The therapeutic pipeline in migraine: What does the future hold?

Large conductance calcium- and voltage-activated potassium channels (BK Ca ) are expressed in neurons, smooth muscle and heart, where they are believed to have a protective function ( Fig. 1 ) [ 50 , 51 ]. When the cell membrane depolarizes and/or intracellular Ca 2+ levels are increased, BK Ca channels open causing K + efflux which leads to hyperpolarization [ 52 ]. They have also been implicated as a downstream signaling mechanism in migraine. In migraine-related structures, BK Ca are expressed in dural and cerebral arteries [ 53 ] where their opening causes vasodilation. In TNC and TG neurons they are known to decrease neuronal excitability [ 54 , 55 ]. Here, one could expect a contrasting approach to this target, BK Ca channels openers would be expected to dilate cranial vasculature, hypothetically ‘triggering’ migraine attacks, but also inhibit neuronal activity in the trigeminal nociceptive system, making them possible ‘anti-migraine’ agents. Fig. 1 Ion flow and endogenous activators of potassium channel-related targets in migraine . ATP-sensitive potassium channels (K ATP ) and large conductance calcium- and voltage-activated potassium (BK Ca ) ion channels are activated as a downstream consequence of cAMP and cGMP production via cell membrane receptors of various migraine provoking agents, including CGRP, PACAP and nitroglycerin, in vascular smooth muscle cells. As a consequence, low intracellular ATP levels open K ATP and high intracellular calcium (Ca +2 ) levels activate BK Ca , both causing potassium (K + ) efflux and hyperpolarizing the cell membrane. This mediates potent vasodilation of the dural and cerebral vasculature and activates trigeminal nociceptors that innervate the dural vasculature [ 36 ]. It has been hypothesized that this may contribute to mechanisms involved in triggering spontaneous migraine. However, centrally this mechanism decreases neuronal excitability, which could represent an anti-nociceptive mechanism. Created with Biorender.com . Fig. 1 Ion flow and endogenous activators of potassium channel-related targets in migraine . ATP-sensitive potassium channels (K ATP ) and large conductance calcium- and voltage-activated potassium (BK Ca ) ion channels are activated as a downstream consequence of cAMP and cGMP production via cell membrane receptors of various migraine provoking agents, including CGRP, PACAP and nitroglycerin, in vascular smooth muscle cells. As a consequence, low intracellular ATP levels open K ATP and high intracellular calcium (Ca +2 ) levels activate BK Ca , both causing potassium (K + ) efflux and hyperpolarizing the cell membrane. This mediates potent vasodilation of the dural and cerebral vasculature and activates trigeminal nociceptors that innervate the dural vasculature [ 36 ]. It has been hypothesized that this may contribute to mechanisms involved in triggering spontaneous migraine. However, centrally this mechanism decreases neuronal excitability, which could represent an anti-nociceptive mechanism. Created with Biorender.com . In preclinical studies, one study showed that BK Ca opening is involved in NTG-mediated, but not CGRP-mediated, dural vasculature vasodilation [ 56 ]. In contrast, another study reported that iberiotoxin (a BK Ca blocker) significantly increased CGRP release in isolated TNC, which was prevented by the BK Ca opener, NS11021. However, NS11021 did not change basal CGRP levels [ 55 ]. In contrast, systemic administration of the BK Ca opener, NS1619, inhibited neurogenically mediated dural vessel dilatation, but failed to inhibit dural-evoked TNC neuronal responses. However, in the same study, locally administrated NS1619 decreased dural-evoked TNC neuronal responses [ 57 ]. Taken together, these data suggest that neuronally targeting BK Ca opening, whether peripherally or centrally, is therapeutic in migraine models, but targeting the vasculature mediates vasodilation. In human studies, administration of MaxiPost (a BK Ca channel opener) in healthy volunteers caused headache with increased intensity when compared to placebo [ 58 ]. This was also associated with an increased diameter of the middle cerebral artery [ 58 ]. In a migraine without aura population, Maxipost (BMS-204352) infusion mediated a migraine attack in 95 % of patients compared to none in the placebo group [ 21 ]. This was also associated with higher headache intensity and increased blood flow in the middle cerebral artery [ 21 ]. In contrast to the preclinical neuronal data, these results support the blockade of BK Ca channels as a potential target for migraine treatment [ 59 , 60 ]. Together, the data for K ATP and BK Ca channel openers provide a potential opportunity for drug development, based largely on their profound ability to trigger headache and migraine attacks in migraine patients. It is argued that this is predominantly at the vascular level; both are present in the smooth muscle of cranial vessels, and they are believed to be involved in the downstream signaling mechanisms related to many migraine provoking agents ( Fig. 1 ). However, targeting these potassium channels has proved less than conclusive. In the case of K ATP , this seems to be due to the relatively non-specific molecules that have been used for screening potential efficacy, with the clear need to develop more selective K ATP blockers that target the SUR2B subunit. For BK Ca it is more confusing. Neuronal BK Ca channel opening inhibits neurogenic dural vasodilation, thought to be by blocking the release of neuropeptides from trigeminal nerve endings, as well as neuronal trigeminal responses, which is a reliable and predictive migraine model [ 61 ]. Yet, this is in direct contrast to their effects on vessels and the clinical data related to migraine triggering. While development for K ATP openers as a therapeutic target continues, for BK Ca progress is somewhat paused. This does highlight the risks and uncertainty with drug development, and how on the surface both targets do appear to have equal potential and yet while one proceeds forward another is strategically stopped. It may also reflect that robust preclinical screening is necessary that factors in both vascular and neuronal models before clinical efficacy can be predicted.

Another TRP target is the transient receptor potential melastatin 3 (TRPM3) ion channel ( Fig. 2 ). TRMP3 is highly expressed in human tissues, especially in sensory neurons [ [85] , [86] , [87] ]. These channels detect thermal, mechanical, and chemical stimuli [ 87 ]. Importantly, TRPM3 is a cation-specific channel activated directly by sex hormones, indicating a potential role in sex-related differences in migraine [ [88] , [89] , [90] , [91] ]. TRPM3 channels are extensively expressed in small-diameter sensory neurons in both mice [ 92 ] and humans [ 93 ], where they are implicated in peripheral sensitization associated with migraine. TRPM3 channels respond to endogenous neurosteroid pregnenolone sulfate (PregS) [ 89 ], elevated temperatures (>37 °C) [ 94 ], and intracellular lipid signaling via phosphatidylinositol 4,5-bisphosphate [ 90 ]. Activation results in increased intracellular Ca 2+ concentrations [ 95 ], which facilitates the release of migraine-associated mediators, such as CGRP and glutamate, thereby augmenting neuronal excitability. Changing Mg 2+ concentration also affects TRPM3 activation state, with elevated intracellular Mg 2+ concentrations inhibiting TRPM3 activity, whereas reduced Mg 2+ levels enhancing it [ 90 ]. TRPM3 is expressed in trigeminal and meningeal afferents [ 89 , 96 ]. Immunohistochemical analyses reveal that TRPM3 is broadly distributed across trigeminal, glossopharyngeal, nodose and vagal ganglia, particularly in neurons projecting to the facial skin, oral mucosa, and dental pulp [ 97 , 98 ]. TRPM3 frequently co-localizes with TRPV1 and CGRP, key molecules involved in pain transmission, and is predominantly observed in unmyelinated, IB4-positive neurons. These findings suggest that TRPM3 plays a significant role in thermosensitive nociception in orofacial regions, highlighting its potential as a therapeutic target. Mechanistic studies have implicated TRPM3 in spinal nociceptive processing independent of its function as a thermoreceptor [ 99 ], via glutamatergic transmission of primary sensory neurons [ 95 ]. In the hemi-skull preparation, the TRPM8 agonist, CIM0216, induces CGRP release in TG and dural tissues in male and female rats [ 96 ]. In the same study in a functional assay, CIM0216 did not mediate periorbital hypersensitivity in either male or female rats when given systemically, suggesting TRPM8 activation is not mediating central sensitization. In relation to functional trigeminal nociceptive processing, using the hemi-skull preparation, activation of TRPM3 channels in peripheral meningeal afferents (using agonists CIM0216 or PregS) caused stronger nociceptive firing in female mice compared to males [ 89 ]. Given that TRPM3 channels are under extensive hormonal regulation and the significant differences in neuronal responses to TRPM3 activity observed between males and females, it was proposed that TRPM3 channels may hold a pivotal role in the mechanisms responsible for the known sex differences in migraine prevalence. TRPM3 has also been linked to the molecular mechanisms of photophobia in migraine via a model of light sensitivity. It was found that the stress-responsive long non-coding RNA, NEAT1, modulates TRPM3 expression through the miR-196a-5p/TRPM3 regulatory axis. Specifically, umbellulone, a migraine trigger that mediates light sensitivity in rodents, induces upregulation of NEAT. This suppresses miR-196a-5p activity via a competing endogenous RNA mechanism, thereby enhancing TRPM3 expression. The resulting increase in TRPM3-mediated calcium influx and neuroinflammatory signaling may contribute to trigeminovascular sensitization and photophobic behavior, potentially linking NEAT1-mediated transcriptional regulation to the pathophysiology of migraine and novel therapeutic target. TRPM3 also plays a crucial role in vascular function, particularly in the regulation of vascular tone. Recent research investigated the vasoactive properties of TRPM3 in human blood vessels, focusing on its expression, localization, and pharmacological behavior. It was demonstrated that activation of TRPM3 by PregS leads to maximal relaxation in human coronary arteries and middle meningeal arteries, with this effect being significantly more pronounced in females [ 100 ]. This vasodilation was blocked by the TRPM3 antagonist isosakuranetin, the NMDA receptor antagonist MK-801, and l -NAME, indicating involvement of both TRPM3 and NMDA pathways. Collectively, these preclinical data position TRPM3 as a promising therapeutic target for neurovascular disorders such as migraine, particularly with its potentially sexual dimorphic effects, with the advantage of minimal adverse coronary effects. This has resulted in the clinical development of BHV-2100, a selective TRPM3 antagonist. Preliminary data from a Phase 2, randomized, placebo-controlled trial to assess the efficacy, safety, and pharmacokinetics of BHV-2100, a novel, selective, oral TRPM3 antagonist, for the treatment of acute migraine has recently been present at the 2025 Annual Meeting of the American Academy of Neurology [ 101 ]. It showed good safety, predictable dose-related pharmacokinetics, and rapid target engagement. If the results are positive, BHV-2100 could advance to Phase 3, and may provide a novel targeted approach via a new pathway for managing pain.

TRPM8 constitutes a component of the eight-member melastatin subfamily within TRP ion channels ( Fig. 2 ). It has been identified in various tissues, including both neuronal and non-neuronal types. The majority of research has focused on its role in sensory neurons, where it is primarily recognized as a sensor for cold temperatures. TRPM8 functions as a multimodal, non-selective cation channel, activated by low temperatures (8–28 °C) and chemical agents such as menthol and icilin [ 63 , 64 ]. TRPM8 has been associated with migraine through clinical genetic studies and preclinical models [ 65 ]. Activation of this channel facilitates the influx of Ca 2+ , resulting in membrane depolarization and the subsequent initiation of a secondary messenger signaling cascade. In the context of migraine, the role of TRPM8 seems to be complex, potentially involving both pro- and anti-nociceptive effects depending on its specific situation and how it is activated [ 66 , 67 ]. Fig. 2 Ion flow and endogenous activators of TRPM channel-related targets in migraine . Transient receptor potential melastatin (TRPM) channels, TRPM3 and TRPM8, are activated endogenously by certain temperature changes and sex hormones. Once open, they promote sodium (Na + ) and Ca +2 influx which causes depolarization of the membrane. This increases release of CGRP and glutamate with possible activation of dural-trigeminal nociceptors. They are also thought to be involved in dural vasodilation and photophobia which may contribute a role in migraine. Created with Biorender.com . Fig. 2 Ion flow and endogenous activators of TRPM channel-related targets in migraine . Transient receptor potential melastatin (TRPM) channels, TRPM3 and TRPM8, are activated endogenously by certain temperature changes and sex hormones. Once open, they promote sodium (Na + ) and Ca +2 influx which causes depolarization of the membrane. This increases release of CGRP and glutamate with possible activation of dural-trigeminal nociceptors. They are also thought to be involved in dural vasodilation and photophobia which may contribute a role in migraine. Created with Biorender.com . Initial evidence implicating TRPM8 in the pathophysiology of migraine emerged from genome-wide association studies (GWAS) that identified several single-nucleotide polymorphisms (SNPs) associated with migraine susceptibility. The SNPs rs10166942 and rs7577262 are associated with migraine with and without aura [ [68] , [69] , [70] , [71] ], although these were not observed in a Chinese cohort [ 72 ], suggesting population-specific genetic differences in migraine susceptibility. In contrast, carriers of the rarer rs10166942 (C) allele have a lower migraine risk and reduced sensitivity to cold pain [ 73 ]. This information not only supports a role for TRPM8 in cold perception but also suggests that reduced TRPM8 activity might confer a protective effect against migraine. Anatomical studies in animal models have furnished insights into TRPM8 within structures related to migraine. Anatomical mapping revealed only 3–4 % of adult dural afferent TG fibers are TRPM8-positive [ 74 ], although it is expressed during prenatal development before undergoing specific axonal pruning postnatally, indicating a transient developmental role [ 75 ]. This supports the notion that TRPM8 contribution to dural nociception is limited and likely indirect. Evidence from migraine models indicates that its expression may be dynamically regulated; for instance, in the NTG model, expression of TRPM8 decreases in the dura mater of rats, suggesting that migraine-related inflammatory or vascular processes may suppress TRPM8 activity under migraine-like conditions [ 76 ]. Functional studies have also assessed the activity of TRPM8 at the level of the dura mater. Local dural application of the TRPM8 agonist, menthol, in the presence of inflammatory mediators causes an analgesic effect on head-directed nocifensive behaviors [ 75 ], whereas another TRPM8 agonist, icilin, on its own mediates facial cutaneous allodynia that was attenuated by a TRPM8 antagonist and sumatriptan [ 77 ]. Menthol has also been demonstrated to dose-dependently cause increased excitability of dural-responsive trigeminal neurons but inhibit neurogenic dural vasodilation [ 78 ]. A possible mechanism to explain the anti-nociceptive effect of TRPM8 agonism is that during meningeal inflammation, TRPV1-positive neurons mediate a cranial pro-nociceptive/allodynic effect. Over time, TRPM8 is upregulated in these neurons, including those that innervate both the dura and the face. In this state, activating facial TRPM8 can inhibit TRPV1-mediated pain, which partly explains how dural TRPM8 activation can produce either pro- or anti-nociceptive effects, depending on the microenvironment [ 79 ]. In NTG and CGRP migraine models, measuring hind paw withdrawal thresholds as a surrogate of migraine-like behavior, the reduction of thresholds normally induced by NTG and CGRP was attenuated and prevented in TRPM8−/− knockout mice or using a TRPM8 antagonist [ 80 ]. These data support a role of TRPM8 in both evoked and spontaneous migraine-like pain, in both acute and chronic settings [ 80 ]. It is thought that this might be by preventing the release of CGRP within the trigeminovascular system [ 81 ]. This aligns with the known properties of TRPM8 as a non-selective cation channel with high calcium permeability, which promotes neurotransmitter exocytosis through calcium influx. This mechanistic role reinforces the genetic links found in human studies and further confirms TRPM8 as a central player in migraine pathophysiology, emphasizing its potential as a promising target for early therapeutic intervention. Lastly, in a surrogate model of migraine aura, using measures of cortical spreading depression (CSD), it was found that increased levels of TRPM8 in the cortex increased susceptibility to CSD, as well as neuroinflammation induced by CSD [ 82 ]. It was also found that TRPM8 on peripheral trigeminal neurons facilitated CGRP release and triggered a neuroinflammatory response, supporting previous studies. This highlights both peripheral and central mechanisms in migraine pathophysiology. Taken together, the evidence points to a role for meningeal, peripheral, and central TRPM8 in migraine-related pain signaling, highlighting the need for further research into its potential as an anti-migraine therapy. While preclinical data are encouraging, the clinical relevance and application of TRPM8 is limited. However, there have been a couple of development programs for TRPM8 blockers for the treatment of migraine. In a phase 1 study, AMG333, a selective TRPM8 antagonist, was tested in both healthy subjects and migraine patients [ 83 ]. The study was completed in 2014, but the results are not published. Also, Elismetrep (K-304), which has previously been screened for diabetic neuropathy is currently undergoing a Phase 2B trial for the treatment of migraine [ 84 ]. At the time of writing, the study is active and therefore data are not available but this represents a promising opportunity for this target in the pipeline of migraine therapeutics.

MGE, ABSA and SA designed the concept of the review. MGE, ABSA, MRR and SA drafted, wrote, revised, and approved the manuscript.

This review was supported by grant funding 10.13039/100000005 within the United States from the US Department of Defense ( W81XWH2110485 ) and 10.13039/100000002 NIH ( NS120930 ). The funding sources did not influence the work.

Oxytocin (OXT) is a neuropeptide secreted principally by the hypothalamic paraventricular and supraoptic nuclei ( Fig. 3 ). It reaches the general circulation through the posterior portion of the pituitary gland and other structures in the CNS and the CSF through axonal and dendritic release [ [102] , [103] , [104] , [105] ]. OXT acts through activation of its receptors (OTR), which are coupled to G q or G i proteins, expressed widely in the CNS and peripheral tissue [ 106 ]. Studies have shown the possible role of oxytocin in inhibition of pain through cortical, descending inhibitory and peripheral mechanisms [ 107 , 108 ]. In a migraine context, OXT and OTR have been reported in many migraine-related areas, including TG and TNC neurons, hypothalamus, periaqueductual gray (PAG), locus coeruleus (LC) and insular cortex [ 106 , 109 , 110 ]. It is not found in dural or cranial arteries. Fig. 3 Production and receptors of hypothalamic peptide-related targets in migraine . Prolactin (PRL) is secreted by the anterior portion of pituitary gland. The prolactin receptor (PRLR) belongs to the cytokine receptor family associated with non-receptor tyrosine kinase Janus kinase 2 (JAK2) and the signal transducer and activator of transcription 5 (Stat5) pathway, considered the primary downstream signaling mechanism for PRLR activation. PRLR activation modulates TRP and K ATP channel activity, which increases neuronal excitability and induces vasodilation, both are related to pronociceptive effects in migraine pathophysiology. Oxytocin (OXT) is secreted from the hypothalamus and can produce ‘anti-migraine’ effects through activation of oxytocin receptor (OTR). This decreases the level of CGRP release and may engage mechanisms of endogenous descending inhibitory modulation. Orexins are neuropeptides secreted by the hypothalamus that include orexin A (OxA) and orexin B (OxB). Orexin receptor 1 (OX1R) is activated by OxA and decreases neuronal excitability and vasodilatation [ 134 , 135 ]. The orexin receptor 2 (OX2R) is activated by both OxA and OxB with a specific effect on cortical spreading depression (CSD) in males only. OxB and OX2R appear to demonstrate sex-specific effects [ 144 ]. OxB activation of OX2R in dural-trigeminal afferents is restricted to males only, with little effect in females, while dual orexin receptor antagonism (DORA) is specifically effective in males, using migraine models. Created with Biorender.com . Fig. 3 Production and receptors of hypothalamic peptide-related targets in migraine . Prolactin (PRL) is secreted by the anterior portion of pituitary gland. The prolactin receptor (PRLR) belongs to the cytokine receptor family associated with non-receptor tyrosine kinase Janus kinase 2 (JAK2) and the signal transducer and activator of transcription 5 (Stat5) pathway, considered the primary downstream signaling mechanism for PRLR activation. PRLR activation modulates TRP and K ATP channel activity, which increases neuronal excitability and induces vasodilation, both are related to pronociceptive effects in migraine pathophysiology. Oxytocin (OXT) is secreted from the hypothalamus and can produce ‘anti-migraine’ effects through activation of oxytocin receptor (OTR). This decreases the level of CGRP release and may engage mechanisms of endogenous descending inhibitory modulation. Orexins are neuropeptides secreted by the hypothalamus that include orexin A (OxA) and orexin B (OxB). Orexin receptor 1 (OX1R) is activated by OxA and decreases neuronal excitability and vasodilatation [ 134 , 135 ]. The orexin receptor 2 (OX2R) is activated by both OxA and OxB with a specific effect on cortical spreading depression (CSD) in males only. OxB and OX2R appear to demonstrate sex-specific effects [ 144 ]. OxB activation of OX2R in dural-trigeminal afferents is restricted to males only, with little effect in females, while dual orexin receptor antagonism (DORA) is specifically effective in males, using migraine models. Created with Biorender.com . In animal models, local OXT administration into the trigeminocervical complex (TCC) decreases dural-evoked neuronal responses [ 111 ]. OXT also attenuated capsaicin-mediated CGRP release from dural nociceptors in ex-vivo models [ 109 ]. Interestingly, OTR expression increased significantly in TG neurons in the chronic NTG migraine model [ 112 ] and under inflammatory conditions or electrical stimulation of cutaneous afferents [ 109 ]. However, in an orofacial pain model, the anti-nociceptive effects on OXT in TG were mediated by vasopressin receptors (V 1A ) rather than OTR [ 113 ]. Projections from the hypothalamic paraventricular nucleus have been shown to modulate TCC neuronal activity, with a possible role of OXT, PACAP and serotonin [ 114 , 115 ]. Finally, repeated intranasal OXT administration in the chronic NTG migraine mouse model attenuated CGRP release in the TNC and blocked mechanisms of central sensitization. These studies support a possible use of intranasal OXT as a preventive treatment in migraine [ 112 ]. In humans, a cross-sectional study showed that patients with migraine, particularly chronic migraine, show a higher response to capsaicin-induced pain during interictal periods, which was associated with increased oxytocin plasma levels when compared to controls [ 116 ]. These increased oxytocin levels could be an endogenous mechanism to mitigate emotional distress during the interictal period, but might help to decrease overall pain during migraine attacks. In a double-blind placebo-controlled pilot study of 80 subjects, a single dose of intranasal OXT (32 IU) did not show significant decrease of pain intensity when compared to placebo in low frequency episodic migraine patients [ 117 ]. However, patients treated with intranasal OXT did report slightly higher satisfaction with regard to other migraine-related symptoms, such as photophobia and nausea. In a phase 2 interventional study in a population of high frequency episodic and chronic migraine female patients, across 12 sites in three countries, intranasal OXT (T-001, 30IU) was given as needed [ 118 ]. The results showed a good safety profile, with a reduction in headache frequency, but the primary endpoint of reduction of migraine headache days compared to a 28-day baseline period before patient enrollment in the study, was not statistically significant compared to placebo [ 117 ]. It was reported that this was largely due to a high placebo rate at one site. A sub-analysis of the sites with a more typical placebo rate did suggest a significant difference between OXT and placebo groups. In a more recent phase-2 RCT, the efficacy of intranasal OXT (TNX-1900, 30 IU) as a low dose (once/day) or high dose (twice/day) for 3 months was tested, compared to placebo in chronic migraine prevention [ 119 ]. The study was completed with 78 participants, predominantly males, and initial results (as reported on ClinicalTrials.gov ) show that 34.5 % of patients who received high dose of TNX-1900 showed ≥50 % reduction in the number of migraine days, compared to 48.4 % in low dose group and 55.6 % in placebo. Comparative statistics are not currently available. Another factor to consider with OXT is that it may be specifically effective as a therapeutic in menstrual migraine, which is the presence of migraine during the perimenstrual period (pure menstrual migraine) [ 120 ]. OXT release and the expression of its receptor are regulated by estrogen. As estrogen levels drops there is also a drop in OXT concentration and trigeminal OTR expression [ 121 ]. This coincides with the development of menstrual migraine and suggests that either migraine triggering is a direct consequence of the drop in estrogen and OXT/OTR level, or these changes lower the threshold to trigger a migraine. Therefore, increasing OXT levels, intuitively, would provide some protection from migraine triggering if OXT is a relevant mechanism. While these studies suggest there may be a signal of a response with intranasal OXT, there is a lack of consistency of responses across studies, with noted high placebo rates. Possible causes could be related to methodological differences, sub-optimal dosing, sex differences, the fact that intranasal administration may produce a high placebo given the invasive nature of delivery, and whether OXT levels inside the nervous system after exogenous use is enough to mediate significant effects [ 104 , 122 ]. Studies have shown that exogenous OXT can reach the CNS through olfactory and trigeminal fibers and despite its short half-life, it increases endogenous OXT secretion both in specific brain nuclei and CSF, which could be responsible for long term actions on neuronal activity [ [123] , [124] , [125] ]. Also, perhaps the wrong patient population is being chosen for these studies, with an effect in the menstrual migraine patient not, as yet, explored. In conclusion, OXT plays an important role in neuronal modulation within many neurological conditions, including migraine, and continues to represent a novel opportunity as a migraine therapeutic. Its position as an advanced pipeline target in migraine remains, although modifications in the clinical approach of using intranasal OXT in migraine management may need to be reassessed to manage optimal outcomes with the most relevant patient population. For a detailed analysis of challenges in the use of intranasal OXT we refer the reader to several recent publications [ 104 , 122 ]. The orexinergic (also called hypocretin) system consists of two G-protein-coupled receptors, OX1R and OX2R, activated by the neuropeptides orexin A (OxA) and orexin B (OxB), derived from prepro-orexin [ 126 , 127 ]. Both receptors activate a transduction pathway that elevates intracellular Ca 2+ levels via Gq-mediated phospholipase C signaling ( Fig. 3 ). Although OX1R is selective for OxA, OX2R binds both peptides, with higher affinity for OxB [ 127 , 128 ]. Orexins, produced exclusively in the lateral hypothalamus, are highly conserved and co-released with glutamate and dynorphin [ 126 , 129 ]. Orexinergic neurons project widely throughout the CNS, targeting regions such as the PAG, spinal cord, hippocampus, cortex, brainstem nuclei, and spinal trigeminal nucleus, and other brain areas [ 130 , 131 ]. This system regulates neuroendocrine and autonomic functions, playing a key role in migraine pathogenesis [ 129 , 132 ]. Preclinical studies in rats report that OX1R and OX2R are primarily localized in neuronal cell bodies of the TG and in laminas I–II of spinal trigeminal nuclei; areas involved in trigeminal nociceptive processing [ 133 ]. Mechanistic studies have demonstrated the active role of these receptors in modulating nociception, suggesting a physiologically relevant function in the sensitization and transmission of craniofacial pain. Systemic OX1R activation via OxA inhibits dural-evoked meningeal vasodilation and trigeminal neuronal firing, with OxB having little effect [ 134 , 135 ]. Local activation of hypothalamic orexin receptors via microinjection of OxA or OxB into the posterior hypothalamus causes differential modulation of dural-trigeminovascular nociceptive processing, with OxA producing an anti-nociceptive effect and OxB being pronociceptive. These results suggest that hypothalamic orexin signaling can centrally modulate head pain [ 136 ]. Similarly, anti-nociceptive effects via OX1 receptor activation were observed following OxA administration into the ventrolateral PAG [ 137 ]. From a therapeutic perspective, a dual orexin receptor antagonist (DORA) reduced neurogenically-mediated dural meningeal vasodilation, trigeminal firing, and susceptibility to CSD-phenomena associated with migraine aura [ 138 ]. Together, these preclinical data support the ability of orexins and their receptors to modulate trigeminal nociception as it relates to migraine headache. In clinical studies, a genetic association between orexin and migraine has been established, with a significant variation in the frequency of rs2271933 polymorphisms within the OX1R gene observed between individuals with migraine and control subjects. The A allele correlates with an increased risk of migraine without aura, suggesting that OX1R may confer a genetic predisposition [ 139 , 140 ]. Furthermore, OxA levels in CSF have been reported to be higher in patients with chronic migraine (CM) and medication-overuse headache (MOH) compared to healthy controls [ 141 ]. Together, the data suggest that orexins and their receptors may represent a relevant therapeutic target. To date, the only clinical trial of this peptide class was conducted using a DORA molecule. In a randomized, double-blind, placebo-controlled pilot trial, 235 patients with 4–14 migraine days per month were treated with either filorexant (10 mg nightly) or placebo for three months. While filorexant was generally well tolerated, there was no statistically significant difference between groups with the primary endpoint of reduction of monthly migraine or headache days, with the conclusion that filorexant did not show clinical efficacy [ 142 ]. Somnolence was reported as the most significant adverse event. While these data suggest that non-selective (or dual) orexin antagonism is not a viable therapeutic approach, more recent data may suggest we need to reassess these findings or at least rethink orexins as a pipeline target. It is noteworthy that the majority of the preclinical studies reported were conducted in male rodents, while the filorexant clinical trial included 85 % females. Several recent studies have demonstrated sexual dimorphism in nociceptor responses [ 143 ], and specifically related to OX2R [ 144 ]. Here, a male-specific mechanism has been identified for OX2R with respect to trigeminal nociceptive processing. Supradural administration of OxB elicited pain behaviors, photophobia, and heightened ERK phosphorylation in the TNC, exclusively in male mice, with elevated OX2R expression and increased neuronal excitability. Inhibition of orexin signaling with a DORA or the application of CRISPR/Cas9 to decrease OX2R expression effectively prevented migraine-like pain in a dural-inflammatory soup model, in males [ 144 ]. These findings open a gateway to explore a sex-specific target for migraine, selective for males only. More research is necessary in the orexin space to validate this male-specific target. Screening responses, with either DORA or selective OX2R antagonism in a platform of migraine-related models, comparing male and female responses, is necessary to drive this research to the next step and drug development. It may also be worthwhile to revisit the clinical data to determine if there are differences in male and female responses to filorexant in the RCT. However, these findings do open a gateway to explore a sex-specific target for migraine. They also underscore the sexual dimorphism in migraine pathophysiology and emphasize the importance and opportunity of independent, sex-specific strategies, in treatment development and clinical research. In conclusion, orexins have taken a new and positive direction in their therapeutic development for migraine, with a renewed focus in their therapeutic pipeline. Prolactin (PRL) is a 23-kDa protein endocrine hormone, synthesized and secreted by lactotroph cells within the anterior pituitary gland and by extrapituitary sources, including mammary glands, ovaries, prostate gland, testes, endothelial cells, and adipose tissue [ [145] , [146] , [147] ]. PRL also interacts within the CNS, influencing pain and migraine mechanisms, stress, and dopamine regulation [ 148 ], having profound sexually dimorphic effects, particularly in females [ 149 ]. PRL binds directly to a transmembrane receptor (PRLR, long and short isoforms), which belongs to the cytokine receptor family ( Fig. 3 ) [ 145 , 150 ]. PRL secretion responds to stimulation by estrogen, vasoactive intestinal peptide, and thyrotropin-releasing hormone. High PRL levels activate tuberoinfundibular dopamine neurons in the hypothalamic arcuate nucleus, leading to dopamine release. This in turn inhibits PRL via D2 receptors through a short-loop feedback mechanism, highlighting PRL as the only pituitary hormone whose secretion is controlled by dopamine [ 145 , 151 ]. In preclinical studies, both PRL and PRLRs are found along the trigeminal pain pathway, including peripheral-TG and central-TNC trigeminal nociceptive nerve terminals [ 152 ]. PRLRs are expressed in Nav1.8-positive TG neurons [ 153 ]. PRL modulates sensory neurons under estradiol control, with PRL expression increasing (∼40-fold) in TG neurons of ovariectomized rats after estradiol replacement. It shows that PRL exists in a releasable pool within sensory neurons and that both PRL and its receptor are responsive to estradiol. Estradiol also modulates PRL-induced trigeminal sensitization of capsaicin-evoked responses in vitro and in vivo , and PRL increases TRPV1 phosphorylation in cultured TG neurons [ 154 ]. These findings show that estradiol precisely controls the signaling components of PRL and their effects on TRPV1 sensitivity in TG neurons, revealing a possible mechanism by which estrogen affects sensitivity and contributes to the development of painful disorders in females. It is noted however that TRPV1 as a therapeutic target in migraine failed in a preclinical screening model [ 155 ] and in a controlled clinical trial [ 156 ]. In migraine-like models, PRL application to the dura mater triggered a nociceptive migraine-like behavioral response that was blocked by a CGRP receptor antagonist, in females only. Similarly, dural-CGRP only elicited migraine-like behaviors in females, which was blocked by a PRLR antagonist [ 157 ]. These data support the cross-talk between prolactin and CGRP to produce migraine-like nociceptive responses, and the sexual dimorphism of PRL, as responses are restricted to females. In another study, glibenclamide, the K ATP channel blocker, significantly attenuates migraine-like behaviors induced by dural prolactin in female mice [ 158 ]. Together, it suggests that the effects of prolactin in sensitizing trigeminal neurons is via CGRP and its subsequent downstream molecular signaling that activates K ATP channels to contribute to migraine-like behaviors. TRPA1 channels have also been implicated in prolactin-induced nociception. In a mouse endometriosis model, elevated circulating PRL contributes to migraine vulnerability by sensitizing trigeminal neurons. Endometriosis increases systemic PRL levels, which in turn upregulates the prolactin receptor on trigeminal neurons. This shift promotes pronociceptive signaling, sensitizes TRPA1 channels, and primes neurons to respond to stimuli that are otherwise non-painful. Subsequent exposure to a low dose of a TRPA1 activator enhances neuronal excitability and triggers migraine-like pain responses [ 159 ]. This supports a link between the nociceptive effects of prolactin and TRPA1 signaling. Prolactin also increases migraine-like behaviors in stressed rodents. Stress-induced prolactin release triggers acute cephalic allodynia in mice, promoting latent sensitization, resulting in long-lasting priming of trigeminal nociceptors. This results in migraine-like pain in response to normally innocuous stimuli [ 160 ]. Here, PRL plays a female-specific role through the activation of the PRLR long isoform on sensory neurons [ 153 ]. A sexual dimorphic effect was also observed in the sumatriptan medication overuse headache model, as a surrogate of chronic migraine. A single injection of sumatriptan increased serum PRL levels in female but not in male mice. However, while repeated sumatriptan administration promoted cutaneous allodynia in both sexes, it downregulated the TG PRLR long isoform, without altering the PRLR short isoform, only in females. Similarly, co-administration of sumatriptan with cabergoline, a dopamine receptor agonist that inhibits circulating prolactin, prevented allodynia and down-regulation of PRLR long isoform, again only in females [ 161 ]. These findings provide a more comprehensive understanding of the potential increased susceptibility of females to developing chronic migraine, attributable to prolactin receptors activated by subthreshold stimuli. In support of these mechanistic studies, research in patients consistently demonstrates that serum prolactin levels are elevated in persons with migraine compared to those without [ [162] , [163] , [164] , [165] , [166] ]. Moreover, the presence of prolactinomas (prolactin-secreting pituitary adenomas) is associated with an increased risk of headache in otherwise healthy individuals and with more frequent or severe migraine episodes in individuals with migraine [ [167] , [168] , [169] , [170] , [171] ]. These findings imply a potential role for prolactin and its receptors in the pathophysiology of migraine. In a treatment study, flunarizine and topiramate were equally efficacious at improving non-headache symptoms, but only flunarizine increased PRL levels, while slightily reducing dopamine, supporting a role in the DA-PRL axis. Topiramate had no effect on hormonal levels [ 172 ]. These data emphasize a potential connection between prolactin and the underlying mechanisms of migraine, especially in elucidating the greater incidence of migraine in females. Targeting PRL or its receptor, or limiting the levels of circulating PRL, represents an innovative and intricate therapeutic pipeline approach. It necessitates a meticulous assessment of both efficacy and safety, given the multifaceted physiological functions of PRL and its interactions with various hormonal and neurotransmitter systems. Reactive nitrogen and oxygen species (reactive nitroxidative species; RNOS) are a group of molecules that produce oxidative, nitrosative, and nitroxidative stress in cells and are considered to be pronociceptive. These include nitric oxide (NO), superoxide (O 2 .- ; SO), and peroxynitrite (ONOO − ; PN), which is the downstream reaction product of NO and SO [ 173 , 174 ]. NO was discussed in detail in a previous special issue of Neurotherapeutics [ 175 ], but is perhaps not considered a pipeline target for migraine anymore. However, SO and PN may represent novel targets for development. Several clinical studies have highlighted that platelet levels of both SO-and PN-related proteins are disrupted in people with migraine [ 176 , 177 ]. This suggests a potentially mechanistic role in migraine pathogenesis. While SO has been demonstrated to play a role in the nitroglycerin model of migraine [ 178 ], it is also an important molecule in learning and memory development [ 179 ], limiting opportunities for direct targeting. Thus, sparing SO, while directly targeting PN, would be considered a more optimal approach for potential drug development [ 180 ]. PN is known to be critical for the development and maintenance of spinal sensitization in preclinical models of persistent neuropathic and inflammatory pain [ [180] , [181] , [182] , [183] , [184] ]. It is also believed to have no known role in acute physiological nociception [ 185 , 186 ]. During persistent pain over-production of PN leads to the post-translational nitration and activation of NMDA receptors. It also causes the nitration and inactivation of manganese superoxide dismutase (MnSOD) and glutamate transporters and glutamine synthetase [ 173 , 180 , 181 ]. The net result is increased levels of glutamate within the synapse and enhanced glutamatergic signaling ( Fig. 4 ), which mediates sensitization of nociceptive spinal neurons and persistent pain. In support of this, targeting PN production both prevents and aborts persistent nociceptive responses. Together, it provides compelling evidence for PN as a novel target for pain management [ 173 , 180 ]. In the context of migraine, in models of dural-trigeminovascular nociceptive activation and sensitization, targeting PN with a decomposition catalyst (PNDC; FeTPPS) inhibits central trigeminal neuronal responses and prevents the development of trigeminal neuronal sensitization and hypersensitive responses to intra and extracranial stimulation [ 187 ]. It also prevents the development of periorbital hypersensitivity [ 187 ]. In a model of trigeminal latent sensitization, using repeated restraint stress that is unmasked with low dose sodium nitroprusside (SNP), PNDCs attenuate the development of periorbital hypersensitivity after unmasking with SNP and accelerates recovery after restraint stress, in both male and female mice [ 188 ]. Lastly, a footprint marker for PN production, 3-nitrotyrosine, was increased in trigeminal ganglia and dura mater of mice post repeated stress and after SNP (14 days post-stress) [ 188 ]. Together, these data support a role of PN in mechanisms related to both chemical and endogenous triggering of migraine and as a potential therapeutic target. Fig. 4 Reactive nitroxidative mechanisms in persistent pain . Nitroxidative species (nitric oxide, NO; superoxide, O 2 .-/ SO; and peroxynitrite, ONOO − /PN) are implicated in persistent pain. PN is the reaction product of NO and SO in both glial cells and neurons. PN has the ability to post-translationally nitrate proteins. It nitrates and inactivates mitochondrial manganese superoxide dismutase (MnSOD), which is involved in tightly controlling SO levels, providing a ‘feed-forward’ mechanism for the production of SO-derived PN [ 173 , 180 , 181 ]. PN subsequently enhances glutamatergic signaling via the nitration and activation of NMDA receptors. PN also nitrates and inactivates glutamate transporters (GLT-1, GLAST, and EAAC1) that remove glutamate from the synapse, as well as glutamine synthetase (GS) that converts glutamate to glutamine that is then taken back up by neurons via the glutamine transporters. The net result is toxic levels of glutamate, enhanced glutamatergic NMDA activation, as well activation of neuroimmune responses. This ultimately leads to sensitization of nociceptive neurons. Fig. 4 Reactive nitroxidative mechanisms in persistent pain . Nitroxidative species (nitric oxide, NO; superoxide, O 2 .-/ SO; and peroxynitrite, ONOO − /PN) are implicated in persistent pain. PN is the reaction product of NO and SO in both glial cells and neurons. PN has the ability to post-translationally nitrate proteins. It nitrates and inactivates mitochondrial manganese superoxide dismutase (MnSOD), which is involved in tightly controlling SO levels, providing a ‘feed-forward’ mechanism for the production of SO-derived PN [ 173 , 180 , 181 ]. PN subsequently enhances glutamatergic signaling via the nitration and activation of NMDA receptors. PN also nitrates and inactivates glutamate transporters (GLT-1, GLAST, and EAAC1) that remove glutamate from the synapse, as well as glutamine synthetase (GS) that converts glutamate to glutamine that is then taken back up by neurons via the glutamine transporters. The net result is toxic levels of glutamate, enhanced glutamatergic NMDA activation, as well activation of neuroimmune responses. This ultimately leads to sensitization of nociceptive neurons. The commercially available PN-targeted molecules are metal-based, which provide limitations for oral bioavailability and many do not spare SO, and therefore are not suitable for clinical translation. However, several development programs have been established for selectively targeting PN, with molecules that are non-metal based, orally active, and some do spare SO, and in preclinical screening these have proven effective in pain models [ 181 , 189 ]. However, these programs do not appear to have progressed forward and so are no longer considered pipeline. However, it has been established that targeting either sphingosine-1-phosphate (S1P)/S1P receptor 1 (S1PR1) or adenosine A3 receptor (A3AR) signaling modulates PN-driven downstream nociceptive mechanisms [ 190 , 191 ] This suggests an alternative therapeutic approach by indirectly modulating PN and its downstream mechanisms through selectively targeting specific G-protein coupled receptors (GPCR). Furthermore, targeting these GPCRs is analgesic in preclinical models of persistent spinal-related neuropathic pain [ [191] , [192] , [193] , [194] , [195] ]. This is significant as ‘functional’ S1PR1 antagonists are already FDA approved for the treatment of relapsing-remitting multiple sclerosis (MS) and psoriasis, with a good safety profile [ 196 , 197 ]. Further, highly selective A3AR agonists are also already in advanced preclinical development for neuropathic pain, and considered safe, based on ongoing clinical trials for non-pain indications [ 198 ]. This is relevant as early studies have suggested that selective modulation of S1PR1 and A3AR is analgesic in models of migraine [ [199] , [200] , [201] , [202] ]. Given the advanced nature of their development in other pain models, if further validation supports these two targets as potentially therapeutic in migraine, it provides a unique opportunity to advance these potentially early preclinical options to more advanced pipeline targets.

The potential of targeting ion channels, hormones, as well as neuropeptides, in migraine therapy has accelerated in the last 10–15 years. These, and others discussed in this special issue, are expressed all along the migraine pain pathway, from the dura mater to peripheral and central trigeminal projections, to brainstem and diencephalic structures thought to be involved in modulating migraine-trigeminal nociceptive mechanisms. They appear to be ideally positioned to modulate dural-nociceptive trigeminal mechanisms, as they relate to migraine. In many cases clinical experimental and subsequent clinical trial data provide further support to validate these pipeline targets as therapeutics in migraine. A further breakthrough is the potential for sex-specific therapeutics. The work with OX2R (male) and PRLR (female) and the dissection of sex-specific nociceptors [ 143 ] illustrates the potential to develop therapeutics that reflect pain processing networks specific to a genetic profile, in this case the genetics of sex. A potential first step in a more personalized medical approach. As noted in our introduction, there have been great successes in drug development over the last 50 years for migraine-specific therapeutics that have ultimately benefitted the migraine patient population. However, it is important to highlight that not every patient is benefiting from these current discoveries and therefore, more is always needed. The pipeline we outline demonstrates that these efforts do continue and a platform and pathway to discover and develop novel pipeline targets has become established. This is based on preclinical efforts to identify potential targets and localizing expression of these targets relative to migraine pathways. Successful development is then built around rigor; testing in a battery of in vitro and in vivo migraine models that represent both noxious and non-noxious migraine symptoms, and using different molecules for screening, ultimately providing different insights into the molecular mechanisms involved. This is followed by translating these findings into human experimental models, before the lead therapeutic molecules enter clinical trial. With this pipeline, one can be optimistic about the future of migraine therapeutic development. That said, more studies are needed to evaluate the efficacy and safety of these targets for migraine treatment (both preclinical and clinical studies), including testing in difficult to treat populations. The hope is that this will lead to effective and well-tolerated therapeutics, which hopefully capture patient populations that still struggle to gain benefit from existing therapeutics. For ‘the therapeutic pipeline of migraine’ , the future holds optimism, with many new opportunities for drug development, based upon the extensive discovery sciences that is being conducted.

Therapeutic development in migraine has seen many important milestones over the past 50 years, but perhaps there are two of most significance when it comes to the impact on patient care. The first was the publication of the clinical trial confirming subcutaneous sumatriptan for the abortive treatment of migraine [ 1 ], followed by its clinical approval in 1991/1992 (Europe/USA). It was the first in the triptan class of migraine specific abortives, primarily targeting serotonin 5-HT 1B/1D receptors. There are now seven triptans approved for therapeutic use in migraine [ 2 ]. The second major milestone is split between 2004 and 2017. In 2004 data was published of the first proof-of-concept study confirming the potential efficacy of targeting the calcitonin gene-related peptide (CGRP) receptor pathway for the treatment of acute attacks of migraine, using BIBN4096 (olcegepant), a non-peptide CGRP receptor antagonist [ 3 ]. In 2017, erenumab, a fully humanized monoclonal antibody (mAb) that binds the CGRP receptor was approved for migraine prevention. These two milestones heralded a new generation in migraine therapies; essentially the culmination of 30+ years of research, going back to 1988 with the first report of the possible role of CGRP in migraine [ 4 ]. There are now four mAbs approved for migraine prevention [ 5 ] and four gepants approved for the acute and/or preventive treatment of migraine. It is argued these milestones have changed the landscape of clinical care for persons with migraine. Despite these successes, one should not stand still with drug development. Indeed, not only is it healthy to have a robust ‘ therapeutic pipeline in migraine’ , it is still crucial. A huge population of patients, up to 50 %, are still unable to use current medicines, either because they are ineffective for them or for safety reasons [ 6 ]. Indeed, recent reports suggest that even newly approved CGRP-targeted treatments have response rates of only ∼50 % [ [7] , [8] , [9] ]. Furthermore, factors that might predict success with these new therapies include previous good response to triptans [ 10 , 11 ]. It suggests that newly approved treatments are not capturing patient populations that already have limited therapeutic options. It illustrates that major gaps still exist in our understanding of migraine pathogenesis and reinforces the need to dissect new disease mechanisms that may help to identify and validate novel therapeutic targets that will hopefully capture these under-served patients. Also, of course, for every success story, there are targets that were part of the therapeutic pipeline that never actually made it to the clinic, for one reason or another [ [12] , [13] , [14] , [15] , [16] ]. Here, we will review some of the current ‘therapeutic pipeline for migraine’ , targets and drugs that represent the future of drug development for the treatment of migraine, with the hope of being the next significant milestone success. We will cover targets that are in early preclinical development having been identified as potentially novel targets, through to those that are in advanced development and have already reached Phase I/II clinical trial. This review is not exhaustive as other targets are being discussed elsewhere in this special issue (including PAR2, MRGPRX2, and adenosine) but it should provide a snapshot of ‘what the future holds .’

ATP-sensitive potassium channels (K ATP ) are hetero-octameric complexes that consist of four inwardly rectifying potassium channels (Kir) and four regulatory sulfonylurea receptor subunits (SUR) [ 23 ]. They are present in cellular and mitochondrial membranes. They allow a small current of potassium to flow into the cell when closed (resting state), but once activated they promote K + efflux leading to hyperpolarization [ 24 , 25 ]. K ATP (Kir6.1 or Kir6.2 coupled to SUR2B) are expressed in vascular smooth muscle, where elevated ATP levels maintain the channels closed state, while low ATP levels promote K ATP channel opening. K ATP channel openers are linked to the vasodilatory mechanism of various migraine triggers [ 26 , 27 ]. K ATP are expressed within the trigeminovascular system, including the meningeal and cerebral vasculature, where their activation induces vasodilatation, and in the trigeminal ganglia (TG), trigeminal nucleus caudalis (TNC) and other brain areas related to migraine mechanisms where they decrease cellular excitability [ 25 , 28 , 29 ]. The overall evidence suggests that K ATP opening could be involved in migraine triggering mechanisms, and therefore their blockade may represent a potential therapeutic target [ 24 , 30 , 31 ]. In in vivo and organ bath models K ATP channels openers, levcromakalim and pinacidil (which preferentially target the SUR2B subunit), are potent vasodilators of rat dural, pial and cerebral arteries [ 32 ]. Using behavioral observations in functional preclinical models, repeated systemic administration of levcromakalim (1 mg/kg, ip) induces migraine-related periorbital hypersensitivity, similar to the persistent NTG model of migraine [ [33] , [34] , [35] ]. Similarly, levcromakalim (1.43 mg/kg via intracarotid infusion) activates a significant proportion of dural-responsive trigeminal ganglion neurons, with firing rates increased significantly at two and 3 h [ 36 ]. While no data related to markers of peripheral sensitization are reported in this study, such as neuronal responses to dural probing, these data point to a potential neuronal mechanism in mediating periorbital hypersensitivity. Levcromakalim- and NTG-induced periorbital hypersensitivity was also prevented in mice with a conditional loss-of-function mutation in the smooth muscle K ATP channel subunit Kir6.1 [ 34 ], further supporting the potential role of K ATP in these mechanisms. Nevertheless, it should be noted that central administration of levcromakalim has an analgesic effect, increasing paw withdrawal responses to the hot plate test [ 34 ]. While K ATP opening has been linked to CGRP-related migraine mechanisms, it does not increase CGRP release in migraine-related areas [ 28 , 37 ]. This provides support to the notion that K ATP opening represents a key downstream cellular response to CGRP receptor activation [ 24 , 28 ]. Together, these data support an extracranial, vascular contribution to the periorbital hypersensitivity mediated by levcromakalim. Experimental clinical studies also support the role of K ATP opening as a possible mechanism in migraine initiation. In healthy individuals that received intravenous infusion of levcromakalim (1 mg over 20 min), 86 % reported mild headache, including 10 % referring to ‘migraine-like attacks’ [ 38 ]. In these studies, dilatation of the middle meningeal artery, but not the middle cerebral artery, was observed during the 5 h observation period [ 38 ]. In another study, every participant with migraine that received levcromakalim (0.05 mg/min) developed a migraine [ 22 ]. Furthermore, in a migraine with aura population, 59 % of patients developed migraine with aura and overall 82 % developed either migraine with or without aura after levcromakalim (0.05 mg/min) [ 20 ]. There are also ongoing studies looking at the possible role of K ATP in migraine pain via activation of meningeal nociceptors and aura development through the induction of cortical spreading depression (CSD) [ 39 ]. Taken together, these data confirm the potency of levcromakalim in triggering migraine in susceptible people. Glibenclamide has been used as a K ATP blocker in many preclinical and clinical studies, to screen for potential efficacy of K ATP as a relevant therapeutic target. It should be noted that glibenclamide has high affinity for the SUR1 subunit, present in the pancreas, with lesser affinity to SUR2B. In experimental studies dosing has to be moderated to avoid SUR1 related side effects, which does limit its applicability as an experimental tool. That said, glibenclamide attenuated vasodilation mediated by levcromakalim and pinacidil [ 32 ]. In preclinical models, glibenclamide (1 mg/kg, ip) successfully reversed periorbital hypersensitivity induced by levcromakalim and NTG. It also inhibited ex-vivo CGRP release in TG and dura mater in a spontaneous trigeminal allodynia model [ 40 ], despite K ATP signaling mechanisms assumed to be downstream of CGRP receptor activation. Finally, glibenclamide attenuated the vasodilatory effects of CGRP and PACAP, but not NTG [ 41 , 42 ]. Additionally, PNU-37883A (K ATP blocker via the Kir6.1 subunit) also attenuated the vasodilatory effect induced by levcromakalim [ 43 ]. In randomized, double-blind, placebo-controlled studies, glibenclamide was unable to prevent the development of levcromakalim-induced headache in healthy subjects [ 44 ]. It did however delay headache onset. In another study, glibenclamide did not reverse vasodilatation induced by levcromakalim [ 45 ]. Similarly, oral glibenclamide was also unable to decrease the incidence of headache induced by CGRP infusion in healthy subjects when compared to placebo [ 46 ]. Finally, oral glibenclamide (10 mg) did not reduce headache after PACAP infusion [ 47 ]. The lack of specificity of glibenclamide on K ATP channels, preferentially targeting the SUR1 subunit, whereas Kir6.1/SUR2B seems to be the subtype involved in triggering migraine attacks with levcromakalim [ 34 , 43 ], may potentially explain the limited effects of glibenclamide in these studies. PNU-37883A, which does selectively inhibit the Kir6.1/SUR2B subunit, did not progress in clinical studies due to possible cardiac inhibitory effects [ 31 , 48 ]. Another selective blocker is rosiglitazone, used in the treatment of diabetes mellitus type 2 and has been related to cardiovascular side effects such as heart failure [ 49 ]. It seems clear that the next step in the pipeline for this target is the development of molecules that selectively target SUR2B, which appears to be specifically relevant in migraine, while sparing other subunits that have cardiovascular or pancreatic indications.

MRR reports consultancy fees/honoraria from Kallyope, Pain Medicine (co-section Editor, Headache & Facial Pain), and Headache Cooperative of the Pacific; and has received research support from US Department of Defense (DoD), and the National Institutes of Health (NIH/NINDS). Dr. Romero is Member at large Board of Directors of the AHS. All financial interests are unrelated to this work. SA reports grant funding from the US Department of Defense and National Institutes of Health, and consultancy fees/honoraria from Septerna, AbbVie, Kallyope, Curie.Bio, Guidepoint, Patent Legal work, American Academy of Orofacial Pain, and Headache Cooperative of the Pacific. Dr. Akerman is also Associate Editor for the Headache journal. All financial interests are unrelated to this work. All authors declared no known competing financial interests or personal relationships with respect to the authorship and/or publication of this article that could have appeared to influence the work reported in the paper.