Pharmacological treatment of migraine: Drug classes, mechanisms of action, clinical trials and new treatments

Migraine is the sixth most prevalent disease globally, a major cause of disability, and it imposes an enormous personal and socio‐economic burden. Migraine treatment is often limited by insufficient therapy response, leading to the need for individually adjusted treatment. In this review, we analyse historical and current pharmaceutical development approaches in acute and chronic migraine based on a comprehensive and systematic analysis of Food and Drug Administration (FDA)‐approved drugs and those under investigation. The development of migraine therapeutics has significantly intensified during the last 3 years, as shown by our analysis of the trends of drug development between 1970 and 2020. The spectrum of drug targets has expanded considerably, which has been accompanied by an increase in the number of specialised clinical trials. This review highlights the mechanistic implications of FDA‐approved and currently investigated drugs and discusses current and future therapeutic options based on identified drug classes of interest.

Migraine is classified into two major clinical types: migraine without aura (MO) and migraine with aura (MA). Each type is partly different in the composition of clinical features and molecular biological susceptibility factors (Pisanu et al., 2017(Pisanu et al., , 2020. Visual, sensory or other CNS symptoms generally precede the headache followed by associated migraine symptoms (Headache Classification Committee of the International Headache Society, 2018). In addition to these two major types of migraine, other subtypes or forms are known, including vestibular, ocular and/or abdominal migraine (Headache Classification Committee of the International Headache Society, 2018). Dependent on the frequency of attacks, migraine is furthermore divided as chronic or episodic.
Sumatriptan was approved in 1992 by the US Food and Drug Administration (FDA). Since then, triptans have been the leading option for treating acute migraines. As for preventive treatment, various classes of medicines, such as beta-blockers (propranolol and metoprolol), calcium channel blockers (flunarizine), tricyclic antidepressants (amitriptyline) and also anti-epileptics (such as valproic acid [VPA] and topiramate), have been used (Parikh & Silberstein, 2019).
Such a heterogeneous group of drugs with varying mechanisms of action (MOAs) reflect the partial understanding of the pathomechanisms underlying migraine. Subsequently, these drugs have various efficacy and tolerability. The recent approval of anti-CGRP therapies is a milestone in the development of migraine therapy, as this new group of therapeutic agents allows an unprecedented targeted therapeutic approach to migraine prophylaxis based on the modulation of the signalling pathways that are known to play a crucial role in migraine (Edvinsson, 2017). Recent preclinical and clinical investigations have provided evidence for the critical role of the trigeminovascular system, particularly as a site of action of CGRP, in migraine initiation (Lassen et al., 2002;Storer et al., 2004).
Of note, several recent publications have focused on providing an overview of a single treatment option and discussing the CGRP system.
We aim to systematically analyse the trends of the development of novel treatment compounds and targets in the field of migraine by assessing the agents approved between 1970 and 2020, as well as the investigational agents in clinical trials during the last 3 years. This review highlights drug development for acute migraine and migraine prophylaxis and provides an update on the latest developments, including the CGRP system, as well as other therapeutic pathways and targets.

| MIGRAINE PATHOGENESIS
A migraine attack can be divided into several pathogenic mechanisms relevant to the different phases of migraine, that is, premonitory phase, the aura phase, the headache phase and the postdrome phase (Charles, 2013), which are briefly described in the following sections.

| Premonitory phase
The premonitory phase of migraine can start as early as 3 days before a migraine attack. This allows approximately 75% of patients to correctly predict a migraine incidence for up to 12 h before its onset (Giffin et al., 2003). The most common premonitory symptoms include fatigue, impaired concentration, neck stiffness, photophobia, yawning and food craving (Karsan et al., 2016). PET studies of triggered (Maniyar et al., 2014) and spontaneous attacks (Denuelle et al., 2007) provided evidence for a hypothalamic activation in the premonitory phase. Activation of this region and its central connections to the limbic system may explain symptoms during the premonitory phase, such as yawning, polyuria, food craving and mood changes (Maniyar et al., 2014), and the fact that migraine is commonly triggered by alterations in homoeostasis (e.g., changes in sleep-wake cycles and missed meals). Moreover, structural and functional imaging studies have demonstrated differences in thalamic and thalamo-cortical activities in patients with migraine and controls during migraine attacks (Coppola et al., 2016). Studies have shown that the thalamus is a key mediator of cutaneous allodynia (Wang et al., 2015) and exacerbation of headache by light (Noseda et al., 2016).

| Aura phase
One of the components of migraine headaches in less than 25-30% of patients is the occurrence of aura (Russel et al., 1996). Migraine aura is defined as a focal neurological disturbance that manifests as visual, sensory or motor symptoms (Headache Classification Committee of the International Headache Society, 2018). The exact pathological mechanism underlying aura is not fully understood. To date, cortical spreading depression (CSD) is considered to be the physiological cause of the aura phase of migraine attacks (Pietrobon & Moskowitz, 2013). CSD is an abnormal phenomenon characterised by a slowly propagating wave of depolarisation of cortical neuronal and glial cells followed by a depression of electrical activity (Leo & Morison, 1945). CSD is associated with toxic alterations in transmembrane ion gradients, which result in a massive influx of Na, Ca and water, and the efflux of K, proton, glutamate and ATP (Harriott et al., 2019). Several metabolic factors are associated with CSD. For instance, hyperglycaemia causes resistance against CSD initiation and hastens CSD recovery, in contrast with hypoglycaemia, which prolongs CSD (Hoffmann et al., 2013).

| Headache phase
The headache phase of migraine is due to the activation of the trigeminovascular system (Burstein et al., 2015;Stankewitz et al., 2011). One possible mechanism underlying this process is the opening of pannexin 1 channels in neurons, which are associated with the P2X7 ligandgated ion channel (Bolay et al., 2002), which lead to the activation of astrocytes that may induce the formation of cytokines and prostanoids that sensitise meningeal nociceptors (Karatas et al., 2013). Sensory transmission of nociceptive signals involves the release of several neurotransmitters, including CGRP, pituitary adenylate cyclase-activating polypeptide-38 (PACAP-38), glutamate and NO (Goadsby, Holland, et al., 2017). The release of CGRP and PACAP-38 also results in cranial vessel dilation and mast cell degranulation, both of which could further activate vascular and meningeal nociceptors and contribute to migraine headache. Several studies have shown that the TRP channels, a large family of non-selective cation channels that are important in pain signalling pathways are particularly involved. TRP channels are considered to play an important role in migraine pain and associated symptoms, including hyperalgesia and allodynia (Benemei et al., 2013). However, Summ et al. (2011) reported that a blockade of the TRPV1 receptor did not have significant effect on trigeminal nociception in rats. CGRP is known to play a role in spontaneous and triggered migraine headaches.
A recent randomised control trial reported that 67% of migraineurs developed migraine attacks after they received CGRP as intravenous infusion (Younis et al., 2019). Additionally, CGRP concentrations have been reported to be increased in migraineurs, in contrast with the concentrations of vasoactive intestinal peptide, substance P and neuropeptide Y, which appeared to be stable during migraine attacks.
This underscores the central and specific role of CGRP in the evolution of migraine headache (Goadsby et al., 1990).

| DATA COLLECTION
The current review collects and systematically analyses FDA-approved drugs and drugs undergoing clinical trials for migraine treatment to provide insights into the historical and novel trends in pharmacological treatment approaches in this field. The data for FDA-approved drugs were collected manually from the Drugs@FDA database; this dataset was compiled by searching for all drugs with the indications including 'migraine', 'migraine acute' and 'migraine prophylaxis', which had been approved by FDA in the interval between 1970 and 2020. Drugs@FDA includes information about drugs, including biological products, approved for humans. The agents undergoing clinical trials were obtained from ClinicalTrials.gov using the following search terms: 'migraine', 'migraine acute', 'migraine, preventive', 'migraine with aura', 'migraine without aura' and 'migraine disorders'. The search results falling within the time frame from January 2018 to April 2020 were considered. ClinicalTrials.  (Table 1). Drugs that have been approved for migraine within the last 50 years belong primarily to two classes: drugs specifically developed for the treatment of acute migraine, including 5-HT (serotonin) receptor agonists, and anti-CGRP acting drugs. The second class includes the more general analgesics such as the COX inhibitors, which are ubiquitously used for the treatment of different pain-related syndromes ( Figure 2). Of note, the four drugs approved between 2018 and 2020 targeted the CGRP and 5-HT 1F receptors, indicating a shift in the drug development strategy for novel drug targets. This shift is based on the new knowledge gained about the role of CGRP and 5-HT 1F receptors in migraine pathogenesis. The novel CGRP receptor antagonists (gepants) and 5-HT 1F receptor agonists (ditans) are considered promising alternatives, given their better efficacy combined with their milder side effect profile than those of older drugs used for acute migraine, although the entire spectrum of pharmacological effects induced by these agents is not yet fully understood. The overall spectrum of drug targets has considerably widened since 2018 when compared with that of the drugs approved between 1970 and 2018, targeting the 5-HT system and its 5-HT 1D/1B receptors ( Figure S1).

| Agents in clinical trials in the interval between 2018 and 2020 for acute migraine-Spectrum of molecular targets
Altogether, 12 drugs that had undergone clinical trials for acute migraine in the interval between 2018 and 2020 were identified (Table 1). Six of these drugs have already been approved by the FDA for treatment of other indications (i.e., not migraine) and are considered repurposed drugs. They primarily act as anaesthetics or analgesics and are currently being tested for their efficacy in treating migraine.
Although studies of five of these agents are conducted through medical centres (dexamethasone, lidocaine and bupivacaine) or are universitydriven (ketorolac and indomethacin), ketamine is tested through an industry driven study. Interestingly, several of these repurposed drugs function via new MOAs for migraine treatment, including targeting the

Ergot alkaloids
Dihydroergotamine (DHE) is indicated for the acute treatment of migraine headaches with or without aura and the acute treatment of cluster headache episodes. DHE binds with a high affinity to 5-HT 1Dα and 5-HT 1Dβ receptors. Two theories have been proposed to explain the efficacy of the 5-HT 1D receptor agonists in migraine: (i) activation of 5-HT 1D receptors located on intracranial blood vessels and subsequent vasoconstriction, although this is of minor importance for therapeutic action, and (ii) activation of 5-HT 1D receptors on sensory nerve endings of the trigeminal system, which results in the inhibition of the release of pro-inflammatory neuropeptides, such as CGRP, and is the pathway of major therapeutic importance (Silberstein & Kori, 2013

Triptans
For approximately 25 years, triptans have been the most effective drugs for treating migraine and thus an important part of migraine therapy. Triptans amplify the 5-HT signal by stimulating 5-HT receptors located in cranial blood vessels and nerve endings and relieve pain by inhibiting the release of peptides such as CGRP and substance P, as well as by acting in other ways that are not yet known.
Based on a better understanding of migraine pathophysiology, the vasoconstriction potential of triptans is no longer considered the main path of the pharmacological action of triptans (Edvinsson et al., 2018).
At the cellular level, triptans alleviate migraine symptoms by binding to 5-HT 1B/1D receptors along with pain signalling circuits and are effective due to high receptor and disease specificity (Ma et al., 2001).
Zolmitriptan is a 5-HT 1 agonist that mainly targets the 5-HT 1B/1D receptors and was first approved by the FDA in 1997. It has two putative mechanisms for therapeutic action: first, inhibiting of the release of vasoactive neuropeptides from the perivascular trigeminal sensory neurons, and second, reduction of pain signal transmission in the trigeminal dorsal horn (Rizzoli, 2014). Zolmitriptan is currently in a Phase 3 clinical trial as a novel formulation as a nasal spray (Zomig) for use in children aged 6-11 years (NCT03275922). Rizatriptan is another triptan that was already approved by the FDA in 1998 and is currently in several clinical studies, including a Phase 2 clinical trial for the treatment of episodic dizziness in vestibular migraine, and in a Phase 3 trial in combination with naproxen or meloxicam for the acute treatment of migraine (NCT04384367). As triptans also have vasoconstrictive effects on peripheral vessels, including cardiac vessels (Ferrari et al., 2002), these drugs are contraindicated in patients with cardiovascular disease, cerebrovascular disease, uncontrolled hypertension or hemiplegic migraine. Emerging evidence suggests that cerebral vasodilation is not the nociceptive stimulus in migraine; and attention has turned to developing new acute migraine treatments, such as the ditans, that target the trigeminal pathways while avoiding the vasoactive 5-HT 1B and 5-HT 1D receptors (Oswald & Schuster, 2018).

Ditans
Ditans belong to a novel drug class that targets the 5-HT 1F receptor for the treatment of acute migraine. One of this class, lasmiditan, a highly selective 5-HT 1F agonist, was approved in 2019. Ditans are structurally different from triptans in that triptans possess an indole structure that closely resembles the 5-HT receptor, whereas ditans replace this indole group with a pyridine-piperidine scaffold.
Triptans bind non-selectively to 5-HT 1B and 5-HT 1D receptors and with varying affinity to 5-HT 1F receptors, causing direct vascular vasoconstriction. In contrast, ditans are selective for the 5-HT 1F receptor, and their MOA is neuronal, without evidence of vasoactive effects (Goadsby & Classey, 2003). Preclinical studies have shown that lasmiditan inhibits CGRP release via binding to 5-HT 1F receptors (Labastida-Ramírez et al., 2020). The release of pain-provoking stimuli (CGRP and substance P) from the presynaptic membrane of the trigeminal system can be indirectly measured by quantifying protein extravasation from the dura mater following painful stimulation (Markowitz et al., 1987). Johnson et al. showed that extravasation of dura plasma proteins in guinea pigs significantly decreased with increasing 5-HT 1F receptor affinity. Lasmiditan has been shown to block plasma protein extravasation (PPE) (Johnson et al., 1997).
However, PPE as a therapeutic target for migraine is still under investigation. Additionally, Shepheard et al. (1999) observed a dosedependent 5-HT 1F response decrease in action potential generation due to electrical dural stimulation, suggesting that 5-HT 1F agonists inhibit second-order neurons in the trigeminal nucleus caudalis.
Lasmiditan is a promising acute antimigraine therapy, in particular for patients with cardiovascular risk factors, contraindications or unwanted side effects to triptans. However, lasmiditan is associated with CNS-related side effects including dizziness, somnolence and paraesthesia (Goadsby et al., 2019). A Phase 3 study reported that the incidence of dizziness with lasmiditan increased with dose (Tepper et al., 2019). Lasmiditan is currently in a Phase 3 trial testing safety and efficacy in acute migraine treatment in children (NCT04396574).

| Drugs targeting the CGRP system-Gepants
Since the late 1980s, studies have shown that CGRP is significantly involved in the pathophysiology of migraine (Goadsby et al., 1990;Shepheard et al., 1999). CGRP can trigger migraine in patients (Younis et al., 2019), and blockade of the canonical CGRP receptor (Lassen et al., 2002) is effective in the treatment of acute migraine (Goadsby et al., 1990). Low MW CGRP receptor antagonists (gepants) have been investigated regarding their effect in both acute migraine treatment and migraine prevention. Several previously studied low MW CGRP receptor antagonists were effective in Phase 2 and 3 trials; however, their development was halted owing to hepatotoxicity, the inability to develop an oral formulation or other unknown reasons (Ho et al., 2014;Marcus et al., 2014). Blocking of CGRP receptors by gepants inhibits both vascular nociceptive transmission and thalamic trigeminal nociceptive activation (Olesen, Diener, et al., 2004). Olesen, Diener, et al. (2004) (Scott, 2020). Ubrogepant is well tolerated with the most common adverse events reported as nausea, somnolence and dry mouth with a frequency rate below 5% . The second drug of this class, rimegepant, was approved by the FDA in February 2020. Rimegepant is given orally and reaches peak plasma concentrations after 2 h (Croop et al., 2018). The side effect profile of rimegepant is mild, with nausea and urinary tract infections being the most commonly occurring events, and it has an overall similar side effect profile comparable with placebo (Croop et al., 2018). However, the therapeutic gain for gepants (rimegepant: 5-7.6%; ubrogepant: 6.4-9.4%) appears to be lower than for sumatriptan (16-21%) or lasmiditan (7.3-17.5%) (Derry et al., 2014). Because gepants do not constrict cranial arteries, they can be used as a first-line anti-migraine treatment in patients at increased risk for cardiovascular events risk or in those with documented cardiovascular disease (Conway et al., 2019). Furthermore, they can be used as a second-line treatment where treatment with triptans has failed (Do et al., 2019).

| Anaesthetics
Ketamine The hypothesis that ketamine may be used as a drug against migraine is not new and has been studied for more than 20 years (Lauritzen & Hansen, 1992). The interest in ketamine in migraine treatment has increased again recently. Ketamine has been previously investigated as an opioid-sparing agent in the treatment of pain (Motov et al., 2015) and as a treatment option for mood disorders (Lapidus et al., 2014). Compelling evidence exists for each of these indications.
In contrast to this, reliable data for the significant effect of ketamine in migraine are still missing. NMDA receptors targeted by ketamine are widely distributed in the CNS. Under resting conditions, the channel pores are blocked by magnesium (Mg 2+ ), which prevents the flow of ions through the receptor channel. During neuronal depolarisation, Mg 2+ leaves pores allowing the influx of Ca 2+ (Parsons et al., 1999). Several preclinical studies suggest the effect of Mg 2+ on NMDA receptors as a migraine mechanism. Studies have shown that glutamate receptors are located on the superficial laminae of the trigeminal nucleus caudalis (TNC) (Tallaksen-Greene et al., 1992). Local stimulation of these receptors with microiontophoretically administered glutamate induces the postsynaptic facilitation of neuronal activity within the TNC (Hill & Salt, 1982). Ketamine is a selective antagonist of the NMDA receptor, which is believed to be its primary Ketamine is currently in a clinical trial for evaluating its efficacy in acute migraine treatment (NCT02697071).

Lidocaine and bupivacaine (GON blockers)
The blocking of the GON has a considerable effect on the trigeminovascular system, which plays a key role in migraine pathophysiology. A rat study reported that the excitability of meningeal afferent input increased via stimulation of the GON and cutaneous C-fibre afferents in response to mustard oil. The study reported that there is a functional coupling between nociceptive meningeal afferent and cervical afferents in the GON (Bartsch & Goadsby, 2002). Two medical centre-driven studies have been investigating the use of lidocaine and bupivacaine as treatment alternatives for acute migraine, testing the effect of injectable and intranasal formulations of the agents. The strength of intranasal lidocaine is its route of administration, rapid effectiveness and rare adverse reactions (Avcu et al., 2017;Maizels & Geiger, 1999). Lidocaine has been studied for over 20 years (Maizels & Geiger, 1999) regarding its usability in acute and chronic migraine treatment (Kashipazha et al., 2014;Puledda et al., 2018). Furthermore, blocking the sphenopalatine ganglion decreases signals to the intracranial nociceptors that innervate migraine pain (Avcu et al., 2017;Maizels & Geiger, 1999). However, the effect of this mechanism on migraine remains controversial (Avcu et al., 2017;Maizels & Geiger, 1999).

| Glucocorticosteroids and non-steroidal antiinflammatory drugs
Dexamethasone Migraine is also characterised by inflammation. Thus, the use of antiinflammatory agents to inhibit the inflammatory cascade may help to relieve migraine headaches and prevent their recurrence, and antiinflammatory agents may provide benefits through the mitigation of inflammation (Khazaei et al., 2019). Glucocorticosteroids and nonsteroidal drugs are the most appealing, economic and familiar agents for the treatment of migraines. The use of glucocorticosteroids in the emergency department may vary, and published reports suggest that corticosteroids are infrequently used for acute severe migraine (Vinson, 2002). A medical centre-driven study currently investigates optimal dosing of dexamethasone in acute migraine (Phase 4) (NCT04112823) and its efficacy to prevent relapse in acute migraine in children (Phase 1) (NCT02903680).

NSAIDs (ketorolac and indomethacin)
NSAIDs are one of the most common classes of pharmacological interventions to treat acute migraine attacks. They are considered non-specific abortive treatments with a long history and demonstrated efficacy for acute migraine treatment in adults (Lipton et al., 2005;Misra et al., 2004). Two repurposed NSAIDs, indomethacin and ketorolac, are in university-driven clinical trials for treatment of migraine. Indomethacin has been used to treat different forms of indomethacin-responsive headaches such as trigeminal autonomic cephalgias (for instance, paroxysmal hemicranias), Valsalva-induced headaches and primary stabbing headaches (ice pick headache or jabs and jolts syndrome) (Dodick, 2004

| The evolution of drug approvals for the treatment of chronic migraine treatment over time-A quantitative and drug class analysis
In total, 10 drugs have been approved by the FDA for the treatment of chronic migraine in the time interval from 1970 to 2020 (Table 2).
As shown in Figure 1 Anti-CGRP agents target a biological pathway that has been shown to play an important role in migraine pathogenesis, offering novel perspectives for therapeutic options in episodic and chronic migraine.

| Agents in clinical trials in the interval between 2018 and 2020 for chronic migraine-Spectrum of molecular targets
Ten drugs were identified in clinical trials for the indication of migraine prophylaxis between 2018 and 2020 ( Table 2) F I G U R E 1 (a) Agents approved by the FDA for the acute and prophylactic treatment of migraine from 1970 to 2020. Since the 1970s, 27 agents have been approved by the FDA for migraine treatments. (b) Agents that have been approved as a new molecular entity specifically for migraine treatment and repurposed agents that have additional approval for migraine treatment. In total, 18 agents have been approved as a new molecular entity specifically designed for migraine management, and seven repurposed agents have been approved for either acute or prophylactic migraine treatment F I G U R E 2 FDA-approved agents and their therapeutic targets for acute and prophylactic treatment of migraine. Fourteen agents that were new molecular entities (NMEs) designed specifically to treat acute migraine have been FDA approved, and the therapeutic targets they target are shown in blue. Three repurposed agents, that is, drugs that have been previously FDA approved for a different indication, have had additional approvals for acute migraine treatment and are shown in dark blue. Four FDA-approved NMEs that were specifically designed for prophylactic migraine treatment and their targets in the CGRP pathway are shown in light orange, and six repurposed drugs have also been approved and with their targets are shown in brown F I G U R E 3 Investigative agents and their therapeutic targets for prophylactic migraine treatment. In total, 11 agents are in clinical trials for treatment of preventive or acute migraine. All of the agents in trials for acute migraine treatment are repurposed drugs that target COX enzymes, the greater occipital nerve or the NMDA receptor (all shown in dark blue). Agents in trials for prophylactic migraine treatment are more diverse, with three new molecular entities (NMEs) (shown in light orange) that target two novel targets-pre-kallikrein mRNA and calcium channels as well as the validated CGRP receptor. Additionally, two repurposed drugs (shown in brown) are also in trials that target the angiotensin AT 1 receptor and acid-sensing ion channels (ASICs) prophylaxis. Drugs act via different molecular targets, including the CGRP receptor, the angiotensin II receptor, ASICs, SNAP-25, calcium channels, prekallikrein mRNA, sodium channels, β-adrenoceptors, GABA transaminase and CGRP ligand ( Figure S2).

| Characterisation of drug classes approved
and those undergoing clinical trials for the treatment of chronic migraine

| Anticonvulsants
Since 1970, three anti-epileptics have been recommended as drugs of the first choice (topiramate and valproate) or third choice (gabapentin) to prevent migraine attacks (Linde et al., 2013a). The use of antiepileptics for the prophylactic treatment of migraine is theoretically warranted by several known modes of action, which relate either to the general modulation of pain systems or more specifically to systems involved in the pathophysiology of migraine (Silberstein et al., 2008). However, it is not, currently, possible to state the particular mode or modes of action of anti-epileptics, relevant to migraine prophylaxis (Linde et al., 2013a).

Topiramate
Topiramate is a sulfamate-substituted monosaccharide (2,3:4,5-bis-O-[1-methylethylidene]-β-D-fructopyranose sulfamate) derived from the naturally occurring sugar D-fructose (Martella et al., 2008). The efficacy of topiramate in migraine appears to be mediated by the interaction with multiple sites of action. It decreases the frequency of action potentials elicited by the depolarising electric current, which leads to a blockade of voltage-dependent Na + channels (McLean et al., 2000).
Topiramate modulates cortical excitability in migraineurs; however, it does not appear that this alone explains its efficacy in migraine prophylaxis (Artemenko et al., 2008). Topiramate inhibits the excitatory activity of glutamate at the receptor subtype for kainate/AMPA. It has been shown to inhibit neurons of the trigeminocervical complex via a GABA-mediated mechanism. Furthermore, topiramate inhibits the release of CGRP from prejunctional trigeminal neurons (Akerman & Goadsby, 2005). An inhibitory effect on high-voltage-dependent (HVA) Ca 2+ channels, particularly in the periaqueductal grey (PAG), is a possible mechanism to explain the therapeutic effect in migraine (Martella et al., 2008). Differential sensitivity to topiramate has been observed for HVA Ca 2+ channels located on cortical neurons and those in the PAG region. Topiramate inhibits N-, P-and L-type channels in PAG neurons, whereas in cortical neurons, it modulates only Pand L-type channels. Chronic, but not acute, treatment suppresses CSD in rats (Edvinsson & Linde, 2010, 2013b. Collectively, these point towards a reduction in excitatory transmission and increase in inhibitory neurotransmission (Linde et al., 2013a;Linde et al., 2013b).

| Botulinum toxin type A Onabotulinumtoxin A (Botox)
Botulinum toxin is a natural product synthesised by the anaerobic bacterium, Clostridium botulinum. The clinical use of botulinum toxin for conditions involving excessive muscle contractions began in the early 1980s. It has become the first-line therapy for many conditions, including dystonia, spasticity, hyperhidrosis and some forms of bladder disturbance. Since then, botulinum toxin has shown pain-relieving properties above what might be expected due to the relief of muscle contractions and its use for the treatment of migraine has developed (Herd et al., 2018). For treating migraine, botulinum toxin is administered intramuscularly to multiple sites around the head and back of the neck, with regular retreatment. One preparation of botulinum toxin, onabotulinum toxin A (Botox), has recently been licensed in the United States and the United Kingdom as an agent for the prevention of chronic migraine (a minimum of 15 migraine daysÁmonth À1 ) based on two large randomised, placebo-controlled trials Diener, Dodick, et al., 2010). Botulinum toxin has been reported to alleviate pain in a variety of conditions; however, the MOA is not well understood. It has been postulated that botulinum toxin acts to reduce pain by inhibiting the release of neuropeptides such as CGRP and substance P, which are involved in the initiation of migraine Diener, Dodick, et al., 2010;Herd et al., 2018).

Gepants and anti-CGRP monoclonal antibodies
CGRP is now a validated target for the treatment of migraine, as it has a crucial role in migraine pathophysiology. Therapeutic agents developed to inhibit the CGRP pathway include gepants and monoclonal antibodies (Edvinsson et al., 2018). The mechanism by which monoclonal antibodies are proposed to prevent migraines is by blocking CGRP transmission in the trigeminovascular system (Manoukian et al., 2019). Monoclonal antibodies that target the neuropeptide CGRP (galcanezumab, eptinezumab and fremanezumab) are proposed to bind and thus deactivate CGRP released by trigeminal sensory nerve fibres (Melo-Carrillo et al., 2017), whereas antibodies that target the CGRP receptor (erenumab) presumably act by preventing access of CGRP to its canonical receptor (Manoukian et al., 2019). There are two known isoforms of CGRP. CGRPα is the predominant form in the CNS and peripheral nervous system implicated in migraine pathology, and it is targeted by monoclonal antagonising antibodies. CGRPβ is mainly found in the enteric nervous system and has not been well studied (Lassen et al., 2002). There are several components to receptors that CGRP bind with: the calcitonin receptor-like receptor (CLR), the calcitonin receptor (CTR) and three receptor activity-modifying proteins (RAMPs). The canonical CGRP receptor consists of the CLR/RAMP1 complex and has high affinity to CGRP (Hay et al., 2018). In addition, the CTR/RAMP1 complex, called AMY 1 , also has high affinity for CGRP and should certainly be considered as a possible player in CGRP action in migraine besides the canonical CLR/RAMP1 CGRP receptor (Hay, 2019). The binding of CGRP to the extracellular binding pocket formed by the heterodimer causes activation of G proteins containing the Gαs subunit bound to CLR, which in turn activates AC and cAMP-dependent signalling pathways that mediate vasorelaxation in blood vessels (Lassen et al., 2002). Much remains to be learned about the physiological and pathophysiological roles of CGRP as this peptide has effects throughout the body (Lassen et al., 2002) and circulating antibodies could affect all peripherally accessible sites where CGRP acts. One concern relates to the role of CGRP as a potent vasodilator throughout the vascular system (Lassen et al., 2002). Consequently, chronic reduction of CGRP has the potential to cause cardiovascular pathophysiologies, such as hypertension, cardiac dysfunction, and episodes of coronary or cerebral ischaemia (MaassenVanDenBrink et al., 2016). To date, no cardiovascular adverse effects or development of hypertension due to use of anti-CGRP and anti-CGRP receptor antibodies have been reported in up to 6 months of Phase 3 clinical testing (Goadsby, Reuter, et al., 2017;Silberstein et al., 2017). In a study published in 2018, erenumab had no effect on patients with angina pectoris (men aged $65 years) who were monitored using electrocardiography, while exercising on a treadmill until they reported chest pain (Depre et al., 2018). Studies have shown how a reduction in CGRP signalling is likely to interfere with a mild rather than a complete block of CGRP-dependent signalling pathway. Hence, this risk for cardiovascular side effects maybe rather low. It should be noted that CGRP actions in both the peripheral nervous system and CNS are well positioned to contribute to migraine symptoms. Although the critical sites of action remain controversial, it appears likely that both central and peripheral sites may be important in the initiation and treatment of migraine (Russo, 2015).

Antagonists of β-adrenoceptors
Propranolol is one of the most regularly used drugs for migraine prophylaxis with many clinical trials consistently proving its efficacy in reducing the frequency of migraine attacks since the 1970s (Diener et al., 2002). The MOAs of propranolol in migraine prophylaxis are not fully understood. It has been proposed that propranolol reduces central hyperexcitability through β 1 -adrenoceptor-mediated inhibition of noradrenaline release, thus reducing central catecholaminergic hyperactivity. Furthermore, propranolol inhibits NO production by blocking inducible NOS and also inhibits kainate-induced currents and is synergistic with NMDA blockers, which reduces neuronal activity and has membrane-stabilising properties (Dakhale et al., 2019). One recent study has shown a significant reduction in frequency, severity and duration of migraine headaches (Dakhale et al., 2019).
Timolol is a non-selective β-adrenoceptor antagonist that is available for both topical and systemic administration. Topical timolol is primarily used to reduce intraocular pressure in patients with open-angle glaucoma and ocular hypertension (Lazreg et al., 2018); however, systemic administration of timolol can be part of a regimen managing hypertension and, importantly, migraine prophylaxis (Durley et al., 1981;Stellar et al., 1984). Timolol has been an effective method for migraine prophylaxis (Tfelt-Hansen et al., 1984) and was approved in 1978, although the exact mechanism of timolol as migraine prophylaxis is unknown. It is likely to exert its effects through a variety of processes. One proposed mechanism is that blockade of β-adrenoceptors decreases the synthesis and release of noradrenaline, an important intermediate in the pathophysiology of migraine. Another pathway that could contribute to its migraine prophylactic properties is that β-adrenoceptor antagonists can regulate the neuronal firing of PAG matter using GABA. Timolol also appears to play a role in regulating the 5-HT system by inhibiting 5-HT, another important mediator in the pathophysiology of migraine. This modulation of the effects of 5-HT also appears to contribute to the ability of β-adrenoceptor antagonists to reduce the sensitivity of the auditory system and thus reduce the frequency of migraine attacks. There is also a hypothesis that β-adrenoceptor antagonists play a significant role in reducing the excitability of the visual system in patients with migraines. However, it should be noted that the vast majority of preventive medications for migraine have the potential to diminish multisensory sensitivities to a certain extent. β-Adrenoceptor antagonists, such as timolol, are also thought to reduce the spread of signals through the brain, including the cortical spreading as well as the excitability of the ventroposteromedial thalamic nucleus (Shields & Goadsby, 2005).

Flunarizine
Flunarizine is a well-recognised migraine preventive agent that has been extensively used in some parts of the world in paediatric and adult populations for over three decades. Flunarizine is a non-selective calcium ion channel and dopamine receptor antagonist, which has 5-HT receptor antagonist and anti histamine (H 1 receptor antagonist) actvityies (Karsan, Palethorpe, et al., 2018). It should be noted that flunarizine is not approved by the FDA owing to concerns over its potential to induce Parkinsonism. Currently, flunarizine is being tested in a Phase 4 study to assess α-lipoic acid as an add-on therapy for the treatment of adolescent migraine (NCT04064814). Louis (1981), for the first time, reported a significant reduction in the frequency of migraine attacks following flunarizine therapy, compared with placebo.
Based on this study, more double-blind placebo-controlled trials were performed (Diener et al., 2002;Sørensen et al., 1986). Calcium antagonists are known to display anti-vasoconstrictive action and to protect tissues from the sequelae of hypoxia and/or ischaemia (Van Reempts et al., 1983). However, the substances have different effects on different calcium channels, different antihypoxic effects on cerebral tissues and several other pharmacological properties. For example, no effect on any type of migraine could be established for the calcium antagonist nimodipine (Ansell et al., 1988). Based on the hypothesis that cerebral hypoxia plays a decisive role in the pathogenesis of migraine (Amery, 1987), the anti-hypoxic activity of flunarizine has been discussed as a MOA in the prevention of migraine. Although hypoxia has been well documented in attacks of classical migraine attacks (Olesen et al., 1982), for cases other than simple migraine, this hypothesis cannot be established. This is particularly true as clinical studies do not show a discrepancy in the effectiveness of flunarizine in simple or classical migraine. Adequate insights into the MOA of flunarizine in migraine prophylaxis will probably not be achieved before sufficient explanations concerning the exact pathophysiological process in migraine disease are available.

Candesartan
Hypertension is recognised as a serious risk factor for the development of migraines (Owada, 2004). Candesartan has been reported to provide migraine prophylaxis since 2003 (Tronvik et al., 2003). Candesartan is an angiotensin II AT 1 receptor blocker that was initially developed as an antihypertensive agent to block the action of angiotensin II (Mizuno et al., 1992). The renin-angiotensin system has the potential to cause organ damage, independent of hypertension (Morgan, 2003), and administration of candesartan is reported to inhibit oxidative stress (Dohi et al., 2003). Based on the migraine mechanisms described earlier, the characteristics specific to candesartan, and the efficacy of treatment in these cases, it can be hypothesised that the mechanism of candesartan migraine prophylaxis involves either or both of the following: (i) inhibition of excessive vasoconstriction due to 5-HT release and (ii) inhibition of neurogenic inflammation. Candesartan inhibits vasoconstriction by blocking of AT 1 receptor stimulation in vascular smooth muscles. This action may break the vicious circle of migraines by inhibiting excessive cerebrovascular constriction (Groth et al., 2003). In addition, recent reports indicate that pretreatment using candesartan acts to regulate the progression of vascular permeability (Groth et al., 2003) and to deactivate the inflammatory nuclear transcription factor NF-κB induced by angiotensin II (Yoshiyama et al., 2001). These findings suggest that candesartan may suppress neurogenic inflammation in the cerebral vasculature, which is conducive to migraine (Mizuno et al., 1992). Candesartan is currently in clinical trials for its efficacy and tolerability in episodic and chronic migraine in adults (NCT04138316 and NCT04574713).
5-HT levels have long been linked to migraine (Wang et al., 2013), and NGF levels are increased in the CSF of patients with chronic headache and a history of migraine (Dussor, 2014). NO donors, one of the most reliable migraine triggers in humans, also potentiate ASIC currents (Cadiou et al., 2007). Holton et al. (2020) reported that the blockade of ASIC-3 inhibits durovascular-evoked and NO-induced sensitisation of trigeminal nociceptive responses in rats. Moreover, decreased pH and ASIC activation may also play a role in the initiation or propagation of CSD (Dussor, 2015), and it is known that there is a decrease in pH within the cortex during CSD as well as in related types of spreading depolarisations (Pietrobon & Moskowitz, 2014), which may occur as CSD can increase metabolic demand without a corresponding increase in blood flow leading to hypoxia of cortical tissue (Takano et al., 2007).
Increased ASIC1a expression has been found in the spinal dorsal horn neurons following peripheral inflammation, and antagonising this channel reduced pain-related behaviour in rodents (Piilgaard & Lauritzen, 2009). However, there is no clear evidence that shows amiloride is an effective choice for migraine treatment. Amiloride is currently recruiting a Phase 2 clinical trial to evaluate its efficacy in the prophylaxis of migraine with aura in adults (NCT04063540).

IONIS-PKKRx
The notion of creating 'antisense' oligonucleotide-based drugs, first enunciated in 1978 (Wu et al., 2004), is seductive as the forces that determine whether an oligonucleotide binds to its cognate RNA sequence, Watson-Crick hybridisation, are well understood and as such drugs should, in principle, be much more specific than conventional low MW drugs. Bradykinin has been shown to produce both excitation and mechanical sensitisation of somatic (Stephenson & Zamecnik, 1978) and meningeal nociceptors (Steen et al., 1992).
Migraine, or vascular, headache has been hypothesised by some investigators to be caused by cerebral arterial dilation and stimulation of pain receptors associated with the vasculature (Strassman et al., 1996).
Based on these investigations, it has been hypothesised that bradykinin may be involved in the development of vascular headache, through the conversion of endogenous brain kininogens to bradykinin and subsequent bradykinin-dependent cerebral arterial dilation and bradykinindependent stimulation of pain fibres (Caviness & O'Brien, 1980

| CONCLUSION
In this article we have systematically analysed the trends of migraine drug development both qualitatively and quantitatively and described each of the drug classes, MOAs, clinical trials and new treatments. We identified two major classes of FDA-approved drugs for the treatment of acute migraine, which include antimigraine and anti-inflammatory acting drugs. Four primary classes for chronic migraine were identified: anti-migraine, anti-convulsant, cardiovascular and neuromodulatory drugs. In comparison, therapeutic classes in clinical studies for acute migraine treatment also include steroidal acting drugs, whereas chronic migraine also includes diuretic and anticonvulsant acting drugs.
Although the first migraine drugs came into the market in the 1970s, the development of migraine therapeutics has significantly intensified significantly during the last 3 years as shown by our analysis of trends in drug development between 1970 and 2020. Overall, the drugs for migraine treatment approved in the last 50 years are a heterogeneous group of therapeutic agents characterised by widely different MOAs. This is also reflected in an increase in our understanding of the many different mechanisms, proposed to be important for migraine.
In this analysis, we can see that there are two time periods Drug approvals in chronic migraine made in the last 3 years are dominated by compounds targeting the CGRP system. The successful outcomes from studies with antibodies targeting CGRP or its receptor have driven pharmaceutical companies to transform the migraine drug developmental landscape towards specific anti-migraine-acting drugs.
The development of new technology platforms to address endogenous soluble ligands such as CGRP as drug targets is likely to enhance this development further (Attwood et al., 2020). Furthermore, we see that the number of trials for chronic migraine is increasing. One clear aim is to widen the indication for these drugs in different demographics, for example, in children and adolescents, as well as certain migraine subtypes and usability as both acute and prophylactic treat- This analysis highlights the trend towards a higher number and wider spectrum of therapeutic agents tested for the treatment of acute and chronic migraines. This also reflects considerable progress in understanding the pathogenic processes underlying migraine that has been gained in recent years. Given that the non-responder rate is relatively high, even for novel anti-migraine therapeutic agents targeting the CGRP system is approximately 30% (Do et al., 2019;Wrobel Goldberg & Silberstein, 2015). Thus, consideration of a wider spectrum of drug targets and classes is important for designing individually adapted therapy for patients with migraine.
6.1 | Nomenclature of targets and ligands