CGRP-Targeting Peptides

Human alpha-CGRP is a 37-amino-acid neuropeptide of about 3.8 kDa, released from trigeminal ganglion neurons. Infusing it into people with migraine reliably triggers an attack, and CGRP levels rise in jugular blood during spontaneous attacks — that pair of observations is why this target is unusually well validated for a pain drug. Two drug strategies exist. Monoclonal antibodies either sequester the peptide (fremanezumab, galcanezumab, eptinezumab) or block its receptor (erenumab); they are large, do not meaningfully cross the blood-brain barrier, and act on the peripheral trigeminal system with half-lives of roughly a month. Small-molecule gepants (ubrogepant, rimegepant, atogepant, zavegepant) block the CGRP receptor with half-lives of hours, and can be used both acutely and preventively. The peptide-adjacent history matters here: olcegepant (BIBN4096BS) was the first proof-of-concept CGRP receptor antagonist, effective intravenously in migraine but abandoned for lack of oral bioavailability, and telcagepant was killed by hepatotoxicity — the current generation of gepants exists because those two failed instructively.

Targets: CGRP receptor (CALCRL + RAMP1), Alpha-CGRP peptide itself, Trigeminal ganglion and trigeminocervical complex

• · Erenumab 70 or 140 mg monthly; galcanezumab 240 mg loading dose then 120 mg monthly; fremanezumab 225 mg monthly or 675 mg every three months. These are milligram-scale antibody doses, not microgram peptide doses — 70 mg is 70,000 mcg.
• · Eptinezumab 100 mg or 300 mg IV quarterly. The IV route gives near-immediate onset, which is its main advantage over the subcutaneous antibodies.
• · Ubrogepant 50 or 100 mg, may repeat once after 2 hours to a maximum of 200 mg in 24 hours; rimegepant 75 mg orally disintegrating, one dose per 24 hours. Zavegepant 10 mg is the intranasal equivalent.
• · Atogepant 10, 30 or 60 mg once daily. Rimegepant 75 mg every other day is the unusual case of one drug approved for both acute and preventive use.

Work out your exact syringe units →

• very commonConstipation— Most pronounced with erenumab and occasionally severe enough to require intervention.
• commonInjection-site reactions— Redness and stinging with the subcutaneous antibodies.
• commonNausea and somnolence— Mainly with the oral gepants.
• commonDysgeusia and nasal discomfort— Specific to intranasal zavegepant.
• uncommonNew or worsened hypertension— A post-marketing signal with erenumab; CGRP is a vasodilator, so blocking it can raise blood pressure.
• uncommonHypersensitivity reactions— Reported with all the antibodies, occasionally delayed by days.
• rareRaynaud phenomenon or worsened peripheral vasospasm— Follows logically from removing a potent endogenous vasodilator.

• Pregnancy — CGRP has roles in placental and fetal vascular adaptation, the antibodies persist for months, and these agents are avoided in anyone who may conceive during or shortly after treatment.
• Poorly controlled hypertension, especially with erenumab.
• Severe hepatic impairment for the gepants, which are hepatically cleared; the class's ancestor telcagepant was abandoned for liver toxicity.
• Strong CYP3A4 inhibitors or inducers alongside gepants — exposure changes substantially in both directions.

• redundantsubstance-p-antagonists — Both target trigeminal neuropeptide signalling; NK-1 blockade failed in migraine where CGRP blockade succeeded, so there is no case for adding it.
• cautionpt-141 — Melanocortin agonists frequently trigger headache and flushing, which can confuse whether a migraine preventive is working.
• synergyara-290-neuropathy — Non-overlapping mechanisms in neuroinflammatory pain; combining them is speculative but not pharmacologically contradictory.

• · A headache diary counting monthly migraine days — this is the entire efficacy endpoint and impressions are unreliable.
• · Blood pressure at baseline and periodically, particularly on erenumab.
• · Bowel habit; constipation is the single most common reason people stop.
• · Liver enzymes if using a gepant long-term with other hepatically cleared drugs.

Approved prescription medicines in the US, EU and most developed markets for migraine prevention and acute treatment. Not controlled substances. Olcegepant and telcagepant were discontinued in development and are not available.

• Goadsby et al. 2017, NEJM — STRIVE phase 3 trial of erenumab in episodic migraine (trial)
• Croop et al. 2019, Lancet — rimegepant for acute migraine (trial)
• Edvinsson et al., CGRP as the target of new migraine therapies — mechanistic review (review)
• Olesen et al. 2004, NEJM — olcegepant (BIBN 4096 BS) for acute migraine, the proof-of-concept trial (trial)

CGRP is the rare pain target where the human causal evidence came first and the drugs came second. Two observations built the case: intravenous CGRP infusion reliably triggers a migraine attack in people who get migraines and does not in people who do not, and CGRP concentration rises in jugular venous blood during spontaneous attacks. That is close to a Koch's-postulates argument for a neuropeptide, and nothing else in pain pharmacology has it. Mechanistically, CGRP is released from trigeminal ganglion neurons whose peripheral branches innervate the dura and meningeal vessels and whose central branches synapse in the trigeminocervical complex. Released peripherally it binds the CGRP receptor — a genuinely odd assembly of the calcitonin receptor-like receptor (CALCRL, a class B GPCR) plus RAMP1, a single-pass accessory protein that is what confers CGRP specificity, plus RCP for signal coupling. Activation is Gs-coupled: adenylyl cyclase, cAMP, PKA, with downstream vasodilation of meningeal arteries, mast cell degranulation and sensitisation of the trigeminal afferents themselves. That sensitisation, not the vasodilation, is now thought to be the part that matters — the old vascular theory of migraine has largely collapsed, and blocking CGRP works even though the vasodilation it prevents is not the pain generator. The critical anatomical fact for understanding this whole class is that neither the antibodies nor, largely, the gepants need to enter the brain. The trigeminal ganglion sits outside the blood-brain barrier. So a 150 kDa IgG that cannot cross into the CNS at all still reaches the target, which is why these are among the best-tolerated preventives ever developed for migraine — no sedation, no cognitive effect, no weight gain, none of the things that make topiramate and amitriptyline miserable. Two mechanistic subtleties are worth carrying. First, erenumab blocks the receptor while fremanezumab, galcanezumab and eptinezumab sequester the ligand; that difference probably explains why constipation is most prominent with erenumab (CGRP receptors in the gut are being continuously blocked rather than the peptide being partially mopped up) and why erenumab carries the clearest hypertension signal. Second, CGRP is a potent endogenous vasodilator with a physiological role in protecting tissue during ischaemia, in wound healing and in placental adaptation. Removing it for years is not obviously free, and the hypertension and Raynaud signals are the first evidence that it is not. The history is also instructive about what does not work: olcegepant proved the mechanism intravenously in 2004 and was abandoned for lack of oral bioavailability, and telcagepant was killed by hepatotoxicity — the current gepants exist because both of those failed in ways that told the field exactly what to fix.

The most common failure is abandoning the drug at week four. These are preventives with a slow ceiling: benefit often appears within the first month but the full effect takes three months, and the guideline decision point is at three months for exactly that reason. The mirror-image failure is staying on one for years without ever having counted. Roughly half of patients do not achieve a 50% reduction; without a diary from before you started, you cannot tell whether you are in that group, and people routinely continue an expensive monthly injection on the strength of a vague impression. The third failure is not switching. Non-response to erenumab does not predict non-response to galcanezumab or fremanezumab, and the receptor-blocker versus ligand-binder distinction is a real one — switching within the class after a genuine three-month trial is worth doing. Fourth is the constipation trap, which is specific to erenumab and specific to the 140 mg dose: it builds gradually, it is easy to attribute to something else, and by the time you decide to stop you have a month of drug still on board. Fifth is blood pressure, which nobody measures because the drug 'has no side effects' — the hypertension signal is post-marketing precisely because the trials did not catch it. Sixth is the CYP3A4 problem with gepants, where a course of clarithromycin or a switch to carbamazepine silently changes the exposure by a factor that matters. And seventh, the one with the longest consequences: the antibodies persist for five to six months, so a pregnancy decision made this month is really a decision made half a year ago. If pregnancy is a possibility, the gepants with their 11-hour half-life are a fundamentally different risk proposition from an antibody with a 28-day one.

1. 70 mgErenumab, months 1-3 — 70 mg subcutaneously once monthly. This is the starting dose and a substantial fraction of responders do fine on it. 70 mg is 70,000 micrograms — antibody dosing is milligram-scale, not the microgram scale of the peptides elsewhere in this class.
2. 140 mgErenumab, month 4 onward if response is partial — 140 mg monthly. Escalate only if the headache diary shows a partial rather than adequate response after three months. Constipation and the hypertension signal both track the higher dose, so do not start here by default.
3. 240 mgGalcanezumab, month 1 — 240 mg loading dose subcutaneously — given as two 120 mg injections at the first visit. The only true loading dose in the class.
4. 120 mgGalcanezumab, month 2 onward — 120 mg monthly maintenance.
5. 10 mgAtogepant, from day 1 — 10 mg once daily orally is the lowest approved preventive gepant dose, and the one to use with a strong CYP3A4 inhibitor on board or with severe renal impairment.
6. 60 mgAtogepant, if response is partial and tolerability fine — 30 mg or 60 mg once daily. Unlike the antibodies, the gepants can be moved up and down quickly because the half-life is hours — a wrong decision here costs you a day, not a season.

• CGRP receptor — CALCRL + RAMP1 + RCP heterocomplex — Erenumab binds the receptor with high affinity as a competitive antagonist; the gepants bind at the CALCRL/RAMP1 interface with sub-nanomolar to low-nanomolar affinity. RAMP1 is what makes the receptor CGRP-selective rather than adrenomedullin- or amylin-selective, and it is the reason a small molecule can achieve selectivity at all against a class B GPCR.

  Blockade prevents Gs coupling, so no adenylyl cyclase activation, no cAMP rise, no PKA signalling. Downstream: no meningeal vasodilation, no mast cell degranulation, and — the part that matters — no sensitisation of trigeminal afferents.

• Alpha-CGRP peptide (ligand sequestration) — Fremanezumab, galcanezumab and eptinezumab bind circulating CGRP itself with picomolar affinity rather than touching the receptor.

  The peptide is mopped up before it reaches the receptor. Functionally equivalent to receptor blockade for migraine prevention, but not identical in side-effect profile — ligand-binding antibodies leave some receptor signalling intact wherever local CGRP concentration is high, which may be why constipation is less prominent with them than with erenumab.

• Trigeminal ganglion and trigeminocervical complex — The anatomical site, not a binding target. Critically, the trigeminal ganglion lies outside the blood-brain barrier.

  This is why drugs that cannot enter the CNS still work, and why the class has almost no central side effects. It is also the strongest argument that migraine prevention here is a peripheral phenomenon.

• CGRP receptors in vascular smooth muscle, gut and skin — Same receptor, no tissue selectivity available.

  The side-effect profile in one line. Blocking a potent endogenous vasodilator raises blood pressure in some people (most clearly with erenumab), worsens Raynaud phenomenon and peripheral vasospasm in a few, and slows gut transit — constipation is the single most common reason people stop erenumab, and it is occasionally severe enough to require intervention.

• STRIVE — erenumab for episodic migraine Phase 3, randomised, double-blind, placebo-controlled · n=955 · 24 weeks · 2017

  Change in mean monthly migraine days from baseline over months 4-6: a reduction of 3.2 days with erenumab 70 mg and 3.7 days with 140 mg, versus 1.8 days with placebo (p<0.001 for both). Note the placebo arm — 1.8 fewer migraine days per month on placebo is a large effect in its own right and is the reason uncontrolled impressions about these drugs are worthless.

• Croop et al. — rimegepant orally disintegrating tablet for acute migraine Phase 3, randomised, double-blind, placebo-controlled · n=1351 · 1 weeks · 2019

  Single 75 mg dose taken during an attack. Freedom from pain at 2 hours: 21% versus 11% on placebo (p<0.0001). Freedom from the most bothersome symptom at 2 hours: 35% versus 27% (p=0.0009). Real, statistically robust, and worth reading the absolute numbers rather than the relative ones — four in five people were not pain-free at two hours.

• Croop et al. — oral rimegepant for preventive treatment of migraine Phase 2/3, randomised, double-blind, placebo-controlled · 12 weeks · 2021

  Rimegepant 75 mg every other day reduced monthly migraine days versus placebo over weeks 9-12. This is the trial that made rimegepant the unusual case of a single agent approved for both acute and preventive use.

• Olesen et al. — BIBN 4096 BS (olcegepant) for acute migraine Randomised, double-blind, placebo-controlled proof-of-concept · n=126 · 1 weeks · 2004

  Single intravenous dose over 10 minutes. The 2.5 mg dose produced a 66% response rate versus 27% for placebo (p=0.001). This is the trial that proved CGRP receptor antagonism aborts migraine in humans, and everything in the class descends from it. Olcegepant was then abandoned because it had no oral bioavailability.

Three different clocks. Acute gepants: rimegepant peaks at about 1.5 hours orally, ubrogepant similarly, and the trials measured pain freedom at 2 hours — so the honest expectation is meaningful relief within one to two hours, in about one attack in five above placebo, and taking it early in the attack matters. Intranasal zavegepant is faster still. Subcutaneous preventive antibodies: erenumab peaks in serum at about 6 days, and clinically a benefit often shows within the first month, with the full effect by month three and serum steady state at about three months of monthly dosing. Intravenous eptinezumab is the outlier — full concentration is achieved at the end of a 30-minute infusion, and benefit has been demonstrated within a day, which is its main advantage over the subcutaneous antibodies and the reason it is used when speed matters. Coming off is equally asymmetric: stop a gepant and it is gone in two days; stop an antibody and you are still substantially exposed for two months and not clear for five to six.

The most important stacking fact in this class is one people get wrong in both directions: gepants and triptans can be combined, and gepants and CGRP antibodies can be combined. Triptans are 5-HT1B/1D agonists and vasoconstrictors; gepants are CGRP receptor antagonists and are not vasoconstrictors, which is precisely why gepants are usable in people with coronary or cerebrovascular disease where triptans are contraindicated. Taking a gepant as rescue when a triptan has failed within the same attack is a recognised strategy. Using an acute gepant on top of a preventive antibody is also standard, and the trials of rimegepant included patients on preventives. What you should not do is combine two preventive CGRP agents — an antibody plus daily atogepant is redundant on the same target with no evidence of additive benefit and an unstudied safety profile. Watch the CYP3A4 axis with the gepants specifically: adding clarithromycin, itraconazole, ritonavir or a large regular grapefruit habit raises exposure meaningfully, while rifampicin, carbamazepine, phenytoin or St John's wort can drop it enough that the drug simply stops working and you conclude, wrongly, that you are a non-responder. The antibodies have no CYP interactions at all. The other important combination is with medication-overuse limits: acute gepants should be capped at roughly 10 treatment days a month, though the evidence so far suggests gepants are less prone to causing medication-overuse headache than triptans or combination analgesics, and rimegepant's every-other-day preventive schedule is direct evidence of that. Finally, the melanocortin peptides such as PT-141 frequently cause headache and flushing, which will thoroughly confound your assessment of whether a migraine preventive is working — do not start both in the same month.

This is by a wide margin the best-evidenced entry in the class and one of the best-evidenced entries on the site: multiple large phase 3 randomised trials for every approved agent, plus human provocation studies establishing that CGRP infusion causes migraine. Set that against the rest of this page — dermorphin and DALDA have no human dosing at all, the enkephalins and endomorphins have none either, ARA-290 has three small phase 2 trials — and the difference in register should be obvious. Against the older migraine preventives it displaces, the CGRP class is not dramatically more effective; topiramate and propranolol achieve broadly comparable reductions in monthly migraine days in trials. What the CGRP agents win on overwhelmingly is tolerability, and therefore adherence: no sedation, no cognitive slowing, no weight change, no paraesthesia, no mood effect. People stay on them, which is most of what makes a preventive work in real life. Against triptans for acute treatment, gepants are somewhat less effective at two-hour pain freedom but are not vasoconstrictors, so they are usable in cardiovascular disease and appear less likely to cause medication-overuse headache. And the honest framing of effect size: 40-50% of patients get a 50% reduction in migraine days. That is a genuine, mechanism-based, replicated benefit for roughly half the people who try it, and it is not a cure. Anyone selling it as one has not read the placebo arms.

$600–$1200/month. US list prices, approximate and not verified in this session — treat as an order-of-magnitude figure. What people actually pay is usually very different: manufacturer copay cards routinely reduce the commercially insured cost to near zero, and most insurers require documented failure of two or three older oral preventives (topiramate, propranolol, amitriptyline, candesartan) before approving one of these. The older preventives cost a few dollars a month, which is the real reason step therapy exists. Outside the US, availability and price vary enormously by national formulary.

• ?This entry is a target class rather than a single compound, so single-valued pharmacokinetic fields are unavoidably misleading. The tmax of 1.5 hours and volume of distribution of 120 L and 96% protein binding are rimegepant's; the 82% bioavailability and 28-day half-life and 90-day time to steady state are erenumab's. Per-agent numbers differ substantially.
• ?Pharmacokinetics for ubrogepant, atogepant, zavegepant, fremanezumab, galcanezumab and eptinezumab were not individually verified in this session; only the erenumab and rimegepant labels were retrieved.
• ?Participant count for the rimegepant preventive phase 2/3 trial is left null; I confirmed the publication but did not resolve the enrolment figure.
• ?The claim that erenumab's constipation and hypertension signals stem specifically from receptor blockade rather than ligand sequestration is a mechanistic inference from the comparative side-effect profiles, not an established finding.
• ?Long-term consequences of blocking a potent endogenous vasodilator for years — cardiovascular, wound healing, ischaemic tolerance — are genuinely unknown. The Raynaud and hypertension signals are the first data points and the follow-up is still short relative to how long people take these drugs.
• ?The 40-50% responder rate figure comes from the Core record and the general trial literature rather than a single source I resolved in this session; per-trial responder rates vary.
• ?Cost figures are approximate US list prices and were not price-checked in this session.

• A Controlled Trial of Erenumab for Episodic Migraine Goadsby PJ, Reuter U, Hallström Y, et al., New England Journal of Medicine, 2017 · PMID 29171821

  STRIVE. 955 patients, six months, the pivotal trial for the first approved CGRP-targeting antibody. The 3.2 and 3.7 versus 1.8 monthly migraine day figures come from here.

• Efficacy, safety, and tolerability of rimegepant orally disintegrating tablet for the treatment of migraine: a randomised, phase 3, double-blind, placebo-controlled trial Croop R, Goadsby PJ, Stock DA, et al., Lancet, 2019 · PMID 31311674

  The acute rimegepant registration trial. 1,351 patients evaluated, 21% versus 11% pain freedom at two hours.

• Oral rimegepant for preventive treatment of migraine: a phase 2/3, randomised, double-blind, placebo-controlled trial Croop R, Lipton RB, Kudrow D, et al., Lancet, 2021 · PMID 33338437

  The trial behind rimegepant's preventive indication, and the reason one drug can be used both as needed and every other day.

• Calcitonin gene-related peptide receptor antagonist BIBN 4096 BS for the acute treatment of migraine Olesen J, Diener HC, Husstedt IW, et al., New England Journal of Medicine, 2004 · PMID 15014183

  The proof-of-concept trial. 126 patients, single IV dose, 66% versus 27% response. Twenty years of drug development followed from this one result, and it is worth reading to see how clean an early mechanistic trial can be.

• AIMOVIG (erenumab-aooe) injection — US prescribing information Amgen, Inc., DailyMed, US National Library of Medicine

  Source of the antibody pharmacokinetics quoted here: tmax approximately 6 days, estimated absolute bioavailability 82%, volume of distribution approximately 3.86 L, effective half-life 28 days, steady state by approximately 3 months, less than two-fold accumulation, and the two-phase (target-mediated plus non-specific proteolytic) elimination.

• NURTEC ODT (rimegepant sulfate) orally disintegrating tablet — US prescribing information Pfizer Laboratories Div Pfizer Inc, DailyMed, US National Library of Medicine

  Source of the gepant pharmacokinetics: tmax approximately 1.5 hours, absolute oral bioavailability approximately 64%, steady-state volume of distribution 120 L, plasma protein binding approximately 96%, elimination half-life approximately 11 hours, CYP3A4-dominant metabolism with minor CYP2C9, excretion approximately 78% faecal and 24% urinary, and the high-fat-meal exposure reduction of 32-38%.