THE PATHOPHYSIOLOGY OF MIGRAINE
`
`StewartJ.Tepper,MD,AlanRapoport,MDandFredSheftell,MD
`
`BACKGROUND– Migraine results from episodic changes in central nervous system physiologic function in hyperexcitable
`brain manifested by abnormal energy metabolism, lowered threshold for phosphene generation, and increased contingent
`negative variation. Human functional magnetic resonance imaging and magnetoencepholography data strongly suggest
`that aura is caused by cortical spreading depression.
`REVIEW SUMMARY– Brain hyperexcitability may be caused by low magnesium levels, mitochondrial abnormalities with
`abnormal phosphorylation of adenosine 59-diphosphate, a dysfunction related to nitric oxide, or calcium channelopathy.
`Low magnesium can result in opening of calcium channels, increased intracellular calcium, glutamate release, and
`increased extracellular potassium, which may in turn trigger cortical spreading depression. Mitochondrial dysfunction has
`been suggested by a low phosphocreatine:Pi ratio and a possible response by migraine patients to riboflavin prophylaxis.
`Nitroglycerine administration results in a delayed migraine-like headache in migraine patients but not in control patients,
`and a nonspecific nitric oxide synthase inhibitor aborted migraine at 2 hours in the majority of tested migraine patients
`compared to controls. Many patients with familial hemiplegic migraine have a missense mutation in the P/Q calcium
`channel, so that this form of migraine, at least, is associated with a demonstrable calcium channelopathy.
`CONCLUSIONS– The generation of migraine occurs centrally in the brain stem, sometimes preceded by cortical spreading
`depression and aura. Activation of the trigeminovascular system stimulates perivascular trigeminal sensory afferent nerves
`with release of vasoactive neuropeptides, resulting in vasodilation and transduction of central nociceptive information.
`There is then a relay of pain impulses to central second- and third-order neurons and activation of brain stem autonomic
`nuclei to induce associated symptoms.
`KEYWORDSmigraine,trigeminovascular,pathophysiology,serotonin
`(THENEUROLOGIST7:279–286,2001)
`
`There are multiple unanswered questions on the patho-
`physiology of migraine. Is there good evidence for the hy-
`perexcitability of brain? What are the causes of the prodrome
`and aura, which may precede migraine pain? Specifically,
`what is the cause of aura and the relationship between aura
`and cortical spreading depression? How do the anticipatory
`events lead to migraine pain, and what generates and main-
`tains migraine pain? What is the genetic susceptibility to
`migraine? What is the anatomy of the migraine system? What
`
`From The New England Center for Headache, Stamford, Connecticut,
`Department of Neurology, Yale University School of Medicine, New
`Haven, Connecticut, and Department of Psychiatry, New York Medi-
`cal College, New York, New York.
`Send reprint requests to Stewart J. Tepper, MD, Director, the New
`England Center for Headache, 778 Long Ridge Road, Stamford, CT
`06902. E-mail sjtepper@aol.com
`
`is the relationship of serotonin, norepeinephrine, and dopa-
`mine receptor physiology to the migraine system? Finally,
`what areas are likely to result in further understanding of
`migraine and in improved therapeutic options for patients?
`
`HYPEREXCITABILITY AND THE BRAIN
`There are multiple studies supporting interictal and ictal
`hyperexcitability in migraine brain. Thomas et al (1) showed
`exaggerated CO2 reactivity with transcranial Doppler. Sakai
`and Meyer (2,3) described cerebral hyperperfusion and ab-
`normal cerebrovascular reactivity during migraine, which
`suggested elevated neuronal activity. Enhanced photic drive
`responses have been found in electroencephalogram studies
`(4). Schoenen and Timsit-Berthier (5) demonstrated in-
`creased contingent negative variation in migraine, an event-
`related slow cerebral potential. Aurora et al (6) and Aurora
`
`279
`
`EX2295
`Eli Lilly & Co. v. Teva Pharms. Int'l GMBH
`IPR2018-01712
`
`1
`
`

`

`280
`
`T H E N E U R O L O G I S T
`
`P A T H O P H Y S I O L O G Y O F M I G R A I N E
`
`and Welch (7) found a lower threshold for, and higher
`incidence of, phosphene generation in migraine patients,
`especially in cases of migraine with aura.
`Two groups have shown abnormal energy metabolism in
`migraine brain, Welch’s group (8) and Barbirolli’s group (9).
`Finally, Welch and Ramadan (10) demonstrated low phos-
`phocreatine, high adenosine 59-diphosphate, and a low phos-
`phorylation PCr:Pi ratio, which can lead to membrane in-
`stability. Thus,
`there is much data that point
`to this
`hyperexcitability.
`It is possible that migraine can occur in anyone. Raskin
`(11) described the precipitation of migraine-like headaches
`when neurosurgeons implanted electrodes in the periaque-
`ductal gray and raphe nuclei of the brain stem. Instead of
`resulting in analgesia,
`the implants produced migrainous
`headaches in the patients.
`Genetic susceptibility may result in a lowered threshold
`to activation of neurons in migraine patients. Presumably, a
`complex multifactorial polygenic inheritance with variable
`penetration results in the protean clinical picture of migraine,
`perhaps even including the presence or absence of aura and
`the differing triggers to migraine activation. The reasons for
`the episodic activation remain obscure.
`
`ANTICIPATORY SYMPTOMS
`Prodrome
`Prodromes consist of a variety of autonomic, neurologic,
`and systemic symptoms. Autonomic symptoms include fluid
`retention, yawning, diaphoresis, diarrhea, constipation, thirst,
`cold feelings, and polyuria. Neurologic symptoms may be
`euphoria,
`irritability, depression, photophonophobia,
`stiff
`neck, disinhibition and hyperactivity, and cognitive dysfunc-
`tion. Systemic symptoms include anorexia, fatigue, and food
`cravings. The nature of prodrome symptomotology suggests
`cortical, limbic, and hypothalamic activation (12).
`Prodrome is not aura, and how often it precedes migraine
`pain is unknown. There are no validated criteria for its
`diagnosis. Nonevolutive prodrome is described as occurring
`more than 6 hours and up to 48 hours before migraine pain,
`whereas evolutive prodrome leads directly into migraine
`pain, with or without aura, within 6 hours.
`
`Aura and Cortical Spreading Depression
`Because aura is most frequently a visual phenomenon,
`Welch et al (13) stated, “The occipital cortex is to migraine
`as the temporal
`lobe is to epilepsy.” Visual aura is most
`frequently a paracentral homonomous scotoma of crenellated
`crescentic shape or arc, with a bright, jagged border, which
`often enlarges, moves across the visual field, and scintillates.
`Lashley (14), in 1941, described his own aura as fortification
`spectra moving at a rate of 2 to 3 mm/min across the cortex.
`The International Headache Society criteria for migraine
`with aura are:
`
`1. Aura consisting of one or more reversible brain
`symptoms (cortical or brain stem)
`2. Gradual development of one aura symptom over 4
`minutes or more
`3. Aura duration less than 60 minutes
`4. Headache follows aura in less than 60 minutes (15).
`
`In 1944, Leao (16), at Harvard, reported that after a
`needle stab to rabbit cortex, there was a neural and glial
`gradual depolarization at a rate of 2 to 5 mm/min. Applying
`potassium or glucose or other trauma to mammalian cortex
`can also induce this cortical spreading depression (CSD).
`Milner (17) suggested the link between aura and CSD in
`1958, but he did not prove causality or equivalence.
`Is there human CSD? It can be induced in human
`hippocampus cells in tissue culture but has never been de-
`scribed in situ. In laboratory animals, CSD is a primary
`neuronal event, with secondary vascular components. The
`initial event of CSD is neuronal activation with hyperoxia,
`and Lauritzen et al (18 –20) found up to a 300% increase in
`blood flow.
`The hyperemia is followed by a wave of neuronal de-
`pressed activity, accompanied by spreading oligemia, at a
`level that does not reach ischemia. The events occur in gray
`matter, are not localized to one vascular distribution, and stop
`at major sulci,
`for example, the parieto-occipital sulcus.
`These features further underline the primarily neuronal,
`rather than the primarily vascular, basis of CSD.
`Direct observations of human aura by a variety of tech-
`niques are closing the circle of CSD and aura. Barkely et al
`(21), working in Welch’s laboratory, described magnetoen-
`cephalographic changes that included spontaneous depolar-
`ization, transient neuronal suppression, and large-amplitude
`waves in humans that parallel animal CSD.
`Moskowitz and colleagues in his laboratory have used
`BOLD, perfusion-weighted imaging, time of flight in func-
`tional magnetic resonance imaging (fMRI), and retinotopy
`with visual stimulation as they attempted to characterize the
`onset, speed of progression, pattern, duration, and recovery
`from human aura. They assumed that blood flow would be a
`useful surrogate to detect neuronal activation.
`The Moskowitz/Cutrer laboratory examined patients
`(including Dr. Cutrer himself ) in both spontaneous and
`precipitated aura and found initial brief vasodilation and
`hyperemia and then hypoperfusion, decreased mean transit
`time and blood flow, and a foveal representation moving
`anteriorly in the cortex. They demonstrated the onset of the
`aura by changes in the detection of luminance and motion,
`with suppression of visual activation, in V3A in the occipital
`cortex, moving slowly to V2, then to V1 (22).
`Welch’s magnetoencephalographic data revealed direct
`current potential shifts at the beginning of aura, also consis-
`tent with the initial activation (21). His laboratory’s fMRI
`data in precipitated left homonymous quandrantanopic aura
`demonstrated initial bilateral hyperoxia, right greater than
`left, and then spreading depression and oligemia involving
`
`2
`
`

`

`V O L. 7 / N O. 5 / S E P T E M B E R 2001
`
`T H E N E U R O L O G I S T
`
`281
`
`the gray matter during aura. The event began in the posterior
`cerebral artery distribution, spread to the middle cerebral
`artery distribution, and did not respect a single vascular
`distribution (23).
`Cao et al (24), in a follow-up study, studied normal
`patients and those with migraine with and without aura and
`stimulated them to precipitate occipital events. Here, too,
`there was suppression of visual activation.
`Moskowitz (25) presented a summary of his findings at
`the American Neurological Association meeting in October
`1999. He listed the characteristics of CSD and pointed out
`that each of the findings of visual aura in human subjects
`matched animal CSD. His impressive list included: initial
`hyperemia lasting about 3 minutes, subsequent hypoperfu-
`sion lasting about 2 hours, suppression of visual activity, a
`spread of the oligemia at a rate of about 3 mm/min, a
`termination at the parieto-occipital sulcus, and an observa-
`tion that the first areas activated in the aura as well as in CSD
`were the first areas that recovered. The evidence is therefore
`very strong that CSD, a primary neuronal process with
`secondary vascular manifestations, is the cause of visual aura.
`
`Corticalspreadingdepression,a
`primaryneuronalprocesswith
`secondaryvascularmanifestations,is
`thecauseofvisualaura.
`
`Woods et al (26) in 1994 published another important
`report bearing on CSD. In this case, a 45-year-old woman
`was in a positron emission tomography (PET) scanner for
`cognitive testing. She developed an evolutive prodrome that
`consisted of visual blurring and difficulty in fixation but not
`a true aura. PET scans were measured at 15-minute intervals.
`Her brain PET demonstrated profound spreading oligemia,
`beginning in the posterior cerebral artery distribution and
`moving anteriorly into the middle cerebral artery distribu-
`tion, not respecting a single vascular distribution. During this
`oligemia, she had no real neurologic symptoms, no ischemic
`symptoms, and it was 45 minutes from the onset of her mild
`visual blurring to the actual migraine headache.
`The Woods paper raises a number of fundamental ques-
`tions. Is migraine without aura just CSD with oligemia in the
`less eloquent or more silent areas of the brain? Is it that the
`neuronal depression is less severe and not appreciated by the
`patients as an aura? Do patients with recognizable anticipa-
`tory symptoms that are not auras, e.g., recognizable pro-
`dromes, really just have migraine with atypical, subtle aura?
`Or are there patients with prodrome who really have
`migraine with aura, and other patients with true migraine
`without aura or CSD before initiation of migraine pain?
`
`Cutrer’s laboratory has studied patients with migraine with-
`out aura without finding antecedent CSD, suggesting that
`there are patients with true migraine without aura (M. Cu-
`trer, personal communication, January 2000).
`It is worth reviewing the prevalence of aura in migraine.
`About 15% to 20% of people with migraine were found to
`have migraine with aura in the older literature (27). How-
`ever, since the advent of the International Headache Society
`criteria, two studies have found much higher prevalence—
`39% in Rasmussen and Olesen’s study (28) and 38% in
`Dahlof and Riman’s study (29).
`Queiroz et al summarized aura information obtained
`from 100 patients at The New England Center for Headache.
`He gave the questionnaire to 120 consecutive patients with
`visual aura and reported that only 19% had visual auras with
`every attack over a lifetime of migraine with aura. Thirty-
`nine percent had visual aura with their first migraine. In 43%
`aura occurred at least part of the time during the migraine
`pain itself, and in 20% the aura occurred or persisted after the
`migraine pain was over (30).
`
`POTENTIAL CAUSES FOR THE HYPEREXCITABILITY
`OF MIGRAINE BRAIN
`A variety of causes for hyperexcitability of the brain in
`migraine have been suggested. These include low magnesium
`levels, mitochondrial abnormalities, a dysfunction related to
`nitric oxide, or a channelopathy.
`
`Low Magnesium and Brain Hyperexcitability
`Magnesium binds to the glutamate NMDA receptor to
`maintain calcium homeostasis. If there is low magnesium, its
`gating function is not as proficient, and the patient might
`have a tendency to experience aura. Low magnesium can
`result in opening of calcium channels, increased intracellular
`calcium, and aspartate and glutamate release. In addition, the
`level of external potassium rises, and this may trigger CSD.
`Finally,
`low magnesium is also associated with increased
`platelet aggregation, increased serotonin, and, therefore, va-
`soconstriction.
`A number of studies have described low magnesium in
`migraine brain. Jain et al (31) reported low magnesium in
`serum and cerebrospinal fluid, Schoenen et al (32) in serum
`and erythrocytes, and Gallai et al (33,34) and Sarchielli et al
`(35) in mononuclear cells, serum, and saliva in both adult and
`juvenile populations of headache patients. Ramadan et al
`(36,37) have reported low magnesium in a magnetic reso-
`nance spectroscopy study and in patients with familial hemi-
`plegic migraine. In addition, Mauskop et al (38) have re-
`ported low serum ionized magnesium levels, but not total
`magnesium, in patients with migraine.
`Trials with magnesium supplementation for migraine
`prophylaxis,
`including those of Facchinetti et al (39) for
`menstrual migraine, Peikert et al (40), and Pffafenrath et al
`(41), have yielded mixed results. One of the most compelling
`
`3
`
`

`

`282
`
`T H E N E U R O L O G I S T
`
`P A T H O P H Y S I O L O G Y O F M I G R A I N E
`
`of the therapeutic trials was that of Mauskop et al (42) who
`showed in a blinded manner that intravenous magnesium
`aborted severe migraine in those patients with low ionized
`magnesium, but not in those patients with normal ionized
`magnesium.
`
`Mitochondrial Dysfunction
`Magnetic resonance spectroscopy studies have shown
`low phosphocreatine, high adenosine 59-diphosphate, and
`low phosphocreatine:Pi in migraine patients (10). Because
`flavinoids are the cofactor for the Krebs cycle, Schoenen et al
`(43) wondered whether this possible mitochondrial dysfunc-
`tion could be corrected by giving high-dose riboflavin to
`migraine patients. Initially in an open-label, and then in a
`small double-blind, placebo-controlled study, he found that
`after 3 to 4 months of daily use, 400 mg of riboflavin
`decreased migraine frequency, severity, and/or duration in
`about half of the users. If confirmed, this certainly raises the
`question of whether there is a subset of migraine patients
`with a mitochondrial energy metabolic defect that could be
`easily treated.
`
`Nitric Oxide Hypothesis
`Olesen and his collaborators have described that nitro-
`glycerin (NTG) can trigger migraine-like headaches, suggest-
`ing a specific role in primary headache pathogenesis. If non-
`migraine patients are given NTG, they develop immediate
`nonspecific headache. However, when NTG is given to
`migraine patients, they develop the early-onset nonspecific
`headache, and then a headache meeting International Head-
`ache Society criteria for migraine hours later. The nonspe-
`cific headaches may be due to cyclic GMP (cGMP) activa-
`tion, while the delayed migraine may be due to the NTG
`acting as a nitric oxide (NO) donor (44,45).
`There is pharmacologic but indirect clinical evidence for
`the NO hypothesis. Sumatriptan can terminate the NTG-
`induced migraine. Calcium channel blockers may work by
`reducing NO synthase activity and thus levels of NO.
`Serotonin2 (5-HT2) receptor activation stimulates NO
`release; 5-HT2 antagonists are used worldwide as migraine
`prophylactic agents,
`including methysergide, cyprohepta-
`dine, and pizotifen.
`The most interesting evidence is from a small study by
`Lassen et al (45) in Olesen’s clinic. They published a double-
`blind, placebo-controlled study on L-methylarginine HCL
`(546C88), a nonspecific NO synthase inhibitor in the treat-
`ment of acute migraine. They found that 10 of 15 patients
`had headache relief at 2 hours compared to 2 of 14 patients
`who received placebo.
`
`Channelopathy
`Familial hemiplegic migraine (FHM) is defined as an
`autosomal dominant disorder with recurrent episodes of mi-
`graine with prolonged aura manifested by severe unilateral
`
`weakness. On chromosomes 19 (46) and 1q (47,48), “mis-
`sense mutations in a neuronal specific P/Q calcium channel
`gene give rise to amino acid substitutions in the pore-forming
`and voltage sensor regions of the alpha 1a subunit of the
`channel. These mutations produce changes in channel ex-
`pression and kinetics that result in a net gain and net loss of
`function” (49,50). Most of these patients respond preven-
`tively to nonspecific calcium channel blockers.
`Upregulation or dysregulation of the calcium channels
`could result in disinhibited NO synthase activity, which in
`turn will liberate NO and induce hyperexcitability or frank
`aura, linking the calcium channelopathy and NO hypotheses.
`Alternatively, the abnormal calcium channel simply destabi-
`lizes the membranes resulting in abnormal ion fluxes, ele-
`vated external potassium concentrations, and initiation of
`CSD.
`As summarized by Hargreaves and Shepheard, “ . . . it is
`possible that the abnormal hyper-responsivity of the brain of
`migraineurs, like FHM, may be a consequence of genetic
`abnormalities in ion channels that regulate neuronal excit-
`ability. This has given rise to the description of migraine as a
`‘channelopathy’” (49,51).
`At the American Headache Society meeting in Montreal
`in June 2000, Professor Jean Schoenen presented evidence
`for more widespread subclinical neurologic dysfunction in
`migraine patients. These abnormalities included asymptom-
`atic cerebellar hypermetria and also increased jitter on single-
`fiber electromyogram. These subtle abnormalities, found by
`physiologic testing, which Professor Schoenen showed in
`patients with migraine without aura, with aura, and with
`prolonged aura, could be accounted for by calcium chan-
`nelopathy (52). The findings represented the first evidence
`for widespread calcium channelopathy in patients without
`FHM, but with more garden-variety migraine types. Also
`they might explain the responsiveness of non-FHM migraine
`patients to calcium channel blocker preventive medication.
`
`GENERATION OF THE MIGRAINE PAIN
`In 1995, Diener and colleagues (53) published a seminal
`report on PET scans of nine patients with acute right-sided
`migraine without aura. Nonspecific cortical activation was
`noted, including activation in the cingulate area, auditory
`association, and in the parieto-occipital cortex. These areas of
`activation may be related to the “emotional response” to
`somatic or visceral pain, possibly related to photophono-
`phobia.
`Most importantly, there was activation in the contralat-
`eral brain stem, including the dorsal raphe and locus ceruleus.
`This was the same area that triggered migraine-like headache
`when instrumented by neurosurgeons and reported by
`Raskin et al (11). Because the unilateral dorsal raphe area
`remained activated even after successful treatment of the pain
`with sumatriptan, Professor Diener speculated that this area
`might be the central generator for the migraine process, but
`not the cause of the pain.
`
`4
`
`

`

`V O L. 7 / N O. 5 / S E P T E M B E R 2001
`
`T H E N E U R O L O G I S T
`
`283
`
`There is not much sensory innervation for the brain itself,
`so explaining how head pain is generated requires some
`reach. Moskowitz has made the analogy to visceral pain,
`because migraine pain is poorly localized, often diffuse, and
`difficult to characterize in quality. He describes the meninges
`as the pain-sensitive capsule, with the brain as the visceral
`organ. Pain can be generated both intracranially from vessels
`and dura and extracranially from the skull and its coverings
`and investments, such as muscle, fascia, and skin.
`For 20 years, Moskowitz has explored the anatomy and
`pathophysiology of head pain and has been joined by labo-
`ratories worldwide. He began by restating the obvious: the
`key to the head pain is V1, the first (ophthalmic) division of
`the trigeminal nerve.
`The trigeminal nerve in this distribution arises from
`pseudounipolar neurons in the trigeminal ganglion, and it
`innervates ipsilateral meninges. It projects peripherally to
`intracranial extracerebral vessels in the meninges and centrally
`through the pons to the trigeminal nuclei in the brain stem
`(trigeminal nucleus caudalis) and upper cervical cord, the
`trigeminocervical complex.
`The trigeminal nerve fires upon repetitive cortical dis-
`charge, and this may offer a clue to the difficult question of
`how aura or prodrome links to migraine pain. Further, the
`trigeminal nerve can be sensitized, offering a possible expla-
`nation for the intensification of migraine pain during an
`attack, and for chronic pain, central sensitization, and allo-
`dynia in chronic headache.
`Tassorelli et al (54) noted that fos expression is seen, in a
`delayed fashion, after nitroglycerin administration in trigem-
`inal nucleus caudalis, locus ceruleus, and periaqueductal gray.
`Thus, the nitroglycerin would be the donor for NO, and the
`NO could cause central trigeminal activation.
`Trigeminal sensory C fibers contain substance P, calci-
`tonin gene-related peptide (CGRP), and vasoactive intestinal
`peptide. Antidromic stimulation of the trigeminal nerve re-
`sults in release of CGRP, substance P, and vasoactive intes-
`tinal peptide from sensory C fibers. This release causes dila-
`tion of meningeal vessels, fenestration of the vessels, and
`plasma extravasation, resulting in sterile, neurogenic inflam-
`mation (55). Goadsby and Edvinsson found an increase in
`CGRP in human jugular during migraine (but not of sub-
`stance P or vasoactive intestinal peptide), and a decrease of
`CGRP after treatment with sumatriptan (56). The work is
`important both because it suggests that CGRP may be the
`most important neuroinflammatory peptide in the migraine
`process and that CGRP does increase in humans, not just in
`experimental animals.
`All or some of this peripheral pathology (vasodilation and
`inflammation of the meninges) stimulates trigeminal sensory
`afferents, which, in turn, transduce this nociceptive informa-
`tion centrally. This arc is referred to as the trigeminovascular
`system.
`Serotonergic, noradrenergic, and dopaminergic pathways
`are involved in the processing of pain centrally. Ascending
`serotonergic pathways are associated with enkephalin in the
`
`second-order neurons, and project to midbrain raphe, thal-
`amus, hypothalamus, and cortex. These pathways are thus
`also intimately associated with pain control, autonomic and
`neuroendocrine function, and sleep.
`
`Calcitoningene-relatedpeptide
`(CGRP)maybethemostimportant
`neuroinflammatorypeptideinthe
`migraineprocess....
`
`Descending serotonergic pathways project to the periaq-
`ueductal gray, from there to the medullary raphe and the
`spinal cord dorsal horn (57,58). Noradrenergic pathways
`connect locus ceruleus to cortex and utilize gamma ami-
`nobutyric acid in the interneurons (59).
`In addition to the attention to serotonergic pathways,
`there has been renewed interest recently in dopaminergic
`pathways, which connect the substantia nigra and red nucleus
`(60). The red nucleus is activated in CSD and aura (23).
`It may be possible to put together the following steps in
`the generation of migraine pain.
`
`1. Activation of a central brain stem generator after
`CSD or without CSD. The generator may be the
`dorsal raphe contralateral to the pain, if the findings
`of Professors Diener and Raskin are accepted, or may
`be the trigeminal nucleus caudalis, activated by NO.
`2. Antidromic stimulation of the trigeminovascular sys-
`tem with resultant meningeal vasodilation, resulting
`in:
`a. Activation of perivascular trigeminal sensory affer-
`ent nerves and transmission of pain impulses cen-
`trally to caudal brain stem nuclei
`b. Release of vasoactive neuropeptides resulting in
`further vasodilation and neurogenic inflammation.
`3. There is then a relay of pain impulses from the
`activated peripheral sensory nerves to central second-
`order sensory neurons in the trigeminocervical com-
`plex. There is sensitization and intensification of
`headache pain as the attack progresses and increased
`sensitivity to convergent sensory stimuli from ex-
`tracranial tissues. This is a “windup phenomenon”
`(49).
`4. Trigeminal nuclei then relay incoming pain signals to
`higher cortical centers to register pain, photophobia,
`and phonophobia. There is activation of adjacent
`brain stem nuclei such as nucleus tractus solitariuus to
`initiating nausea, vomiting, and dysautonomia (61).
`
`5
`
`

`

`284
`
`T H E N E U R O L O G I S T
`
`P A T H O P H Y S I O L O G Y O F M I G R A I N E
`
`5-HT RECEPTORS
`Triptans and ergots are serotonin receptor agonists and
`activate, in particular, 5-HT1B and 5-HT1D receptors. The
`profile of receptor stimulation differs somewhat between
`triptans, which are more specific, and ergots, which activate
`a broad range of serotonin and nonserotonin receptors.
`
`Triptansandergotsareserotonin
`receptoragonistsandactivate,in
`particular,5-HT1B and5-HT1D
`receptors.
`
`Recent information has changed yet again and expanded
`our view of the key 5-HT1 receptors. Hargreaves and Shep-
`heard (49) have summarized these findings recently, added
`some provocative comments, and divided them up as follows:
`1. Vasoconstrictor 5-HT1B receptors occur on human
`meningeal blood vessel smooth muscle, and, in far
`fewer numbers, on human coronary arteries. Activa-
`tion produces vasoconstriction.
`2. Vasodilatory 5-HT1B and 5-HT7 receptors occur on
`human vascular endothelium also in meningeal and
`coronary arteries. Activation produces vasodilation.
`a. The presence of 5-HT1B vasodilating receptors
`raises the question of whether the initiation of
`migraine pain could come from activation of these
`receptors. Further, they could explain the para-
`doxical responses some patients complain of when
`they state that a triptan or ergot caused worsening
`of their headache.
`b. If stimulation of 5-HT7 receptors causes chronic
`vasodilation, could a 5-HT7 antagonist be devel-
`oped as a migraine preventive medication?
`c. Could preferential activation of the endothelial
`5-HT1B and 5-HT7 receptors be involved in mi-
`graine recurrence? That is, as the 5-HT1B/1D ag-
`onist plasma concentrations fall, the smooth mus-
`cle vasoconstriction is reduced, but the endothelial
`vasodilation is maintained, resulting in return of
`migraine pain and associated symptoms.
`d. There appear to be 5-HT1B receptors located
`centrally in human trigeminal ganglia, but the
`associated trigeminal nerves do not project to the
`periphery as do the trigeminal nerves served by
`the central 5-HT1D receptors. Thus, the central
`5-HT1B receptors are of unknown clinical signif-
`icance.
`3. Trigeminal inhibitory 5-HT1D receptors are located
`prejunctionally on trigeminal nerves projecting pe-
`
`ripherally to dural vasculature and modulating neu-
`rotransmitter release, inhibiting activated trigeminal
`nerves, and normalizing blood vessel caliber. They
`are also located centrally on trigeminal ganglia and
`interrupt nociceptive signals to brain stem second-
`order neurons initiated by peripheral blood vessel
`dilation, possibly neurogenic inflammation, and sen-
`sitized sensory nerve endings. The 5-HT1D receptors
`are also located on nucleus tractus solitariuus and may
`inhibit nausea, vomiting centrally (61).
`4. Trigeminal inhibitory 5-HT1F receptors are in loca-
`tions similar to the 5-HT1D receptors and may have
`similar peripheral action. However, the 5-HT1F re-
`ceptors have a wider distribution in brain, and their
`activity is not fully elucidated. Activation of these
`receptors, like the 5-HT1D receptors, is without vas-
`cular effect. Lilly developed and tested a 5-HT1F
`receptor agonist (LY334370), and it worked in abort-
`ing migraine, but the program was discontinued after
`toxicity in laboratory animals (49,62).
`
`AREAS FOR FURTHER RESEARCH
`The fundamental question is whether the hyperexcitabil-
`ity of the migraine brain, which must be regarded as the
`cardinal problem, can be corrected. Now that CSD appears
`the likely cause of migrainous aura, a variety of approaches
`can be taken, given the level of understanding of the process,
`to abort or prevent CSD genesis.
`
`Thefundamentalquestioniswhether
`thehyperexcitabilityofthemigraine
`brain,whichmustberegardedas
`thecardinalproblem,canbe
`corrected.
`
`The CSD may be initiated by NO, which may also
`activate the brain stem directly to generate the migraine pain.
`So one unifying therapeutic approach may be to inhibit
`synthesis of NO or block its effects.
`NO synthase (NOS) inhibitors and the NO pathways
`may be a key in both aborting and preventing migraine, the
`latter as a means to reducing the hyperexcitability. There
`remains debate as to which isoform of NOS is most impor-
`tant, probably inducible NOS (iNOS) or neuronal NOS
`(nNOS) but not endothelial NOS (eNOS).
`At a mitochondrial
`level, a study should be possible
`matching mitochondrial dysfunction by clinical
`laboratory
`
`6
`
`

`

`V O L. 7 / N O. 5 / S E P T E M B E R 2001
`
`T H E N E U R O L O G I S T
`
`285
`
`testing (e.g., phophorylation index) to flavinoid responsive-
`ness.
`Many pharmaceutical companies and clinics are collect-
`ing genetic material during clinical research. Further eluci-
`dation of channelopathies and their clinical treatment and
`correction may be helpful, and positional cloning with link-
`age and association has already linked FHM, CADASIL
`(cerebral autosomal dominant arteropathy with subcortical
`infarcts and leukoercephalopathy), and some of the episodic
`cerebellar ataxias.
`With respect to calcium channel blockers, the most ef-
`fective nonspecific calcium channel blocker for migraine
`prophylaxis is flunarazine, and usage is limited by depression
`and weight gain. Flunarazine is not available in the United
`States, but dotarazine is under study. More specific calcium
`channel blockers, especially P/Q blockers, may be even more
`effective and await development.
`A question remains as to whether vasoconstriction is
`necessary to abort acute migraine. A pure vasoconstrictor
`agonist, working on meningeal vessels, has not yet been
`tested. CGRP antagonists may prevent or reverse meningeal
`artery vasodilation. On the other hand, a number of seroto-
`nin-active medications that are not vasoactive have been
`studied, and, at the least, one 5-HT1F agonist was clinically
`useful in aborting migraines. Clearly, the study of medica-
`tions that are inactive on vessels will help further elucidate the
`mechanism of the generation of migraine pain and symptoms
`and may yield great clinical benefit.
`Pure inhibition of peripheral neurogenic inflammation
`has so far been relatively unsuccessful clinically, by way of
`5-HT1D agonists and SP/NK1 receptor antagonists, al-
`though the 1D agonist tried may have lacked sufficient
`efficacy at the human 1D receptor or may not have been
`given in the correct doses.
`Leukotrienes are released in neuro