Humanin — a mitochondrial-derived peptide (MDP)

General Characteristics

Humanin (HN) is a short mitochondrial-derived peptide (MDP) 24 amino acid residues long (sequence MAPRGFSCLLLLTSEIDLPVKRRA; empirical formula approximately C119H204N34O32S2, average molecular mass ≈2687 Da). It is encoded by a small open reading frame (sORF) nested within the region of the mitochondrial 16S ribosomal RNA gene (MT-RNR2), that is, the peptide originates not from a classical protein-coding nuclear locus, but from a region of the mitochondrial genome that for decades was considered purely structural (rRNA, with no translational output of its own). Humanin was the first identified MDP and remains the prototype of an entire class of short peptides that the cell apparently uses as signals for mitochondrial-nuclear and intercellular communication ^[1][4]. It is essential not to confuse humanin with another well-known MDP, MOTS-c: this is encoded by an entirely different region of the mitochondrial genome (12S rRNA, gene MT-RNR1), has a different sequence, and a different leading mechanism of action (predominantly AMPK-mediated metabolic signaling via de novo purine synthesis) ^[17]. Both peptides belong to the same conceptual class of MDPs, but they are distinct molecules with different loci, structures, and pharmacology, and conflating them in a research context is incorrect.

In addition to the mitochondrial MT-RNR2 gene, at least a dozen nearly identical homologous sequences have been described in the human nuclear genome, paralogs MTRNR2L1-MTRNR2L12, which arose from the insertion of mitochondrial DNA fragments into the nuclear genome (so-called NUMTs, nuclear mitochondrial DNA segments) over the course of evolution. The functional status of most of these paralogs has not been conclusively established, but for at least one of them (designated HN1) the ability to bind the pro-apoptotic protein Bax and suppress apoptosis in a manner analogous to mitochondrially encoded humanin has been experimentally confirmed, meaning cytoprotective activity is likely duplicated across several genetic loci simultaneously, which complicates the interpretation of knockout and overexpression experiments in model systems. This material is strictly for reference and research purposes (research-use-only, RUO): it describes chemistry, molecular mechanisms, and published research findings, and contains no recommendations regarding use in humans; no established human therapeutic dose exists.

History and Discovery Context

Humanin was discovered in 2001 by Japanese researchers led by Y. Hashimoto in the laboratory of Ikuo Nishimoto (Keio University, Tokyo), during a functional screen of a cDNA library for factors capable of protecting neurons from death caused by familial Alzheimer's disease genes and β-amyloid. The library was derived from preserved occipital lobe tissue of an Alzheimer's disease patient, a region relatively less affected by the pathological process. Within this library, the researchers identified a clone that rescued neuronal cells from apoptosis induced by mutant presenilin-1/2, the amyloid precursor protein (specifically the V642I-APP variant), and Aβ itself, and named it "humanin" ^[1]. Notably, humanin's protective action proved relatively specific: in the same experimental systems, it did not rescue cells from death modeled by Huntington's disease or amyotrophic lateral sclerosis genes, indicating not a universal anti-apoptotic "panacea," but a somewhat context-specific survival factor ^[1].

Initially, the humanin sequence was assumed to be the product of a nuclear gene, since it was identified in a cDNA library. Further analysis showed that it corresponds with high precision to a region of the mitochondrial 16S rRNA gene (MT-RNR2). This became the first direct evidence that the mitochondrial genome, beyond the classical respiratory chain proteins and structural RNAs, is capable of encoding biologically active secreted peptides. This discovery made humanin the first example of an MDP and launched a new field of research into mitochondrial-nuclear signaling. The following decade of work by the groups of P. Cohen and C. Lee (University of Southern California) shaped the broader concept of mitochondrial-derived peptides, in which humanin emerged as the "harbinger" of an entire family ^[4][11]: in 2015, this same research tradition led to the discovery of MOTS-c ^[17], and in 2016, to the systematic description of six related peptides, SHLP1–6 ^[8]. In parallel, it was established that some of the homologous sequences are integrated into the nuclear genome as NUMT paralogs (MTRNR2L1–12), adding an additional layer of complexity to the discovery history of humanin regarding its origin and expression regulation.

Structure and Physicochemistry

Canonical humanin is a linear 24-amino-acid peptide with a cluster of hydrophobic leucines in the central portion of the sequence (…CLLLL…), which gives the molecule a propensity for self-association and formation of an amphipathic structure, as well as two cysteine residues and a cluster of basic arginines at the C-terminus. There are two natural lengths of the peptide, determined by the site of translation: if translation occurs directly within the mitochondrion (on mitochondrial ribosomes), the product is 21 amino acid residues long; if translation occurs in the cytoplasm (from an exported mitochondrial transcript or from one of the nuclear NUMT paralogs), the full-length 24-amino-acid form is produced, which is the form featured in the vast majority of published studies ^[1]. This duality of translation site is a rare example of how a single peptide sequence can be produced by two cellular compartments simultaneously.

Position 14 (serine) proved to be a critical point for activity based on structure-function studies: the S14G substitution yields the synthetic analog [Gly14]-humanin (HNG), which is approximately 1000 times more potent than native humanin in standard biological assays ^[13]. The active core of the peptide has been mapped to residues 3–19, with position 14 as the determinant of potency ^[13]. The solution NMR structure of HNG showed that the peptide forms an α-helical region from Phe6 to Thr13, while the Ser14→Gly14 substitution disrupts the helix and confers increased conformational mobility and disorder on the C-terminal portion of the molecule compared to the wild type. Paradoxically, it is precisely this structural destabilization that correlates with the sharp increase in biological activity ^[13]. Separately, a chaperone-like activity has been described for HNS14G and the related peptide SHLP2: both are capable of binding and stabilizing partially unfolded client proteins in vitro via a mechanism reminiscent of small heat-shock proteins, hinting at an additional, receptor-independent proteostatic level of action that complements the classical anti-apoptotic pathways ^[22].

Molecular Mechanism

Humanin is described primarily as a cytoprotective, anti-apoptotic peptide with several parallel, partially independent mechanisms of action: both intracellular and receptor-mediated.

At the intracellular level, humanin directly binds the pro-apoptotic Bcl-2 family protein Bax and prevents its conformational activation and translocation from the cytosol to the outer mitochondrial membrane; this blocks the subsequent release of cytochrome c and initiation of the caspase cascade ^[2]. Suppression of endogenous humanin expression by small interfering RNA sensitizes cells to Bax-mediated death and enhances Bax translocation to membranes, direct evidence that endogenous humanin functions as a physiological buffer of Bax activity, not merely as a pharmacological artifact of excess dosing ^[2]. Later work expanded this picture: humanin and HNG inhibit the membrane association and oligomerization of Bax, as well as similar behavior of the related pro-apoptotic protein Bid, reducing mitochondrial outer membrane permeabilization (MOMP) through a composite mechanism rather than a single target ^[15].

A second intracellular partner is IGFBP-3 (insulin-like growth factor-binding protein 3), identified via yeast two-hybrid screening as a specific, high-affinity binding partner of humanin; the interaction has been mapped to the C-terminal heparin/glycosaminoglycan-binding domain of IGFBP-3 and does not alter IGFBP-3's ability to bind insulin-like growth factor-1 itself, meaning humanin's action at this node is IGF-independent ^[3]. Functionally, humanin blocks the pro-apoptotic activity of IGFBP-3 in glioblastoma cells, and the complex's interaction has been confirmed by co-immunoprecipitation in vivo from mouse testicular tissue ^[3].

For extracellular (receptor-mediated) signaling, at least two distinct systems have been described. The first involves the G-protein-coupled chemotactic receptors FPRL1 (formyl peptide receptor-like 1, now FPR2) and FPRL2 (FPR3): humanin, and especially its N-formylated form, is a potent ligand for these receptors and induces chemotaxis of cells expressing them; N-formylated humanin acts on FPRL1 with a significantly lower EC50 than the non-formylated form ^[14]. The second system is the heterotrimeric cytokine receptor complex CNTFRα/WSX-1(IL-27Rα)/gp130: humanin induces heterooligomerization of these three subunits on the surface of neuronal cells, triggering phosphorylation and activation of STAT3, a signal necessary for the peptide's neuroprotective effect; an alternatively spliced soluble isoform of WSX-1 (sWSX-1) has also been described, which is itself capable of mediating anti-Alzheimer's activity in neurons, pointing to a possible alternative receptor complex configuration ^[9]. Together, these intracellular and receptor-mediated pathways align with humanin's role as a multilevel buffer of cellular stress, combining direct inhibition of the apoptotic machinery with active pro-survival signaling through STAT3 ^[6].

Signaling and Downstream Effects

At the whole-organism level, humanin behaves more like a systemic metabolic and cytoprotective signal than a narrowly neuronal factor. In rodents, central (intracerebroventricular, ICV) administration of humanin increased overall insulin sensitivity, and this effect depended on activation of STAT3 in the hypothalamus: pharmacological inhibition of hypothalamic STAT3 abolished the peptide's insulin-sensitizing action. Peripheral (intravenous) administration of potent humanin analogs reproduced the effect of central administration, showing that for systemic metabolic action the peptide need not act exclusively through the central nervous system ^[5]. Humanin levels in the hypothalamus, skeletal muscle, and cortex declined with age in rodents, and circulating levels declined with age in both humans and mice, giving grounds to describe humanin as a "central regulator of peripheral insulin action" ^[5].

In pancreatic β-cells, humanin suppressed apoptosis induced by serum starvation and pro-inflammatory cytokines via a STAT3-dependent mechanism in the NIT-1 cell line and in primary islets; specific inhibition of STAT3 completely abolished this protective effect, confirming the causal role of the signaling pathway. In a NOD-mouse model (spontaneous autoimmune type 1 diabetes), six weeks of humanin administration normalized glucose tolerance and was accompanied by reduced lymphocytic infiltration and reduced severity of insulitis in pancreatic tissue, while twenty weeks of administration delayed and attenuated the development of overt diabetes ^[9]. Separately, the potent analog HNGF6A directly enhanced glucose-stimulated insulin secretion (GSIS) in isolated islets and β-cell lines, an effect that was independent of K-ATP channels and instead was linked to enhanced glucose metabolism within the β-cell itself, i.e., acting through a mechanism distinct from classical sulfonylurea or incretin pathways ^[16].

In the human vascular wall, humanin is expressed in the endothelial layer of internal thoracic arteries, atherosclerotically altered coronary arteries, and segments of the great saphenous vein; humanin mRNA has also been detected in cultured human aortic endothelial cells (HAEC). Pretreatment of endothelial cells with humanin (on the order of 0.1 μM) reduced oxidized-LDL-induced reactive oxygen species formation by approximately half, and apoptosis by roughly the same magnitude, providing grounds to regard endogenous vascular humanin as a local protective factor against early atherogenic oxidative injury ^[7].

Skeletal muscle is simultaneously both a source and a target of humanin: acute high-intensity interval exercise raises humanin levels in plasma and muscle tissue in men within minutes after exercise, and twelve weeks of resistance training increases humanin protein levels in skeletal muscle in men with impaired glucose tolerance (prediabetes). The proposed downstream mechanism includes increased expression of PGC-1α and associated mitochondrial biogenesis programs, allowing humanin to be regarded as a candidate "exercise mitokine," conceptually similar to MOTS-c though with different targets ^[18]. In the retina, in models of oxidatively stressed retinal pigment epithelium (RPE), including transmitochondrial ARPE-19 cybrids relevant to age-related macular degeneration (AMD), HNG and SHLP2 sustain mitochondrial bioenergetic status, biogenesis, and autophagy, and reduce oxidation-induced cellular damage, generating interest in the MDP axis as a potentially modifiable direction in AMD research ^[19].

Preclinical Research Data

The main body of evidence on humanin is preclinical: cell cultures and rodent models. Neuroprotection has been demonstrated in primary neurons and cell lines against death induced by familial Alzheimer's disease genes and Aβ, with relative specificity regarding Huntington's disease and ALS models (where the protective effect was not reproduced) ^[1][2]. Metabolic effects have been characterized in models of insulin resistance, in hypothalamic ICV studies, and in the NOD model of autoimmune diabetes, including direct enhancement of glucose-stimulated insulin secretion by the analog HNGF6A ^[5][9][16]; vascular effects in models of oxidized-LDL-induced endothelial oxidative stress ^[7]; retinal effects in transmitochondrial ARPE-19 cybrid models relevant to age-related macular degeneration ^[19]; and muscle-physiological effects in human studies of acute and chronic exercise, although these border on clinical-level evidence since they were performed on human volunteers without therapeutic intervention ^[18].

Humanin and related MDPs have also been shown to be age-dependent regulators of apoptosis, insulin sensitivity, and inflammatory markers; circulating humanin levels decline with age in rodents and humans ^[5][8]. A notable exception is the naked mole-rat, a species exhibiting a "negligible senescence" model, in which humanin levels remain relatively stable throughout life, unlike the typical age-related decline seen in other species, reinforcing the hypothesis that humanin is implicated in interspecies differences in aging rate ^[12]. Within this same research direction, the SHLP1–6 family has been systematically characterized: the best-studied SHLP2 and SHLP3 reduced apoptosis and reactive oxygen species production and improved mitochondrial metabolic parameters in vitro, with predominant expression in spleen, kidney, and liver ^[8]. Comparative context is added by MOTS-c: this peptide, through an AMPK-dependent mechanism (with more than a twentyfold increase in endogenous AICAR via the de novo purine synthesis pathway), prevented diet-induced obesity and insulin resistance and reversed age-dependent muscle insulin resistance in mice, an independent line of evidence supporting the general concept of MDPs as a class of metabolic regulators, even though MOTS-c and humanin are distinct molecules ^[17]. It should be emphasized that all of the findings listed here belong to the preclinical level of evidence and do not constitute proof of clinical efficacy in humans.

Clinical and Human (Observational) Data

No controlled interventional clinical trials of humanin as a drug have been conducted. The available human dimension is predominantly observational: measurements of circulating humanin levels in plasma and tissues and their correlation with age, metabolic status, and specific diseases ^[8][10]. One of the most reproducible findings concerns exceptional longevity: in children of centenarians, circulating humanin levels were found to be approximately threefold higher than in age-matched control subjects without a family history of longevity, a pattern consistent with the hypothesis that humanin is part of a "protective mitokine" profile associated with exceptional healthy aging ^[12].

A mitochondrial-wide association study (MiWAS) identified a single-nucleotide polymorphism in the region of the mitochondrial gene encoding humanin (rs2854128), associated with reduced circulating humanin levels and with accelerated cognitive aging in an independent elderly human cohort. In parallel, in mice, several months of sustained humanin administration improved performance on the rotarod test and Barnes maze compared to untreated aging controls, indirect but consistent translational evidence supporting the human genetic association ^[20]. These data show that endogenous humanin is a physiologically present and age-variable genetic factor, but do not establish either the efficacy or the safety of exogenous administration in humans.

An important nuance and caveat for interpretation: humanin is not unambiguously a "good" molecule in every context. In human biopsy material, humanin has been found overexpressed in gastric and bladder cancer tissue compared to adjacent unaffected tissue, and in rat pituitary adenoma cells compared to normal pituitary tissue; in an experimental model of triple-negative breast cancer, humanin promoted tumor progression and chemoresistance ^[21]. This is a reminder that a cytoprotective, anti-apoptotic peptide can, in the context of malignant growth, act against the organism's interests: where cell death is therapeutically desirable, "protecting the cell from apoptosis" ceases to be an unambiguously beneficial property. Therefore, any generalized claims about the "benefit" of humanin outside a specific research context would be an oversimplification, and no established human dosing regimen exists; the human-data direction here serves only as a pointer for future research, not as guidance for use.

Pharmacokinetics and Metabolism

As a short linear peptide of moderate hydrophobicity, native humanin is expected to undergo rapid proteolytic clearance in plasma, by analogy with other native peptide hormones of comparable size, whose circulating half-life is typically measured in minutes. This practical limitation is a key reason why the vast majority of mechanistic and in vivo studies in rodents use not native humanin but the stabilized, orders-of-magnitude more potent analog HNG ([Gly14]-humanin), or other engineered variants (e.g., HNGF6A). Use of the native form in animal experiments typically requires continuous (e.g., intracerebroventricular) administration precisely because of its short window of activity ^[5][6]. No validated human ADME profile (bioavailability by various routes of administration, tissue distribution, metabolite excretion pathways) has been described in the peer-reviewed literature; the available pharmacokinetic information pertains exclusively to animal models and cell systems.

Related Compounds and Analogs

Humanin opened up a broader class of mitochondrial-derived peptides, and today at least three categories of "relatives" are recognized. MOTS-c is a 16-amino-acid MDP encoded in a different region of the mitochondrial genome (12S rRNA, MT-RNR1), which partially translocates to the nucleus in response to metabolic stress to regulate nuclear gene expression, and activates AMPK through a sharp (more than twentyfold) increase in endogenous AICAR via the de novo purine synthesis pathway; in mouse models it reduced diet-induced obesity and insulin resistance and reversed age-dependent muscle insulin resistance ^[17]. SHLP1–6 (small humanin-like peptides) are six peptides identified in silico within the same 16S rRNA region as humanin; the best-characterized, SHLP2 and SHLP3, demonstrate cytoprotective activity (reduced apoptosis and reactive oxygen species, improved mitochondrial metabolism) with predominant expression in spleen, kidney, and liver, and SHLP2 additionally shows retinoprotective activity in age-related macular degeneration models ^[8][19]. The synthetic superagonist HNG ([Gly14]-humanin) remains the most widely used research tool owing to its approximately thousandfold increase in potency relative to native humanin upon Ser14→Gly14 substitution ^[13], while HNS14G exhibits additional chaperone-like activity together with SHLP2 ^[22]. Common to the entire class is mitochondrial origin, short length, and a role as signals of mitochondrial-nuclear and systemic communication that modulate metabolism, stress resistance, and apoptosis ^[11][12].

Handling, Reconstitution, and Storage

For research purposes, humanin is supplied as a lyophilized powder (synthetic peptide, purity confirmed by HPLC (high-performance liquid chromatography)). The lyophilizate should be stored sealed, in a dry environment, protected from light; for long-term storage, freezing (−20 °C or below) is recommended. For experiments, the peptide is reconstituted in bacteriostatic or sterile water; due to the central hydrophobic leucine cluster, individual batches may require additional solvent (a small amount of DMSO or a weakly alkaline buffer) for complete dissolution. Reconstituted solutions should be aliquoted into small portions immediately after dissolution and stored frozen to minimize proteolytic and chemical degradation; repeated freeze-thaw cycles of a single aliquot should be avoided, as they accelerate aggregation of the hydrophobic leucine cluster and loss of activity in biological testing. For further detail, see the accompanying reference material on handling and reconstitution. Everything listed here pertains strictly to laboratory reagent-handling chemistry and is not guidance for use in humans or animals.

Research Applications and Model Systems

In research practice, humanin is used primarily as a probe tool for studying mitochondrial-nuclear signaling, cytoprotection, and metabolic regulation, rather than as a therapeutic candidate. Typical models and systems include: primary neurons and neuronal cell lines to assess anti-apoptotic action against AD genes and Aβ, with dedicated controls in Huntington's disease and ALS models to test specificity ^[1][2]; biochemical systems with purified Bax, Bid, and IGFBP-3 for mapping molecular interactions ^[2][3][15]; cellular systems with reconstituted FPRL1/FPRL2 and CNTFRα/WSX-1/gp130 receptors for studying receptor pharmacology and STAT3 signaling ^[9][14]; models of central (ICV) and peripheral insulin sensitivity, islet and β-cell preparations, including the NOD model of autoimmune diabetes, for metabolic and immunometabolic questions ^[5][9][16]; cultures of human endothelial cells with oxidized LDL for vascular protection studies ^[7]; acute and chronic exercise protocols in human volunteers to study humanin as a muscle mitokine ^[18]; transmitochondrial RPE cybrid models (ARPE-19) for age-related macular degeneration questions ^[19]; cohort measurements of circulating peptide levels and mitochondrial-genome-wide association studies (MiWAS) for gerontological and cognitive hypotheses ^[12][20]; and, as a separate important direction, tumor models and human biopsy material for studying the context-dependent, potentially pro-oncogenic role of humanin ^[21]. The research value of humanin today lies not in clinical application, but in its status as a well-characterized model molecule for testing the concept of mitochondrial-derived peptides as regulators of stress resistance, metabolism, and aging, with all due caveats regarding the context-dependence of its effects.