Kisspeptin (KP-10 and KP-54)

One gene, four peptides, two numbering systems

The KISS1 gene encodes a precursor protein of 138 residues in UniProt's current canonical entry, Q15726. Processing yields a family of carboxy-terminally amidated peptides that share the same final ten residues: kisspeptin-54, also called metastin, spanning precursor residues 68 to 121; kisspeptin-14 at 108 to 121; kisspeptin-13 at 109 to 121; and kisspeptin-10 at 112 to 121, sequence YNWNSFGLRF. UniProt records the terminal phenylalanine as amidated and lists a cleavage site at residue 118 attributed to matrix metalloproteinases 2 and 9, which sits inside the decapeptide itself.

Two numbering conventions are in circulation and they do not agree. Registry entries from Massachusetts General Hospital list the intervention as kisspeptin 112-121, counting from the start of the precursor. Chemical catalogues and much of the older literature call the same decapeptide metastin 45-54, counting from the start of the 54-residue peptide. PubChem's synonym list for CID 25240297 carries metastin 45-54 and a third label, kisspeptin-13 (4-13), but not the 112-121 form; that convention appears instead in the registry entries and in the primary literature, where Liu 2013 writes the molecule as kisspeptin-112-121 or metastin 45-54 (PMID 23524040). All three name YNWNSFGLRF.

Identity for the decapeptide verifies cleanly. PubChem CID 25240297 gives C63H83N17O14, molecular weight 1302.4, CAS 374675-21-5, UNII FS1N52VS3S, ChEMBL376756 and NSC 741805, and recomputing the formula from the amidated sequence returns the same string. The 54-residue peptide is less tidy. PubChem CID 71306396 gives C258H401N79O78 at 5857 and CAS 388138-21-4, and that formula reconciles only if precursor position 81 is arginine. UniProt's canonical sequence carries proline there and records arginine as natural variant rs4889. The two forms differ by 59 daltons.

One registry record does not survive checking. The FDA Global Substance Registration System entry for kisspeptin, UNII 2VG74ZW2K5, carries CAS 388138-21-4 and a calculated molecular weight of 5857, both correct for kisspeptin-54, alongside a 54-residue sequence beginning CRHKPMRTVTNF that is typed as a complete human hormone sequence. That sequence is not from the KISS1 precursor. It matches residues 68 to 121 of KISS1R, the receptor, in UniProt Q969F8: the right coordinates applied to the wrong protein. The same record also stores a molecular formula, C290H436O69N68S6, whose six sulfur atoms settle the question, since the precursor segment contains no cysteine and no methionine while the receptor segment contains four and two. Recomputed, the stored sequence would weigh about 6189 daltons and the stored formula about 6171, not the 5857 the same record displays.

From a metastasis screen to the reproductive axis

KiSS-1 arrived as a cancer gene. Lee and colleagues, at the Jake Gittlen Cancer Research Institute in Hershey, Pennsylvania, used subtractive hybridisation to compare messenger RNA from metastatic human melanoma cells against non-metastatic hybrids carrying an introduced chromosome 6, and recovered a novel complementary DNA they designated KiSS-1. Expression was detectable only in the non-metastatic lines and was highest in placenta. Transfection of the full-length complementary DNA into C8161 melanoma cells suppressed metastasis in an expression-dependent manner. The 1996 paper reported a 164-amino-acid protein.

That length was wrong. A 1997 erratum in the same journal reported that GC compression in the original sequencing gels had caused five contiguous cytosine residues in the coding region to be read as three, that the deduced amino acid sequence in Figure 2A was incorrect, and that the corrected protein contains 145 residues. The GenBank accession was updated. Metastasis suppression itself held, the authors noted, having by then been reproduced in a second melanoma line and a breast carcinoma line. UniProt's canonical entry now gives 138 residues. The figure 164 is still repeated in secondary material.

Five years later a group at Takeda isolated the gene product. Ohtaki and colleagues purified from human placenta a carboxy-terminally amidated 54-residue peptide, identified it as the endogenous ligand of an orphan receptor then called hOT7T175, and named it metastin. In that paper metastin inhibited chemotaxis and invasion of receptor-transfected CHO cells and attenuated pulmonary metastasis of receptor-transfected B16-BL6 melanoma in mice. The receptor is now KISS1R, previously GPR54.

The reproductive role was established from the loss-of-function direction in 2003. De Roux and colleagues mapped a consanguineous family with five affected siblings to chromosome 19p13 and found that every affected member carried a homozygous 155-nucleotide deletion in GPR54 spanning the intron 4 to exon 5 splice acceptor. Seminara and colleagues reported the same association with idiopathic hypogonadotropic hypogonadism in humans alongside a Gpr54-deficient mouse. Both papers describe absent puberty in people who cannot signal through the receptor. Neither administered a peptide to anyone.

What single doses did in people

Administration to humans began with a study of six men, and it remains the source of the only human pharmacokinetic figures located for this page. Dhillo and colleagues gave six male volunteers a 90-minute intravenous infusion of kisspeptin-54 at 4 pmol per kilogram per minute, and saline on a separate day, in a double-blind crossover. Mean 90-minute luteinising hormone was 10.8 plus or minus 1.5 units per litre against 4.2 plus or minus 0.5 on saline. Mean 180-minute testosterone was 24.9 against 21.7 nanomoles per litre. Plasma half-life was calculated at 27.6 plus or minus 1.1 minutes.

Kisspeptin-10 was first given to men six years later. George and colleagues in Edinburgh found that intravenous boluses produced a dose-dependent rise in luteinising hormone with maximal stimulation at 1 microgram per kilogram, from 4.1 to 12.4 units per litre at thirty minutes in six men. The next dose up, 3 micrograms per kilogram, produced a smaller response than 1. A 22.5-hour infusion at 4 micrograms per kilogram per hour in four men raised testosterone from 16.6 to 24.0 nanomoles per litre. That infusion is the measurement behind most statements that the decapeptide raises testosterone.

Direct comparison came in 2015. Jayasena and colleagues infused vehicle, kisspeptin-10, kisspeptin-54 and gonadotropin-releasing hormone for three hours at 0.1, 0.3 and 1.0 nmol per kilogram per hour, five men per group. At the top dose, mean area under the curve for serum luteinising hormone was 10.81 for kisspeptin-10, 14.43 for kisspeptin-54 and 34.06 for GnRH, in hour-units per litre. The two isoforms were similar to one another and both were weaker than GnRH. The authors named the small sample as a limitation.

A null result constrains how far any of this transfers. Jayasena and colleagues gave kisspeptin-10 to healthy women in the follicular phase by intravenous bolus to 10 nmol per kilogram, subcutaneous bolus to 32 nmol per kilogram, and intravenous infusion to 720 pmol per kilogram per minute, four to five per group. No change in serum gonadotropins was observed by any route. The same intravenous bolus raised both gonadotropins in the preovulatory phase, and men had responded at 0.3 nmol per kilogram.

Continuous exposure works against itself

Repeated exposure does not repeat the acute result. Jayasena and colleagues gave ten women with hypothalamic amenorrhoea twice-daily subcutaneous kisspeptin-54 at 6.4 nmol per kilogram, or saline, for two weeks, five per arm, randomised and double-blinded. On the first injection day the mean maximal rise in luteinising hormone within four hours was 24.0 plus or minus 3.5 units per litre. On the fourteenth it was 2.5 plus or minus 2.2, and the follicle-stimulating hormone rise fell from 9.1 to 0.5 on the same measure. Responsiveness to GnRH survived, placing the desensitisation upstream of the pituitary.

A shorter course in a different population did not collapse the same way. Five healthy women with regular cycles received twice-daily subcutaneous kisspeptin-54 at 6.4 nmol per kilogram on menstrual days 7 to 14, each also receiving saline in a separate cycle in a single-blinded one-way crossover, so the study is five participants rather than ten. Cycles persisted in all of them, acute responsiveness to both kisspeptin-54 and GnRH was retained, and the treated cycles were shorter, 26.8 plus or minus 3.1 days against 28.6 plus or minus 1.4. Highest recorded luteinising hormone and the progesterone rise each arrived about two days earlier.

Work published in August 2026 separates the two patterns explicitly. Yeung and colleagues studied fifteen healthy men on subcutaneous kisspeptin-10 against twelve controls. Acute eight-hour infusions from 1.25 to 10.0 nmol per kilogram per hour raised luteinising hormone, follicle-stimulating hormone and testosterone dose-dependently. Five days of continuous infusion at 180 nmol per hour left gonadotropins at vehicle levels while testosterone stayed elevated. Twelve days of intermittent infusion, eight hours on and sixteen off at 150 nmol per hour, sustained the gonadotropin rise, plus 1.68 units per litre on day 1 and plus 1.14 on day 12.

Sustained receptor occupancy has also been pursued deliberately in the suppressive direction. TAK-448, a kisspeptin receptor agonist, entered a phase 1/2 study in nine men with prostate cancer to lower testosterone, and later two Takeda phase 2 studies: seventeen middle-aged and older men with low testosterone, and fifteen men with hypogonadotropic hypogonadism. Both phase 2 studies were terminated by the sponsor, the registry records stating that the primary endpoint and the primary efficacy objective respectively were not met.

The in vitro fertilisation programme

The densest human dataset is a single Imperial College programme, registered as NCT01667406, which ran from June 2012 to October 2016, enrolled 175 participants and has posted results. Its first publication used kisspeptin-54 as a trigger for oocyte maturation. Fifty-three women received one subcutaneous injection at 1.6, 3.2, 6.4 or 12.8 nmol per kilogram after standard stimulation, with retrieval at 36 hours. Maturation was observed at every dose tested. Fertilisation and transfer occurred in 49 of 53 patients. Biochemical and clinical pregnancy rates were 21 of 53 and 12 of 53.

A second stage restricted enrolment to women at high risk of ovarian hyperstimulation syndrome, the complication the trigger was meant to avoid. Sixty women received a single injection at 3.2 to 12.8 nmol per kilogram under adaptive dose allocation. Maturation occurred in 95 per cent. Across fifty-one transfers, biochemical, clinical and live-birth rates were 63, 53 and 45 per cent. No woman developed moderate, severe or critical hyperstimulation. The trial was open-label, and the authors called for direct comparison against established triggers.

Whether a second dose would help was then asked under randomisation. Sixty-two women at high risk all received kisspeptin-54 at 9.6 nmol per kilogram 36 hours before retrieval, then were randomised one to one to a second dose or saline ten hours later. The proportion reaching an oocyte yield of at least 60 per cent was 21 of 31 with the second dose against 14 of 31 with saline, an absolute difference of 26 percentage points, confidence interval 2 to 50, p equal to 0.042. One case of moderate hyperstimulation occurred, in the single-dose arm.

None of this has produced an approved product. ClinicalTrials.gov returned 45 records naming kisspeptin when queried on 18 August 2026, of which 32 are interventional; 36 name kisspeptin as an intervention, and 25 of those are interventional studies. The openFDA drug approvals endpoint returned no match for kisspeptin as a generic name. Development has shifted toward a receptor agonist rather than the native peptide: in nine healthy women in the follicular phase, MVT-602 produced a luteinising hormone rise of similar amplitude to kisspeptin-54 but peaked at 21.4 hours against 4.7, with correspondingly greater exposure (Abbara 2020, PMID 33196464).

The sexual-desire trials, and what they measured

Four studies from one group underlie everything written about kisspeptin and desire, and all four share a design worth stating plainly. Each administered kisspeptin-54, not the decapeptide, by intravenous infusion at 1 nmol per kilogram per hour for 75 minutes, in a gelatin-based colloid vehicle, inside a scanner, against a rate-matched placebo in crossover. Comninos and colleagues ran 29 healthy young men in 2017 and reported enhanced limbic activity to sexual and couple-bonding stimuli. Yang and colleagues ran 33 men on the same protocol in 2020, testing olfactory and visual cues of attraction.

Thurston and colleagues randomised 40 premenopausal women with hypoactive sexual desire disorder, of whom 32 completed both visits, mean age 29.2 years. Modulations were reported in the left inferior frontal gyrus, the right postcentral and supramarginal gyrus and the right temporoparietal junction, with correlations between hippocampal response to erotic video and baseline distress about sexual function. The primary outcome was brain activity, not desire. The report records no adverse effects.

Mills and colleagues ran the male counterpart: 37 men randomised, 32 completing, mean age 37.9 years. The primary outcome was again whole-brain activity on functional imaging, reported as a mean absolute change of 0.81, confidence interval 0.41 to 1.21, p equal to 0.003. Penile tumescence in response to sexual stimuli was a secondary outcome, increased by up to 56 per cent more than placebo, mean difference 0.28 units, confidence interval 0.04 to 0.52. That 56 is the origin of a figure now quoted without its measurement.

Selectivity has been checked once, narrowly. Serum from five healthy women given seven days of twice-daily subcutaneous kisspeptin-54 at 6.4 nmol per kilogram was analysed post hoc for growth hormone, prolactin and thyroid-stimulating hormone; these are the same five women whose menstrual cycles were followed in the crossover described earlier, and the pituitary analysis is a post hoc reading of that one cohort. No change was found against vehicle, acutely or across the week, and growth hormone pulse frequency and amplitude did not change. The authors wrote that they could not exclude an effect. Five women, one dose, one week is the whole of the published pituitary-selectivity record located here.

What is not known

No kisspeptin product is approved for human use anywhere, and the openFDA drug approvals endpoint returned no match for kisspeptin as a generic name when queried on 18 August 2026. Beyond that, the gap that matters most is one of isoform and route. The molecule sold as a research chemical is the decapeptide; the human record on sexual function, in vitro fertilisation triggering and hypothalamic amenorrhoea was generated almost entirely with kisspeptin-54, delivered by intravenous infusion, by pump, or as a single weight-based subcutaneous injection under supervision. No published study administered kisspeptin-10 as a fixed subcutaneous dose on a repeating daily schedule, and no study of either isoform measured a behavioural outcome outside a scanner or a psychometric instrument administered during infusion. Exposure durations are short: the longest published continuous or intermittent human exposures located are twelve days for kisspeptin-10 and fourteen days for kisspeptin-54, with no follow-up beyond the treatment window and no long-term safety data in any population. Sample sizes run from four to sixty-two. Selectivity against other pituitary hormones has been examined once, in five women over one week, and as a post hoc reading of a cohort assembled to answer a different question. No head-to-head comparison exists against established oocyte-maturation triggers, or against any therapy for male hypogonadism; the two registered phase 2 trials of a kisspeptin receptor agonist in men were terminated for failing their primary endpoints. Male and female responses diverge sharply and by menstrual-cycle phase, so figures from one sex or phase do not transfer. Identity data carry unresolved discrepancies: the molecular weight of the 54-mer depends on which allele sits at precursor position 81, and the FDA substance record for kisspeptin currently stores a receptor sequence, and a receptor-derived molecular formula, in place of the ligand.

Questions