Emerging Treatment Targets for Migraine and Other Headaches

Abstract

Migraine is a complex disorder that is characterized by an assortment of neurological and systemic effects. While headache is the most prominent feature of migraine, a host of symptoms affecting many physiological functions are also observed before, during, and after an attack. Furthermore, migraineurs are heterogeneous and have a wide range of responses to migraine therapies. The recent approval of calcitonin gene related peptide (CGRP) based therapies has opened up the treatment of migraine and generated a renewed interest in migraine research and discovery. Ongoing advances in migraine research have identified a number of other promising therapeutic targets for this disorder. In this review we highlight emergent treatments within the following biological systems: pituitary adenylate cyclase activating protein (PACAP), two non-mu opioid receptors that have low abuse liability – the delta and kappa opioid receptors, orexin, and nitric oxide-based therapies. Multiple mechanisms have been identified in the induction and maintenance of migraine symptoms; and this divergent set of targets have highly distinct biological effects. Increasing the mechanistic diversity of the migraine tool box will lead to more treatment options and better patient care.

Introduction

Headache has a high global prevalence and has been identified as a leading cause of disability worldwide^{1, 2}. However, until recently the only pharmacotherapy specifically targeted for migraine was the triptans. The recent approval of anti-calcitonin gene related peptide (CGRP) therapies appear to be a breakthrough in migraine prevention, and they have created a great deal of excitement in the field. The development of CGRP exemplifies how translational research can result in targeted strategies specific for the treatment of migraine. The success of CGRP based therapies has also spurred the investigation for new therapies to encourage the expansion of the migraine toolbox. Increasing the variety of migraine-selective treatments is especially beneficial as not all patients respond to CRGP-based treatments, and patients dosed on CGRP antibodies might still need rescue therapies. In addition, the consequences of long acting CGRP based antibodies on pregnancy are still unclear³ thus further supporting the need for mechanistically diverse treatment strategies.

Migraine is a brain disorder that produces a wide array of neurological and systemic symptoms. The pathophysiology of migraine is correspondingly complex and includes changes in the vasculature, central and peripheral pain processing, and inflammation. Migraine can be treated by a number of different drug classes which supports the idea that divergent mechanisms can regulate the migraine brain state⁴. Anti-convulsants, beta blockers, tricyclic antidepressants, and botulinum toxin have all been used as migraine preventives with varying levels of efficacy^{5, 6}. Prior to the approval of anti-CGRP therapies, most of these therapies were poorly tolerated or ineffective, and over 50% of chronic migraine patients were dissatisfied with their treatment⁷. Migraineurs are a heterogeneous population, and have shown a wide range of response rates to all of these therapies, including CGRP targeting drugs^{6, 8}. Ongoing advances in migraine have provided novel insight into the diverse mechanisms regulating this disorder. In this review we will highlight the following emerging therapies: pituitary adenylate cyclase activating peptide (PACAP), two non-μ opioid receptors - δ and κ receptors, orexin, and nitric oxide (NO) based therapies. All of these targets have compelling evidence supporting their role in migraine; and their physiological diversity highlights the range of mechanisms that can regulate headache disorders. The authors used PubMed as their primary source of information, and only peer review articles were considered. For data on clinical trials, ClinicalTrials.gov and PubMed were utilized.

Pituitary Adenylate Cyclase Activating Peptide

Pituitary adenylate cyclase activating peptide (PACAP) has garnered attention in recent years as a key peptide in migraine pathophysiology^9–11. PACAP is a vasoactive intestinal peptide (VIP) and as such is part of the glucagon-growth hormone releasing factor-secretin superfamily¹². PACAP is found in two forms, the 38 amino acid peptide, PACAP38; and a truncated PACAP27¹³. PACAP38 is the predominant form and accounts for 90% of the PACAP in mammalian tissue¹⁴. PACAP38 is commonly found in both the peripheral and central nervous system. In humans, PACAP38 is expressed in the otic and sphenopalatine ganglia¹⁵. These neurons in turn project to the perivascular parasympathetic nerves to the cranial circulation¹⁴. PACAP38 is also expressed in a number of other regions that regulate headache, including the pituitary, trigeminal ganglia, trigeminal nucleus caudalis, dorsal horn of the spinal cord, brain stem, hypothalamus, and cortex¹⁶. Given its wide spread distribution, PACAP38 has been implicated in a variety of physiological functions including vasodilation, neurotransmission, neuroprotection, and neuromodulation¹². As part of the VIP family PACAP38 can activate three different Gs G protein coupled receptors (GPCRs): VPAC1, VPAC2, and PAC1; and activation of these receptors results in stimulation of adenylyl cyclase and increased levels of cyclic adenosine monophosphate (cAMP). PACAP and its related peptide, VIP, can bind with the same affinity to both VPAC1 and 2, but PACAP alone has up to a 1000X higher affinity for the PAC1 receptor^{17, 18}.

The first clear link between PACAP and headache was made through human provocation studies. In an initial study, infusion of PACAP38 into healthy volunteers resulted in 83% of participants developing headache¹⁹, which is in line with the expression of the PACAPergic system in head pain processing regions and its vasodilatory effects. Subsequent studies compared the effects of PACAP on migraine without aura patients versus healthy controls. Although PACAP38 evoked headache in most participants, it specifically evoked a delayed migraine attack in 58% of migraineurs²⁰. Furthermore, PACAP38 infusion resulted in mild to moderate dilation of both the middle cerebral artery (MCA) and the superficial temporal artery (STA)²⁰. In a follow up study, the same group used high resolution magnetic resonance angiography to examine the middle meningeal artery (MMA) and found significant and sustained dilation after PACAP38 that coincided with headache in healthy volunteers²¹. Interestingly, when patients were administered sumatriptan, PACAP38-induced dilation of the MMA was reduced as was headache severity²¹. This work demonstrates that PACAP38 is sufficient to produce vasodilation of meningeal arteries and evoke headache.

Although there is considerable physiological overlap between PACAP and VIP, including their affinity for VPAC receptors, PACAP appears to be more closely tied to migraine pathophysiology. In human provocation studies comparing the two peptides, the infusion of PACAP38 produced migraine in 73% of migraine without aura subjects as compared to 18% by VIP²². Interestingly both PACAP38 and VIP produced dilation of the MMA²², and this study highlights that vasodilation alone is not sufficient to produce migraine. The effects of PACAP38 and VIP on brain connectivity was also examined using resting state functional magnetic resonance imaging. Migraine without aura participants showed altered functional connectivity of the salience, sensorimotor, and default networks following PACAP38 infusion, effects not observed with VIP²³. Interestingly, infusion of PACAP38 into migraineurs does not result in increased blood levels of calcitonin gene related peptide (CGRP)²⁴, suggesting that PACAP38 can produce migraine independent of this endogenous pro-migraine neuropeptide. Together, these human provocation studies support the role of PACAP38 in migraine development.

There is also evidence that PACAP38 levels are also altered in spontaneous headache conditions. PACAP38 plasma levels were found to be upregulated in migraineurs during the ictal period²⁵. A separate group found a similar increase during migraine, and showed that treatment with sumatriptan correspondingly decreased PACAP38 levels²⁶. Furthermore, elevated PACAP38 levels were also observed during the attack phase in episodic cluster headache patients²⁷. Interestingly, PACAP38 levels were also found to be decreased during the interictal period relative to a healthy control group^{25, 27}, suggestive of a broader dysregulation of the PACAPergic system in headache sufferers.

Preclinical studies have further explored the mechanism by which PACAP38 may be modulating headache. Infusion of PACAP38 in mice produces photophobia, meningeal dilation, and increased c-fos expression (a marker of cellular activation) in trigeminal nucleus caudalis²⁸. More broadly, PACAP also appears to regulate the photophobia and meningeal effects induced by the human migraine trigger, nitroglycerin, as these endpoints are reduced in PACAP knockout mice²⁸. In rats, PACAP38 but not VIP resulted in delayed activation and sensitization of central trigeminovascular neurons¹⁰. In this study PACAP38 caused an increase in spontaneous firing and hypersensitivity to both central and peripheral somatosensory stimulation, which occurred at a distinctly different time point from the meningeal dilation¹⁰. Further, dural meningeal vasodilation induced by PACAP38 was mediated through VPAC2 receptors, while the delayed activation and sensitization of trigeminocervcial neurons was mediated by PAC1¹⁰; another example of nociceptive and vascular effects being regulated separately. PACAP38 was also shown to produce degranulation of peritoneal and dural mast cells which in turn could be responsible for PACAP-induced vasodilation of MMA^{29, 30} , and this effect may be mediated by a non-canonical PACAP signaling pathway³¹.

In addition to the role of PACAP in migraine, PACAP38 and PAC1 are also heavily implicated in stress and emotional processing. In rodents, administration of PACAP38 increases behaviors associated with anxiety and depression³²; and correspondingly, PACAP knockout results in decreased anxiety and depression-related phenotypes^{33, 34}. In addition, studies in humans indicate that polymorphism of the PAC1 gene confers increased risk of post-traumatic stress disorder (PTSD)³⁵, and increased PACAP38 levels were also associated with PTSD³⁶. Considering the high co-morbidity between headache and emotional disorders makes the PACAPergic system appears to be particularly attractive as a novel therapeutic target.

Currently there is no clinically available PACAP-specific treatment. However, a Phase IIa randomized, double blind, placebo-controlled trial is underway to examine the efficacy of AMG-301, a PAC1 receptor targeting antibody, on migraine prevention³⁷ (Table 1). In preclinical studies, this antibody inhibited nociceptive activity in the trigeminocervical complex to the same extent as sumatriptan³⁸. Although it has not yet entered clinical trial, a monoclonal antibody targeting PACAP38, ALD1910, is also in development for the treatment of migraine patients ³⁹.

Table 1:

Clinical Trials and Pipeline for Emerging Therapeutic Targets for Migraine

Target	Drug Name	Mechanism	Condition	Status	Phase	Primary Endpoint Met?	Manufacturer	ClinicalTrials.gov identifier
PACAP	AMG-301	Monoclonal Antibody PAC1 Receptor	Migraine	Completed in 2019	Phase II	Results not available	Amgen
	ALD1910	Monoclonal Antibody PACAP38	Migraine	Ongoing	Pre-clinical	N/A	Alder Biopharmaceutics	Company website
Delta Opioid Receptor	ADL5859	δ receptor agonist	Osteoarthritis of the knee	Completed in 2010	Phase II	No	Adolor Corporation; Pfizer
	ADL 5747	δ receptor agonist	Osteoarthritis of the knee and postherpetic neuralgia	Completed in 2010	Phase II	No	Adolor Corporation; Pfizer
	AZD7268	δ receptor agonist	Major Depressive Disorder	Completed in 2010	Phase II	No	AstraZeneca
	AZD2327	δ receptor agonist	Major Depressive Disorder	Completed in 2011	Phase II	No	AstraZeneca
	TRV 250	δ receptor agonist	Migraine	Completed in 2018	Phase I	Yes	Trevena	Company website
Kappa Opioid Receptor	CR845	peripherally restricted κ agonist	Hysterectomy Bunionectomy Osteoarthritis	Completed in 2012 2013 2017	Phase II	Yes Yes Not available	Cara Therapeutics
	CERC501 (LY-2456302)	κ-opioid receptor antagonist	Anxiety Disorder	Completed 2017	Phase II	Yes	Eli Lilly
Orexin	Filorexant	Dual orexin receptor antagonist	Migraine preventative	Completed in 2013	Phase II	No	Merck Sharp & Dohme Corp.
	Suvorexant (DORA-12)	Dual orexin receptor antagonist	Post-Traumatic Stress Disorder	Ongoing	Phase IV	Results not available	Merck Sharp & Dohme Corp
Nitric Oxide Synthase (NOS)	GW274150	iNOS inhibitor	Migraine	Completed in 2010	Phase II	No	GlaxoSmithKline
	NXN-188	nNOS inhibitor +5HT1B/1D agonist	Migraine with aura	Completed in 2010	Phase II	Results not available No	NeurAxon
	NXN-188	nNOS inhibitor +5HT1B/1D agonist	Migraine without aura	Completed in 2010	Phase II	Yes	NeurAxon

Non-Mu Opioid Receptors - Delta Opioid Receptor

Opioid analgesics in current use (e.g. morphine, hydrocodone, oxycodone, meperidine) act primarily at the μ opioid receptor⁴⁰ and are still commonly prescribed for headache⁴¹. However, while opioids may provide acute relief, regular use of μ receptor agonists can result in increased headache severity, and progression of headache from an episodic to a chronic state⁴². This form of medication overuse headache⁴³ is not only difficult to treat, but can also result in opioid dependence and abuse which has resulted in a significant public health crisis^44–46. Evidence from preclinical studies indicate that other members of the opioid receptor family may provide effective alternatives to μ-based approaches. The opioid receptor family is comprised of μ, δ, κ opioid receptors, and the nociceptin/ORL-1 receptor. All four receptors share homology but are encoded by different genes and produce highly distinct physiological effects. Opioid receptors are activated by endogenous opioid peptides. Although these peptides can bind across opioid receptors there is some selectivity; and enkephalins have the highest affinity for δ receptors, endorphins for μ receptors, dynorphins for κ receptors, and nociceptin for nociception/ORL1 receptor⁴⁷. The δ and κ receptors have been identified as novel therapeutic targets for migraine^{48, 49} and are discussed below.

δ receptors are expressed in a number of regions involved in headache, including trigeminal and dorsal root ganglia, trigeminal nucleus caudalis, and the cortex^50–52. δ receptors are found on a small subset of dorsal root ganglia that express the pro-migraine peptide CGRP⁵³. Further, its expression in limbic regions such as the striatum, hippocampus, and amygdala support its role in emotional regulation^{50, 54}. Relative to μ receptors, δ receptors have low abuse liability. In rodent and non-human primate models, δ agonists do not facilitate rewarding behaviors, nor do they cause physical dependence^55–58. In addition, δ agonists produce less respiratory depression and have fewer effects of gastrointestinal transit^{59, 60} relative to μ agonists. In animal model, δ receptor agonists can produce pro-convulsant effects. However, not all δ agonists show this adverse effect, and it may be possible to separate out this consequence from pain-relieving effects through drug development⁶¹.

Although δ agonists are not highly effective in acute pain, they have shown efficacy in a number of chronic pain states, including models of inflammatory and neuropathic pain⁶². δ agonists appear to be particularly effective in models of headache⁴⁸. In a behavioral model of migraine-associated pain, nitroglycerin, a known human migraine trigger⁶³, caused severe mechanical hyperalgesia in mice that was prevented by three different δ agonists⁴⁸. Another lab also found that heat hyperalgesia induced by nitroglycerin was also blocked by the δ agonist, SNC80⁶⁴. Further, more recent evidence suggests that δ agonists could be effective for a number of different types of headache. SNC80 effectively inhibited cephalic allodynia in mouse models of chronic migraine, and medication overuse headache associated with triptans⁶⁵. In a model of post-traumatic headache, acute SNC80 blocked established cephalic allodynia⁶⁵, and long term treatment with this δ agonist also prevented the chronification of post-traumatic headache⁶⁶. In a model of opioid-induced hyperalgesia, chronic morphine treatment produced a severe cephalic and peripheral allodynia that was inhibited by treatment with a δ agonist⁶⁵. This latter finding particularly supports the distinction between the μ and δ receptors, as chronic μ receptor stimulation by morphine resulted in a correlate of mediation overuse headache which could still be blocked by δ receptor activation. Preclinical studies indicate that δ opioid receptors are functionally upregulated during chronic peripheral pain states^{62, 67–69}, and may serve as an endogenous protective mechanism to counteract increased pain signaling. Ongoing studies are underway to determine if this upregulation is also observed in headache models. Furthermore, moderate to high expression of δ receptors has been observed in trigeminal ganglia, trigeminal nucleus caudalis⁵² and the dura⁷⁰ suggesting that it may be particularly important for the regulation of headache specifically.

The δ receptor was also found to be effective in preclinical studies examining other symptoms of migraine. Cortical spreading depression (CSD) is the slowly propagated wave of depolarization followed by inhibition of brain activity thought to be the underlying physiologic correlate of migraine aura⁷¹. Migraine preventives have been shown to decrease CSD^{72, 73}, and it is therefore used as a screening tool for novel migraine therapies. The δ agonist, SNC80, significantly decreased the number of CSD events evoked by KCl⁷⁴ suggesting that δ receptors may also have potential as a preventive treatment. Chronic δ activation has also been examined for the development of medication overuse headache; and compared to sumatriptan or morphine, SNC80 produced a limited form of overuse allodynia⁶⁵, further supporting its development as a headache therapy.

The δ receptor has been shown to play an important role in emotional regulation⁵⁴, and δ agonists have been developed for the treatment of anxiety and depression. Knockout of the δ receptor results in anxiogenic and depressive-like behaviors⁷⁵, and δ agonists have anxiolytic and anti-depressant effects in animal models^{54, 76}. This positive effect of emotional modulation may be particularly beneficial considering the high co-morbidity between headache and emotional disorders⁷⁷. In a study of migraine-associated negative affect, nitroglycerin treatment was paired with a specific chamber which resulted in conditioned place aversion, and this aversion was prevented by SNC80⁷⁴. This result suggests that δ agonists may also improve the negative emotional state associated with migraine.

δ agonists have been developed and tested for the treatment of pain, and anxiety and depression⁷⁸ (Table 1). Two δ agonists, ADL5859 and ADL5747 were investigated for safety and efficacy in osteoarthritis of the knee, and ADL5747 was also tested in patients with postherpetic neuralgia. The results have not been published in a peer-reviewed journal, however, the posted results from ClinicalTrials.gov show that neither study met its primary endpoint, and the latter study was terminated early due to lack of efficacy. Although disappointing, these pain conditions are mechanistically distinct from migraine, and these trials may not reflect efficacy of δ agonists in headache. Two other δ agonists, AZD7268⁷⁹ and ADZ2327⁸⁰, have been investigated in phase I and phase IIa studies for safety and potential efficacy in major depressive disorder and anxiety. Again, neither study reached primary endpoints but secondary analysis indicated benefit in anxiety. It is important to note that in all of the clinical trials for δ agonists, all 4 compounds were generally well tolerated with limited to no adverse events . A more recent phase 1 clinical trial was performed to test TRV250 for acute migraine. This δ agonist was specifically developed to avoid seizure liability, while maintaining pain-relieving effects. Data from this study has only been disclosed in a press release from the company, and indicates that in this healthy volunteer study TRV250 showed safety, tolerability and a pharmacokinetic profile that would support subsequent trials for efficacy. The δ receptor appears to be a promising therapeutic target for migraine, and future studies are ongoing to determine its clinical utility.

Non-Mu Opioid Receptors – Kappa Opioid Receptor

The κ opioid receptor has also been implicated in a number of physiological processes, including the regulation of pain, reward, and stress. The κ receptor is distributed throughout the brain and high expression is observed in regions related to mood, motivation, and cognitive function such as the nucleus accumbens (NAc), cerebral cortex, hippocampus, and hypothalamus; as well as in pain processing regions such as the periaqueductal grey, spinal cord, and dorsal root ganglia^{51, 81–83}. Activation of the κ receptor is pain relieving and agonists have been investigated for the treatment of pain⁸⁴. However, κ agonists are also found to produce psychotomimesis⁸⁵ and dysphoria⁸⁴ in humans, which has limited their development for clinical use. Peripherally restricted κ agonists have been developed for pain management. ADL10-0101, was found to be effective for pancreatitis in a small study, but its effects in females were sensitive to menstrual cycle, and it also produced headache⁸⁶. Several clinical trials have been performed to investigate the efficacy of CR845 in acute (hysterectomy and bunionectomy) and chronic (osteoarthritis) pain conditions with generally positive results (Table 1). However, to the best of our knowledge there are no published studies showing the effectiveness of κ agonists in headache models.

Inhibition of the κ receptor has emerged as a potential therapeutic strategy for headache disorders. Upregulation of the endogenous κ opioid peptide, dynorphin, has been identified as a marker of stress⁸⁷; and for this reason κ antagonists are being investigated for the treatment of anxiety, depression, and drug abuse and relapse⁸⁸. Stress has frequently been cited as the most common migraine trigger⁸⁹, but despite its prevalence, the pathophysiology of how stress actually triggers migraine is unclear. Xie et al. used a preclinical model of stress-induced headache to examine the effect of κ antagonists on cephalic pain⁴⁹. Rats were treated with chronic sumatriptan to model medication overuse headache, which in turn made them hypersensitive to a bright light stress cue. Stress produced an increase in plasma CGRP and cephalic and peripheral allodynia, which was blocked by systemic injection of two different κ antagonists⁴⁹. Further, this hypersensitivity appears to be mediated through an upregulation of dynorphin in the amygdala, and intra-amygdala injection of κ antagonists were also anti-allodynic⁴⁹. The novel kappa opioid receptor antagonist, CERC-501 (formerly known as LY2456302) is in clinical trials for mood and anxiety disorders and was shown to be safe and well-toleranated⁹⁰ (Table 1). Given these early promising clinical results κ receptor antagonists may serve as promising therapies for headache.

Orexin

Orexins, also known as hypocretins, have been widely implicated in homeostatic regulation; and dysfunction of the orexinergic system in migraine has recently been explored^91–93. The orexin family is comprised of two neuropeptides, orexin A (OXA) and orexin B (OXB), which are derived from the precursor peptide, prepro-orexin⁹⁴. Orexin A is 33 amino acids in length, while orexin B is 28 residues, and mature versions of these peptides are entirely (OXA) or highly (OXB) conserved in mammals⁹⁵. Although these peptides are synthesized exclusively in the dorsolateral hypothalamus, orexinergic neurons project throughout the CNS, including to other hypothalamic nuclei, thalamus, cortex, trigeminal nucleus caudalis, and spinal cord⁹¹. Both orexin peptides bind to orexin receptor 1 (ORX1) and orexin receptor 2 (ORX2), both of which are Gq GPCRs, and as such activation results in increased intracellular Ca²⁺⁹⁴. OXA has equal affinity for both orexin receptors, while OXB has a ten-fold increased affinity for OXR2⁹⁴. The orexin receptors are broadly expressed in the brain, and usually correspond with orexin peptide expression^{91, 93}.

The orexinergic system is involved in the dynamic regulation of a number of physiological processes; including feeding, sleep/wake cycle, stress, and motivation⁹⁶. Many of these processes are disrupted during migraine, particularly in the premonitory phase. For example, changes in appetite are often observed during headache, and disruption to normal feeding patterns are considered a potential migraine trigger. Orexins may regulate this aspect of headache disorders, as they play an important role in appetite and food consumption⁹⁶. Orexin peptide and receptor levels are upregulated following food deprivation or fasting, and activation of the arcuate nucleus by orexins promotes feeding behavior^{97, 98}. Correspondingly, orexin levels are downregulated when blood glucose levels are high^{99, 100}. Changes in orexins and other hypothalamic peptides that regulate energy and metabolism could contribute to the perturbation in appetite and feeding associated with headache disorders.

Orexins levels are tightly regulated with circadian rhythm, and is thought to facilitate the transition from sleep to wakefulness. Peak orexin levels are observed during wakefulness, and orexinergic neurons are most active during wake with frequent firing just before waking^{101, 102}. This role of orexin is most starkly demonstrated in narcolepsy. Knockout of the orexin peptides or mutation of the ORX2 receptor results in murine¹⁰³ and canine¹⁰⁴ narcolepsy, and human narcoleptics show decreased circulating levels of orexin¹⁰⁵. Consistent with the role of orexin in migraine, narcoleptic patients have a significantly higher prevalence for migraine, and 64% of female and 45% of males suffer from migraine compared to 30% and 8%, respectively, in the general population^{93, 106}. More generally, sleep disturbances are known to increase susceptibility to migraine; and orexin may act at the interface between changes in circadian rhythm and headache disorders. A recent study found that following traumatic brain injury, mice showed decreased orexin positive neurons in the hypothalamus; and this may be involved in altered sleep patterns associated in TBI¹⁰⁷ and subsequent PTH. Disturbances in the orexinergic system could serve as a link between headache disorders and sleep dysregulation.

The orexinergic system has also been implicated in pain processing. Knockout of orexin results in increased pain sensitivity in inflammatory pain models, and decreased stress induced hyperalgesia¹⁰⁸. Orexins have also been shown to be modulatory in models of cephalic pain. Microinjection of orexin A or B into the posterior hypothalamus decreases or increases cephalic pain sensitivity, respectively¹⁰⁹. Similarly, orexin A but not B can inhibit neurogenic dural vasodilation¹¹⁰ and trigeminal neuron firing¹¹¹, and it appears that orexin A through activation of ORX1 signals distinctly from orexin B.

A role for the orexinergic system in headache pathophysiology has been supported by studies measuring orexin levels in various headache disorders. Significantly decreased levels of CSF orexin was observed in both episodic and chronic cluster headache patients relative to controls¹¹². In contrast, in chronic migraine and medication overuse headache patients, orexin A levels were found to be increased¹¹³. Although seemingly contradictory, these studies support the notion that the orexin system is dysregulated in headache.

Dual orexin receptor antagonists (DORAs) have been developed for clinical use, and suvorexant (DORA-12) has been approved for the treatment of insomnia. In a preclinical model of temporomandibular joint inflammation, DORA-12 acutely attenuated facial allodynia¹¹⁴. Furthermore, this antagonist also blocked electrophysiological signaling in the trigeminocervical complex in response to dural stimulation¹¹⁵. Interestingly, DORA-12 also increased the threshold for CSD, a correlate of migraine aura¹¹⁵. These studies suggest that antagonism of the orexin system could be an effective migraine therapy. There has been one small clinical trial to test the effect of filorexant, another dual orexin receptor antagonist, as a migraine preventive (Table 1). In this randomized, double-blind, placebo controlled trial patients were treated with 10 mg nightly for three months. There was no statistically significant difference between filorexant treated and placebo controls¹¹⁶, which may be explained by the dosing regimen or half-life of the drug⁹³. Although disappointing, there is still hope for targeting the orexinergic system in migraine. Preclinical studies indicate that ORX1 and ORX2 can have divergent behavioral effects^109–111; which is supported by the finding that certain CNS regions only express one receptor (eg locus coeruleus and ORX1, and the rostroventromedial medulla and ORX2). Further studies are required to clearly define how changes in the orexinergic system correspond to headache susceptibility and symptomatology.

Nitric Oxide Pathway

Nitric oxide (NO) is an endogenous gaseous signaling molecule that is important for several biological processes including vascular tone, neurotransmission and immune defense. NO is ubiquitously expressed in almost every mammalian cell type¹¹⁷. NO production and signaling has been heavily implicated in migraine pathophysiology^{118, 119}. The known human migraine trigger, nitroglycerin, is a NO donor and has been used extensively in human migraine provocation⁶³ studies and as an animal model of migraine^{120, 121}. Nitroglycerin is a potent vasodilator, and it produces a rapid headache in both migraineurs and healthy controls. However, migraine patients go on to develop a delayed migraine attack several hours after this vasodilatory effect¹²² again suggesting that vascular alterations and induction of migraine are not always coupled.

Endogenous NO is produced by the oxidation of L-arginine into NO and L-citrulline¹²³. This reaction is catalyzed by three isoforms of nitric oxide synthase (NOS). Each of the NOS family members corresponds to the tissue type that they were initially discovered in and where they are largely expressed¹²³. Neuronal NOS (nNOS) is expressed in neurons and can be found in both the central and peripheral nervous system¹¹⁷. Endothelial NOS (eNOS) was originally discovered in vascular endothelial cells, but is also expressed in platelets, cardiomyocytes, and the brain¹¹⁷. NO produced by eNOS appears to largely regulate vascular tone¹²⁴. Both nNOS and eNOS are constitutively activated and rely on increase of intracellular calcium and subsequent binding to calmodulin¹¹⁷. The third NOS isoform, inducible NOS (iNOS), is expressed in macrophages, glia, and neurons. iNOS is unique because unlike the other two isoforms it is not constitutively active, but is induced by infection and pro-inflammatory cytokines and serves as part of the host immune defense system¹²³. iNOS can rapidly produce up to 1,000X more NO than nNOS or eNOS¹¹⁷.

NO is highly membrane permeable and can activate signaling cascades both within the cell it was produced as well as in neighboring cells. The only known receptor for NO in the body is soluble guanylyl cyclase (sGC)¹²³, which upon activation converts guanosine-5’-triphosphate (GTP) to cyclic guanosine monophosphate (cGMP). Signaling through this pathway includes activation of cGMP-specific ion channels, alterations in Ca²⁺ homeostasis, and activation of the transcription factor CREB^125,126. Several lines of evidence indicate that the sGC pathway has an important function in migraine etiology and maintenance. cGMP is broken down to GMP by phosphodiesterase 5 (PDE5). The PDE5 inhibitor, sildenafil, can produce a headache in healthy control patients along with a migraine like attack in migraineurs¹²¹. In a preclinical study direct activation of the sGC pathway by the novel sGC stimulator, VL-102, produced cephalic allodynia which was blocked by known migraine treatments¹²⁷. In addition, a sGC inhibitor, ODQ, completely abolished migraine-associated allodynia induced by nitroglycerin¹²⁷. Interestingly, in a chronic migraine model, ODQ was also able to inhibit established allodynia in the absence of an exogenous NO donor, which suggests that upregulation of the sGC pathway may be a key mediator for the maintenance of the chronic migraine state. This study strengthens the connection between the NO-sGC pathway and headache, and open the possibility for inhibitors or negative modulators of sGC as novel migraine therapeutics.

Although the sGC pathway is seen as the primary NO signaling mechanism, NO can also produce significant oxidative stress through the formation of the reactive nitrogen species peroxynitrite¹¹⁷. The role of peroxynitrite in migraine is still unlcear¹²⁸, but migraine is commonly observed in patients with mitochondrial diseases¹²⁹, and antioxidants such as riboflavin and coenzyme Q10 can be effective migraine preventives^{130, 131}. Future studies will determine if peroxynitrite scavengers could be promising migraine therapies¹³².

The blockade of endogenous NO production by NOS inhibitors (NOSi) has been extensively investigated for the treatment of migraine, and has been summarized in a recent review¹¹⁸. In a clinical study, the non-selective NOS inhibitor, L-N^G -methylarginine hydrochloride (546C88) provided significant headache relief during spontaneous migraine attacks relative to placebo¹³³. A different non-selective NOSi, N^G L-arginine hydrochloride (L-NMMA) was more efficacious than placebo in chronic tension-type headaches¹³⁴. These studies provide proof of concept for the targeting of NO for the treatment of migraine. However, non-selective NOSi produce significant cardiovascular effects and increased blood pressure which limits their therapeutic potential¹¹⁸. Similarly, eNOS is found predominantly in the vascular endothelium and is considered to be fundamental for healthy cardiovascular function¹³⁵ and therefore is also not a highly desirable target for migraine.

iNOS has been explored for the treatment of headache disorders. In a rodent model of migraine, nitroglycerin results in an upregulation of iNOS mRNA and protein in dural meningeal macrophages ^{136, 137}. Similarly, in humans blood levels of iNOS are increased during a migraine attack, and levels lower after the attack had subsided¹³⁸. Despite the evidence, targeting iNOS for migraine has proven to be less effective than hoped. Two clinical trials investigated the effect of the iNOS inhibitor, GW274150, to prevent or treat migraine attacks, and neither reached primary endpoints (Table 1)^{139, 140}. These negative results suggest that targeting iNOS might not be sufficient to inhibit migraine, and this treatment strategy has been largely abandoned¹¹⁸.

nNOS is expressed throughout peripheral and central regions that regulate headache pain, and therefore could serve as an ideal candidate for drug development¹⁴¹. Similar to iNOS, nitroglycerin administration in rodents evoked increased nNOS expression in two key pain processing regions, the trigeminal ganglia and the trigeminal nucleus caudaulis^{142, 143}. A nNOS inhibitor decreased meningeal blood vessel diameter and inhibited neurogenic dural meningeal blood vessel dilation¹⁴⁴. In a separate model the selective nNOS inhibitor NXN-413 inhibited the elevation in CGRP by KCl¹⁴⁵. Similarly, in a dural inflammation model NXN-413 inhibited increased vasodilation¹⁴⁵. In another preclinical model of MOH a different nNOS inhibitor, NXN-323, was able to reverse established periorbital and hind paw allodynia¹⁴⁶. Clinically, NXN-188, a mixed nNOSi-triptan, has been investigated in two clinical trial studies for the treatment of migraine with and without aura. Preliminary studies suggest that this molecule may be effective in migraine treatment^{147, 148}, but there has been limited movement on this strategy since 2010. nNOS alone, or in combination with a 5-HT1B/1D agonist, may still prove to be a novel and effective strategy for the treatment of migraine.

Conclusions

It is an exciting time in migraine research and patient care. The success of CGRP based therapies reveals the power of a targeted treatment strategy for migraine. It has also generated increased interest in basic and translational headache research. A few of the discussed targets were tested in clinical trials, but did not reach primary endpoints. An emphasis on understanding the dominant signaling pathways underlying different types of headache could result in more directed clinical trials with a greater chance of success. Considering the prevalence and heterogeneity of migraine the development of mechanistically diverse targets will lead to greater treatment options. There are a number of promising targets that have been identified as migraine regulators and future work will hopefully lead to the next breakthrough drug.