Calcitonin gene-related peptide (CGRP) is a 37-amino-acid signaling peptide found particularly in sensory neurons. It is one of the most potent endogenous dilators of blood vessels and participates in pain transmission, local tissue responses, and communication between the nervous and immune systems. Humans produce two closely related forms, α-CGRP and β-CGRP, encoded by different genes. Its name reflects the discovery that α-CGRP and the hormone calcitonin can arise from different processing of transcripts from the same gene, rather than indicating that the two peptides have identical physiological functions.
CGRP became especially important in medicine through research on migraine. Evidence from human provocation experiments, pharmacology, and clinical trials established CGRP signaling as a therapeutically actionable component of migraine biology. Drugs that neutralize CGRP or block its receptor are used for migraine prevention, and some receptor antagonists are also used for acute treatment. Beyond headache disorders, CGRP provides a model for understanding how sensory nerves actively regulate blood vessels, epithelial barriers, and immune responses rather than merely conveying information to the brain.
Discovery, molecular identity, and gene organization
CGRP was identified in 1982 during studies of the rat calcitonin gene by Susan Amara, Michael Rosenfeld, and colleagues. Their work showed that a single gene could generate distinct messenger RNAs through tissue-specific processing. In thyroid tissue, its expression predominantly produced calcitonin; in neural tissue, a different transcript encoded CGRP. This became an influential early example of how alternative splicing and alternative RNA processing expand the functional output of a gene without requiring a separate gene for every mature signaling molecule.
In humans, α-CGRP is encoded by CALCA, located on chromosome 11. The calcitonin-producing transcript principally uses exons 1–4 and a proximal polyadenylation site, whereas the CGRP-producing transcript joins exons 1–3 to exons 5–6 and uses a more distal processing pathway. These choices change the coding sequence delivered to the ribosome. The resulting products have different structures, receptors, and dominant biological roles: calcitonin participates in calcium regulation, whereas CGRP is particularly associated with sensory signaling and vascular control.
Human β-CGRP is encoded by the nearby CALCB gene. The mature α and β peptides differ at three of their 37 amino-acid positions and have broadly similar pharmacological activities. Their expression patterns are not identical: α-CGRP is prominent in many sensory neurons, while β-CGRP has an important presence in the enteric nervous system. This distinction is not absolute, and both peptides may occur in the same organ. Experimental antibodies and assays do not always distinguish the isoforms, so a reported measurement of “CGRP” may represent their combined immunoreactivity.
The mature peptide has a molecular mass of approximately 3.8 kilodaltons. A disulfide bond between cysteine residues 2 and 7 creates a ring near the amino terminus, and the carboxyl-terminal phenylalanine is amidated. Both features contribute to biological activity. The amino-terminal region is especially important for receptor activation, while the carboxyl-terminal portion contributes strongly to receptor binding. The peptide also adopts an α-helical conformation over part of its sequence when associated with appropriate molecular environments.
CGRP belongs to the calcitonin peptide family, which also includes amylin, adrenomedullin, and intermedin, also called adrenomedullin 2. These peptides share structural features and use overlapping receptor components, but they are not interchangeable. Their relatedness helps explain why receptor classification requires more than identifying a single receptor protein. It also provides a molecular basis for cross-reactivity at sufficiently high peptide concentrations, an important consideration when interpreting laboratory experiments.
Biosynthesis, neuronal distribution, and release
CGRP is synthesized as part of a larger precursor rather than directly as the mature 37-residue peptide. After translation, a signal sequence directs the precursor into the secretory pathway. Proteolytic processing, disulfide-bond formation, and carboxyl-terminal amidation generate the active peptide. The processing machinery includes prohormone-cleaving enzymes and peptidylglycine α-amidating monooxygenase. Mature CGRP is packaged into dense-core secretory vesicles, which are transported along the axon to release sites.
A major source is the small- and medium-diameter sensory neuron population in dorsal root and trigeminal ganglia. Many CGRP-containing fibers are unmyelinated C fibers or thinly myelinated Aδ fibers, although peptide expression does not map perfectly onto conduction class. These neurons innervate skin, joints, muscle, viscera, and the coverings of the brain. Within the central nervous system, their terminals project into sensory processing regions such as the superficial dorsal horn of the spinal cord and the spinal trigeminal nucleus.
CGRP is also present in intrinsic neurons of the gastrointestinal tract and in several neuronal populations within the brain. At the neuromuscular junction, motor neurons provide another source, illustrating that CGRP is not exclusively a pain peptide. Its distribution must therefore be described at the level of cell type and projection rather than inferred from a whole-organ measurement. A tissue rich in CGRP-containing fibers may receive the peptide from neurons whose cell bodies lie far outside that tissue.
Release commonly follows membrane depolarization and calcium entry through voltage-gated channels. Activation of the heat- and chemical-sensitive TRPV1 channel, for example by capsaicin, can trigger CGRP secretion from sensory endings. Other stimuli include tissue acidosis, inflammatory mediators, and mechanical or chemical irritation, depending on the neuronal population. Because dense-core vesicle release has different activation requirements from the release of small-molecule transmitters, CGRP secretion often becomes especially prominent during sustained or intense neuronal activity.
A sensory neuron can release CGRP both centrally and peripherally. Central release can modify information transfer in the spinal cord or brainstem, while peripheral release can change vascular tone and local cell behavior. In an axon reflex, activity spreads through branches of a sensory axon and promotes mediator release from neighboring peripheral endings without requiring a complete reflex circuit through the central nervous system.
Unlike many classical transmitters, released CGRP is not chiefly cleared by rapid reuptake into the nerve terminal. Its effects are limited by dispersion, receptor regulation, and enzymatic degradation. Its circulating lifetime is on the order of minutes, with measured estimates depending on experimental conditions. Consequently, plasma concentration is an imperfect proxy for the transient, potentially much higher concentration near an active nerve ending.
Receptor architecture and intracellular signaling
The canonical CGRP receptor is a molecular complex whose principal components are calcitonin receptor-like receptor, commonly abbreviated CLR, and receptor activity-modifying protein 1, or RAMP1. CLR is encoded by CALCRL and belongs to the class B family of G protein-coupled receptors. RAMP1 is a single-pass membrane protein that helps determine receptor trafficking and ligand recognition. Receptor component protein, an intracellular accessory protein, has also been implicated in efficient coupling to downstream signaling.
The importance of RAMPs became clear in 1998 through work by Linda McLatchie and colleagues. Association of CLR with RAMP1 produces the characteristic CGRP receptor phenotype, whereas association with RAMP2 or RAMP3 produces receptors preferentially activated by adrenomedullin. Thus, receptor identity emerges from a protein partnership rather than solely from the sequence of the seven-transmembrane receptor. Differences in the relative abundance of these components help create tissue-specific pharmacology.
At the canonical receptor, CGRP predominantly activates the stimulatory G protein , which stimulates adenylyl cyclase. This enzyme converts adenosine triphosphate into cyclic adenosine monophosphate, or cAMP. A simplified signaling sequence is
Increased cAMP activates protein kinase A and other cAMP-sensitive effectors. The eventual response depends on the cell: relaxation in vascular smooth muscle, altered excitability in neurons, or changes in mediator production in immune and stromal cells. Coupling to additional pathways can occur in particular cellular or experimental settings.
Ligand recognition follows the broad two-domain mechanism characteristic of many class B peptide receptors. The peptide’s carboxyl-terminal region engages the receptor’s extracellular binding apparatus, including contributions from RAMP1, while the amino-terminal region interacts with the transmembrane receptor core to promote activation. Cryo-electron microscopy has resolved active CGRP receptor complexes and helped explain both peptide recognition and the binding sites of receptor-blocking drugs. Several small-molecule antagonists bind near the interface between CLR and RAMP1.
CGRP can also activate the AMY1 receptor, formed by the calcitonin receptor and RAMP1. This creates a significant interpretive issue: a biological response to CGRP does not automatically establish that CLR–RAMP1 mediated it. Receptor expression, antagonist selectivity, and peptide concentration all matter. The peptide fragment CGRP(8–37), widely used as an experimental antagonist, is useful but does not provide a universally selective test under every condition.
Receptor activation is followed by regulatory processes that include phosphorylation, recruitment of arrestin proteins, and internalization. Internalized receptors may recycle to the cell surface or undergo degradation, and signaling can sometimes persist from intracellular compartments. These processes affect the duration of a response and the development of desensitization, explaining why receptor abundance alone cannot fully predict tissue sensitivity.
Vascular actions and cardiovascular physiology
CGRP is a powerful vasodilator in many vascular beds. Sensory fibers containing the peptide run alongside arteries and arterioles, often in the outer vessel wall. When released, CGRP can act locally on smooth muscle to reduce contractile tone. In suitable experimental preparations, measurable relaxation occurs at very low concentrations, although potency varies with species, vessel size, receptor expression, and the initial state of contraction.
The principal smooth-muscle mechanism involves cAMP-dependent signaling. Protein kinase A can alter the activity of potassium channels, calcium-handling systems, and proteins that determine the contractile apparatus’s sensitivity to calcium. The resulting membrane and biochemical changes favor relaxation. Depending on the vessel, CGRP may also recruit endothelial mechanisms involving nitric oxide. Endothelial participation is not obligatory in all vascular beds, so describing every CGRP response as nitric-oxide dependent would be inaccurate.
Local release can increase tissue blood flow without producing a large change in systemic arterial pressure. This distinction is central to its physiology. A high concentration confined to the microenvironment of an activated nerve can influence nearby resistance vessels, while only a small fraction reaches the general circulation. Intravenous administration, by contrast, exposes many vascular beds at once and can cause flushing, a fall in blood pressure, and a compensatory increase in heart rate.
In skin, CGRP contributes to the flare component of the response to sensory nerve activation. Its vascular action can interact with other mediators released from nerves or immune cells. The related sensory peptide substance P is especially associated with increased vascular permeability in several experimental systems, whereas CGRP is particularly effective at increasing blood flow. Their combined actions helped establish the concept of neurogenic inflammation, but the relative contribution of each mediator differs between organs and species.
Experimental studies have also examined CGRP as a protective response during cardiovascular stress. In animal models, sensory-neuron-derived CGRP can oppose vasoconstriction, support tissue perfusion, and modify responses to ischemia or elevated blood pressure. These findings establish plausible physiological roles, not a demonstrated therapeutic benefit of administering CGRP to people with cardiovascular disease. Systemic delivery is additionally constrained by the peptide’s short lifetime and its broad hemodynamic effects.
The vascular functions of endogenous CGRP are relevant to the safety evaluation of drugs that inhibit its pathway. Blocking the receptor is not equivalent to directly constricting a blood vessel, and CGRP-targeted migraine drugs lack the direct vasoconstrictor mechanism of triptans. Nevertheless, the possibility that blockade might alter compensatory responses under particular cardiovascular conditions remains a distinct research question. Evidence from clinical trials and postmarketing surveillance must therefore be interpreted separately from findings in isolated vessels or genetically modified animals.
Nociception and the biology of migraine
CGRP participates in nociception, the neural processing of potentially damaging stimuli, but its presence does not mean that a neuron signals pain exclusively. In the spinal dorsal horn and trigeminal brainstem pathways, CGRP can facilitate synaptic transmission and interact with glutamate-dependent signaling. Persistent activity in these circuits contributes to sensitization, in which neurons become more responsive to incoming stimulation. Pain itself is a perceptual experience produced by distributed neural systems, not a direct readout of peptide concentration.
Migraine research focuses particularly on the trigeminovascular system, comprising trigeminal sensory neurons and their relationships with cranial blood vessels, meninges, and central processing pathways. Many meningeal sensory fibers contain CGRP. Their activation releases the peptide near vascular and other peripheral targets, while signaling through the trigeminal ganglion and brainstem conveys activity toward higher brain regions. Interactions among neurons, satellite glial cells, vascular cells, and inflammatory mediators can influence the responsiveness of this network.
Several lines of human evidence connected CGRP to migraine. In 1990, Peter Goadsby, Lars Edvinsson, and Rolf Ekman reported increased CGRP in external jugular venous blood during migraine attacks. Subsequent studies found changes in CGRP-associated measures during some attacks and after effective treatment, although results have varied with sampling site and assay methodology. These observations were important historically, but blood CGRP measurement has not become a standardized diagnostic test for migraine.
Provocation experiments provided a different form of evidence. Infusion of CGRP can produce headache in healthy volunteers and delayed migraine-like attacks in susceptible people with migraine. Conversely, drugs that inhibit the peptide or its receptor can reduce migraine attacks or relieve an ongoing attack. The convergence of provocation and intervention evidence is stronger than a simple association between a measured peptide level and symptoms.
CGRP does not explain every aspect of migraine. Migraine involves sensory processing abnormalities, hypothalamic and brainstem networks, inherited susceptibility, and, in some attacks, cortical spreading depolarization, the physiological process associated with typical aura. The success of CGRP-targeted treatments does not establish that dilation of cranial blood vessels is the sole cause of pain. Modern accounts distinguish the peptide’s vascular effects from its broader actions on neural and non-neural components of the trigeminovascular system.
CGRP is also implicated in cluster headache, but therapeutic findings differ by disorder and disease subtype. Galcanezumab demonstrated preventive efficacy in episodic cluster headache and received a United States indication for that use in 2019; a trial in chronic cluster headache did not establish comparable efficacy. This contrast illustrates why biological involvement in a headache disorder does not guarantee that blocking the same pathway will work equally across all clinical categories.
Gastrointestinal, immune, and tissue-regulatory functions
The gastrointestinal tract contains CGRP both in extrinsic sensory fibers and in intrinsic enteric neurons. β-CGRP is particularly associated with enteric expression, although the distribution of the two isoforms overlaps. CGRP can alter intestinal secretion, smooth-muscle activity, mucosal blood flow, and sensory signaling. Its net effect on transit depends on the anatomical region, receptor-bearing cells, and whether an experiment examines an isolated tissue or an intact neural circuit.
In gastric and intestinal mucosa, sensory CGRP participates in responses that help preserve local perfusion during chemical or mechanical stress. Experimental exposure to irritants can activate sensory endings and trigger peptide release, increasing blood flow near the affected tissue. Animal studies also link CGRP signaling to mucus-related defenses and epithelial responses. These findings concern coordinated local physiology; they do not imply that increased CGRP is uniformly beneficial in every gastrointestinal disorder.
CGRP is an important mediator of neuroimmune communication. Its receptors occur on several immune-cell populations, and cAMP signaling can change the production of cytokines, antigen-presenting activity, cell migration, and effector functions. In a number of experimental settings, CGRP suppresses production of selected pro-inflammatory mediators by macrophages or modifies the function of dendritic cells. Effects on T cells and other lymphocytes depend on the activation state, tissue environment, and accompanying signals.
These actions cannot be reduced to a single label such as “anti-inflammatory.” Suppression of one immune pathway may protect tissue in one setting while impairing microbial defense in another. Conversely, sensory-neuron activity can contribute to local inflammatory disease by promoting vascular changes or changing immune-cell behavior. Studies of bacterial infection, allergic inflammation, and epithelial barriers have therefore produced context-dependent outcomes. Distinguishing direct action on immune cells from secondary effects on blood flow or tissue damage is essential.
CGRP also participates in skeletal biology. Sensory fibers containing CGRP innervate bone and periosteum, and experimental evidence links the peptide to osteoblast activity, osteoclast regulation, and fracture repair. In several animal systems, loss or inhibition of CGRP signaling impairs aspects of bone formation or healing. These observations support a role for sensory innervation in skeletal maintenance, but they do not establish CGRP-based treatment as a clinical method for repairing fractures or preventing osteoporosis.
Additional experimental work concerns skin wound healing, airway responses, and interactions with reproductive tissues. The recurring principle is that a sensory nerve can directly influence the tissue it monitors. Such functions are especially important when interpreting systemic pathway inhibition: a migraine drug targets a molecule with roles well beyond cranial pain. However, a physiological role demonstrated in a laboratory model does not by itself establish a clinically significant adverse effect when that pathway is inhibited in humans.
Therapeutic targeting and clinical development
Two major drug classes inhibit CGRP signaling: small-molecule receptor antagonists, commonly called gepants, and monoclonal antibodies directed against either the peptide or its receptor. Their pharmacological differences are substantial. Gepants have relatively short exposure times and some are suitable for treating individual attacks. Antibodies persist for weeks and are principally preventive treatments. Neither approach requires nonspecific suppression of neuronal activity throughout the brain.
Early proof of concept came from olcegepant, an intravenously administered receptor antagonist that showed efficacy in an acute migraine trial published in 2004. Telcagepant subsequently demonstrated that an oral antagonist could be effective, but its development was discontinued after liver-enzyme elevations emerged during repeated-exposure studies. This history shaped the development and monitoring of later gepants. The hepatic findings associated with particular early compounds are not evidence that every molecule blocking the receptor has the same toxicity.
In the United States, ubrogepant was approved for acute migraine treatment in 2019. Rimegepant received an acute-treatment approval in 2020 and a preventive indication for episodic migraine in 2021. Atogepant was approved for prevention of episodic migraine in 2021, with its indication expanded to chronic migraine in 2023. Zavegepant, administered as a nasal spray, was approved for acute treatment in 2023. Indications, formulations, and authorization dates differ among jurisdictions.
Among antibodies, erenumab targets the canonical CGRP receptor, whereas fremanezumab, galcanezumab, and eptinezumab target the peptide. The first three received United States approvals for migraine prevention in 2018, followed by eptinezumab in 2020. Erenumab, fremanezumab, and galcanezumab are administered subcutaneously; eptinezumab is administered intravenously. Their prolonged persistence reflects the pharmacokinetics of immunoglobulin G antibodies, not a permanent alteration of the patient’s CGRP-producing neurons.
Large antibodies cross the intact blood–brain barrier only to a limited extent. Their efficacy supports the importance of targets accessible outside that barrier, including structures associated with the meninges and trigeminal ganglion. It does not prove that migraine originates exclusively outside the brain. Peripheral interruption can modify activity throughout a connected neural system even when a drug does not reach central sites in substantial concentrations.
Efficacy is evaluated using outcomes such as monthly migraine days, acute-medication use, and the proportion of participants achieving a specified reduction in attack frequency. Acute-treatment trials commonly assess freedom from pain and from the most bothersome associated symptom at two hours. Not every participant responds. Reported adverse effects vary by agent and include injection-site reactions, constipation, nausea, and hypersensitivity; postmarketing surveillance has also identified hypertension and Raynaud-related concerns with CGRP-pathway antagonism. Trial populations, real-world populations, and individual drug labels provide different kinds of safety information and are not interchangeable.
Experimental methods and unresolved questions
CGRP research combines molecular genetics, anatomical mapping, physiological recording, analytical chemistry, and human clinical experimentation. Immunohistochemistry reveals peptide-containing fibers and cell bodies, while in situ hybridization identifies cells expressing CALCA, CALCB, or receptor-component transcripts. These methods answer different questions: detecting stored peptide does not show that release is occurring, and detecting receptor messenger RNA does not establish that a functional receptor complex is present at the cell surface.
Circulating CGRP is usually measured with immunoassays, but reported concentrations vary substantially across methods. Sample handling, degradation before analysis, extraction procedures, calibration, antibody specificity, and recognition of peptide fragments can all affect results. Blood obtained from a peripheral arm vein also samples a different physiological compartment from blood draining the head. More selective analytical approaches, including mass spectrometry, can address some identification problems, although low peptide abundance and sample complexity remain technical challenges.
Experimental causality is investigated using peptide infusion, receptor antagonists, antibodies, gene deletion, and selective manipulation of sensory neurons. Each method has limitations. A lifelong gene deletion can produce developmental compensation, whereas an acute antagonist tests a shorter period of inhibition. Some genetic tools based on Calca expression label neuronal populations more broadly than the cells actively releasing mature CGRP during an experiment. Manipulating those neurons therefore does not necessarily isolate the action of the peptide itself.
Species differences are particularly important in receptor pharmacology. Some small-molecule antagonists have substantially different affinities at human and rodent receptors, partly because of sequence differences in RAMP1 and associated binding regions. A dose that effectively blocks human receptors may not produce equivalent receptor occupancy in a mouse. Differences in vascular anatomy, immune responses, and headache-model behavior introduce additional barriers to direct translation.
Several questions remain about therapeutic response. There is no routinely established CGRP concentration threshold that identifies a person who will respond to a CGRP-targeted drug. Migraine diagnosis remains clinical, and treatment response cannot be inferred from a single peripheral peptide measurement. Researchers are investigating whether combinations of clinical features, sensory testing, genetics, or molecular measurements can provide more reliable predictive biomarkers, but these approaches require validation across independent populations.
Other unresolved issues concern the relative importance of canonical CGRP and AMY1 receptors, the contribution of α- versus β-CGRP in different organs, and the balance between neuronal, vascular, glial, and immune targets. Long-term pharmacovigilance is especially relevant for populations underrepresented in pivotal randomized controlled trials, including pregnant people and individuals with certain major cardiovascular conditions. Tissue-specific receptor mapping and experiments that separate local signaling from circulating peptide are central to resolving these questions.