OUP user menu

PACAP38 induces migraine-like attacks in patients with migraine without aura

Henrik Winther Schytz, Steffen Birk, Troels Wienecke, Christina Kruuse, Jes Olesen, Messoud Ashina
DOI: http://dx.doi.org/10.1093/brain/awn307 16-25 First published online: 3 December 2008

Summary

Experimental studies have shown that infusion of vasoactive neurotransmitters may trigger headache or migraine-like attacks in man. Pituitary adenylate cyclase activating peptide-38 (PACAP38) is a strong vasodilator found in trigeminal sensory and parasympathetic perivascular nerve fibers. We therefore hypothesized that infusion of PACAP38 would cause headache in healthy subjects and migraine-like attacks in migraine patients. Twelve healthy subjects and 12 migraine patients were examined in two separate studies. All subjects were allocated to receive 10 pmol/kg/min PACAP38 and placebo in a randomized, double-blind crossover study design. Headache was scored on a verbal rating scale (VRS) during hospital (0–2 h) and post-hospital (2–12 h) phases. Mean blood flow velocity in the middle cerebral artery (VMCA) by transcranial Doppler (TCD) and diameter of the superficial temporal artery (STA) by high resolution ultrasonography were recorded during hospital phase in migraineurs. PACAP38 infusion caused headache in all healthy subjects and 11 out of 12 migraine patients. Seven migraine patients experienced migraine-like attacks after PACAP38 and none after placebo (P = 0.016). Most of attacks (6 out of 7) occurred during the post-hospital phase [mean time 6 h (range 2–11)]. Two healthy subjects reported migraine-like attacks after PACAP38 during the hospital phase and none during the post-hospital phase. In the hospital phase, the area under the curve (AUC) for headache score was larger during PACAP38 infusion compared to placebo in healthy subjects (P = 0.005) and tended to be larger in migraineurs (P = 0.066). In the post-hospital phase, the AUC for headache was larger after PACAP38 infusion compared to placebo in both healthy subjects (P = 0.005) and migraine patients (P = 0.013). In migraine patients, PACAP38 caused a peak decrease of 16.1% in VMCA and a 37.5% increase in STA diameter at 20 min after start of infusion. In conclusion, PACAP38 infusion caused headache and vasodilatation in both healthy subjects and migraine patients. In migraine sufferers, PACAP38 caused delayed migraine-like attacks. The findings stimulate further investigation of the neuronal and vascular mechanisms of PACAP38.

  • PACAP38
  • migraine without aura
  • transcranial Doppler
  • cerebral vessels
  • parasympathetic nervous system

Introduction

It is still a matter of intense debate if vasodilatation contributes to migraine pain per se or is just an epiphenomenon during migraine attacks (May and Goadsby, 1999; Welch, 2003; Olesen and Goadsby, 2006; Schoonman et al., 2008). In the last 15 years we have systematically investigated both the migraine eliciting and hemodynamic effects of different vasoactive neurotransmitters found in perivascular nerve fibers (Thomsen, 1994; Lassen et al., 2002; Rahmann et al., 2008). The studies have shown, with one exemption of sildenafil (Kruuse et al., 2003), that all known inducers of migraine-like attacks dilate cephalic vessels (Thomsen, 1994; Lassen et al., 1995, 2002; Kruuse et al., 2006). However, we also observed that vasodilatation is not a sufficient factor for induction of migraine-like attacks (Rahmann et al., 2008). Given that cephalic vessels might also be dilated during a migraine attack (Iversen et al., 1990; Friberg et al., 1991), simultaneous recording of headache and vasodilatation in migraine patients after infusion of neuropeptides are of key importance.

Cephalic vessels are innervated by sensory, parasympathetic and sympathetic nerve fibers (Edvinsson et al., 2001; Jansen-Olesen et al., 2004; Edvinsson and Uddman, 2005). We have recently shown that infusion of the parasympathetic neuropeptide vasoactive intestinal polypeptide (VIP) induces a marked dilatation of cephalic arteries (Hansen et al., 2006; Rahmann et al., 2008). However, VIP induces only mild headache and no migraine-like attacks (Rahmann et al., 2008). Pituitary adenylate cyclase activating peptide-38 (PACAP38) is a 38 amino-acid neuropeptide (Miyata et al., 1989) belonging to the same secretin/glucagon superfamily as VIP. Both VIP and PACAP activate the VPAC1 (Hosoya et al., 1993) and VPAC2 (Lutz et al., 1993) receptors, but interestingly a third receptor, PAC1, is selectively activated by PACAP. Infusion of PACAP38, which is the most prominent form (Arimura and Shioda, 1995), induced vasodilatation in healthy subjects of a similar magnitude but longer lasting than VIP (Birk et al., 2007). The headache characteristics after PACAP38 have not been previously described in healthy subjects or migraine patients. Given that PACAP is found in perivascular nerve fibers implicated in migraine pathogenesis, we hypothesized that PACAP38 infusion would induce headache in healthy subjects and migraine-like attacks in migraine patients. In addition, we aimed to describe the dilatation of cephalic vessels in migrainers after PACAP38 infusion. We therefore performed two double-blind placebo controlled crossover studies to record possible headache and migraine eliciting effects of PACAP38 in healthy subjects and migraine patients without aura.

Methods

The study consisted of two separate studies. In the first study we recruited 12 healthy subjects. The systemic and cerebral hemodynamic effects of PACAP38 in the healthy subjects have been previously published (Birk et al., 2007). None of the healthy subjects had been in similar studies in the past. Exclusion criteria were: a history of migraine or any other type of headache (except episodic tension type headache <5 days per month); any daily medication apart from oral contraceptives; serious somatic or psychiatric disease.

In the second study we recruited 12 otherwise healthy patients diagnosed with migraine without aura (MO) according to the International Headache Society (IHS). One of the migraineurs had previously been participating in a similar study with infusion of VIP (Rahmann et al., 2008). Exclusion criteria were as in the healthy study with the exception of the migraine diagnosis.

The Ethics Committee of the County of Copenhagen approved both studies (healthy subjects: KA02139 and migraine patients: KA20060087). In addition, the migraine study was approved by the Danish Medicines Agency (Eudract nr: 2006-003774-94), which was monitored by Copenhagen University Hospital GCP-unit and registered at Clinicaltrials.gov (ID: NCT00380263). All subjects gave informed consent to participate, and both studies were undertaken in accordance with the Helsinki Declaration of 1964, as revised in Edinburgh in 2000.

Experimental design

In a double-blind, placebo-controlled, crossover design, the subjects were in a balanced order randomly allocated to receive 10 pmol/kg/min PACAP38 (Calbiochem®, Darmstadt, Germany) or placebo (isotonic saline) over 20 min on 2 days separated by a least 1 week. Before the experiment each subject underwent a general physical examination. All subjects reported to the laboratory 08.30 h headache free. The experiment was postponed, if the subject had a migraine attack within 5 days or tension type headache 48 h before the start of the study. The intake of coffee, tea, cocoa or other methylxanthine-containing foods or beverages was not allowed for the last 8 h prior the study. All procedures were performed in a quiet room at a temperature of 25°C. The subjects were placed in the supine position and a venous catheter was inserted into the right antecubital vein for drug infusion. The subject then rested for at least 30 min before baseline measurements of blood pressure, heart rate (HR) and ECG were performed, and the infusion started using a time and volume controlled infusion pump. Headache intensity, VMCA, STA diameter, end-tidal partial pressure of CO2 (PetCO2), adverse events and vital signs were recorded before and then every 10 min until 90 min after the beginning of infusion. The subjects were discharged from the hospital after finishing the measurements and asked to complete a headache diary every hour until 12 h after start of infusion. The diary included headache characteristics and accompanying symptoms, any rescue medication and adverse events. Migraine patients also reported if headache was believed to mimic their usual migraine attack. Subjects were allowed to treat headache with over the counter rescue medication and their usual migraine treatment.

Headache and migraine-like attack criteria

Headache intensity was recorded repeatedly on a verbal rating scale (VRS) from 0 to 10 [0, no headache; 1, a very mild headache (including a feeling of pressing or throbbing—pre-pain); 10, worst imaginable headache] (Iversen et al., 1989). Headache characteristics and associated symptoms were also recorded to determine the quality and type of the headache.

Headache induced experimentally by infusion of a neuropeptide can not fulfill strict IHS criteria for migraine without aura (IHS, 2004). First, the migraine-like attacks reported are induced by pharmacological substances and can therefore not be spontaneous, though they often phenotypically mimic spontaneous migraine attack in the majority of patients (Thomsen, 1994; Lassen et al., 2002). Secondly, most spontaneous migraine attacks develop in a matter of hours and often go through a phase where they phenomenologically only fulfill the criteria for tension-type headache before the headache gets worse, becomes unilateral and has the associated symptoms required for migraine. Thirdly, patients in experimental provocation studies cannot be denied treatment of the induced attacks and often treat before all migraine criteria are fulfilled.

Based on these circumstances we have used the following two criteria for a migraine-like attack induced 0–12 h after infusion of an experimental drug:

Migraine-like attack attacks fulfilling either (1) or (2):

  1. Headache fulfilling criteria C (Moderate to severe pain intensity is considered ≥4 on VRS.) and D for migraine without aura (IHS, 2004).

  2. Headache described as mimicking usual migraine attack and treated with a triptan.

Middle cerebral artery blood flow velocity

VMCA was recorded bilaterally with transcranial Doppler (TCD) with hand-held 2 MHz probes (Multidop X; DWL, Sippelingen, Germany). Fixed probes were avoided, because they may cause discomfort and even headache (Thomsen, 1995). Four recordings were taken and averaged at each time point. One recording is a time-averaged mean over 4 s or approximately four cardiac cycles. Identification of the MCA were done as previously described (Thomsen and Iversen, 1993). Every TCD recording was performed by the same trained physician (HWS). End-tidal CO2 (PetCO2) was recorded simultaneously with TCD recordings using an open mask that caused no respiratory resistance (ProPaq Encore®; Welch Allyn Protocol, Beaverton, OR, USA). PACAP38 is known to increase CBF due to a decrease in PetCO2 caused by increased ventilation (Birk et al., 2007) and according to Dahl et al. (1989), the changes in diameter (Δd) of MCA can be calculated if CBF is unchanged. Thus, after correcting VMCA with e0.034 for each mmHg change in PetCO2 (Markwalder et al., 1984) the change in MCA diameter can be estimated as: Embedded Image

Diameter of the superficial temporal artery

Diameter of the frontal branch of the superficial temporal artery (STA) was measured by a high resolution ultrasonography unit (Dermascan C; Cortex Technology, Hadsund, Denmark: 20 MHz, bandwidth 15 MHz) as previously described (Iversen et al., 1990; Kruuse et al., 2003).

Vital signs

HR and blood pressure were measured every 10 min using an auto-inflatable cuff (ProPac Encore®; Welch Allyn Protocol). ECG (Cardiofax V; Nihon-Koden, Shinju-ku, Tokyo, Japan) was monitored on a LCD screen and recorded on paper every 10 min.

Data analysis and statistics

All values are presented as mean values ± SD, except headache scores which are presented as median values. Peak mean vascular variables are presented with 95% confidence intervals. Baseline was defined as T0 before the start of infusion of each dose. We calculated median peak headache score and median time to peak headache after PACAP38 and placebo.

The mean plasma half life of PACAP38 has been shown to be 3.5 min (Birk et al., 2007). We therefore defined an immediate phase as 0–30 min, a post-infusion phase from 30 to 90 min after start of infusion and a post-hospital phase 2–12 h after start of infusion. The two studies were analyzed separately and no direct statistical analysis was applied for comparing the healthy subjects and migraine patients.

In the healthy subject study the primary end-points were: the difference in incidence of headache or migraine-like attacks between PACAP38 and placebo; the difference in area under the curve (AUC) for headache score during infusion (0–30 min), post-infusion (30–90 min) and post-hospital phase (2–12 h) between PACAP38 and placebo.

In the migraine patient study the primary end-points were: the difference in incidence of headache or migraine-like attacks between PACAP38 and placebo; the difference in AUC for headache score during infusion (0–30 min), post-infusion (30–90 min) and post-hospital phase (2–12 h) between PACAP38 and placebo. The secondary end-points were difference in AUC for VMCA, STA diameter, HR, MAP, systolic blood pressure, diastolic blood pressure and PetCO2 between PACAP38 and placebo day.

Incidence of headache, migraine-like attacks and adverse events were analyzed as binary categorical data with McNemar test. We calculated AUC according to the trapezium rule (Matthews et al., 1990) to obtain a summary measure and to analyze the differences in response between PACAP38 and placebo. Baseline was subtracted before calculating AUC to reduce variation between sessions within subject. Analysis of AUC values were performed with a paired two-way t-test, except headache scores where data were tested with Wilcoxon signed rank test. We tested for period and carry-over effects for all baseline variables with Mann-Whitney test and independent t-test.

Furthermore, to explore possible changes in MAP, systolic and diastolic blood pressure, we analyzed for changes over time for each treatment day separately with univariate analysis of variance (ANOVA) with the fixed factors volunteer and time. To reduce mass significance the following time points were selected for analysis (T20, T30, T60 and T90). If overall differences were found, Dunnett's test was applied to characterize which time points were different from baseline.

All analysis was performed with SPSS for windows 15.0 (Chicago, IL, USA). Five percent (P < 0.05) was accepted as the level of significance.

Results

Twelve healthy subjects (7F/5M, mean age 25, range 20–31) and 12 migraine patients (11F/1M, mean age 31, range 20–47 years) completed the study on both study days. The migraine patients had an attack frequency ranging from 1 to 4 attacks per month. There were missing data on VMCA, MAP, HR and PetCO2 values on the placebo day of one migraine patients due to a technical error.

For the vascular variables investigated in migraine patients we found no differences at baseline (Table 1), except for MAP, which, for unknown reasons, was higher on the placebo day (P = 0.031) due to a higher diastolic blood pressure (P = 0.042). There was no carry-over or period effect for baseline values of MCA, STA, MAP, HR or PetCO2.

View this table:
Table 1

Baseline values (± SD) in 12 migraine patients subjects on the two trial days: mean velocity of flow in the middle cerebral artery (VMCA), end-tidal partial pressure of CO2 (PetCO2) measured during transcranial doppler measurements, mean arterial blood pressure (MAP), heart rate (HR), diameter of the superficial temporal artery (STA)

PACAP38PlaceboP-values
VMCA74.2 ± 11.873.9 ± 10.70.923
PetCO2 (mmHg)35.2 ± 3.734.9 ± 3.50.741
MAP (mmHg)77.6 ± 10.382.0 ± 7.90.031
Heart rate63.7 ± 7.965.3 ± 5.70.305
STA1.15 ± 0.191.14 ± 0.230.906
  • P-value: paired t-test. VMCA = Mean velocity middle cerebral artery; PetCO2 = End-tidal partial pressure of CO2; STA = Superficial temporal artery.

Headache in healthy

All healthy subjects experienced headache (0–12 h) after PACAP38 infusion compared to two after placebo (McNemar test, P = 0.002) (Fig. 1). Two subjects reported a migraine-like attack occurring 10 and 50 min after start of PACAP38 infusion compared to zero subjects after placebo (McNemar test, P = 0.500) (Table 2). The median peak headache score was 3.5 (range 2–4) after PACAP38 and 0 (range 0–1) after placebo. The median time to peak headache was 5.5 h after PACAP38. Four subjects took paracetamol and three of these responded well to the medication.

Fig. 1

Individual and median headache scores on a VRS from time 0–90 min and 2–12 h after PACAP38 infusion in healthy subjects and migraine patients. Thick lines are median headache scores. In healthy subjects there were higher headache responses after PACAP38 infusion compared to placebo during the infusion (P = 0.005), post-infusion period (P = 0.008) and post-hospital phase (P = 0.005). In migraine patients there were higher headache responses after PACAP38 infusion compared to placebo during the post-infusion period (P = 0.009) and post-hospital phase (P = 0.013). Seven migraine patients reported migraine-like attacks after start of PACAP38 infusion compared to zero after placebo (McNemar test, P = 0.016).

View this table:
Table 2

Headache characteristics after PACAP38 in healthy subjects

SubjectPeak headache (duration)Headache characteristicsAssociated symptomsMigraine- like attack (onset)Treatment (time)/efficacy
1a10 min (10–60 min)Bilat/1/throb/yes+/−/−Yes (30 min)
b6 h (5–12 h)Bilat/3/pres/yes+/+/−No
2a10 min (10–50 min)Bilat/1/pres/yes−/−/−No
b5 h (4–6 h)Bilat/2/throb/yes−/−/−NoParacetamol (5 h)/yes
3a20 min (20–50 min)Bilat/1/pres/yes−/−/−No
b5 h (4–9 h)Bilat/3/pres/yes−/−/−No
4a60 min (10 min to 12 h)Bilat/3/pres/no−/−/−No
b7 h (10 min to 12 h)Bilat/4/pres/no−/−/+NoParacetamol (8 h)/yes
5a10 min (10 min to 8 h)Bilat/1/pres/no−/−/−No
b3 h (10 min to 8 h)Bilat/4/pres/no−/−/−NoParacetamol (4 h)/yes
6a30 min (20 min to 11 h)Bilat/2/pres/no−/−/−No
b2 h (20 min to 11 h)Bilat/1/pres/yes+/−/−NoParacetamol (2 h)/no
7a80 min (10 min to 12 h)Bilat/4/throb/yes−/−/−No
b2 h (10 min to 12 h)Bilat/2/pres/yes−/−/−No
8aNone
b6 h (5–10 h)Bilat/4/pres/no+/−/−No
9aNone
b6 h (4–12 h)Bilat/3/throb/no−/−/−No
10a10 min (10–30 min)Bilat/1/pres/no−/−/−No
b9 h (3–5 h)Bilat/4/throb/yes−/−/−No
11a10 min (10 min to 12 h)Bilat/1/throb/no−/−/−No
b7 h (10 min to 12 h)Bilat/2/throb/no−/−/−No
12a40 min (10–60 min)Bilat/4/throb/yes+/−/−Yes (30 min)
b4 h (3–12 h)Bilat/3/pres/no−/−/−No
  • a = PACAP38 day 0–90 min; b = PACAP38 day 2–12 h; c = placebo day 0–90 min; d = placebo day 2–12 h. Headache characteristics: localisation/intensity/quality (throb = throbbing; pres = pressing)/aggravation by movement. Associated symptoms: nausea/photophobia/phonophobia.

During the hospital phase the AUC was larger after PACAP38 compared to placebo both during the immediate (AUC0–30 min, P = 0.005) and the post-infusion periods (AUC30–90 min, P = 0.008). During the post-hospital phase we found that AUC was larger after PACAP38 compared to placebo (AUC2–12 h, P = 0.005).

Headache in migraine patients

In total 11 migraine patients experienced headache (0–12 h) after PACAP38 infusion compared to three after placebo (McNemar test, P = 0.021) (Fig. 1). Seven migraine patients reported migraine-like attacks after start of PACAP38 infusion compared to zero after placebo (McNemar test, P = 0.016) (Table 3). One of the migraine-like attacks occurred during the hospital phase 50 min after start of PACAP38 infusion, whereas six attacks occurred during the post-hospital phase [mean 6 h (range 2–11 h)]. In two out of seven migraine-like attacks the subjects took a triptan and described headache as mimicking a usual migraine attack before all headache characteristics for a usual migraine attack were fulfilled. The median peak headache score was 2.5 (range 0–10) after PACAP38 and 0 (range 0–4) after placebo. The median time to peak headache occurred at a median of 4.0 h (range 0–12).

View this table:
Table 3

Headache characteristics after PACAP38 in migraine patients

SubjectPeak headache (duration)Headache characteristicsAssociated symptomsMimics migraine attackMigraine-like attack (onset)Treatment (time)/efficacy
1a10 min (10–50 min)Bilat/2/throb/yes+/+/−YesYes (20 min)
b3 h (3–12 h)Bilat/2/pres/yes+/−/−NoNoSumatriptan (10 h)/yes
2a50 min (50 min to 12 h)Left/1/pres/no−/−/−NoNo
b12 h (50 min to 12 h)Left/3/pres/no−/−/−YesYes (11 h)Sumatriptan (12 h)/yes
3a80 min (80 min to 10 h)Bilat/5/pres/no−/−/−NoNo
b7 h (80 min to 10 h)Bilat/7/pres/yes+/−/−YesYes (2 h)Sumatriptan (7 h)/yes
4a70 min (70 min to 12 h)Left/1/pres/no−/−/−NoNo
b6 h (70 min to 12 h)Left/5/pres/yes+/+/−YesYes (6 h)Sumatriptan (6 h)/yes
5a50 min (50–70 min)Bilat/1/pres/yes−/−/−NoNo
b4 h (90 min to 12 h)Bilat/10/pres/yes+/+/−YesYes (3 h)Sumatriptan (4 h)/yes
7a50 min (50–60 min)Bilat/1/throb/yes−/−/−NoNo
b9 h (6–12 h)Bilat/2/pres/yes−/−/−NoNo
8a10 min (10–30 min)Bilat/1/throb/no−/−/−NoNo
bNone
9a40 min (40 min to 3 h)Left/1/throb/no−/−/−NoNo
b2 h (40 min to 3 h)Left/1/throb/no−/−/−NoNo
10a70 min (20 min to 7 h)Bilat/2/pres/yes−/+/−NoNo
b4 h (20 min to 7 h)Bilat/3/throb/yes−/+/−NoNoNSAID (4 h)/yes
11a60 min (20 min to 12 h)Bilat/2/pres/yes−/+/−NoNo
b2 h (20 min to 12 h)Bilat/1/pres/no−/−/−YesYes (11 h)Sumatriptan (12 h)/yes
12a40 min (40–50 min)Bilat/1/pres/no−/−/−NoNo
b5 h (4–12 h)Bilat/4/throb/yes+/+/+YesYes (4 h)
  • a = PACAP38 day 0–90 min; b = PACAP38 day 2–12 h; c = placebo day 0–90 min; d = placebo day 2–12 h. Headache characteristics: localisation/intensity/quality (throb = throbbing; pres = pressing)/aggravation by movement. Associated symptoms: nausea/photophobia/phonophobia. Migraine-like attack according to criteria described in method section.

During the hospital phase AUC was larger after PACAP38 compared to placebo during the post-infusion phase (AUC30–90 min, P = 0.009), though it only tended to increase during the immediate phase (AUC0–30 min, P = 0.066). In the post-hospital phase AUC was larger after PACAP38 compared to placebo (AUC0–12 h, P = 0.013).

Middle cerebral artery

During the hospital phase VMCA corrected for PetCO2 decreased after PACAP38 compared to placebo both during the immediate (AUC0–30 min, P = 0.001) and the post-infusion phase (AUC30–90 min, P = 0.007) (Figs 2 and 3).

Fig. 2

Individual and mean flow velocities in the middle cerebral arteries (VMCA) assessed by transcranial doppler ultrasonography. Thick lines show mean values. Infusion of PACAP38 (filled square) in migraine patients resulted in a decrease in VMCA during the infusion (P = 0.001) and post-infusion (P = 0.005) period compared with placebo (open square).

Fig. 3

Hemodynamic and respiratory changes from baseline in migraine patients after PACAP38 in comparison with median headache score on a VRS.

PACAP38 decreased PetCO2 compared to placebo during the immediate (AUC0–30 min, P = 0.001) and post-infusion phases (AUC30–90 min, P = 0.005). After correcting VMCA for PetCO2 changes, the diameter of MCA was calculated to dilate 9.5% (95% CI: 6.3–12.7) compared to baseline at T20 after onset of PACAP38 infusion. The peaks of all vascular data after PACAP38 are shown in Table 4.

View this table:
Table 4

Mean peak hemodynamic variables 0–90 min after PACAP38 infusion in healthy subjects and migraine patients

Peak time after onset of PACAP38 (min)Mean change from baseline after PACAP38 (95% CI)
VariableHealthy subjectsMigraine patientsHealthy subjectsMigraine patients
MCA diameter20208.5% (6.2–10.8)9.5% (6.3–12.7)
Velocity in MCA1020−12.8% (−21.2 to −4.4)−16.1% (−18.2 to −10.6)
STA diameternm30nm37.5% (28.3–46.6)
Heart rate202061.9% (47.7–76.1)70.1% (57.2–83.0)
Systolic blood pressure30308.3% (3.7–12.9)6.7% (1.6–11.7)
Diastolic blood pressure2020−12.2% (−19.4 to −5.0)−14.4% (−20.4 to −8.4)
  • MCA = middle cerebral artery; STA = superficial temporal artery; nm = not measured in study. Data from healthy subjects have previously been published elsewhere (Birk et al., 2007).

STA

There was a significant increase in STA diameter on PACAP compared to placebo during the immediate (AUC0–30 min, P < 0.001) and post-infusion phases (AUC30–90 min, P < 0.001). (Figs 3 and 4).

Fig. 4

Individual and mean diameters in the superficial temporal artery (STA) assessed by high-resolution ultrasonography. Thick lines show mean values. Infusion of PACAP38 (filled square) in migraine patients resulted in an increase in STA during the infusion (P < 0.001) and post-infusion (P < 0.001) period compared with placebo (open square).

HR and mean arterial blood pressure

We found an increase in HR on PACAP38 compared to placebo during both the immediate (AUC0–30 min, P < 0.001) and the post-infusion phases (AUC30–90 min, P < 0.001) (Fig. 3 and Table 4).

No difference in MAP was found between PACAP38 and placebo during the immediate (AUC0–30 min, P = 0.366) and the post-infusion (AUC30–90 min, P = 0.104) phases. In addition, no difference in systolic blood pressure was found between PACAP38 and placebo during the immediate (AUC0–30 min, P = 0.051) or the post-infusion phase (AUC30–90 min, P = 0.529). There were also no difference in diastolic blood pressure between PACAP38 and placebo during the immediate (AUC0–30 min, P = 0.056) and the post-infusion phases (AUC30–90 min, P = 0.056). ANOVA showed that systolic blood pressure increased over time after PACAP38 (P = 0.001). A post-hoc test revealed a maximum increase from baseline found at T30 (Table 4). Further, diastolic blood pressure decreased over time after PACAP38 (P < 0.001), where the maximum decrease from baseline was found 20 min after start of infusion. The systolic blood pressure also changed over time during the infusion of placebo (P = 0.023), but post-hoc analysis revealed no statistical changes. The diastolic blood pressure changed over time after the placebo infusion (P = 0.024), where the maximum decrease from baseline was found at T20 (−6.0%, 95% CI: −9.6 to −2.4) (Table 4).

Adverse events

Adverse events were recorded and reported during immediate and post-infusion phases in both studies (Table 5). Heat sensation, palpitations and flushing were more often reported on PACAP38 than on placebo in both healthy subjects and migraine patients. One migraine patient experienced uncontrolled crying during PACAP38 infusion. The patient described this episode as very peculiar, since no pain or shift in emotions was experienced. All subjects experienced flushing, especially on the face and trunk, which lasted up to 24 h. One migraine patient described facial puffing 24 h after PACAP38 infusion and also right sided headache (VRS 2), which was very uncomfortable. An oral antihistamine (8 mg acrivastin) almost completely abolished the facial puffing and completely terminated the headache after 2.5 h.

View this table:
Table 5

Number of healthy subjects and migraine patients reporting adverse events after PACAP38 and placebo infusion from 0 to 90 min

Healthy subjects Symptoms (0–90 min)PlaceboPACAP38P
Heat sensation4110.016
Palpitation0110.001
Flushing012<0.001
Nausea150125
Headache0100.002
Migraine patients Symptoms (0–90 min)PlaceboPACAP38P
Heat sensation012<0.001
Palpitation1120.001
Flushing1120.001
Nausea020.500
Headache1110.002
  • PACAP38 and placebo compared with McNemar test.

Discussion

The major finding of the present study was that PACAP38 induced mild to moderate headache in both healthy subjects and migraine patients. Furthermore, 50% of the migraine patients reported migraine-like attacks several hours (mean 6 h) after start of PACAP38 infusion. Increased flushing, palpitations and heat sensation were more often reported after PACAP38 than after placebo as a sign of autonomic activation. The adverse events after PACAP38 infusion may have compromized blinding, since they could be clearly noticed by both the investigator and subject after infusion of the experimental drug. The adverse events were however caused by the physiologic response to PACAP38 and could not be avoided. Though imperfect, the present double-blind approach is therefore the best possible way of coping with methodological error.

Distribution of PACAP38 relevant for headache

PACAP immunoreactive nerve fibers surrounding cerebral vessels have been demonstrated in the cat (Uddman et al., 1993; Jansen-Olesen et al., 1994) and rat (Edvinsson et al., 2001). PACAP have been identified in human sensory (Tajti et al., 1999), sympathetic and parasympathetic (Uddman et al., 1999) ganglia and in the trigeminal nucleus caudalis (Uddman et al., 2002). PACAP receptors are found in human smooth muscle cells of cerebral arteries (Knutsson and Edvinsson, 2002), sensory, parasympathetic and in sympathetic ganglia with perivascular nerve fiber projections (Knutsson and Edvinsson, 2002). Hence, PACAP can be defined as a sensory, a parasympathetic and a sympathetic neuropeptide, which can modulate both vessels and nerve fibers through its receptors. The headache induced by PACAP38 could thus originate from several anatomical locations.

PACAP induced vasodilatation in comparison with VIP

PACAP38 elicits vasodilatation in both animal (Uddman et al., 1993; Jansen-Olesen et al., 1994; Seki et al., 1995; Dalsgaard et al., 2003; Boni et al., 2008) and human (Jansen-Olesen et al., 2004; Birk et al., 2007) cerebral arteries. Animal studies have suggested that the VPAC1 receptor is primarily responsible for vasodilatation (Fahrenkrug et al., 2000; Boni et al., 2008), but the precise receptor mechanism of PACAP-induced vasodilatation is not yet fully understood and may thus vary between species and from tissue to tissue.

The mean peak decrease in VMCA in the healthy volunteers after PACAP38 was, as previously published (Birk et al., 2007), −12.8% (95% CI: −21.2 to −4.4), which is close to the −16.1% (95% CI: −18.2 to −10.6) mean peak decrease found in the migraine sufferers (Table 4). In comparison, infusion of the parasympathetic neuropeptide VIP caused a −14.4% (95% CI: −17.4 to −11.4) (Hansen et al., 2006) decrease in healthy subjects and −16.3% (95% CI: − 19.9 to −12.6) (Rahmann et al., 2008) in migraine patients. Thus, it seems that migraine patients are not hypersensitive to the activation of the shared VIP/PACAP VPAC1 receptor that most likely is responsible for vasodilatation (Fahrenkrug et al., 2000; Boni et al., 2008). Furthermore, maximal vasodilatation in all vessels peaked 20 min after start of infusion and then very slowly declined while the median time to peak headache score in migraine patients occurred at a median of 4.0 h after start of infusion. This indicates that vasodilatation per se is not causative for induction of migraine-like attacks in the post-hospital phase. However, we cannot exclude that a secondary vasodilatation occurred during the post-hospital phase or that the initial and long lasting vasodilatation induced a cascade of events leading to the delayed headache. Interestingly, VIP infusion did not induce migraine-like attacks (Rahmann et al., 2008) and only mild headache (Hansen et al., 2006). Therefore, the shared VIP/PACAP receptors do not seem to be causal for induction of migraine-like attacks after PACAP38 infusion. Instead the PAC1 receptor and the physiological processes initiated by its activation might be the primary target for the difference in headache sensitivity after VIP and PACAP38 infusion.

PACAP induced headache

Both healthy subjects and migraine patients in the present study appeared to experience headache with almost the same intensity (Fig. 1). However, the main difference between the healthy group and the migraine group was that no healthy subjects experienced delayed migraine-like attacks in the post-hospital phase after PACAP38 compared to six migraine patients. Obviously, the headache experienced by healthy subjects cannot mimic a usual migraine attack and they do not use triptans. Still, by excluding the two migraine patients, who only described the headache as mimicking usual migraine attack and treated it with a triptan, it appears that PACAP38 induced more migraine-like attacks in the post-hospital phase among migraine patients compared to healthy subjects (4 versus 0 attacks).

The peak median headache score after PACAP38 was 2.5, which is less than compared to other neuropeptides known to induce migraine attacks (Table 6). This difference may be attributed to a more frequent intake of triptans compared to previous studies (Thomsen, 1994; Lassen et al., 2002). Hence, Lassen et al. reported two migrainers taking sumatriptan, while the study by Thomsen et al. took place before triptans were widely available. Interestingly, all studies report a similar median time to peak headache score, which indicate that the neuropeptides all induce changes, which take hours to manifest. The migraine patients developed migraine-like attacks after a mean of 6 h after PACAP38. Given that the plasma half-life of PACAP38 is only 3.5 min (Birk et al., 2007), it seems most likely that PACAP38 acts as a initiator of a cascade of events that eventually lead to a migraine-like attack. In the following we would like to discuss the possible cascade of events.

View this table:
Table 6

Median peak headache score and median time to peak headache score of migraine patients in the present study using PACAP38 and previous studies of GTN and CGRP (Lassen et al., 2002; Thomsen, 1994)

Median peak headache scoreMedian time (h) to peak headache score
PACAP382.5 (0–10)4 (0–12)
GTN5.5 (0–10)5.5 (3–10)
CGRP4 (1–6)5 (1–12)

Possible mechanisms behind PACAP38 induced migraine-like attacks

In principle, at least three different mechanisms could be involved in PACAP38's induction of migraine-like attacks: (i) Sensitization of peripheral sensory trigeminal fibers; (ii) mast cell degranulation secondarily causing activation of peripheral sensory trigeminal fibers; (iii) facilitation of pain by direct sensitization of central, second order trigeminal neurons.

Activation of VPAC and PAC1 receptors elevates cellular cyclic adenosine monophosphate (cAMP) (Dickson et al., 2006). Calcitonin gene-related peptide (CGRP) and cilostazol also increase cAMP and cause a delayed headache that develops hours after the experimental drugs are administrated (Lassen et al., 2002; Birk et al., 2006). Animal models in both rat and guinea pig (Ingram and Williams, 1996) have shown that trigeminal neurons can be sensitized through elevation of cAMP. Interestingly, approximately half of healthy subjects and migraine patients in the present study reported a throbbing headache quality, which has been related to sensitization of meningeal nociceptors (Strassman et al., 1996; Burstein et al., 1998). At present we do not know if activation of VPAC or PAC1 receptors can actually mediate sensitization of trigeminal neurons. PACAP38 has been reported to stimulate adenylate cyclase activity at least 1000 times greater than that of VIP in cultured neural cells (Miyata et al., 1989). Furthermore VIP only induces a minimal headache and no migraine-like headache in migrainers (Hansen et al., 2006; Rahmann et al., 2008). It is possible that the VPAC receptors, in contrast to the PACAP selective PAC1 receptor, are not involved in sensitization of trigeminal neurons.

Degranulation of mast cells may be involved in migraine pathophysiology (Theoharides et al., 2005). Mast cells surround cerebral (Edvinsson et al., 1976) and dural vessels (Ottosson and Edvinsson, 1997; Rozniecki et al., 1999) in close apposition to both parasympathetic and sensory nerve fibers (Rozniecki et al., 1999). Furthermore, mast cell degranulation caused neuronal activation in C-fibers innervating the dura mater (Levy et al., 2007). PACAP and VIP injected intradermally in healthy subjects caused a rapid flare, which became erythematous after 5 min (Warren et al., 1992). Maximum increase in skin blood flow occurred 15 min after PACAP injection and the increase lasted approximately 6 h compared to only 2 h after VIP injection. Furthermore, VIP only induces a 10% increase in degranulation of histamine from dural mast cells. PACAP38 has a higher uptake rate into the brain compared to VIP (Dogrukol-Ak et al., 2004) and CNS represents a relatively protected space for PACAP38 in comparison with its degradation in blood (Dogrukol-Ak et al., 2004). If headache/migraine-like attack is caused by nociception within the brain, these pharmacokinetic findings could explain why PACAP38 might be more potent than VIP in causing headache/migraine-like attack either directly or via mast cell degranulation. It would be highly relevant to compare the potency of PACAP and VIP in degranulating mast cells in future experimental models.

Experimental animal models have proposed that PACAP might have a role in central pain transmission (Hashimoto et al., 2006). Capsaicin can elevate PACAP in cerebrospinal fluid (Zhang et al., 1997) suggesting that PACAP might be released from activated C-fibers in the spinal cord. Furthermore, the PAC1 receptor antagonist, PACAP 6-38, effectively attenuates nociception in animal models of chronic inflammatory and persistent pain (Davis-Taber et al., 2008; Ohsawa et al., 2002) after intrathecal administration. In PACAP−/− gene knockout mice inflammatory pain disappears (Mabuchi et al., 2004), and PACAP38 promotes late-onset, transcriptional-dependent, activity dependent central sensitization (Ji et al., 2003; Mabuchi et al., 2004), which takes hours to manifest and lasts for prolonged hours or days. Based on these data, it would be plausible to suggest that exogenously administrated PACAP38 could have a facilitatory effect on second-order trigeminal neurons contributing to headache/migraine.

Conclusion

PACAP38 induced both headache and migraine-like attacks. Dilatation of cephalic arteries seems unlikely to be the direct cause of the migraine-like attacks occurring in the delayed phase after PACAP38 infusion. Possible mechanisms of PACAP38 induction of headache and migraine-like attacks are peripheral sensitization of trigeminal sensory neurons, mast cell degranulation secondarily leading to activation of peripheral sensory trigeminal fibers and facilitation of pain by sensitization of central second order trigeminal neurons. Regardless of this, the effect is probably mediated by PAC1 receptor activation. Therefore a PAC1 receptor antagonist might be a future target for the treatment of migraine.

Acknowledgements

The authors wish to thank lab technicians Winnie Grønning Nielsen, Kirsten Brunsgaard and Lene Elkjær for their excellent and dedicated assistance. The study was supported by the Mauritzen La Fontaine Foundation, the Cool Sorption Foundation and the Lundbeck Foundation via the Lundbeck Foundation Center for Neurovascular Signalling (LUCENS).

Footnotes

  • Abbreviations:
    Abbreviations
    PACAP38
    Pituitary adenylate cyclase activating peptide-38
    STA
    superficial temporal artery
    TCD
    transcranial Doppler
    VIP
    vasoactive intestinal polypeptide
    VRS
    verbal rating scale

References

View Abstract