and the rate-limiting enzyme, NAMPT, that refills the tank.
Station 01 — Sirtuins (SIRT1–SIRT7)
Sirtuins are NAD+-dependent deacylase enzymes that regulate metabolism, stress resistance, DNA repair and, in model organisms, lifespan [5]. Mechanistically, each deacylation reaction cleaves a molecule of NAD+ — so sirtuin activity is gated by how much NAD+ is available, making the coenzyme a rate-setting input rather than a passive bystander [9]. The mitochondrial sirtuin SIRT3 is a concrete example of the dependency: in mice, the mitochondrial NAD+ transporter SLC25A47 supplied the NAD+ that sustained SIRT3 (and AMPKα) activity, and removing that NAD+ supply impaired mitochondrial and hepatic lipid handling [13]. The landmark Science review frames sirtuins as one of the two major classes of NAD+-consuming signaling enzymes whose function tracks the cell's NAD+ concentration as it changes with age [8].
Station 02 — PARP1 (the DNA-repair drain)
PARP1 (poly(ADP-ribose) polymerase 1) is a DNA-damage-response enzyme that consumes large amounts of NAD+ when it is activated [9]. When DNA is damaged, PARP1 fires and poly(ADP-ribosyl)ates target proteins — attaching long chains of ADP-ribose built directly from NAD+ — drawing heavily on the pool, which is exactly why a burst of DNA-repair activity can transiently starve the sirtuins of their cosubstrate [5]. PARPs are named, alongside sirtuins, as the cosubstrate-consuming enzymes that make NAD+ a shared and competitive resource rather than a private fuel for any one pathway [8]. The practical upshot is that NAD+ is not spent on one process in isolation; DNA repair, deacylation and the CD38 reaction all bid for the same molecules. This competition reframes what "low NAD+" means: it is not only a shortage of energy currency but a shortage of the substrate three maintenance systems need at once, so whichever process is most active — a wave of DNA damage activating PARP1, or rising CD38 with age — can leave the others under-supplied [5][8].
Station 03 — CD38 (the age-rising consumer)
CD38 is an NAD glycohydrolase (NADase) — a cell-surface and intracellular enzyme that breaks NAD+ down to generate the second messenger cADPR, consuming NAD+ in the process and limiting its availability to PARPs, ARTs and sirtuins [11]. CD38 is the pivotal enzyme for the age story: it is the principal NAD+-consuming enzyme whose activity rises with age, and CD38-knockout mice are protected against the age-related fall in tissue NAD+, retaining SIRT3 activity and better mitochondrial function [2]. In other words, much of the age-related NAD+ decline is not a failure to make NAD+ — it is increased consumption by the CD38 NAD-consuming enzyme outcompeting the sirtuins for the shared pool [2][11].
The refill: NAMPT and the salvage pathway
Against three consumers stands the salvage pathway, the dominant route that recycles nicotinamide back into NAD+ through the rate-limiting enzyme NAMPT (nicotinamide phosphoribosyltransferase) [10]. NAMPT sets the ceiling on how fast the pool refills, and its output is dynamically regulated — it is induced by exercise and follows a circadian rhythm via the CLOCK–SIRT1 loop, so NAD+ naturally oscillates over 24 hours [7]. That circadian control is itself tissue-specific: in mice, deleting NAMPT erased most circadian transcripts in brown and white fat (89% and 77% respectively) while skeletal muscle stayed completely refractory despite the same NAD+ drop — direct evidence that the salvage-NAD+ axis is wired differently into different tissues [7]. NAMPT also has a darker role: it governs the proinflammatory senescence-associated secretory phenotype (SASP) in senescent cells through an HMGA–NAMPT–NAD+ axis that activates NF-κB, which is why one mechanistic study explicitly cautions that NAD+ augmentation should be administered with precision in aging populations rather than maximized indiscriminately [6]. The refill, in other words, is not a simple tap you can open wider for benefit everywhere — it is a regulated, context-dependent system.
What is studied about NAD+ therapy
"NAD+ therapy" is a marketplace label, not an approved treatment, and the cited evidence is uneven across routes. The oral-precursor route is the best characterized: NR and NMN dose-dependently raise blood NAD+ and were well tolerated in randomized trials [4][3]. The IV/injectable route — the most aggressively marketed as "NAD+ therapy" — has the weakest controlled evidence: infused NAD+ is rapidly cleared from plasma [12], fast infusions can cause discomfort, and a compounded injectable NAD+ product drew an FDA Class I recall for endotoxin contamination. So when the literature is read straight, raising the NAD+ pool to feed sirtuins is biologically coherent, but the leap from "blood NAD+ rose" to a therapeutic outcome remains, in the words of a 2025 review, preliminary [14]. This digest characterizes those findings; it does not prescribe a protocol.