Abstract
Objective
Migraine displays clinical heterogeneity of attack features and attack triggers. The question is whether this heterogeneity is explained by distinct intracellular signaling pathways leading to attacks with distinct clinical features. One well-known migraine-inducing pathway is mediated by cyclic adenosine monophosphate and another by cyclic guanosine monophosphate. Calcitonin gene-related peptide triggers migraine via the cyclic adenosine monophosphate pathway and sildenafil via the cyclic guanosine monophosphate pathway. To date, no studies have examined whether migraine induction mediated via the cyclic adenosine monophosphate and cyclic guanosine monophosphate pathways yields similar attacks within the same patients.
Methods
Patients were subjected to migraine induction on two separate days using calcitonin gene-related peptide (1.5 µg/min for 20 minutes) and sildenafil (100 mg) in a double-blind, randomized, double-dummy, cross-over design. Data on headache intensity, characteristics and accompanying symptoms were collected until 24 hours after drug administration.
Results
Thirty-four patients were enrolled and 27 completed both study days. Seventeen patients developed migraine after both study drugs (63%; 95% CI: 42–81). Eight patients developed migraine on one day only (seven after sildenafil and one after calcitonin gene-related peptide). Two patients did not develop migraine on either day. Headache laterality, nausea, photophobia and phonophobia were similar between drugs in 77%, 65%, 100%, and 94%, respectively, of the 17 patients who developed attacks on both days.
Conclusion
A majority of patients developed migraine after both calcitonin gene-related peptide and sildenafil. This supports the hypothesis that the cyclic adenosine monophosphate and cyclic guanosine monophosphate intracellular signaling pathways in migraine induction converge in a common cellular determinator, which ultimately triggers the same attacks.
Introduction
Migraine is a complex neurological disease with remarkable clinical heterogeneity of attacks and attack triggers (1–3). A range of pharmacological headache-inducing substances are established and commonly used in human migraine models for the study of migraine attacks and their underlying pathophysiological mechanisms (4). These triggers produce migraine attacks by stimulating different cellular signaling pathways (4).
Studies in migraine patients have revealed at least two signaling pathways causing migraine attacks. The first pathway is mediated via cyclic adenosine monophosphate (cAMP), which can be activated by calcitonin gene-related peptide (CGRP) infusion (5). The second pathway is mediated via the cyclic guanosine monophosphate (cGMP) pathway (6). It has been suggested that sildenafil triggers migraine attacks via inhibition of intracellular phosphodiesterase-5 which normally breaks down cGMP (6). This is in contrast to CGRP, which exerts its effects from the extracellular space, binding to its G-protein coupled receptor and resulting in upregulation of cAMP (5).
The question is whether clinical heterogeneity of migraine attacks may be explained by distinct signaling pathways. Another explanation is that a common, and more downstream, determinator of cAMP- and cGMP signaling pathways of migraine ultimately conjoins the two cascades and leads to initiation of the same attacks. In that case, CGRP and sildenafil would be capable of inducing migraine attacks with similar clinical features in the same patients. To date, no studies have explored this facet of the migraine pathophysiology using mediators of two different cellular signaling pathways of attacks in a head-to-head comparison. For this purpose, we administered CGRP and sildenafil in a paired, randomized, double-blinded, double-dummy, two-way cross-over study design.
Methods
Recruitment
We recruited patients with migraine without aura via a Danish recruitment website (www.forsogsperson.dk) and advertisements at hospitals and educational institutions. Patients were eligible for inclusion if they were male or female with a diagnosis of migraine without aura based on the IHS criteria (7), aged 18–50 years, weighed 50–100 kg and experienced migraine attacks at least once every other month. Exclusion criteria included inconsistent headache laterality, chronic migraine, any other primary headache disorder (apart from tension-type headache), daily medication intake (apart from oral contraceptives), being pregnant or breastfeeding females, daily smoking, a history of serious somatic or psychiatric disease, and hypo- or hypertension (systolic blood pressure >150 mmHg or <90 mmHg and/or diastolic blood pressure >90 mmHg or <50 mmHg). This study was part of a larger parent study with specific exclusion criteria including contraindications for MRI. Patients were enrolled from August 2017 to November 2018.
All patients underwent a medical examination and pregnancy tests were performed for female patients.
All patients provided written informed consent in agreement with the Declaration of Helsinki of 1964 with revisions until 2013.
Approval and data availability
The study was approved by the Ethical Committee of the Capital Region of Denmark (H-15019063) and registered at www.clinicaltrials.gov (NCT03143465). Data from the study can be made available upon reasonable request to the corresponding author. Other parts of the parent study have been and/or will be published elsewhere.
Design and data acquisition
Patients reported to the clinic headache free for at least 48 hours, having fasted for 4 hours; coffee, tea, alcohol, cocoa and tobacco were not allowed for 12 hours prior to study start. All patients were randomly allocated to receive a tablet of 100 mg sildenafil (TEVA pharmaceutical industries Ltd., Petah Tikva, Israel) or CGRP (Tocris Bioscience, Bristol, United Kingdom) as an IV infusion (1.5 µg/min for 20 minutes) on two separate study days. Randomization was balanced between drugs. Double blinding was obtained by administering both tablet and IV infusion on both days, with either tablet or IV formulation containing the active substance.
Patients were instructed about possible side effects of both CGRP and sildenafil on both study days. Before study start, the patients were informed about the possible outcomes of headache on both, one, or none of the study days, independent from each other.
Patients rested in the supine position until 90 minutes after drug administration and were monitored with a purpose-developed headache interview along with blood pressure, heart rate, blood oxygenation and respiratory frequency every 10 minutes. After the monitoring period, patients stayed in the clinic until 8 hours after study start and were discharged thereafter with a headache questionnaire (Figure 1). During this period, an MRI scan was performed from 6 hours to 7 hours (for another part of the parent study). Information obtained in the headache interview and questionnaire included intensity of headache pain using the numerical rating scale (NRS) from 0–10 (‘0’ denoting no pain and ‘10’ maximum imaginable pain) as well as headache characteristics and associated symptoms (pain location, aggravation by physical activity, nausea/vomiting, sensitivity to light and sound and mimicking usual migraine). The headache interview and questionnaire contained data on the same parameters and were merged upon study completion. Information on habitual migraine triggers was obtained from each patient before study start.
Overview of study design with depiction of headache evaluation intervals. Study design was double-blind, randomized, double-dummy, two-way cross-over. Drug: Calcitonin gene-related peptide and sildenafil.
Data analysis
Headache data were evaluated by SY and CEC, who were blinded to study drug randomization. Diagnosis of migraine-like attacks (hereinafter referred to as migraine attacks) was based on the following criteria (8):
Migraine attack fulfilling either (i) or (ii):
Headache fulfilling criteria C and D for migraine without aura according to the IHS criteria: C. Headache has at least two of the following characteristics: Unilateral location; pulsating quality; moderate or severe pain intensity (≥4 on NRS); aggravation by cough (in-hospital phase) or causing avoidance of routine physical activity (out-hospital phase); D. During headache at least one of the following: Nausea and/or vomiting; photophobia and phonophobia. Headache described as mimicking the patient’s usual migraine attack and treated with acute migraine medication.
Statistical analyses
Our main outcome parameter was incidence of migraine attacks, reporting the proportion of patients with attacks on both days. We calculated the binomial proportion confidence interval with comparisons to placebo responses in previous human migraine studies (9–11).
Pain intensity (NRS) over time was compared between the CGRP and sildenafil days using the Wilcoxon signed-rank test on the area under the curve using the trapezoidal rule. Area under the curve for pain intensity was calculated until 7 hours, as patients were allowed rescue medication from that time point. Time to onset of migraine and difference in peak NRS are reported as median with range and compared between the CGRP and sildenafil days using the Wilcoxon signed-rank test. In explorative analysis, binomial proportion confidence interval of patients who reported the same clinical characteristics on both days were calculated using sign test due to sample size. Mean arterial blood pressure is reported as mean with standard error from time 0 to 90 min. Sample size was determined based on other measurements, which were part of the larger parent study.
R (Version 3.5.2) was used to conduct the statistical analyses. P values are reported as two-tailed with a 5% level of significance.
Results
Migraine incidence and intensity
We recruited 34 patients, of whom 27 completed both study days (25 females; mean age 25 years, range 20 to 47) (Figure 2). Median time between the two study days was 14 days.
Flowchart of inclusion process and distribution of migraine attacks in each group separated into study drug. Reason for drop-out for those who developed an attack after the first study day: Withdrawal of consent (n = 1), study termination (n = 2) and exclusion due to daily medication after day 1 unrelated to study (n = 1). Reason for drop-out for those who did not develop an attack: Withdrawal of consent (n = 1) and lost to follow-up (n = 2).
Seventeen patients developed migraine attacks on both study days (63%; 95% CI: 42–81). Eight patients developed an attack on one day (30%) (seven after sildenafil and one after CGRP) (Figure 2) and finally, two patients did not develop an attack on either day (7%) (Figure 3). Twenty-four of 27 patients (89%) developed an attack after sildenafil and 18 (67%) after CGRP (Figure 3).
Number of patients who developed migraine attacks after calcitonin gene-related peptide (CGRP) and sildenafil. Overlap represents number of patients who developed attacks on both days. Twenty-five of 27 patients developed migraine attacks on either or both study days.
Area under the NRS curves for CGRP and sildenafil yielded no difference between drugs ( Individual red and blue lines represent headache pain intensity by time after calcitonin gene-related peptide (CGRP) and sildenafil for patients who completed both study days (n = 27). Solid horizontal black lines show median pain intensity. Dashed vertical lines with time and range annotation depict median time to onset of migraine.
Clinical characteristics and symptoms
The characteristics of headache after CGRP and sildenafil are presented in Table 1. Migraine symptoms were compared between drugs for the 17 patients who experienced an attack on both study days and comparability in attack characteristics is reported in Table 2 and depicted in Figure 5. Mean arterial pressure on each study day is visualized in Figure 6.
Number of patients who developed each migraine characteristic during attacks after calcitonin gene-related peptide (CGRP) and sildenafil. Overlaps represent patients who developed the same characteristic after both study drugs. Figure includes patients who experienced an attack on both days (n = 17). Relative change in mean arterial pressure (MAP) after calcitonin gene-related peptide (CGRP) and sildenafil from time of study drug administration (0 min). Values depicted as median ± standard error. Clinical characteristics of headache and associated symptoms after calcitonin gene-related peptide (CGRP) and sildenafil. The criteria for a migraine attack are described in ‘Methods'. Treatment efficacy: 50% decrease of headache intensity within 2 hours. Headache characteristics: Localization/intensity/quality/aggravation. Associated symptoms: Nausea/photophobia/phonophobia. Bilat: bilateral; Uni: unilateral; T: throbbing; P: pressing; NS: not specified (missing data). Spon: spontaneous migraine attack; M: male; F: female; Y: years of age. Diclodan: diclofenac 100 mg (suppository); I: ibuprofen 400 mg; Kodipar: codeine 30.6 mg + paracetamol 500 mg; P: paracetamol 1 g. S1: sumatriptan (tablet); S2: sumatriptan 12 mg (injection); AC: aspirin 500 mg + caffeine 50 mg; O: ondansetron 16 mg (suppository). Does not fulfill the associated symptoms criteria, but the attack is reported to mimic the patient’s usual attack. No data from > 24 hours. Went to sleep. No side preference. Shifted from left to right at 7 hours. Clinical features of migraine attacks after calcitonin gene-related peptide and sildenafil in patients who developed attacks on both days. Includes patients who experiences bilateral location, pressing quality or no nausea as same features on both days. CI: confidence interval.
Discussion
The major finding of the present study was that the majority of patients experienced a migraine attack after both CGRP and sildenafil with similar clinical features of attacks.
Incidence and clinical characteristics of migraine attacks
We found that 17 patients (63%; 95% CI: 42–80) developed migraine attacks after both CGRP and sildenafil. In comparison, the largest previously reported placebo rate for migraine induction was 16.7% (11). If CGRP and sildenafil attacks were mutually exclusive in patient subgroups; that is, if some patients were only responsive to CGRP and others only to sildenafil, we would expect the overlap to be similar to the previous placebo response. However, 16.7% is far from the boundaries of our confidence interval. These findings suggest that the two different signaling pathways, activated by CGRP and sildenafil respectively, both cause migraine attack induction in migraine patients in general, challenging the notion of distinct dominant signaling pathways between patients (4).
Migraine attacks are heterogenous across all clinical characteristics such as location of headache, severity and quality of pain (throbbing and/or pressing) and associated symptoms (photophobia, phonophobia and nausea) (1,12). A majority of patients commonly report that attacks can be triggered by at least one non-pharmacological trigger, such as stress and mental tension, menstruation, alcohol consumption and sleep disturbances (1,13). Diversity of the inter-individual features of migraine attacks and sensitivity to natural migraine triggers are reflected in our study sample as well (Table 1). We showed that characteristics of attacks were alike across the two study drugs (Figure 5). Most of the patients reported that the pharmacologically induced attacks mimicked their usual migraine attacks and were treated by their usual abortive medication (Table 1). Based on these observations, it is highly unlikely that variability of clinical heterogeneity of migraine attacks between patients is elucidated by different signaling pathways.
Cellular mechanisms of CGRP and sildenafil
The effects of infused CGRP are fostered via binding to its receptor, which actives the intracellular signaling pathways including formation of cAMP leading to activation of protein kinase A (5,14). Sildenafil, as a lipophilic drug, directly inhibits intracellular phosphodiesterase-5, and thus cancels the breakdown of cGMP, which in turn activates protein kinase G (6). Altogether, our findings of a large overlap of migraine induction after both CGRP and sildenafil as well as similarity of those attacks within patients support the existence of a more downstream common determinator where the two signaling pathways converge, ultimately leading to initiation of the same migraine attacks (15).
Preclinical studies have demonstrated that sildenafil is capable of inhibiting degradation of cAMP via increased cGMP levels (16). Furthermore, studies in animal coronary and pulmonary arterial tissues have shown that activation of cAMP can also promote the cGMP pathway (17,18). Collectively, this indicates interaction in both directions between the cAMP and cGMP pathways as an alternative explanation to the common outcome of migraine attack after CGRP and sildenafil. Thus, cAMP may drive the key signaling pathway of migraine attacks, noting that other well-established headache-inducers such as PACAP-38 and cilostazol carry their effects via the cAMP signaling pathway as well (19,20). Pharmacological headache-inducers are believed to act on the trigeminovascular system. CGRP, when infused, is unable to cross the blood brain barrier, and thus most likely activates peripheral CGRP receptors, which are available in both trigeminal neurons (e.g. ganglion, fibers) and cranial vascular cells (21,22). On the other hand, sildenafil freely crosses cellular membranes due to its lipophilic properties, and may therefore exert central effects, in addition to the well-known peripheral vascular effects (23,24).
Seven of the eight patients who developed an attack on one study day only reported the migraine attack after sildenafil administration, yielding migraine induction rates of 67% after CGRP and 89% after sildenafil. Mean induction rates across previous studies, in varying cohorts, amount to similar results with 64% for CGRP (8–10,25,26) across five studies and 83% for sildenafil (11). The relatively higher migraine attack induction after sildenafil is indicative of a more potent migraine-inducing ability. This might be attributed to its more downstream effects, thus being closer to the common determinator compared to CGRP in the migraine initiating cascade.
Due to the study design, there was no placebo control in addition to the active drugs, and the patients were not exposed to the same agent twice. However, we aimed to reduce expectation bias by informing the patients about possible outcomes, which would be independent of study day. As median time between the two study days was 14, and half-lives of CGRP and sildenafil were ∼7–10 min (27) and ∼4 h (28), the risk of carry-over effect was minimum. Furthermore, previous studies found similar induction rates both in placebo-controlled and open-label studies alike (8–11,25,26). The consistent reproducibility of CGRP and sildenafil migraine induction rates further suggests that false positive induction and day-to-day variation is unlikely to influence overlap. As CGRP induction rates are lower than for sildenafil, the overlap of patients developing attacks after both agents will be limited by CGRP as the lowest denominator.
Time to onset of migraine
We found that CGRP induced migraine faster than sildenafil (Figure 4). These differences in initiation of migraine might be explained by the fact that CGRP was administered intravenously, while sildenafil was administered as an oral formulation with tmax ∼ 1.5 hours (28). With this quicker access to target cells and subsequent faster activation of protein kinase, CGRP initiation of the migraine cascade might reach the proposed common determinator faster than sildenafil; however, the end result is the same.
The difference in “time-to-effect” is also demonstrated in studies showing immediate vascular responses to CGRP infusion (29), while response to oral sildenafil treatment occurs within 30 minutes of administration (30). The same effect is visualized in Figure 6, showing an immediate response after CGRP. The middle cerebral artery is dilated after CGRP during migraine attacks (26), while previous study showed no dilation of the middle cerebral artery after 100 mg sildenafil during attacks (31). However, this was assessed using transcranial Doppler (31) in contrast to MRI (26).
Conclusion
A majority of migraine patients developed attacks after both CGRP and sildenafil, with similar attack characteristics that mimicked their habitual attacks and were largely treatable by their usual abortive medication. A possible explanation of these findings is that migraine could be initiated by a common cellular determinator in which the cGMP and cAMP pathways converge. Future studies should focus on this commonality of migraine attack initiation as it could prove a prospective cellular target for new preventive therapeutics.
Clinical implications
Calcitonin gene-related peptide induces migraine via the cyclic adenosine monophosphate pathway, and sildenafil via the cyclic guanosine monophosphate pathway, as two distinct cellular signaling pathways mediating migraine mechanisms. Calcitonin gene-related peptide and sildenafil induced the same attacks within the majority of patients, despite attack initiation via distinct intracellular signaling pathways. Our findings suggest that modulation of migraine initiation via the two signaling pathways ultimately converge in a common cellular determinator.
Footnotes
Acknowledgements
The authors gratefully thank lab technicians Lene Elkjær and Winnie Grønning and research assistant Marius Lendal for assistance with data collection and handling. The authors also thank Drs Anne Luise Haulund Vollesen and Nita Wienholtz for their assistance in study drug randomization.
Author contributions
SY and CEC: Study concept and design, acquisition, analysis, and interpretation of data, and drafting of manuscript. NMT and TS: Acquisition and processing of data. FMA, AH and MA: Study concept and design, interpretation of data and overall supervision of study.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MA is a consultant or scientific advisor for Allergan, Amgen, Alder, Eli Lilly, Novartis and Teva, and principal investigator for Alder, Amgen, ElectroCore, Novartis and Teva trials. MA has no ownership interest and does not hold stock in any pharmaceutical company. MA serves as co-editor of the
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Lundbeck Foundation (R155-2014-171 and R249-2017-1608) and the Research Foundation of Rigshospitalet (E-23327-02). The funding parties had no influence on study design, patient inclusion or data interpretation.