GH Secretagogues: Current Landscape of Somatotropic Axis Research

Growth hormone (GH) secretion from anterior pituitary somatotrophs is not driven by a single stimulus. It is modulated by the interaction of at least three hypothalamic signals: GH-releasing hormone (GHRH), which stimulates transcription and release; somatostatin (SST), which tonically inhibits it; and endogenous ghrelin, acting on the GHS-R1a receptor expressed both in the pituitary and in the hypothalamic arcuate nucleus. The physiological output is a pulsatile pattern with nocturnal peaks and daytime troughs.

This three-lever architecture explains why secretagogue drug classes are not redundant. GHRH analogs amplify the stimulatory arm; ghrelin mimetics (GHS-R1a agonists) act in parallel and additionally function as functional somatostatin antagonists. In preclinical models the combination of both classes —a GHRH analog plus a GHS-R agonist— produces a synergistic effect on the GH peak that is greater than the sum of the individual effects.

For research this has a practical consequence: compound choice depends on which fragment of the axis you want to interrogate. Studying somatotrope sensitivity to GHRH is not the same as evaluating somatostatin tone, and modeling a physiological pulse is not the same as sustaining elevated plasma IGF-1 over long periods.

Sermorelin is the synthetic fragment of the first 29 amino acids of human GHRH —the biologically active portion of the native 44-residue molecule. It retains affinity for the GHRH receptor but also inherits its Achilles' heel: a very short plasma half-life (on the order of minutes), attributed to rapid proteolysis by DPP-4 (dipeptidyl peptidase 4) at the N-terminus. In animal models it produces acute, clean GH pulses, which is useful when the goal is to study the intact somatotrope response.

CJC-1295 introduces two modifications over the GHRH(1-29) sequence: four amino acid substitutions that make it resistant to DPP-4 and —in the DAC (Drug Affinity Complex) variant— a maleimidopropionyl linker that covalently binds the peptide to plasma albumin after injection. This albumin binding dramatically extends half-life relative to sermorelin. The result is a sustained elevation of IGF-1 rather than discrete pulses, which conceptually separates it from the physiological pattern.

Tesamorelin is the full GHRH(1-44) sequence with a trans-3-hexenoic group anchored to the N-terminal tyrosine residue. The modification sterically blocks the DPP-4 cleavage site without altering receptor binding. It is the only GHRH analog with regulatory approval in the U.S. (for HIV-associated lipodystrophy), which translates into the most solid body of controlled clinical evidence in the class. For research, this makes it the reference comparator when published human data are needed.

GH-releasing peptides (GHRPs) are the first generation of synthetic ghrelin receptor agonists. GHRP-6 was the historical prototype —a hexapeptide that stimulates GH but also activates appetite via arcuate nucleus circuitry and raises cortisol, ACTH and prolactin at higher doses. GHRP-2 has somewhat greater potency on the GH peak but retains the broad hormonal profile. Both remain useful pharmacological tools precisely because their lack of selectivity allows interrogating multiple axes at once.

Ipamorelin is a pentapeptide designed to solve the selectivity problem. In the seminal work by Raun and colleagues (1998), ipamorelin stimulated GH release at doses several-fold above the ED50 without significantly elevating ACTH or cortisol relative to baseline GHRH stimulation. That signature —clean GH pulse, HPA axis respected, no marked orexigenic effect— made it the preferred option when experimental design requires isolating the somatotropic pathway.

Hexarelin is a potent hexapeptide producing higher acute GH peaks than ipamorelin at equimolar doses in animal models, but also elevating prolactin, cortisol and ACTH. The preclinical literature additionally describes direct cardiovascular actions of hexarelin mediated by CD36 and cardiac GHSR receptors, independent of GH release —a relevant point when interpreting any study where cardiac function is among the outcomes.

MK-677 is a small molecule, orally bioavailable GHS-R1a agonist. Its conceptual importance lies in demonstrating that ghrelin receptor activation can be achieved outside the peptide space, with pharmacokinetics that allow once-daily dosing. In murine aging models, preservation of lean mass and grip strength has been described. In published human studies with older adults, MK-677 at 25 mg/day for periods of up to 12 months produced sustained IGF-1 elevations into ranges observed in young adults.

For 2026 research, MK-677 offers three useful features: a half-life compatible with chronic oral dosing; selective activation of the ghrelin arm without touching the GHRH receptor; and a literature body covering aging, sarcopenia and metabolism. The limitations are also well characterized: fluid retention, increased appetite via orexigenic circuit activation, and elevations in fasting glucose in some cohorts —all observations to keep in mind when designing preclinical protocols.

Three questions guide compound choice. First: preserve the physiological pulsatile pattern or sustain a flat IGF-1? The answer points toward short half-life peptides (sermorelin, ipamorelin, GHRP-2) or long-acting compounds (CJC-1295 with DAC, MK-677). Second: how critical is isolating the effect on GH from the rest of the endocrine axis? If it is, ipamorelin and GHRH analogs are the clean option; if coupled-axis study is the goal, GHRP-2/6 and hexarelin are legitimate tools.

Third: is the inquiry about aging, sarcopenia or metabolism? Here the densest preclinical evidence covers MK-677 and tesamorelin —the former for its oral profile and data in older adults, the latter for its regulatory dossier on body composition. For any design, the basic rule applies: these compounds are research tools (Research Use Only). Doses, vehicles and endpoints must be adjusted to the model and the predefined outcomes, and reading must be done on the data generated, not on extrapolation across classes.