Humanin: Mitochondrial-Derived Peptide Evidence Review

Humanin is a mitochondrial-derived peptide (MDP) first identified in 2001 inside the cDNA of neurons that had survived in an Alzheimer’s disease patient’s brain — its name a reflection of the hope that it encoded a “rescue factor” keeping those cells alive ^[1]. Unlike most peptides encoded in the cell nucleus, humanin is translated from a short open reading frame embedded in the 16S ribosomal RNA gene (MT-RNR2) of the mitochondrial genome ^[1], placing it in the same emerging MDP family as MOTS-c, which is encoded in the adjacent 12S rRNA region. Both peptides represent a newly characterized class of mitochondria-to-nucleus signals that researchers believe help explain why cellular communication degrades with aging. That genuinely interesting biology has also attracted a research-peptide market selling humanin and its more potent analog HNG as injectable compounds promoted for fat loss, insulin sensitivity, neuroprotection, and longevity. This article reviews all the evidence honestly: what is established in cell and animal models, what the human observational data shows and does not show, and what is plainly absent — specifically, any randomized trial of administered humanin in people for any metabolic, weight-loss, or longevity endpoint.

What humanin is — a peptide encoded inside the mitochondrial genome

Humanin was discovered when Hashimoto and colleagues screened a cDNA library from preserved neurons in the occipital visual cortex of an Alzheimer’s disease patient, looking for factors that could rescue neurons from familial-AD-gene-driven and amyloid-beta-driven cell death ^[1]. The peptide they isolated is 21 amino acids long and encoded within a short open reading frame in the MT-RNR2 locus — a region of the mitochondrial genome that canonically codes for 16S ribosomal RNA and had long been assumed to carry no protein-coding function. This makes humanin one of the founding members of the mitochondrial-derived peptide class: small peptides whose genes have historically been overlooked because they reside inside RNA-coding regions.

The most studied close relative is MOTS-c, encoded in the 12S rRNA region of the same mitochondrial genome. Both peptides are cytoprotective signaling molecules, both decline with aging, and both are under investigation as potential therapeutic targets ^[4][6]. Their receptor systems, primary tissue effects, and bodies of evidence differ, however. Humanin was initially characterized as neuroprotective and anti-apoptotic in Alzheimer’s disease cell models ^[1][2]; MOTS-c was initially characterized for metabolic and exercise-mimetic effects in mice. Today both are described as having broad cytoprotective and metabolic functions, but that convergence rests on preclinical data and should not be read as clinical equivalence or therapeutic interchangeability.

How humanin works: three receptor pathways

Humanin’s effects are mediated through at least two distinct receptor systems characterized in preclinical models. First, it binds the formyl peptide receptor 2 (FPR2, also called FPRL-1), a G-protein-coupled receptor expressed on neurons, immune cells, and other tissues ^[8]. A 2022 cryo-EM structural study resolved how humanin occupies the FPR2 binding pocket and competes with amyloid-beta peptide (Aβ42) to suppress neurotoxic signaling — a mechanistic insight that explains the neuroprotective selectivity of the peptide ^[8]. Second, humanin independently signals through a tripartite cytokine-receptor complex consisting of CNTF receptor alpha, WSX-1, and gp130 ^[2]. This trimeric assembly is shared with ciliary neurotrophic factor and neuropoietic cytokines, placing humanin within a signaling axis governing neuronal survival, energy homeostasis, and inflammatory responses.

A third mechanistic arm concerns insulin sensitivity. Muzumdar and colleagues demonstrated in mice that humanin administered centrally — directly into the cerebral ventricles — acted as a central regulator of peripheral insulin action, suppressing hepatic glucose production and improving peripheral glucose uptake through a neuroendocrine pathway distinct from the peripheral receptor systems above ^[3]. This central-to-peripheral metabolic effect is pharmacologically important context, but it derives entirely from animal (mouse and rodent) experiments. No equivalent human dosing or pharmacodynamic study has been conducted.

The evidence landscape — preclinical, observational, and the gap

Understanding humanin requires clearly separating three evidence tiers that are routinely collapsed together in research-peptide marketing:

• Preclinical (cell and animal) evidence — strong. Humanin’s neuroprotective effect was established in the 2001 discovery paper using neuronal cell-death models and validated in mouse models of familial Alzheimer’s disease ^[1]. The receptor mechanism through CNTFRα/WSX-1/gp130 was characterized in cell culture ^[2], and the FPR2 binding structure was resolved by cryo-EM ^[8]. Insulin-sensitizing central effects were demonstrated in mice ^[3]. A 2024 review covers the antidiabetic and metabolic preclinical evidence for humanin and related MDPs comprehensively ^[9]. These are real, peer-reviewed results — in cells and mice.
• Human observational (biomarker) evidence — real but not interventional. Circulating humanin has been measured in people, and the findings are consistent: it declines with age ^[4], rises in response to exercise — acute high-intensity interval training increases humanin in both plasma and skeletal muscle biopsies from healthy men ^[7] — and associates with healthspan and cognitive-age markers in cross-sectional human studies ^[5][6]. Studies including longevity-associated cohorts have found that higher circulating humanin correlates with longevity phenotypes, including centenarian-associated subjects ^[6]. These are association data: observational, correlational, and not establishing causal benefit of administered humanin.
• Human interventional (administered humanin) evidence — absent. A PubMed search returns zero randomized controlled trials, zero open-label human dosing studies, and zero pharmacokinetic studies of exogenously administered humanin or HNG. No approved product exists. No active IND (Investigational New Drug application) for humanin appears in public FDA records. No human dose-finding or safety trial has been published.

Human observational data — what it shows and what it doesn't

The human observational data on circulating humanin is meaningful and should be characterized accurately. Cobb and colleagues (2016) measured humanin in a cross-sectional human cohort alongside MOTS-c and found that both MDPs decline with age and that humanin levels correlate with markers of apoptosis, insulin sensitivity, and inflammation in people ^[4]. This was one of the first demonstrations that humanin circulates measurably in human plasma and changes with biological aging.

Yen and colleagues (2018) extended this to cognition: in a study combining mouse experiments with human cross-sectional data, they found that higher humanin levels were associated with better cognitive-age metrics in human participants ^[5]. The mouse arm showed that administered humanin reduced age-related cognitive decline; the human arm was observational. The same group’s 2020 study, drawing on multiple organisms including C. elegans, mouse, and longevity-associated human cohorts, described humanin as a regulator of lifespan and healthspan ^[6] — again with the human component based on circulating levels in centenarian-associated subjects rather than a treatment intervention.

The exercise study by Woodhead and colleagues (2020) provides the clearest human physiology data: acute HIIT increased humanin in both plasma and skeletal muscle biopsies from healthy men ^[7]. This confirms humanin as an exercise-responsive biomarker in living people — consistent with the broader MDP literature and with the framing of mitochondrial-derived peptides as metabolic signals that respond to physical stress. But confirming that exercise raises your circulating humanin is not the same as showing that injecting humanin delivers the benefits of exercise.

Evidence summary table

Humanin in context — MOTS-c, epitalon, and NAD+ pathways

Humanin and MOTS-c are frequently grouped as the two founding mitochondrial-derived peptides, and the comparison is useful. MOTS-c (12S rRNA region) is primarily characterized for its metabolic and exercise-mimetic effects in mice — protecting against diet-induced obesity and improving insulin sensitivity. Humanin (16S rRNA region, MT-RNR2) was initially characterized for neuroprotective and anti-apoptotic activity, with metabolic effects described secondarily ^[1][3]. Both lack human interventional data. Both circulate as biomarkers of aging and respond to exercise in humans. Neither has an FDA-approved form or a published human dose-finding safety study.

Epitalon, a synthetic tetrapeptide derived from the pineal gland extract epithalamin, occupies a different mechanistic space — telomerase activation and melatonin regulation — but faces the same evidence-level limitation: strong preclinical literature, some human observational data, no FDA approval, and no large human randomized trial for longevity outcomes. Similarly, NAD+ precursors (NMN, NR) target mitochondrial energy substrate rather than mitochondrial-derived signaling peptides, but they share the same translational gap: compelling biology, most human trials limited to short-term biomarker changes, and no large randomized trial evidence proving supplementation extends human healthspan. The longevity-biology landscape shares this common structure: genuine mechanistic science, genuine human biomarker associations, and a consistent absence of large randomized trial evidence proving that supplementation or injection delivers clinical longevity outcomes.

HNG, humanin analogs, and regulatory status

Researchers have developed humanin analogs with improved potency. The most studied is HNG (Humanin Glycine), in which serine at position 14 is substituted with glycine — a change that increases biological activity by approximately 1,000-fold in cell-based assays ^[9]. HNG is the form most commonly sold alongside native humanin in the research-peptide market. Neither humanin nor HNG has an FDA approval for any human indication. No active IND (Investigational New Drug application) for either compound appears in publicly available FDA records, meaning no human pharmacokinetic, safety, or efficacy trial has been formally authorized in the United States.

In the current regulatory environment, humanin and HNG are sold by research-chemical suppliers under disclaimers stating the products are “for research use only, not for human consumption.” Unlike compounds that hold a prior FDA approval — sermorelin, for example, has a prior-NDA pathway that preserves some compounding eligibility — humanin has no prior NDA to anchor any compounding framework. Any product currently injected by users is an unvalidated preparation of unknown purity, potency, and sterility. The pharmacokinetics of subcutaneously injected humanin in humans, the duration of receptor engagement, the risk of receptor desensitization, and the effects of chronic apoptosis-pathway modulation in healthy adults are all uncharacterized.

The honest clinical bottom line

Humanin is a genuine and important piece of mitochondrial biology. Its discovery filled a conceptual gap — confirming that protein-coding sequences can reside inside the mitochondrial genome’s RNA loci — and its receptor pharmacology, cytoprotective cell effects, and circulating behavior in humans are all real and reproducible. The observational human data showing that levels decline with age and associate with favorable aging and cognitive metrics are consistent across independent research groups ^[4][5][6][7].

What does not exist is any published randomized controlled trial, open-label dosing study, or human pharmacokinetics study of administered humanin or HNG. There is no FDA-approved form, no IND on record, and no human safety data for any route of administration. Injecting humanin purchased from a research-chemical supplier means dosing an unknown-purity preparation via an unknown pharmacokinetic profile to engage receptor systems — including the apoptosis-signaling axes ^[1][2] — that have never been characterized after exogenous peptide administration in a living human. The preclinical and observational biology is compelling and warrants continued clinical research; the evidence basis for injecting it today is nonexistent.