Nicotinamide mononucleotide, a potential future treatment in ocular diseases

This 2024 review article, published in Graefe's Archive for Clinical and Experimental Ophthalmology by a group at Keio University School of Medicine (Tokyo), summarises the preclinical evidence for NMN as a candidate treatment across a range of ocular diseases. It is the first synthesis of this literature in one place. The reviewed evidence covers ischemic retinopathy (retinal ischemia-reperfusion, ocular ischemic syndrome from carotid artery occlusion), retinal detachment, photoreceptor degeneration, retinal pigment epithelium senescence relevant to age-related macular degeneration, corneal denervation and epithelial defect, and dry eye disease with meibomian gland dysfunction.

The proposed mechanisms cluster around three effects that recur across models: (1) restoration of tissue NAD+ levels in ocular tissue depleted by the disease insult; (2) activation of downstream SIRT1 signalling, which contributes to anti-oxidant and anti-senescence effects; and (3) activation of the NRF2 pathway, which regulates antioxidant gene expression. In multiple murine models, systemic intraperitoneal NMN (typically 100–500 mg/kg/day) preserved retinal function measured by electroretinography, reduced inflammatory cell recruitment, and prevented tissue morphological damage. Topical NMN eye drops in an aged-mouse model of meibomian gland dysfunction improved gland morphology, suggesting local delivery is also feasible.

The review closes with several honest caveats worth reading directly rather than glossed. The authors note that most work uses systemic intraperitoneal dosing, so distinguishing direct ocular effects from whole-body effects (metabolic, cardiovascular, inflammatory) has not been done rigorously. Ideal dosing for ocular indications is not established. Bioavailability studies for the eye specifically are absent. And — notably for a review by proponents of NMN — the authors explicitly acknowledge the SARM1-mediated axon degeneration concern raised by Figley et al. 2021 and recommend more investigation of whether NMN could influence injury-induced SARM1 activation in ocular contexts. No published human RCT of NMN for any ocular endpoint had reported results at the time the review was written.

The review's key claims about the underlying evidence, condition by condition:

• Retinal ischemia/reperfusion (mouse). Intraperitoneal NMN 500 mg/kg/day preserved ERG b-wave amplitude and reduced inflammatory cell recruitment; retinal NAD+ was restored (Lee et al. 2022, IJMS 23:11228).
• Ocular ischemic syndrome (mouse UCCAO model). Intraperitoneal NMN 500 mg/kg/day preserved a- and b-wave amplitudes, suppressed reactive gliosis and Ccl2/Ccl12 chemokines, and restored retinal Nrf2 and NAD+ (Lee et al. 2022, IJMS 23:14711).
• Retinal detachment (mouse). Intraperitoneal NMN 250 and 500 mg/kg/day dose-dependently reduced photoreceptor apoptosis (TUNEL), suppressed CD11b+ inflammatory cell accumulation, increased HO-1, and reduced oxidative stress; SIRT1 upregulation was proposed as a downstream mechanism (Chen et al. 2020, Aging 12:24504).
• Photoreceptor Nampt conditional knockouts (mouse). Intraperitoneal NMN 150 mg/kg preserved a- and b-wave in Nampt−rod/−rod and Nampt−cone/−cone mice; 300 mg/kg improved function in a light-induced retinopathy model (Lin et al. 2016, Cell Reports 17:69).
• Aging (wild-type mouse, 12-month administration). Long-term intraperitoneal NMN 100 and 300 mg/kg reduced age-associated weight gain, improved a-wave amplitude, improved tear production, and suppressed accumulation of subretinal microglia and macrophages (Mills et al. 2016, Cell Metab 24:795 — the same paper that anchors much of the systemic-supplementation NMN literature).
• RPE senescence / early AMD-relevant biology (mouse and ARPE-19 cells). Intraperitoneal NMN 300 mg/kg restored NAD+, reduced canonical senescence markers (SA-β-gal, p21, p16), and reduced CD11b+/F4/80+ inflammatory recruitment in a NaIO3 oxidative-toxicity model; the effect appeared SIRT1-dependent (Ren et al. 2022, Oxid Med Cell Longev 2022:5961123).
• Corneal denervation (mouse). Subconjunctival NMN 251 ng/eye improved corneal epithelial defect healing after nerve injury; also reversed FK866-induced human corneal epithelial cell death in vitro, with activation of AKT, CREB, and SIRT1 (Li et al. 2019, IOVS 60:3538).
• Hyperosmolar corneal stress (in vitro). NMN dose-dependently restored NAD+ and reduced hyperosmolar-induced cell death, IL-17A secretion, ROS, and macrophage activation in a corneal epithelial cell / macrophage co-culture (Meng et al. 2021, J Inflamm Res 14:479).
• Dry eye / meibomian gland dysfunction (aged mouse). Topical NMN eye drops (5% in PBS, six times daily, two weeks) improved 3β-HSD activity and ameliorated meibomian gland atrophy — the first evidence that local delivery is feasible for ocular indications (Sasaki et al. 2022, Nat Aging 2:105).
• Ex vivo human RPE from AMD donors. NMN improved ATP production in RPE from human donors with age-related macular degeneration (Ebeling et al. 2020, Redox Biol 34:101552) — the only ex vivo human evidence discussed.

Readers should note the following in weighing this review's conclusions:

1. Narrative synthesis, not systematic review. The Methods section explicitly states the paper was assembled from "our own recent reports as well as a search of the related literature." There is no PRISMA reporting, no pre-registered protocol, no formal risk-of-bias assessment, and no meta-analysis. This shapes what the review can and cannot conclude — it is a useful map of the field written by insiders, not an independent evidence assessment.
2. Substantial author-group overlap with primary literature. Several of the most-featured primary studies are authored by the review team, including the retinal I/R paper (Lee et al. 2022), the ocular ischemic syndrome UCCAO paper (Lee et al. 2022), and the Kurihara-group contribution to the Sasaki et al. 2022 meibomian gland paper. This is common in specialised subfields with few active groups, but readers should be aware that the review is not independent of the literature it summarises.
3. Predominantly preclinical evidence. The reviewed efficacy data are almost entirely from mouse models and cell lines. Only one small ex vivo dataset uses human tissue (Ebeling 2020 RPE from AMD donors). No published human clinical trial of NMN for any ocular endpoint had reported results at the time of writing. The human data cited elsewhere in the paper are safety trials (MIB-626 and others), not efficacy trials in ocular disease.
4. Systemic dosing dominates the literature. The authors acknowledge that most reviewed studies used intraperitoneal injection at doses of 100–500 mg/kg/day in rodents. Distinguishing direct ocular effects from whole-body effects (metabolic, cardiovascular, immune) has not been done rigorously in the underlying literature. Interspecies dose translation to human oral supplementation is also unaddressed.
5. Bioavailability to ocular tissue is not established. The authors state directly: "There is no clear pharmacokinetic study to define the bioavailability of NMN through oral or topical administration. Thus, the bioavailability of NMN could still be a limiting factor for its use." This is the correct honest framing but it means the pathway from oral supplementation to ocular NAD+ elevation in humans is not yet demonstrated.
6. The SARM1 concern is explicitly acknowledged. In the Discussion, the authors state: "sterile alpha and toll/interleukin-1 receptor motif-containing 1 (SARM1) was suggested to be activated by an increase in the ratio of NMN to NAD+, and its alteration could be linked to trigger axon degeneration. As NMN has a possibility to influence injury-induced SARM1 activation and axon destruction in particular experimental cases, more investigations are recommended." Retinal ganglion cells are CNS neurons with long axons — the same cell type whose Wallerian degeneration is the paradigm case for the SARM1 mechanism documented by Di Stefano 2015 and Hou 2022. This concern is not resolved by the ocular preclinical protective findings, since the mechanistic contexts differ.
7. Industry funding disclosed but no direct conflict declared. One funding source is Novo Nordisk Pharma Ltd., a pharmaceutical company. The authors declare no conflict of interest. Novo Nordisk does not to our knowledge market an NMN product, but the industry funding is worth noting for transparency.
8. US regulatory context has shifted since publication. The paper references (in mid-2023) then-emerging FDA action to restrict NMN supplement marketing. The regulatory situation continues to evolve and readers should consult current sources for the state of NMN's status as a dietary supplement versus investigational drug.