Can nitric oxide induce migraine in normal individuals?

Abstract

Introduction

For many years, scientists have debated the possibility that an individual “migraine threshold” determines the likelihood with which individuals may express migraine attacks.

Discussion

Recent discoveries provided evidence for both genetic and environmental influences on individual migraine expression. The question is whether any person may express a migraine attack given a sufficiently strong stimulus or provocation. Here, we reviewed and discussed the ability of nitric oxide to induce migraine-like attacks in normal individuals.

Conclusion

Experimental data show that normal individuals may develop a migraine-like attack and that the human data point to different ways of further developing existing animal and human models.

Introduction

For decades it has been presumed that an individual “migraine threshold” determines the likelihood with which individuals may express migraine attacks. Arguments for such a threshold were already intellectually satisfying decades ago, but evidence was lacking. Recent hard-core science from genetics and environmental studies has, however, provided solid data supporting this view. Migraine with typical aura and migraine without aura have complex inheritance, and twin studies demonstrated a heritability ranging between 40% and 60% (1,2). Recent studies of the influence of environmental factors have identified a few such factors but the major part of environmental risk remains unexplained (3). Since the etiology of migraine is multifactorial, the next conclusion is that migraine is a quantitative trait, for which individual susceptibility varies within and between individuals.

These discoveries lead to a new important question: Can any person express a migraine attack given a sufficiently strong stimulus/provocation? The question is currently of immense importance because increasing efforts throughout the world focus on fast-track drug development programs, and these requires valid animal models. If migraine can be induced in normal individuals, then it is more likely that rats and other experimental animals can express changes typical of migraine.

Nitric oxide (NO) model of headache

Over the last 20 years we and others have used a human provocation model to study the effect of signaling molecules in migraine (4). In particular, the role of NO in primary headaches has been extensively investigated using intravenous glyceryl trinitrate (GTN) administration over 20 minutes or oral administration of the long-acting NO-releasing drug 5-isosorbide-mononitrate (5-ISMN). The headache response is recorded in the laboratory for two to three hours. After that, the experimental participants record headache characteristics and associated symptoms relating to migraine for the subsequent 12 hours at home using a questionnaire. Headache is scored on a 0–10 categorical scale and associated symptoms on a 0–3 categorical scale. The 0–10 scale is translated for classification purposes to a 0–3 scale: 1–3 mild headache, 4–7 moderate headache and 8–10 severe headache. The International Classification of Headache Disorders, second edition (ICHD-2) criteria for migraine without aura (5) cannot be applied to experimental studies. Firstly, the attacks are provoked and therefore in a formal sense, secondary to the provoking substance even if they are typical of migraine attacks. Secondly, only one attack is provoked and not five as required for a migraine diagnosis. Thirdly, in experimental provocation studies patients require treatment as soon as the headache becomes unpleasant. It usually takes time before a migraine attack manifests all the symptoms necessary to fulfill ICHD-2 diagnostic criteria for migraine. Therefore, treatment in experimental studies is often given before symptoms are fully developed.

We have proposed the following two criteria for an induced migraine-like attack 0 to 12 hours after infusion of an experimental drug:

An induced migraine-like attack must fulfill (1) and/or (2):

(1) headache fulfilling criteria C and D of the ICHD-2 for migraine without aura

(2) headache described by the patients as mimicking their usual migraine attacks and treated with a triptan.

In addition to patients fulfilling these relatively strict criteria for a migraine-like attack, it is also relevant to look at the so-called delayed headache. The majority of the substances discussed in the following induce only an immediate headache attack in normal patients but they induce a delayed headache in migraine sufferers defined as a headache occurring two hours or more after the provoking substance has been given. The picture is complex because some migraine patients develop a migraine-like attack even during the infusion or shortly after but the majority develop the attacks after the two-hour period. This also depends on the substance used because some substances manifest a very clear separation between immediate and delayed headache while other substances do not.

In addition to the scoring of subjective symptoms, a range of hemodynamic studies have usually also been conducted but these are not relevant for the present purpose and will not be discussed here. In previous provocation studies a thorough history of headache and migraine has been taken in all normal participants to rule out individuals with even very rare migraine attacks. This has, however, not generally been performed in first-degree relatives of experimental patients, which makes it impossible to relate finding to a hypothetical genetic load. Given that the NO model is the best-studied model to trigger migraine-like attacks, we review papers in which individuals received an infusion of GTN or the oral long-acting NO-releasing drug 5-ISMN, resulting in at least one participant developing a migraine-like attack. We have conducted many provocation studies using other triggers but they are not part of the present review.

Results

In Table 1 we present the previous provocation studies using NO donors in normal individuals conducted in our lab. The studies are not 100% comparable because the definition of a migraine-like attack has varied somewhat over the years. It appears from the table that provocation with a single substance generally has resulted in approximately 15% of normal experimental participants who have developed a migraine-like headache. As mentioned above, it remains unknown unfortunately whether these individuals were those with first-degree relatives affected by migraine. It also remains unknown whether these migraine-like headaches responded to a triptan since normal individuals have not received take-home medication after the experiment, unlike migraine patients. The conclusion from these studies is that it is possible using provocation with a single substance to induce a migraine-like headache in a minority of normal individuals.

Table 1.

Previous provocation studies using nitric oxide donors in normal individuals.

Substance (number of participants)	Formulation/dosage/time	Migraine-like attack	Rescue medication or intervention with antimigraine drug	Reported headache data	Reference
GTN or placebo (n = 6)	0.5 µg/kg/min over seven hours	Three	N/A	Headache developed in all patients during infusion, whereas no individual experienced headache during placebo. Median maximal headache intensity was medium (headache score 5) during first five hours.	Iversen et al. Eur J Clin Pharmacol 1993 (6)
GTN or placebo (n = 16)	0.5 µg/kg/min over 20 minutes	One	Three took rescue medication during delayed headache (two had family history)	During infusion: Nine out of 16 reported headache without associated symptoms (maximum 1.5. at 20 minutes). After discharge: Six out of 16 reported delayed headache. On placebo day: None of participants reported headache	Ashina et al. Brain 2000 (13)
GTN (n = 14)	0.5 µg/kg/min over 20 minutes a1 g of Az or placebo was injected intravenously over five minutes, and after 30 minutes an infusion of GTN	Three (one was adopted, and two with no family history)	Seven of 11 individuals with delayed headache (including the three with migraine-like attack) took rescue medication after Az, compared with two of eight with delayed headache after placebo	Period I (0–115 minutes): 13 of 14 experienced headache on both study days. After discharge: 11 of 14 reported delayed headache after GTN on Az day and eight out of 14 on placebo day	Daugaard et al. Cephalalgia 2009 (9)
5-ISMN or placebo (n = 10)	5-ISMN 15, 30, 60 mg and placebo dose were administered orally on four separate study days	Three (15 mg), two (30 mg) and three (60 mg)	Seven took rescue medication after 15 mg Six took rescue medication after 30 mg Five took rescue medication after 60 mg	All but one patient experienced headache during all active doses. The median time to onset of headache was 75 minutes (range 20–140 minutes) after 15 mg, 50 min (range 10–180 minutes) after 30 mg and 40 minutes (range 10–120 rain) after 60 mg	Iversen. Eur J Clin Pharmacol 1992 (9)
5-ISMN or placebo (n = 11)	Oral 30 mg three times daily for seven days	Five	Eight took rescue medication after initial 5-ISMN dose. During the following days, 10 individuals occasionally took analgesics. During placebo, only two patients took analgesics	5-ISMN induced more headache than placebo. Headache was experienced by eight of 11 individuals at four hours after the first dose of 5-ISMN (median 2, range 0–6), and after second dose all patients experienced headache, the intensity of which had increased to a median of 4 (range 2–7). Headache still present (median 3, range 0–8) in all but one individual on fourth day	Christiansen et al. Cephalalgia 2000 (10)

GTN: glyceryl trinitrate; 5-ISMN: isosorbide-5-mononitrate; Az: acetazolamide.

In the following we comment specifically on studies in which provocation has been performed with prolonged use of a migraine-inducing substance or with more than one substance. Iversen et al. (11) gave a seven-hour lasting intravenous infusion of GTN to six healthy volunteers in a placebo-controlled crossover experiment. The induced headaches increased after four hours, and two participants discontinued the experiment. Three of six individuals became nauseated and two vomited and withdrew. After five hours the mean intensity increased to severe. Thus, headache fulfilled the criteria for a migraine-like attack in three out of six individuals. In a further study, three doses of the long-acting NO-releasing drug 5-ISMN were compared to placebo in a four-way crossover experiment (12). Headache intensity gradually increased over the first four hours. Three out of 10 participants had a headache fulfilling diagnostic criteria for migraine on the low dose, two on the intermediate dose and three on the high dose (60 mg). Christiansen et al. (10,11) studied the development of tolerance to 5-ISMN by giving 30 mg three times daily for seven days to 11 healthy individuals in a double-blind, randomized, placebo-controlled crossover design. In 10 patients the headache fulfilled the pain subcriteria for migraine and in five individuals all diagnostic criteria for migraine without aura were fulfilled between four and 24 hours. This number gradually decreased as tolerance developed. Long-lasting influence of an NO donor may thus induce migraine-like headache in almost half of all patients. In another study pretreatment was given to 14 healthy volunteers with saline or acetazolamide followed on both days by infusion of GTN 0.5 mg/kg/min over 20 minutes (8,11). GTN combined with acetazolamide induced significantly more delayed headache than GTN alone. Furthermore, a migraine-like headache was induced in three volunteers by the combination and these three individuals did not develop migraine after GTN plus placebo. Although numbers are too small to conclude anything with confident evidence, the results indicated that pretreatment with acetazolamide can augment the effect of GTN and can result in the induction of migraine-like attacks in normal volunteers.

Discussion

This compilation of results from previous human migraine provocation studies strongly suggests that it is possible for non-migraine sufferers, i.e. normal individuals, to develop a migraine-like attack given a sufficiently strong provocation. This is a result that could be anticipated on the basis of current knowledge about migraine genetics. Migraine without aura has complex inheritance. Today 13 common genetic variants are known that increase the risk of migraine, usually around a 10% increase in risk (12,13). It is expected that there may be 50 or even 100 more risk variants. Similarly, protecting variants are likely to occur in large numbers but so far none have been identified. The sum of susceptibility and protective genes will determine the migraine threshold. It can be higher or lower not only in patients manifesting migraine but also in normal individuals. In other words, chances of inducing a migraine attack are likely to vary markedly from one normal individual to another. In our studies approximately 15% of normal individuals challenged with a single infusion or a single tablet developed a migraine-like attack.

A dose-response relationship exists between NO donors and the likelihood of normal individuals to develop a migraine-like attack. Even more obvious is the importance of the duration of the provocation. In the strongest and most long-lasting challenge procedure using ISMN-5 30 mg three times daily for a week, almost half of normal individuals developed a migraine-like attack between four and 24 hours. The number then started to decline as tolerance developed. It is notable that there were no migraine-like headaches in the first four hours. This corresponds to the delayed onset of migraine after infusion of GTN.

Another way of increasing the likelihood of migraine provocation seems to be a combination of different provocative agents. In the experiment in which normal participants were pretreated with placebo or acetazolamide followed by a GTN infusion, three individuals out of 14 after acetazolamide developed migraine but none after placebo. These are small numbers but the results indicate a possible future way of increasing migraine responses by the combined use of offending agents.

The present paper gives some clues as to the future development of animal models of migraine. Lesson number one is that individual strains of animals are highly likely to be more responsive than other strains or other species. A research program should examine the sensitivity of different experimental animal species and different strains of rat and mice to GTN. The next clue is that long-lasting nitro preparations should be studied in animals and may prove to give more robust migraine-like responses than a short testing infusion. Thirdly, it would be valuable to study the effect of various sensitizing procedures in animals. A model has already been developed using repeated dosing of sumatriptan to sensitize rats, apparently with positive results (14). Other ways of sensitizing could be stressing the animals or studying specific phases of the menstrual cycle. Finally, combining two or more migraine-provoking agents may have a greater effect than either agent may.

In human experimentation, it would be interesting to compare normal individuals with a first-degree relative affected by migraine to participants with no family history of migraine. For comparison to the animal model using sumatriptan to sensitize, it would be interesting to study normal individuals but this should be very difficult. Finally, a combination of provoking agents should be studied.

In summary, existing migraine provocation studies do indicate that a minority of normal individuals can be provoked to develop a migraine-like attack. This supports the validity of animal experimental models. The human data also point to different ways of further developing existing animal and human models.

Clinical implications

Can any person express a migraine attack given a sufficiently strong stimulus/provocation?

15% of healthy volunteers report migraine-like attacks after administration of nitric oxide donors.

There is a dose-response relationship that exists between nitric oxide donors and the likelihood of normal individuals to develop a migraine-like attack.

The likelihood of migraine provocation seems to be increased if different provocative agents are combined.

Optimization of human and animal models of migraine may provide new opportunities in fast-track drug development programs.