Research paper

Nicotinamide mononucleotide increases muscle insulin sensitivity in prediabetic women

Mihoko Yoshino, Jun Yoshino, Brandon D. Kayser, Gary J. Patti, Michael P. Franczyk, Kathryn F. Mills, Miriam Sindelar, Terri Pietka, Bruce W. Patterson, Shin-Ichiro Imai, Samuel Klein

Science 2021 Volume 372, issue 6547, pages 1224–1229

Tier 2 · Single well-powered human RCT or comprehensive human reviewMetabolic

Source identifiers

DOI 10.1126/science.abe9985
PubMed PMID 33888596
Funding NIH grants DK56341, DK104995, AG037457, AG047902; NIH SIG grant S10 OD026929 (UC San Diego IGM Genomics Center); the Tanaka Fund; the Foundation for Barnes-Jewish Hospital. Placebo and NMN capsules provided by Oriental Yeast Co., Ltd. (Tokyo, Japan).
Declared competing interests S.I. receives a part of patent-licensing fees from MetroBiotech (USA) and Teijin Limited (Japan) through Washington University; also serves as Invited Chief Scientist at the Institute for Biomedical Research and Innovation in Kobe, Japan (not a competing interest per authors). J.Y. is an inventor of a patent application related to NMN (#20180228824). S.K. receives research support from Janssen Pharmaceuticals and serves on a scientific advisory board for Merck Sharp & Dohme Corp. Other authors declared no competing interests.

Study snapshot

Design
Randomized, placebo-controlled, double-blind, parallel-group trial (10 weeks)
Population
Postmenopausal women with prediabetes (ADA criteria) who were overweight or obese (BMI 25.3–39.1 kg/m²)
Sample
n=25 completed (NMN n=13; placebo n=12)
Intervention
Oral NMN 250 mg/day vs. matched placebo for 10 weeks
Endpoints
Insulin-stimulated glucose disposal rate (hyperinsulinemic-euglycemic clamp); Body composition (DXA, MRI); Hepatic and adipose insulin sensitivity; Muscle NAD⁺ and NMN metabolite content; PBMC NAD⁺ content; Skeletal muscle insulin signaling (pAKT Ser473, pAKT Thr308, pmTOR Ser2448); Skeletal muscle global gene expression (RNA-seq); Muscle mitochondrial respiratory capacity; Muscle physical function (grip strength, torque, fatigability)

What the study found, in plain language

This is the first randomized controlled trial in humans to test whether oral nicotinamide mononucleotide (NMN) — a precursor the body uses to make NAD⁺ — affects how the body handles insulin. Twenty-five postmenopausal women with prediabetes who were overweight or obese took either 250 mg of NMN daily or an identical placebo for 10 weeks. Neither the participants nor the researchers knew which pill each person received until the trial ended.

After 10 weeks, women taking NMN showed a 25% improvement in how effectively their muscles took up glucose in response to insulin, measured with a hyperinsulinemic-euglycemic clamp — the reference method for assessing insulin sensitivity. Women taking placebo showed no such change. Muscle biopsies confirmed that the insulin-signaling machinery inside the muscle cells (the AKT and mTOR pathways) was working harder after NMN. Gene expression analysis showed that NMN activated hundreds of genes involved in muscle repair and remodeling, including the PDGF signaling pathway.

The improvement was specific to muscle. NMN did not change body weight, body fat, liver fat, blood pressure, fasting glucose, fasting insulin, cholesterol, or how insulin acted on the liver or fat tissue. Muscle strength and endurance did not improve during the 10-week window. NMN was well tolerated, with no adverse events reported and standard blood tests unchanged.

Key findings

  • Muscle insulin sensitivity increased 25 ± 7% after 10 weeks of NMN (p<0.01); no change with placebo. Effect size is comparable to what is typically seen after ~10% weight loss or 12 weeks of the insulin-sensitizing drug troglitazone.
  • Skeletal muscle insulin signaling was enhanced. Phosphorylation of AKT (Ser473 and Thr308) and mTOR (Ser2448), plus total AKT and mTOR protein, were higher after NMN during insulin infusion; no change with placebo.
  • PBMC NAD⁺ rose. Basal peripheral blood mononuclear cell NAD⁺ increased with NMN but not placebo. The 240-minute PBMC NAD⁺ AUC was 43% greater in the NMN group than placebo (p<0.01).
  • Muscle NAD⁺ content did not change in either group, but muscle NMN metabolites (N-methyl-nicotinamide, 2PY, 4PY) increased with NMN, suggesting increased NAD⁺ turnover rather than accumulation.
  • Muscle transcriptome shifted substantially. During insulin infusion, NMN produced 308 differentially expressed genes vs. 5 in placebo — a 60-fold difference.
  • PDGF binding was the top-enriched pathway. PDGFRβ was significantly upregulated (p=0.014), along with pericyte markers CD90 (p=0.015) and CD109 (p=0.018), and downstream collagen genes COL1A1 (p=0.003), COL5A1 (p=0.010), COL6A1 (p=0.018).
  • No effect on other metabolic domains: body composition, intra-abdominal fat, intrahepatic triglyceride, blood pressure, HbA1c, fasting glucose, insulin, free fatty acids, triglycerides, HDL-cholesterol, adiponectin, leptin, or basal glucose and palmitate kinetics.
  • No effect on muscle mitochondrial oxidative capacity or physical function (handgrip strength, knee extensor/flexor torque, fatigability, recovery).
  • Safety: No adverse events; no abnormalities on standard blood tests. Compliance 99.6% (placebo) and 100% (NMN).

What this study can and cannot tell us

  • Small sample. Only 25 women completed the trial (13 NMN, 12 placebo). Effect estimates carry wide confidence intervals and the finding requires replication in larger trials.
  • Narrow population. All participants were postmenopausal women with prediabetes who were overweight or obese. The results cannot be generalized to men, premenopausal women, people without prediabetes, or people at healthy body weight.
  • Baseline imbalance in liver fat. Intrahepatic triglyceride was 14.8 ± 2.0% in the placebo group vs. 6.3 ± 1.2% in the NMN group at baseline (p=0.003) despite randomization. A published Comment in Science (Brenner, 2021; doi:10.1126/science.abj1696) argued this imbalance complicates interpretation.
  • Mechanism not fully resolved. Muscle NAD⁺ content did not change even though muscle NMN metabolites rose and insulin sensitivity improved. The authors attribute the effect to increased NAD⁺ turnover and PDGFRβ signaling, but the causal chain is inferred rather than demonstrated.
  • Single dose and single duration. Only 250 mg/day for 10 weeks was tested. Dose–response and long-term safety and efficacy are unknown.
  • No clinical endpoints. The trial measured surrogate markers (insulin sensitivity, gene expression, signaling proteins), not diabetes incidence, cardiovascular events, or functional outcomes. Muscle strength and endurance were assessed and did not improve during the 10-week window.
  • Sponsor and inventor involvement. NMN capsules were supplied by Oriental Yeast Co., Ltd. Several authors hold NMN-related patents or patent-licensing arrangements, disclosed in the competing interests statement.