Glycine

Abstract

Glycine (aminoacetic acid) is the smallest proteinogenic amino acid and the only one that is not chiral at the α-carbon. Humans make it mainly from serine by serine hydroxymethyltransferase, with additional input from choline, hydroxyproline, and, less certainly, threonine; they degrade it through the glycine cleavage system and through the reverse of the same hydroxymethyltransferase reaction (Wang and Wu, 2013). The carbon skeleton feeds collagen, glutathione, heme, purines, creatine, and bile-acid conjugation. In the central nervous system the same molecule is a strychnine-sensitive inhibitory transmitter at glycine receptors and an obligatory co-agonist at the GluN1 site of NMDA receptors (Johnson and Ascher, 1987; Breitinger and Becker, 1998).

Conditional indispensability is real in named high-demand states. Indicator amino-acid oxidation shows glycine to be conditionally indispensable in late human pregnancy (Rasmussen et al., 2021). Pregnant adolescents cannot maintain glycine flux in the third trimester as adult pregnant women do (Hsu et al., 2016). Jackson’s 5-oxoproline work showed that glycine can be driven limiting by benzoate or by catch-up growth after malnutrition (Jackson et al., 1987; Persaud et al., 1996). Those findings do not establish that ordinary adult omnivory is glycine-deficient. Adults switched to 0.75 g protein kg⁻¹ d⁻¹ maintained glycine flux even as whole-body turnover fell (Gibson et al., 2002). Meléndez-Hevia’s stoichiometric argument that biosynthetic capacity cannot meet collagen demand is a model, not a clinical survey (Meléndez-Hevia et al., 2009).

Human free-glycine trials are small. Three grams at bedtime reduced next-day fatigue after acute sleep restriction in one randomized study from the Bannai group (Bannai et al., 2012b). A 2024 systematic review judged the healthy-adult sleep evidence small and at high risk of bias (Soh et al., 2024). Five grams daily for three months lowered HbA1c versus placebo in type 2 diabetes (Cruz et al., 2008); fifteen grams daily for three months lowered TBARS and systolic pressure in metabolic syndrome (Díaz-Flores et al., 2013). Low plasma glycine tracks insulin resistance (Adeva-Andany et al., 2018); that association is not a demonstrated causal deficit, and glycine supplementation worsened glucose intolerance in diet-induced obese mice by increasing gluconeogenesis (Alves et al., 2022).

The glutathione-aging literature that is spent as “glycine” is mostly glycine plus a cysteine donor. Two-week cysteine-plus-glycine restored erythrocyte glutathione synthesis in elderly adults and in uncontrolled type 2 diabetes (Sekhar et al., 2011a, 2011b). Later GlyNAC open-label and randomized work from the same laboratory reports broad aging-related improvements (Kumar et al., 2023). An independent randomized trial of 7.2 g GlyNAC in healthy older adults did not meet its primary glutathione-redox endpoints (Lizzo et al., 2022). The NIA Interventions Testing Program found that an 8% glycine diet extended UM-HET3 mouse lifespan by 4–6% in both sexes (Miller et al., 2019). That is a real animal result. It is not a human longevity trial, and it is not GlyNAC.

Table. Five interrogations this article is written to answer. Full resolutions are in Part Six; the grades are not slogans.

Glycine is aminoacetic acid, C₂H₅NO₂, molecular weight 75.07. The side chain is a hydrogen atom. That is why glycine is the only proteinogenic amino acid that is not chiral at the α-carbon, why it packs into collagen’s every-third-residue slot, and why it is a poor steric obstruction in enzymes and channels. Nutrition inherits that chemistry. It does not get to treat glycine as a vague “calming amino acid” first and a metabolite second.

At physiological pH the molecule is a zwitterion. It is transported by several systems, including the GlyT1 and GlyT2 glycine transporters that dominate synaptic control (Zafra et al., 2017) and by more general amino-acid carriers in gut and kidney. Free glycine in plasma is not the glycine in collagen, and it is not the glycine residue inside glutathione. Those three pools are chemically the same monomer only after hydrolysis. Until then they are different experiments.

02 Synthesis and the serine junction

The principal biosynthetic route in mammals is the reversible conversion of serine to glycine by serine hydroxymethyltransferase (SHMT), transferring a one-carbon unit to tetrahydrofolate to yield 5,10-methylene-THF (Wang and Wu, 2013). Cytosolic SHMT1 and mitochondrial SHMT2 are not interchangeable; the mitochondrion is a major site of glycine and one-carbon production. Additional carbon can enter from choline (via sarcosine) and from hydroxyproline derived from collagen turnover. Threonine has been proposed as a source; its quantitative contribution in adult humans remains less secure than the serine route (Wang and Wu, 2013; Holeček, 2025).

Degradation is not a single waste pipe. The glycine cleavage system (GCS) oxidatively decarboxylates glycine to CO₂ and NH₃ while charging methylene-THF. SHMT running in reverse consumes glycine to make serine. Peroxisomal D-amino acid oxidase can form glyoxylate. GCS is the dominant degradative path in animals (Wang and Wu, 2013). Non-ketotic hyperglycinemia, a GCS defect, is a disease of too much glycine in the brain, not a model of oral supplementation. It is mentioned here only to forbid using a catastrophic inborn error as evidence that more glycine is generally desirable.

03 One-carbon metabolism

Every glycine made from serine is a one-carbon unit made as well. Every glycine cleaved by GCS is a one-carbon unit made from glycine itself. The molecule is therefore both a product and a substrate of the folate cycle. That coupling is established biochemistry. It is also the first place popular writing over-promises. Folate, B12, B6, and methionine cycle activity set the fate of those one-carbon units. Giving glycine does not, by itself, “turn on methylation” in a direction that a longevity brochure can name.

Glycine N-methyltransferase (GNMT) methylates glycine to sarcosine, clearing S-adenosylmethionine. Johnson, Cuellar, and Miller (2023), reviewing glycine and aging, argue that GNMT — methionine disposal — is a more coherent gerontology reading of glycine than collagen or glutathione slogans. In flies, Gnmt is required for some dietary-restriction and insulin-signalling longevity effects; overexpression can extend life. That is animal genetics, recorded as a competing mechanism, not as a human aging endpoint.

04 Collagen, glutathione, heme, purines, creatine

Collagen is about one-third glycine by residue. The triple helix cannot pack if that slot is occupied by a larger side chain. Li and Wu (2018) review dietary glycine, proline, and hydroxyproline as collagen substrates in animals and argue that food protein may not supply enough glycine for maximal collagen synthesis in growing animals. That is a nutrition-of-production argument, strongest in poultry and livestock. de Paz-Lugo, Lupiáñez and Meléndez-Hevia (2018) reported that high glycine concentrations increased collagen synthesis by articular chondrocytes in vitro and inferred that acute glycine deficiency could be an important cause of osteoarthritis. That inference is in vitro plus a leap. It is not a randomized demonstration that free glycine treats human osteoarthritis.

Glutathione is γ-glutamyl-cysteinyl-glycine. γ-Glutamylcysteine ligase makes the first peptide bond; glutathione synthetase adds glycine. Cysteine availability is the textbook limitation on GSH synthesis. Glycine can co-limit when cysteine is supplied, which is why the Sekhar designs always include a cysteine donor. McCarty, O’Keefe and DiNicolantonio (2018) argued that dietary glycine is rate-limiting for glutathione and may have broad health potential. That article is a hypothesis piece, cited here as a claim-source, not as a trial.

Heme biosynthesis begins with glycine and succinyl-CoA, condensed by 5-aminolevulinate synthase. The purine ring takes C4, C5 and N7 from glycine. Creatine begins with arginine:glycine amidinotransferase (AGAT); AGAT deficiency is a treatable creatine-synthesis disorder, which proves the pathway’s necessity and does not prove that extra oral glycine raises creatine in AGAT-sufficient adults. These fates are established. They are not interchangeable with sleep scores, HbA1c, or lifespan.

Table. Metabolism/pathway ledger. Chemistry is established; the right-hand column is the standing forbid for later parts.

Plasma glycine is low in obesity and type 2 diabetes, and higher glycine at baseline associates with lower incident diabetes risk in prospective series (Adeva-Andany et al., 2018; Alves et al., 2019). Branched-chain amino acids move in the opposite direction. Alves and Morio (2023) argue that glycine metabolism now has to be read together with BCAA metabolism, and that strategies which lower BCAAs may restore glycine more coherently than glycine pills. That is a metabolic-network claim, strongly supported as association, not a demonstrated causal hole that oral glycine fills.

Gannon, Nuttall and Nuttall (2002) gave healthy adults 1 mmol glycine per kilogram lean body mass, 25 g glucose, both, or water. Glycine and glucagon rose after glycine, as expected; insulin rose slightly after glycine alone. The study asked whether glycine explained gelatin’s insulinotropic effect in diabetes. It is an acute nine-person experiment, not a metabolic-syndrome outcome trial. It does show that ingested free glycine is metabolically visible within two hours.

06 Protein glycine is not free glycine

Dietary glycine arrives mostly inside proteins. Collagen and gelatin are glycine-rich; muscle proteins are not. A collagen hydrolysate delivers di- and tripeptides (often Pro-Hyp, Hyp-Gly) whose receptors, clearance, and trial literature are distinct from the monomer (Li and Wu, 2018). This article does not spend skin or joint hydrolysate trials as free-glycine evidence. The reverse error is equally forbidden: a free-glycine sleep trial is not evidence that bone broth is a hypnotic.

Endogenous production is large. Wang and Wu (2013) review inter-organ synthesis involving liver and kidney and conclude that under ordinary feeding glycine is not always made in amounts that meet all demands in animals, especially when growth or disease raises collagen and GSH use. Holeček (2025) restates glycine as conditionally essential relative to serine. Conditional language is doing real work in those papers. It is not a diagnosis of the healthy adult supermarket shopper.

07 When glycine is conditionally limiting in humans

Jackson, Badaloo, Forrester, Hibbert and Persaud (1987) gave 4–10 g sodium benzoate to six healthy adults to deplete the glycine pool by hippurate formation. Urinary 5-oxoproline rose within three hours in five of six. The finding is that glycine can be driven limiting in normal people by a conjugation load, and that 5-oxoproline is an index of that squeeze. It is not a finding that habitual diets fail.

During recovery from severe childhood malnutrition, 5-oxoproline excretion rose with rapid weight gain and fell when glycine was supplemented in children gaining less than 17 g kg⁻¹ d⁻¹ (Persaud et al., 1996). That is catch-up growth, a collagen- and GSH-intensive state. Pregnant adolescent girls could not maintain glycine flux in late pregnancy because synthesis from serine fell; birth length associated with third-trimester flux (Hsu et al., 2016). Rasmussen and colleagues (2021), using indicator amino-acid oxidation in healthy women, found no statistically significant glycine-intake response at mid-gestation but a late-gestation pattern consistent with conditional indispensability: lower glycine intakes raised indicator oxidation. Strongly supported: late pregnancy, adolescence in pregnancy, malnutrition recovery, and experimental benzoate load.

Healthy adults are a different experiment. Gibson, Jahoor, Ware and Jackson (2002) reduced protein from 1.13 to 0.75 g kg⁻¹ d⁻¹. Nitrogen balance recovered after two days; whole-body turnover fell; glycine and tyrosine fluxes were maintained despite less amino acid coming from proteolysis. Jackson, Gibson, Lu and Jahoor (2004), in the same intake range, found that erythrocyte glutathione fractional synthesis fell on days 3 and 10 of the “safe” protein intake, with 5-oxoproline up, even though GSH concentration recovered by day 10. The two papers are not a contradiction if they are read as different readouts: glycine flux can be defended while GSH synthesis is strained. Neither paper is a free-glycine pill trial in protein-replete omnivores.

08 The stoichiometric insufficiency claim

Meléndez-Hevia and de Paz-Lugo (2008) and Meléndez-Hevia, de Paz-Lugo, Cornish-Bowden and Cárdenas (2009) argued that SHMT stoichiometry ties glycine production to one-carbon disposal, creating a “weak link”: the organism cannot make glycine without making methylene-THF that must be used elsewhere, so biosynthetic capacity may not meet collagen demand. The 2009 paper states that metabolic capacity for glycine biosynthesis does not satisfy the need for collagen synthesis. That is a theoretical accounting plus kinetic argument. de Paz-Lugo et al. (2018) then showed that raising glycine in chondrocyte culture increased collagen output.

The article grades the biochemistry as plausible and the human-deficiency conclusion as speculative. A weak-link model can be true in a flux-balance sense and still fail as a description of free-living adults who recycle collagen glycine, eat mixed protein, and adapt synthesis as Gibson et al. (2002) observed. Johnson et al. (2023) offer a different mechanistic centre of gravity — GNMT and methionine clearance — which would make glycine a methyl-sink rather than a collagen shortfall. Both mechanisms can operate. Neither has been converted into a human aging endpoint by a glycine-only trial.

The strychnine-sensitive glycine receptor is a pentameric chloride channel concentrated in spinal cord and brainstem. It is the receptor of inhibitory glycinergic synapses, not the GluN1 co-agonist site. Breitinger and Becker (1998) reviewed it as a “therapeutic orphan”: rich pharmacology, few licensed drugs. Oral glycine at a few grams is a poor way to target that channel in a regionally specific manner, because plasma and CSF glycine are under transporter control (Zafra et al., 2017). This article therefore does not read bedtime glycine trials as spinal GlyR experiments.

10 NMDA co-agonism

Johnson and Ascher (1987) showed that glycine potentiates NMDA currents in cultured mouse neurons at concentrations as low as 10 nM, through a strychnine-insensitive site, as an increase in NMDA-channel opening frequency. In 1992 they estimated an apparent K_D near 150 nM (about 130 nM after correction for contaminating glycine) and kinetics consistent with a single glycine site (association ~1.2×10⁷ M⁻¹ s⁻¹; dissociation ~1 s⁻¹). The site is obligatory: without a co-agonist the NMDA receptor does not open. D-serine shares the site in many forebrain synapses.

Whether the site is saturated in vivo is an old argument (Wood, 1995). Nanomolar K_D in a dish does not mean every synapse sits in a saturating bath. GlyT1 locally depletes glycine around NMDA receptors (Zafra et al., 2017). High-dose glycine and GlyT1 inhibitors have been tried in schizophrenia as NMDA co-agonist strategies; that literature is a psychiatric pharmacology experiment, not a sleep or longevity experiment, and Soh et al. (2024) note that longer-term glycine in psychiatric populations is where nervous-system signals were strongest — a different population and usually a different dose band from 3 g bedtime. D’Souza et al. (2000) measured plasma and CSF amino acids after intravenous glycine and oral D-cycloserine in healthy humans; those data belong to occupancy and safety discussion, not to a consumer sleep claim.

Established: glycine is an NMDA co-agonist. Speculative: ordinary oral glycine saturates GluN1 sites throughout the human brain in a way that explains longevity or cognition slogans.

11 Sleep

Bannai and Kawai (2012a), reviewing their own work, reported that glycine ingestion before bedtime improved subjective sleep quality in people with insomniac tendencies, and that in rats oral glycine raised plasma and CSF glycine, increased cutaneous blood flow, and lowered core temperature. Bannai, Kawai, Ono, Nakahara and Murakami (2012b) then restricted healthy volunteers to 25% less than usual sleep for three nights and gave 3 g glycine or placebo at bedtime. Visual-analogue fatigue fell; sleepiness trended down; psychomotor vigilance improved. Plasma melatonin and SCN clock-gene expression in a parallel rat experiment were not significantly shifted; some SCN neuropeptides were. The human trial is randomized and placebo-controlled. It is also acute, small, and from the same laboratory that owns the sleep narrative.

Kawai et al. (2015) dissected mechanism in rats. Oral glycine induced NREM sleep and shortened NREM latency while lowering core temperature. Cutaneous blood-flow increases were blocked by NMDA antagonists, not by strychnine. SCN microinjection of glycine or D-serine raised blood flow; SCN ablation abolished the sleep and hypothermic effects. Animal mechanism, strongly supported in the rat: NMDA receptors in the SCN shell, peripheral vasodilatation, heat loss. Not a human receptor occupancy study.

Soh, Cheah and Stout (2024) reviewed glycine administration across eleven physiological systems in adult humans. In healthy people, interventions lasted up to 14 days; sleep was among the reported positives, but those healthy sleep studies were judged small and at high risk of bias. The review asked for larger, longer, better-designed trials before glycine can be treated as a geroprotective sleep intervention. That is the right epistemic class: emerging, modest, incompletely independently replicated.

Table. Sleep trial matrix. Clinical meaning is modest next-day function after restriction, not hypnotic-drug equivalence.

12 Cognition

Cognition appears in this corpus mainly as a GlyNAC secondary endpoint (Kumar et al., 2020, 2021, 2023) or as an NMDA-psychiatry question. Those are not free-glycine cognition trials in healthy aging. Kumar, Osahon and Sekhar (2023) reported that GlyNAC reversed age-associated cognitive decline in old mice with parallel brain GSH and mitochondrial measures. Animal, combination precursor. Human cognition claims for glycine alone remain speculative in this article. Schizophrenia high-dose glycine is a separate literature and is not converted here into a consumer nootropic claim.

Cruz et al. (2008) randomized 74 people with type 2 diabetes (mean fasting glucose 175.5 mg/dl, HbA1c 8%) to 5 g/day glycine or 5 g/day placebo for three months. HbA1c was significantly lower after glycine than after placebo; TNF-receptor I fell. The paper’s title also reports an interferon-γ increase; cytokine packages in a single trial should be read as exploratory. The result is emerging: a real randomized marker change, not a microvascular or event outcome, and not replicated at scale in this corpus.

Díaz-Flores et al. (2013) gave 15 g/day glycine or placebo for three months to 60 volunteers with metabolic syndrome. TBARS fell about 25% versus placebo; superoxide-dismutase specific activity fell about 20%; systolic blood pressure decreased. Oxidative-stress markers and clinic blood pressure are not coronary events. Grade: emerging for those markers.

Low circulating glycine in insulin resistance is strongly supported as association (Adeva-Andany et al., 2018; Alves et al., 2019). Alves et al. (2022) then did the experiment popular writing skips: glycine in drinking water (300 mg/kg during the last six weeks of a high-fat, high-sucrose diet) did not repair liver mitochondria-associated membrane integrity in mice and worsened fasting glycemia and the response to pyruvate, with a higher hepatic oxaloacetate/acetyl-CoA ratio — a gluconeogenic signature. In hepatocytes, 5 mM glycine had supported MAM integrity; the intact obese mouse did not follow the dish. Animal contradiction to a simple repletion story.

Rom et al. (2020, 2022) reported glycine-based treatments that improved NAFLD and atherosclerosis readouts in mice, with glutathione and microbiome involvement. Those are animal disease models, not human event trials. They are listed so the mechanistic literature is not hidden. They are not spent as human cardioprotection.

Table. Metabolic trial matrix. Marker trials remain marker trials.

14 Glutathione-related claims, with the cysteine donor attached

Sekhar et al. (2011a) compared eight elderly and eight younger adults with [ ²H₂]glycine infusions. Elderly RBC glycine was 218 versus 487 μmol/L; cysteine 19.8 versus 26.2 μmol/L; glutathione 1.12 versus 2.08 mmol/L RBCs; fractional synthesis 45.8 versus 83.1%/day. Two weeks of cysteine plus glycine in the elderly raised glutathione concentration 94.6%, fractional synthesis 78.8%, and absolute synthesis 230.9%, and lowered oxidative-stress markers. Sekhar et al. (2011b) found a parallel GSH synthesis defect in uncontrolled type 2 diabetes (GSH 1.65 versus 6.75 μmol/g Hb) that likewise responded to cysteine plus glycine. Nguyen, Hsu, Jahoor and Sekhar (2014) reported improved mitochondrial fuel oxidation and insulin sensitivity in older GSH-deficient HIV-infected men after the same dual precursors.

These studies are strongly supported as dual-precursor GSH restoration in the states tested. They are not glycine-only trials. McCarty et al. (2018) and Razak et al. (2017) are the reviews that blur that line. This article does not.

15 Oxidative stress and inflammation as endpoints

TBARS, F₂-isoprostanes, TNF-RI, and “reactive oxygen metabolites” appear throughout this literature. They are not clinical events. Cruz and Díaz-Flores moved some of them with free glycine. Sekhar moved some of them with two precursors. Inflammation in HIV after cysteine plus glycine is a letter-scale report (Sekhar, Liu and Rice, 2015). The article treats oxidative and inflammatory marker movement as emerging when randomized, and as not interchangeable with disease outcomes in every case.

16 Collagen biology versus collagen products

Li and Wu (2018) remain the load-bearing nutrition review for glycine, proline, and hydroxyproline in collagen synthesis and animal growth. Human collagen-hydrolysate trials — skin elasticity, joint pain — exist outside this core set and are not imported. Free glycine might, in principle, raise substrate for collagen; de Paz-Lugo et al. (2018) saw that in chondrocytes. A human demonstration that grams of the monomer reproduce hydrolysate endpoints has not been presented in the records cited here. Forbidden leap: collagen peptides = free glycine.

Miller et al. (2019), in the NIA Interventions Testing Program, fed genetically heterogeneous UM-HET3 mice a diet with 8% glycine. Lifespan rose a small but statistically significant 4–6% in both males (p = 0.002) and females (p < 0.001), with an increase in maximum lifespan. Pooled across sex, the increase was present at each of three sites (p = 0.01, 0.053, 0.03). Females were lighter; males were not. End-of-life necropsy suggested fewer pulmonary adenocarcinoma deaths (p = 0.03). No incidental pathology of 40 varieties was significantly increased. In the same cohort, TM5441, inulin, and two aspirin doses did not help. The paper frames glycine as support for the idea that modulating dietary amino acids can extend healthy mouse life, against a background in which methionine restriction is the older, larger effect.

That is established as a mouse ITP finding. The diet is 8% glycine by weight, a pharmacological food, not a 3 g capsule. The effect size is modest next to classic methionine restriction. Site p = 0.053 is the honest middle of a three-site replication, not a hidden failure.

18 Methionine, glycine, and the overstated mechanism

Diets low in methionine extend rodent life. Glycine can mitigate methionine toxicity; a small older rat study had suggested glycine itself might extend life, which is why ITP tested it (Miller et al., 2019). Johnson et al. (2023) put GNMT at the centre: glycine as a methyl acceptor that clears SAM and lowers methionine, with sarcosine as the product. That reading makes glycine a methionine-restriction mimic by chemistry, not a collagen pill. It is a coherent plausible mechanism. It is also untested as a human longevity intervention. No randomized glycine or methionine-restriction trial in people has a hard aging endpoint in this corpus.

Maternal low-protein diets in rats raise offspring blood pressure; glycine supplementation of the dam reverses that rise, unlike alanine or urea nitrogen (Jackson et al., 2002). Brawley et al. (2004) found glycine repletion improved maternal mesenteric endothelial function in the same model. Those are developmental-programming experiments, not adult longevity trials, and not human pregnancy-supplementation prescriptions.

19 GlyNAC is a different title

Kumar, Liu, Hsu, Chacko, Minard, Jahoor and Sekhar (2021) gave GlyNAC for 24 weeks to eight older adults in an open-label design; benefits receded after withdrawal. Kumar et al. (2023), NCT01870193, randomized 24 older adults to GlyNAC or isonitrogenous alanine for 16 weeks (n = 12 each); 12 young adults received GlyNAC for two weeks as a reference. The paper reports that GlyNAC, not placebo, improved GSH deficiency, oxidative stress, mitochondrial function, inflammation, insulin resistance, physical function, and aging hallmarks. Sample size is small; endpoints are many; the intervention is two molecules.

Lizzo et al. (2022) randomized healthy older adults to GlyNAC (7.2 g) or placebo. Primary endpoints were the reduced-to-oxidized glutathione ratio and total glutathione. End-of-study GSH-F:GSSG was 12.65 versus 12.49 (p = 0.739); GSH-T 959.6 versus 903.5 mg/L (p = 0.278). Older adults were more oxidized at baseline than a young reference group, so the redox problem the trial aimed at was real. The combination did not move the primary GSH measures. Post-hoc subsets with high MDA and low baseline GSH were reported as more responsive. A failed primary endpoint in an independent RCT is not erased by an in-lab RCT with broader positive secondaries.

Kumar, Osahon and Sekhar (2022) reported that GlyNAC-supplemented C57BL/6J mice lived 24% longer than controls, with improved GSH and mitochondrial indices. Cieslik et al. (2018) found that old-mouse diastolic function improved on NAC+glycine, not on NAC alone. Combination biology can be real. It still is not glycine.

Sekhar (2021, 2022) reviews and a T2D GlyNAC pilot belong to the sibling article SBL-41/SP-GLYNAC. They appear here only as a boundary: this title will not inherit their benefits.

20 Aging claims that remain after the cuts

What survives for free glycine and aging is modest. A heterogeneous-mouse ITP diet study with a 4–6% lifespan increment. A set of dual-precursor human GSH studies that should not be cited as glycine. An independent GlyNAC RCT that missed its GSH primary. Marker trials in diabetes and metabolic syndrome. A sleep series that is not an aging trial. Razak et al. (2017) listed multifarious benefits; listing is not grading. Johnson et al. (2023) is the most careful recent mechanistic review and still does not convert glycine into a human geroprotector.

Table. Aging / species matrix. Species, precursor, and endpoint stay named. The 24% mouse figure is not an ITP glycine result.

Table. Glycine / GlyNAC distinction ledger. Lizzo et al. (2022) is the independent GlyNAC RCT that missed primary GSH endpoints.

22 Adversarial resolutions

Q1. Conditionally limiting in normal humans? Limiting in named high-demand or experimentally depleted states (benzoate, malnutrition catch-up, late pregnancy, pregnant adolescents). Not established for ordinary adult eating. Gibson et al. (2002) maintained glycine flux on 0.75 g protein kg⁻¹ d⁻¹. Jackson et al. (2004) showed GSH synthesis can still strain on that intake. Those are compatible if the readout is named. Grade for general adult conditional deficiency: speculative.

Q2. Sleep replicated and clinically meaningful? Replicated inside a small same-lab series; Soh et al. (2024) flag bias and size. Meaningful as modest next-day fatigue and vigilance after restriction. Not a hypnotic, not an insomnia-disease trial. Grade: emerging, modest.

Q3. Rodent longevity translate? ITP glycine is a real 4–6% UM-HET3 result. Translation to human lifespan is speculative and presently untested. GlyNAC mouse +24% is out of scope as a glycine result.

Q4. Glutathione independent of NAC? Not in the human aging-GSH experiments that dominate the claim. Lizzo et al. (2022) further weakens a simple GlyNAC-GSH slogan. Open gap: glycine-only GSH kinetics in older adults at Sekhar doses.

Q5. Methionine/glycine arguments overstated? Overstated as human outcomes. Not overstated as rodent chemistry (GNMT, Met restriction, ITP). Meléndez-Hevia is overstated when rewritten as a demonstration that shoppers are collagen-deficient.

23 What this title is not

It is not the GlyNAC article (SBL-41/SP-GLYNAC), not the glutathione article, not a collagen-hydrolysate review, not creatine monohydrate, not a schizophrenia glycine-augmentation manual, and not medical advice. Amino-acid nutrition as a class is covered in SBL-41/SP-AMINO-ACID-NUTRITION; glycine is the monomer that that title flagged as an open quantitative question. Project 06 was searched read-only and was not used as a scientific authority. Peptide News 07 hits were discovery noise (often the word glycine inside unrelated peptides) and are not cited.

24 Evidence that would change the grades

An adequately powered, independently conducted RCT of bedtime glycine for chronic insomnia with polysomnography and daytime function would move Q2. A glycine-only (no cysteine donor) GSH kinetic study in older adults would move Q4. A human methionine-restriction or high-glycine diet trial with a pre-specified hard aging or cardiometabolic event endpoint would move Q3 and Q5. A clinical survey that measures collagen synthesis against glycine intake, rather than a stoichiometric model, would move Q1. Until those exist, the grades above stand.

Peer-reviewed identifiers were taken from NCBI records retrieved on 20 August 2026. In-prose citations are author–year. The numbered list is sorted by first-author surname. Study type is named in the reporting sentence. Animal and cell findings are not rewritten as human outcomes. GlyNAC is labelled a two-precursor experiment wherever it appears. Project 06 and 07 Peptide News were searched as read-only discovery layers; they are not cited as scientific authorities. HOUSE_STYLE section 8a (peptide bioregulators) does not apply.

26 Scope relative to sibling articles

This title is the free-glycine article in the SBL-41 series. It is not SBL-41/SP-GLYNAC, which owns the combination of glycine plus N-acetylcysteine. It is not the glutathione article, not a collagen-hydrolysate review, not creatine monohydrate, and not the amino-acid nutrition framework article, which orients the larger class. Schizophrenia high-dose glycine and GlyT1-inhibitor trials are mentioned only to forbid their use as sleep or longevity evidence.

27 References

Fifty peer-reviewed records below were verified against NCBI. In-prose citations use author–year; two 2011 Sekhar papers and two 2012 Bannai papers are distinguished in the text by journal.