Mitochondrial-derived peptide

Humanin

In mice, the humanin analogue HNG had a plasma half-life of approximately 30 minutes after injection into the abdominal cavity; the app lists 0.5 hours for humanin.1

Emerging evidence (the app's grade, not a conclusion of the sources)Get the app

Level after one dose

Humanin after one dose: 50% at ~0.5 h, 25% at 1 h, 12.5% at 1.5 h.

Drawn from the half-life in the app's model, ~0.5 h. The model's figures

Humanin is a 24-amino acid peptide transcribed from an open reading frame within the mitochondrial 16S rRNA.1 Chin and colleagues write that it was discovered in 2001 and cloned independently by 3 separate groups, with Hashimoto and colleagues first identifying it by screening a cDNA library from the brain of an Alzheimer's disease patient.1 A 2023 systematic review by Coradduzza and colleagues describes it as a peptide "playing a cytoprotective role to preserve mitochondrial function and cell viability under stressful and senescence conditions".2 The same review states that "Both senescence and humanin-related mechanisms are complex processes that have not been fully clarified yet."2

The app places humanin in the metabolic category as a mitochondrial peptide, and its own note calls the data almost entirely preclinical or from animals. The app grades the evidence as emerging. The three rodent papers described below (Chin, Muzumdar, Kuliawat) each report National Institutes of Health grants among their funding, and the Chin paper also names a Glenn Foundation Award and a Clinical and Translational Science Institute grant.1,3,4

2

How it works

Chin and colleagues write that humanin was shown to protect neurons from amyloid β by binding to a cytokine-receptor complex composed of glycoprotein 130, ciliary neurotrophic factor receptor α, and IL-27 receptor (WSX-1); the passage does not state the species.1 Chin and colleagues state that humanin "also binds with high affinity to IGFBP-3".1 In a mouse experiment, IGFBP-3 binding changed how fast an analogue cleared from plasma, which is covered under pharmacokinetics below.1 Two analogues appear throughout the animal work. HNGF6A carries a second substitution (Ala for Phe at position 6) and "does not bind to IGFBP-3".1 The review by Coradduzza and colleagues tabulates animal findings for the HNG analogue: in middle-aged APPswe/PS1dE9 mice it reports that "Treatment with HNG significantly improves spatial learning and memory deficits, reduces Aβ plaque accumulation and insoluble Aβ concentration".2 It also tabulates a pig study in which "A single dose of HNG, administered at the time of reperfusion, diminishes cardiac infarction size and inhibits cardiomyocyte apoptosis".2 These are animal results from an analogue, not results in people and not results for unmodified humanin.

3

Half-life and pharmacokinetics

Values from the Milligram app's model.

Half-life
~0.5 h
Bioavailability
95%
Typical dose
1–5 mg/day
Route
SubQ
Time to peak
~0.5 h

The only pharmacokinetic study of humanin identified here used rodents and analogues, so the table sets the app's figures beside the rodent figures; no human pharmacokinetic figure was identified.

Table 3.1Half-life and pharmacokinetics4 rows
ParameterApp's figureReported in the sources
Half-life~0.5 h (30 minutes)Mice, HNG injected into the abdominal cavity at 2 mg/kg: "a short half-life of approximately 30 minutes"; rats, HNG at 5 mg/kg: "appears to be greater than 4 hours"; HNGF6A in mice, and HNG in mice lacking IGFBP-3, cleared more slowly than HNG in wild-type mice1
Time to peak~0.5 hMice: peak plasma levels at 10 minutes after injection into the abdominal cavity; rats: "The peak plasma HN level was 14.6 ng/mL at 4 hours"1
Bioavailability95%The rodent paper reports plasma levels after injection into the abdominal cavity and gives no bioavailability figure1
RouteSubQThe rodent study used intraperitoneal injection; other rodent work cited below used intracerebroventricular and intravenous routes1,3

Deeper analysis: Chin and colleagues, Endocrinology 2013

Chin and colleagues at the University of Southern California and collaborating centres reported the study in Endocrinology in 2013.1 The design used 11-week-old male wild-type mice and mice lacking IGFBP-3, each given 2 mg/kg of an analogue by injection into the abdominal cavity, with blood taken at 10, 30, 60, 100 and 180 minutes (n = 5 per time point).1 A second experiment gave adult male Sprague Dawley rats HNG and sampled plasma and tissue at 30 minutes, 90 minutes, 4 hours and 24 hours (n = 3 to 5 per time point).1 Peak plasma levels in wild-type mice were "139 ng/mL ± 3.8 and 146 ng/mL ± 27.3" for HNG and HNGF6A, against a significantly lower peak in IGFBP-3-deficient mice given HNG (63 ng/mL ± 11.5, P < .05).1 In rats, mean plasma levels were 6.3, 4.7, 8.5 and 2.7 ng/mL at the four sampling times, and the authors report levels "highest in plasma, present in liver, and undetectable in brain or heart".1 The authors' own caveats include that "the etiology of these interspecies differences remains unclear" and that the rats "did receive a higher dose of HN treatment (5 mg/kg) compared with mice (2 mg/kg)".1 The rat figure legends give 4 mg/kg rather than the 5 mg/kg stated in the methods.1 The assay was a laboratory-made ELISA, and the authors report that "the measured HNG and HNGF6A were approximately 16-fold higher than their actual concentration".1 They also write that "it is also reasonable to assume that the biological half-life of HN is longer than the kinetic half-life".1 Funding came from a Glenn Foundation Award and National Institutes of Health grants.1 The disclosure reads: "Dr Cobb is a stock holder and employee of CohBar, Inc. Dr Cohen is a consultant and stock holder of CohBar, Inc."1

4

Human studies

No controlled trial of humanin in people was identified in PubMed or in the one ClinicalTrials.gov record read in full; the other 6 records returned were not read.5 The ClinicalTrials.gov search returned 7 studies; the one read in full is an observational cohort study of blood and heart-tissue humanin levels, with no dosing, and the other 6 were not read.5

Table 4.1Human studies2 rows
Study and yearDesignPeopleDose and lengthMain result
Muzumdar and colleagues, PLoS ONE 2009 (human plasma measurement)3Plasma humanin measured by ELISA across age in humans, and in young and old mice; the paper's other experiments were in rodents3Humans across ages; the number and ages are not stated in the text quoted here3None given; no humanin was administered to people3"circulating levels of HN were decreased with age in humans and mice"3
University of Tartu, NCT03431844, registry entry (completed 2019)5Prospective observational cohort; blood samples before the operation, one hour after aortic cross-clamp removal and the next morning, plus a right atrial appendage tissue sample5106 adults with ischemic heart disease and a planned on-pump coronary artery bypass graft5No intervention; follow-up of 30 days for mortality, myocardial infarction, acute kidney injury and stroke5The entry shows "Has Results: undefined"; no result is quoted here5

Deeper analysis: Muzumdar and colleagues, PLoS ONE 2009

Muzumdar and colleagues reported mostly rat and mouse experiments, with the human plasma measurement as one part.3 In the authors' words, "Continuous infusion of HN intra-cerebro-ventricularly significantly improved overall insulin sensitivity", and the effect was "negated by co-inhibition of hypothalamic STAT-3".3 In Zucker diabetic fatty rats, a single 100 ug HNGF6A bolus (n = 5), described as an intravenous injection in the results and as infused into the carotid artery in the methods, against 200 uL saline (n = 4) was followed by glucose that "decreased by nearly 50%" while saline-treated rats were "unaltered".3 The authors state that humanin was not seen to act directly on liver cells: this was "consistent with the lack of a direct effect of HN on primary hepatocytes".3 The funding statement lists National Institutes of Health grants and adds that "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript".3 The competing-interests statement says that "The authors have declared that no competing interests exist".3 Two of its authors, Cobb LJ and Cohen P, are the authors whose CohBar, Inc. ties are disclosed in the 2013 Chin paper; the 2009 statement lists no competing interests.3,1 No separate limitations paragraph in the full text was located for quotation, and the animal results come from humanin or analogues given into the brain ventricle or intravenously, not from injection under the skin.3

Deeper analysis: Kuliawat and colleagues, FASEB Journal 2013

Kuliawat and colleagues, from Albert Einstein College of Medicine, studied HNGF6A and insulin secretion in rats, isolated mouse islets and a mouse beta cell line.4 Their report is that "Sprague-Dawley rats (3 mo old) that received HNGF6A required a significantly higher glucose infusion rate and demonstrated higher insulin levels during hyperglycemic clamps compared to saline controls".4 The same paper reports that HNGF6A "did not significantly increase first-phase insulin secretion in vivo in Sprague-Dawley rats" (P=0.34), and that in C57BL6/N mice the increase in first-phase insulin levels in the HNGF6A group over the 30-min period was not significant (P=0.51).4 The disclosure states "N.B. is a consultant and a shareholder of CoHBar."4

5

Doses reported in studies and on labels

No label and no human dosing study of humanin was identified; the table lists only what the rodent papers and the app give.

Table 5.1Doses reported in studies and on labels6 rows
SourceSpecies, compound and routeDoseLength and comparator
Chin 20131Male mice, HNG or HNGF6A, injection into the abdominal cavity2 mg/kgSingle injection, blood at 10 to 180 minutes, n = 5 per time point; vehicle-injected controls
Chin 20131Male Sprague Dawley rats, HNG, injection into the abdominal cavity5 mg/kg in the methods; 4 mg/kg in the figure legendsSingle injection, samples to 24 hours, n = 3 to 5 per time point; vehicle-injected controls
Chin 2013, citing 14 rodent papers1Rodent disease models, analogues, injection into the abdominal cavity, brain ventricle, vein or nose4 μg/kg/d to 4 mg/kg/d, "typically given once daily or every other day"Not stated in a single comparator
Muzumdar 20093Zucker diabetic fatty rats, HNGF6A, bolus described as an intravenous injection (results) and as infused into the carotid artery (methods)100 ugSingle dose, 4 hours of sampling; against 200 uL saline, n = 5 and n = 4
Kuliawat 20134C57BL6/N mice, HNGF6A, intravenous2 mg/kgSingle dose 10 minutes before a glucose load, n = 6 per group; against vehicle
The appHumanin, subcutaneous; no label or human study identified1–5 mg/dayNot applicable; the app gives no comparator
6

Bloodwork and markers

The app lists no health markers to monitor for humanin. In mice, Chin and colleagues report that after HNG injection "plasma IGFBP-3 levels in WT mice decreased over time, falling to 80% of baseline after 180 minutes", and that IGF-1 "decreased by approximately 30% in WT mice injected with HNG" while it stayed unchanged after HNGF6A.1 Muzumdar and colleagues report plasma humanin that declines with age in people and mice, measured by ELISA.3

7

What is not known

No human pharmacokinetic study, dose-finding study or controlled efficacy study of humanin was identified in the searches described above. A DailyMed search for "humanin" returned 0 labels on 16 November 2026, and the searches did not examine the WADA Prohibited List or non-US regulators. The app's 95% bioavailability and 1–5 mg/day dose have no matching figure in the rodent papers cited above, which report intraperitoneal, intravenous and brain-ventricle routes in rodents. Chin and colleagues write that the cause of the mouse and rat differences "remains unclear", and the rat half-life was given only as greater than 4 hours.1 Several authors of the rodent pharmacokinetic and related papers disclose a consulting or shareholder tie to CohBar, Inc.1,4

8

Frequently asked questions

What is the half-life of humanin?

In mice, the analogue HNG had a half-life of approximately 30 minutes, and the app lists 0.5 hours.1 In rats, the same paper says the half-life "appears to be greater than 4 hours".1 No human half-life figure was identified in the sources used.

How long does humanin take to reach its peak level?

In mice given an analogue by injection into the abdominal cavity, peak plasma levels came at 10 minutes, and the app lists about 0.5 hours.1 In rats the peak plasma level of 14.6 ng/mL came at 4 hours.1

What is humanin?

Humanin is a 24-amino acid mitochondrial-derived peptide.1 It was discovered in 2001.1

What is HNG?

HNG is an analogue of humanin made by substituting glycine for serine at position 14.1 Chin and colleagues state that it has "substantially greater cytoprotective action than HN".1

Does humanin decline with age?

Muzumdar and colleagues report that circulating humanin was lower with age in humans and mice.3 The paper's text quoted here gives no number of people.3

Has humanin been tested in people?

No controlled trial of humanin in people was identified in PubMed or in the one ClinicalTrials.gov record read in full; the other 6 records returned were not read.5 The one registry entry read in full, NCT03431844, is an observational cohort of 106 adults having coronary artery bypass surgery.5

Does humanin have a drug label?

A DailyMed search for "humanin" returned 0 labels on 16 November 2026.

Does humanin reach the brain or heart?

In rats given HNG by injection into the abdominal cavity, humanin was "undetectable in brain or heart" at the times tested.1 The authors suggest it may act "on intermediary receptors or penetrates in low concentrations at a later time in these organs".1

10

References

  1. Chin YP, Keni J, Wan J, et al. Pharmacokinetics and tissue distribution of humanin and its analogues in male rodents. Endocrinology 2013;154(10):3739-44. PMID 23836030
  2. Coradduzza D, Congiargiu A, Chen Z, et al. Humanin and its pathophysiological roles in aging: a systematic review. Biology (Basel) 2023;12(4):558. PMID 37106758
  3. Muzumdar RH, Huffman DM, Atzmon G, et al. Humanin: a novel central regulator of peripheral insulin action. PLoS One 2009;4(7):e6334. PMID 19623253
  4. Kuliawat R, Klein L, Gong Z, et al. Potent humanin analog increases glucose-stimulated insulin secretion through enhanced metabolism in the β cell. FASEB J 2013;27(12):4890-8. PMID 23995290
  5. University of Tartu. Humanin isoforms in cardiac muscle and blood plasma and major complications after cardiac operation. ClinicalTrials.gov NCT03431844. NCT03431844