Vascular Actions of Peripheral CGRP in Migraine-Like Photophobia in Mice

Abstract

Background:

CGRP is recognized as a key player in migraine, yet the mechanisms and sites of CGRP action remain unknown. The efficacy of CGRP-blocking antibodies as preventative migraine drugs supports a peripheral site of action, such as the trigeminovasculature. Given the apparent disconnect between the importance of vasodilatory peptides in migraine and the prevailing opinion that vasodilation is an epiphenomenon, the goal of this study was to test whether vasodilation plays a role in CGRP-induced light aversive behavior in mice.

Methods:

Systemic mean arterial pressure and light aversive behavior were measured after intraperitoneal administration of CGRP and VIP in wild-type CD1 mice. The functional significance of vasodilation was tested by co-administration of a vasoconstrictor (phenylephrine, endothelin-1, or caffeine) with CGRP to normalize blood pressure during the light aversion assay.

Results:

Both CGRP and VIP induced light aversion that was associated with their effect on mean arterial pressure. Notably, VIP caused relatively transient vasodilation and light aversion. CGRP-induced light aversion was still observed even with normalized blood pressure. However, two of the agents, endothelin-1 and caffeine, did reduce the magnitude of light aversion.

Conclusion:

We propose that perivascular CGRP cause light aversive behavior in mice by both vasomotor and non-vasomotor mechanisms.

Introduction

Over the past few decades there has been intense debate on whether vasodilation plays a role in the genesis of migraine and the current view is that migraine is a neural, not vascular, disorder (reviewed in (1–4). The theory that migraine pain was due to vessel relaxation was first articulated by Willis at the end of the 17^th century (5) and supported by early observations from Wolff linking the intensity of migraine to the pulsations of carotid arteries (6). Drugs such as sumatriptan that treat migraine symptoms also induced vasoconstriction on vessels that were dilated during migraine attacks (7–11). Additionally, later studies reported that pharmacologically-provoked headaches coincided with an enlarged middle meningeal artery (12), which could be reversed by sumatriptan (10). However, spontaneous and pharmacologically provoked migraines have been associated with only slight intracranial vessel dilation (13, 14), or no dilation (15). Hence, a causal contribution of vasodilation in migraine pathophysiology is still debated and new evidence often contradicts previous conclusions (1–4).

Neuropeptides such as calcitonin gene-related peptide (CGRP), pituitary adenylate cyclase-activating polypeptide (PACAP), and vasoactive intestinal peptide (VIP) have been implicated in the pathogenesis of migraine (16). Receptors for these peptides are found on the vasculature and the peptides induce arterial vasodilation and release of inflammatory mediators (1, 2, 17–20). While CGRP and PACAP have been shown to induce migraine-like headaches in people and sustained vasodilation (12, 14, 21, 22), VIP only induces a mild headache and transient vasodilation (23). All three peptides are present in the trigeminal nerve: CGRP is in ~50% (18, 20, 24), PACAP in ~29%, and VIP in ~7% of trigeminal neurons (25, 26). The vascular smooth muscle of both cranial and extracranial vasculature is heavily populated with CGRP receptors (18, 27) and their activation has been shown to induce a potent dilation of these vessels (10, 28). VIP and PACAP share receptors and these receptors are also found on cerebral and peripheral blood vessels (29, 30). As with CGRP, PACAP/VIP receptors are mainly found on the smooth muscle of arterioles and arteries (29, 31) and activation can induce marked cephalic vasodilation (23, 32).

The therapeutic efficacy of CGRP monoclonal antibodies strongly points to the importance of a peripheral site of action of CGRP (33, 34). We have speculated that CGRP actions on the peripheral vasculature might be sufficient to induce light aversion, a surrogate for photophobia in a mouse model of migraine (35). Photophobia is a common, often debilitating, sensitivity to usually non-painful levels of light. Both central and peripheral CGRP administration is sufficient to cause light aversion in wild-type mice exposed to bright light following treatment (35, 36). This light sensitivity is seen even with dim light in mice that overexpress the human receptor activity-modifying protein 1 (hRAMP1) subunit of the CGRP receptor in the nervous system (37, 38). Interestingly, this enhancement was seen only after central, but not peripheral, CGRP administration (35). These data suggest that, despite similar phenotypes, central and peripheral CGRP have distinct sites of actions and that in the periphery the limiting site of CGRP action is at a non-neural site, possibly the vasculature.

The goal of this study was to investigate whether vasodilation contributes to the actions of peripheral CGRP to induce migraine-like photophobia. Using outbred CD1 mice, we demonstrate that CGRP and VIP induce hypotension and light aversion, albeit with differing kinetics. Importantly, pharmacological inhibition of CGRP-induced vasodilation with vasoconstrictors phenylephrine, endothelin-1, and caffeine, all of which have different mechanisms of action, either only partially attenuated or failed to prevent light aversion.

Materials and Methods

Animals

Wild-type CD1 (Charles River, USA) mice were used. Equivalent numbers of adult male and female mice, aged 10–20 weeks, were used in all experiments. Data from both sexes were combined for all studies. While generally similar results were observed in male and female mice, the experiments were not powered to detect subtle sex differences. Mice were housed in groups of 3 to 5 per cage, unless otherwise indicated, on a 12 h light cycle with food and water ad libitum. All behavioral experiments were performed between 8:00 AM and 2:30 PM. For all telemetry experiments, blood pressure was measured between 9:00 AM and 11:30 AM. For all experiments, investigators were blinded to drug treatment and animals were randomly selected to each treatment group prior to commencement of experiments. Animal procedures were approved by the University of Iowa Animal Care and Use Committee and performed in accordance with the standards set by the National Institutes of Health and the ARRIVE guidelines.

Drug administration

All drugs were diluted with the vehicle, Dulbecco PBS (Hyclone, GE Healthcare Life Science, USA) and administered by intraperitoneal (i.p.) injection. Drugs were rat α-CGRP (Sigma-Aldrich, USA), human, porcine, rat VIP (Sigma-Aldrich, USA), phenylephrine (West-Ward Pharmaceuticals, USA), human endothelin-1 (Sigma-Aldrich, USA), caffeine (Sigma-Aldrich, USA). All drugs were administered at 10 μl/g bodyweight with a 30 g × 0.5 needle. All injections were performed by BNM, A-SW, LKB, AK, and WJK. Animals were gently held but not anesthetized during injection. For telemetry experiments, following injection mice were placed back in their home cage and blood pressure recordings started immediately. For the behavioral experiments, mice were allowed to recover for 30 min in their home cages before testing as previously described (35), unless tested immediately or at 60 min timepoints, as noted.

Light aversion and motility assays

The light/dark and motility data were collected using Activity Monitor version 7.06 from twelve chambers as previously described (35, 36). Mice were pre-exposed to the chamber twice to reduce exploratory drive, then tested with bright light (27,000 lux) (35, 36). This light intensity is comparable to that of full daylight in the shade (not direct sun) (39, 40). Data were collected for 30 min and analyzed in sequential 5 min intervals. The time in light was reported as the mean +/− SEM of all the mice at each interval (left panels in figures) and as the mean +/− SEM of the average time per interval for each individual mouse (right panels in figures).

Motility data were collected during the light aversion assay. Resting was calculated as the percentage of time spent not moving (not breaking any new beams). Vertical beam breaks (rearing) was measured by the animal crossing the beam at 7.3 cm height. Resting and rearing data in the light and dark zones were normalized to time spent in each zone.

Open-Field Assay

Mice were placed in the center of the chamber and tested for 30 min as described (35, 36). The periphery was defined as 4.22 cm from the border with the remaining 18.56 ×18.56 cm area as the center. Light intensity was the same as in the light/dark assay (27,000 lux). Treatments with vehicle and VIP were done in independent experiments from treatments with vehicle, CGRP and constrictors, which were all done at the same time.

Animal surgery and blood pressure measurements

Aged matched male and female CD1 mice (10–20 weeks old) were anesthetized with i.p. ketamine (91 μg/g) and xylazine (9.1 μg/g) for implantation of a radiotelemetry probe (TA11PA-C10, Data Sciences International) into the left common carotid artery through an anterior neck incision to allow measurement of mean arterial pressure in unrestrained mice. The telemetry transmitters were implanted subcutaneously into the left flank of each animal. Following surgery, each mouse was housed in individual cages in the Animal Care Facility and provided with buprenorphine or meloxicam for pain management. All surgeries were performed by WJK. Mice were also monitored daily to ensure animals were healthy. A week after telemeter implantation, changes in mean arterial pressure were measured for approximately 2.5 h at a sampling rate of 500 Hz every 20 s using Data Sciences International Acquisition software. A baseline mean arterial blood pressure is recorded for 30 min prior to injection. The recording is paused and then all mice are injected as a group prior to beginning the recording after treatment. The mean arterial pressure values were averaged in 5 min blocks for each treatment group. Differences in mean arterial pressure can be observed as a result of changes in different parameters (e.g. age, temperature, time of the year, stress, metabolism, etc). For this reason we chose to measure blood pressure in the same animals for all treatments within an experiment. Therefore, statistical comparisons are between groups within the same experiment.

Statistics

Sample size was estimated using the online software clincal.com and based on averages and standard deviations obtained from preliminary experiments. In order to observe a decrease of 50% of the time spent in light, using an error of 0.05 and a power of 80%, it was dertermined that 16 animals per group were needed. Data are reported as mean ± SEM. Data were analyzed using GraphPad Prism 8.4 software (RRID: SCR_002798). The majority of the collected data was normally distributed, with a few groups showing a slight deviation from normality. In all figures, only 6 treatment groups were non-normally distributed. The statistics used in the present paper are parametric for 3 reasons: 1) most of our data is analyzed using a 2-way ANOVA, which has no clear equivalent in non-parametric testing; 2) most of our groups contain a large number of animals; 3) most of the data is normally distributed and running parametric statistics everywhere allows consistency in our analysis.

When data are plotted as a function of time (line graphs), a two-way repeated measure ANOVA was performed (factors time and treatment) including all the time-points presented in the figures. When needed, a Tukey multiple-comparison test was performed to compare the effect of each treatment at each time point, and symbols on the figure indicate the difference of each treatment group compared to the control group, at each time-point. When data are plotted as averages for each treatment (scatter plot graphs), a one-way ANOVA was performed. If the treatment factor was significant, then a Tukey multiple-comparison test was performed to compare the effect of each treatment. All statistics are reported in Table 1.

Figure #	Analysis	Statistics (symbol on Figure)
Figure 1A	Two-way repeated measure ANOVA
	Interaction factor	F(69,552)=3.769, p<0.0001
	Treatment factor	F(3,24)=10.79, p<0.0001
	Time factor	F(2.84,68.29)=9.730, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1B Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,300)=2.355, p=0.0109
	Treatment factor	F(2,60)=25.52, p<0.0001
	Time factor	F(3.519,211.1)=0.080, p=0.9820
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1B Right panel	One-way ANOVA for treatment	F(2,60)=25.52, p<0.0001
	Tukey’s multiple comparisons
	- Veh vs. CGRP	p<0.0001 (****)
	- Veh vs. VIP	p<0.0001 (####)
	- CGRP vs. VIP	p=0.3160
Figure 1C Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,380)=0.5729, p=0.8955
	Treatment factor	F(3,76)=3.003, p=0.0356
	Time factor	F(4.079,310)=6.672, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1C Right panel	One-way ANOVA for treatment	F(3,76)=3.003, p=0.0356
	Tukey’s multiple comparisons between treatments
	- Veh vs. CGRP	p=0.0389 (*)
	- Veh vs. VIP (0.1)	p=0.9922
	- Veh vs. VIP (0.3)	p=0.9999
	- CGRP vs. VIP (0.1)	p=0.0584
	- CGRP vs. VIP (0.3)	p=0.0791
	- VIP (0.1) vs. VIP (0.3)	p=0.9908
Figure 1D Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,160)=1.341, p=0.2128
	Treatment factor	F(2,32)=0.7783, p=0.4677
	Time factor	F(3.456,110.6)=4.903, p=0.0019
Figure 1D Right panel	One-way ANOVA for treatment	F(2,32)=0.7783, p=0.4677
Figure 1E Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,300)=5.404, p<0.0001
	Treatment factor	F(2,60)=25.61, p<0.0001
	Time factor	F(3.548,212.9)=27.80, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1E Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,189)=1.597, p=0.1099
	Treatment factor	F(2,60)=4.147, p=0.0206
	Time factor	F(4.220,159.5)=18.50, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1F Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,300)=2.469, p=0.0075
	Treatment factor	F(2,60)=6.658, p=0.0002
	Time factor	F(3.237,194.2)=6.463, p=0.0002
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 1F Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(10,189)=1.093, p=0.3697
	Treatment factor	F(2,60)=1.390, p=0.2571
	Time factor	F(3.223,121.8)=1.776, p=0.1514
Figure 2A	Two-way repeated measure ANOVA
	Interaction factor	F(92,736)=5.110, p<0.0001
	Treatment factor	F(4,32)=7.199, p=0.0003
	Time factor	F(5.071,162.3)=7.485, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 2B Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,365)=4.007, p<0.0001
	Treatment factor	F(3,73)=7.031, p=0.0003
	Time factor	F(3.871,282.6)=8.755, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 2B Right panel	One-way ANOVA for treatment	F(3,73)=7.031, p=0.0003
	Tukey’s multiple comparisons
	- Veh vs. CGRP	p=0.0360 (*)
	- Veh vs. CGRP+PE	p=0.0149 (†)
	- Veh vs. PE	p=0.0001 (###)
	- CGRP vs. CGRP+PE	p=0.9864
	- CGRP vs. PE	p=0.2193
	- PE vs. CGRP+PE	p=0.3727
Figure 2C Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,365)=3.318, p<0.0001
	Treatment factor	F(3,73)=11.90, p<0.0001
	Time factor	F(4.076,297.6)=101.1, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 2C Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,255)=0.3885, p=0.9813
	Treatment factor	F(3,73)=2.022, p=0.1183
	Time factor	F(3.774,192.5)=17.33, p<0.0001
Figure 2D Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,365)=1.755, p=0.0394
	Treatment factor	F(3,73)=16.16, p<0.0001
	Time factor	F(4.235,309.1)=13.65, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 2D Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,255)=0.9754, p=0.4816
	Treatment factor	F(3,73)=11.08, p<0.0001
	Time factor	F(4.582,233.7)=9.355, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 3A	Two-way repeated measure ANOVA
	Interaction factor	F(69,621)=9.595, p<0.0001
	Treatment factor	F(3,27)=23.04, p<0.0001
	Time factor	F(3.605,97.34)=4.582, p=0.0028
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 3B Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,220)=1.155, p=0.3095
	Treatment factor	F(3,44)=6.132, p=0.0014
	Time factor	F(2.909,128.0)=1.143, p=0.3338
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 3B Right panel	One-way ANOVA for treatment	F(3,44)=6.132, p=0.0014
	Tukey’s multiple comparisons
	- Veh vs. CGRP	p=0.0010 (***)
	- Veh vs. CGRP+ET-1	p=0.0776
	- Veh vs. ET-1	p<0.0166 (#)
	- CGRP vs. CGRP+ET-1	p=0.1487
	- CGRP vs. ET-1	p=0.7596
	- ET-1 vs. CGRP+ET-1	p=0.7203
Figure 3C Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,220)=0.6800, p=0.8029
	Treatment factor	F(3,44)=5.178, p=0.0038
	Time factor	F(3.492,153.6)=20.25, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 3C Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,138)=0.7304, p=0.2635
	Treatment factor	F(3,44)=1.373, p=0.2635
	Time factor	F(3.262,90.03)=8.430, p<0.0001
Figure 3D Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,215)=0.2895, p=0.9960
	Treatment factor	F(3,43)=3.120, p=0.0357
	Time factor	F(3.129,134.5)=1.33, p=0.2667
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 3D Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,139)=0.8798, p=0.5878
	Treatment factor	F(3,44)=1.902, p=0.1432
	Time factor	F(3.217,89.43)=1.174, p=0.3255
Figure 4A	Two-way repeated measure ANOVA
	Interaction factor	F(69,368)=7.226, p<0.0001
	Treatment factor	F(3,16)=12.12, p=0.0002
	Time factor	F(3.729,59.67)=14.16, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 4B Left panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,385)=1.999, p=0.0144
	Treatment factor	F(3,77)=7.903, p=0.0001
	Time factor	F(4.04,311.1)=3.539, p=0.0075
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 4B Right panel	One-way ANOVA for treatment	F(3,77)=7.903, p=0.0001
	Tukey’s multiple comparison
	- Veh vs. CGRP	p=0.0005 (***)
	- Veh vs. CGRP+Caf	p=0.3987
	- Veh vs. Caf	p=0.9958
	- CGRP vs. CGRP+Caf	p=0.0525
	- CGRP vs. Caf	p=0.0003 (‡‡‡)
	- Caf vs. CGRP+Caf	p=0.2892
Figure 4C Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,390)=3.654, p<0.0001
	Treatment factor	F(3,78)=22.17, p<0.0001
	Time factor	F(3.642,284.1)=52.25, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 4C Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,246)=2.418, p=0.0027
	Treatment factor	F(3,78)=1.586, p=0.1996
	Time factor	F(3.987,196.2)=34.21, p<0.0001
Figure 4D Upper panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,394)=1.072, p<0.3807
	Treatment factor	F(3,79)=9.586, p<0.0001
	Time factor	F(3.570,281.3)=6.469, p=0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Figure 4D Lower panel	Two-way repeated measure ANOVA
	Interaction factor	F(15,292)=2.664, p=0.0960
	Treatment factor	F(3,79)=0.3465, p=0.7917
	Time factor	F(1.329,77.62)=2.664, p=0.0960
Suppl. Fig. 1A	Two-way repeated measure ANOVA
	Interaction factor	F(69,598)=3.319, p<0.0001
	Treatment factor	F(3,26)=4.188, p=0.0152
	Time factor	F(2.917,75.84)=8.191, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Suppl. Fig. 1B	Two-way repeated measure ANOVA
	Interaction factor	F(69,598)=6.543, p<0.0001
	Treatment factor	F(3,26)=8.642, p=0.0004
	Time factor	F(4.79,124.5)=3.781, p=0.0037
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Suppl. Fig. 1C	Two-way repeated measure ANOVA
	Interaction factor	F(69,621)=11.92, p<0.0001
	Treatment factor	F(3,27)=20.47, p<0.0001
	Time factor	F(3.675,99.21)=5.314, p=0.0009
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Suppl. Fig. 1D	Two-way repeated measure ANOVA
	Interaction factor	F(69,368)=8.709, p<0.0001
	Treatment factor	F(3,16)=10.80, p=0.0004
	Time factor	F(3.196,51.14)=14.16, p<0.0001
	Tukey’s multiple comparisons between treatments
	*p<0.05 compared to vehicle.
Suppl. Fig. 2	Simple linear regression
	Equation	Y = 1.4*X – 68
	Goodness of fit	R2=0.76

Exclusions were applied to the dataset for the following reasons: never leaving the light zone during 30 min of testing, mice had an overall (both in the light and in dark compartments) resting time >90%, or mice were considered statistical outliers according to the GraphPad Prism criteria (ROUT method, Q=0.1%). For all experiments, 3 mice were excluded for never leaving the light zone, 3 were excluded for resting >90%, and 3 mice were considered statistical outliers. Throughout this study, a total of 9 mice were excluded.

Results

Association of vasodilation caused by CGRP and VIP with light aversion in mice

To ask whether vasodilation correlates with light aversion, we used two vasodilatory peptides, CGRP and VIP. As an indicator of vascular tone, we monitored the systemic mean arterial pressure from the left common carotid artery via radiotelemetry in freely moving mice. Mean arterial pressure was measured for a 30 min baseline prior to i.p. injection of vehicle or peptide and then for an additional 90 min post-injection. The mice were first injected with vehicle, then after a 24 h recovery, a new baseline was recorded prior to injection of VIP (0.1 mg/kg) in the same cohort of mice. This protocol of baseline, injection, recovery was used for subsequent injections of VIP (0.3 mg/kg), then CGRP (0.1 mg/kg). The baseline mean arterial pressures had no significant differences (p=0.08). Before injection of vehicle it was 125 mm Hg, before VIP (0.1 mg/kg), 117 mm Hg, before VIP (0.3 mg/kg), 109 mm Hg, and before CGRP, 116 mm Hg (Fig. 1A). The results obtained when measuring diastolic blood pressure, which is less influenced by cardiac output, show a similar time course (Suppl. Fig. 1).

Figure 1. CGRP and VIP induce light aversion coinciding with their vasoactive periods.

A. Mean arterial pressure (MAP) was measured by telemetry from the carotid artery of CD1 mice at baseline and following i.p. injection of vehicle (n=5), CGRP (0.1 mg/kg, n=5), and VIP (0.1 mg/kg, n=10; 0.3 mg/kg, n=10); 2 experiments. B. Light aversive behavior immediately after i.p. injection. Mice were treated with vehicle (n=21), CGRP (0.1 mg/kg, n=22), or VIP (0.1 mg/kg, n=23), and light aversion was tested immediately after peripheral administration. The mean ± SEM time spent in the light compartment every 5 min over a 30 min period is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. C. Light aversive behavior 30 min after i.p. injection. Mice were treated with vehicle (n=25), CGRP (0.1 mg/kg, n=13), VIP (0.1 mg/kg, n=29), and VIP (0.3 mg/kg, n=13); 2 experiments. The mean ± SEM time spent in the light compartment every 5 min over a 30 min period is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. D. Light aversive behavior 60 min after i.p. injection. Mice were treated with vehicle (n=13), CGRP (0.1 mg/kg, n=9), or VIP (0.1 mg/kg, n=13), and light aversion was tested 60 min after injection; 2 experiments. The mean ± SEM time spent in the light compartment every 5 min over a 30 min period is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. E. Resting time in the dark and light zones immediately after i.p. injection. Results taken from the same experiment as in panel B. The percent ± SEM time spent resting in the dark (upper panel) and the light (lower panel) compartments every 5 min over a 30 min period is shown. F. Rearing in the dark and light zones immediately after i.p. injection. Average time spent rearing in the dark (upper panel) and the light (lower panel) per 5 min intervals following treatments. Statistics are described in Table 1.

During the immediate 30 min after injection, there was a significant drop in mean arterial pressure compared to vehicle or baseline with CGRP (0.1 mg/kg, 88 mm Hg) and both doses of VIP (0.1 mg and 0.3 mg/kg, 82 and 87 mm Hg, respectively) (Fig. 1A). Injection of vehicle did not significantly affect mean arterial pressure (133 mm Hg). Importantly, during the next 30 to 60 min period after injection, which corresponds to the light/dark testing period we used in our prior study with i.p. CGRP (35), mean arterial pressure remained significantly lower in the CGRP (88 mm Hg) cohort compared to baseline or vehicle (115 mm Hg) (Fig. 1A). In contrast, VIP treated mice had normal or only slightly lower pressure (0.1 mg/kg, 114 mm Hg; 0.3 mg/kg, 108 mm Hg) during this period. During the 60–90 min period, mean arterial pressure in CGRP treated mice returned close to vehicle, but did remain slightly lower (Fig. 1A).

Light aversion was measured as both a function of time over a 30 min testing window and in order to show individual mice in a scatter plot, the data are also represented as the average time spent per 5 min during the test period, as previously described (35, 36). When tested during the 30 min immediately after injection, both CGRP and VIP produced a light aversive phenotype (Fig. 1B). The average time spent in the light per 5 min interval was 46 s for VIP and 39 s for CGRP compared to 118 s for vehicle-treated mice.

However, when tested 30 min after injection, there was no significant difference in the time spent in light between vehicle and either dose of VIP, while CGRP still caused significant light aversion, as previously reported (35) (Fig. 1C). This correlates with the return of VIP-induced drop in mean arterial pressure to baseline about 30 min after injection.

During the subsequent 60 to 90 min period, neither CGRP or VIP induced significant light aversion, although there was a trend towards decreased time spent in light in the CGRP-treated mice (60 s) compared to vehicle (84 s) (Fig. 1D). This correlates with the return of CGRP-induced drop in mean arterial pressure to close to baseline levels approximately 60 min after administration. These results reveal an apparent association between the timing of changes in vascular tone and light aversion. A simple linear regression was done to compare the mean arterial pressure at every 5 min from 5 to 90 min after vehicle, CGRP, or VIP treatment with time spent in the light at those same time points (Suppl. Fig. 2). This analysis gave a 76% probability (r²=0.76) of a linear correlation between vascular tone and light aversion.

In addition to light aversion during the initial 30 min period, we also evaluated the effect of VIP on resting and vertical rearing during this period. VIP increased resting and decreased rearing only in the dark, not the light zone (Fig. 1E, F). Comparable changes were seen after injection of CGRP, with increased resting and decreased rearing predominantly in the dark zone (Fig. 1E, F). Changes in motility following VIP injection were not observed in the later 30 to 60 min period (data not shown).

Finally, there was no significant effect of VIP in the open field test measure of anxiety immediately after administration. The mean +/− SEM time spent in the center by mice treated with vehicle was 24% +/− 3% and with VIP (0.1 mg/kg) was 20% +/− 3% (p=0.42) (Table 1).

Inhibition of CGRP-induced vasodilation with phenylephrine does not attenuate light aversion

To address whether vasodilation plays a causative or just a correlative role in CGRP-induced light aversion, we used vasoconstrictors to pharmacologically counteract the vasodilatory effects of CGRP. We first tried phenylephrine to inhibit CGRP-induced vasodilation. Phenylephrine is a potent and selective agonist of α-adrenergic receptors that constricts vascular smooth muscle, along with other actions mediated by the sympathetic nervous system (41). Mean arterial pressure was measured for 30 min prior to treatments to obtain baseline values, then for 90 min post-injections. The mice were first injected with vehicle, then after a 24 h recovery, a new baseline was recorded prior to injection of CGRP plus phenylephrine in the same cohort of mice. This sequential protocol of baseline, injection, recovery was then used for subsequent injections of CGRP, then phenylephrine (1 mg/kg) alone. As expected, injection of CGRP caused a drop in pressure (63 mm Hg) and phenylephrine increased pressure (127 mm Hg) during the light/dark testing period (Fig. 2A). Co-injection of CGRP and phenylephrine significantly attenuated both the CGRP-induced hypotension and the phenylephrine-induced hypertension during this period. The mean arterial pressure following CGRP plus phenylephrine (92 mm Hg) was comparable to the vehicle-treated mice (97 mm Hg) (Fig. 2A). A higher phenylephrine dose (2 mg/kg) combined with CGRP caused the pressure to be a little higher (104 mm Hg) than vehicle, so we chose to go with the lower dose of phenylephrine for the behavioral tests. Of note, the diastolic blood pressure follows the same time-course as the mean arterial blood pressure (Suppl. Fig. 1).

Figure 2. Lack of attenuation of CGRP-induced light aversion by co-administration of phenylephrine.

A. Mean arterial pressure (MAP) was measured by telemetry from CD1 mice at baseline and following i.p. injection vehicle (n=9), CGRP (0.1 mg/kg, n=7), phenylephrine (PE) (1 mg/kg, n=6), and CGRP+PE (n=8); 2 experiments. Data are mean ± SEM. B. Light aversive behavior after CGRP and phenylephrine co-administration. Mice were treated with vehicle (n=18), CGRP (0.1 mg/kg, n=21), CGRP+PE (1 mg/kg, n=21), or PE (1 mg/kg, n=17), and light aversion was tested 30 min after i.p. injection; 2 experiments. The mean ± SEM time spent in the light compartment every 5 min over a 30 min period is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. C. Resting in the dark and light zones. Percentage time spent resting in the dark (upper panel) and the light (lower panel) per 5 min intervals during the same experiment as in panel B. Data are the mean ± SEM. D. Rearing in the dark and light zones. Average time spent rearing in the dark (upper panel) and the light (lower panel) per 5 min intervals following treatments. Data are the mean ± SEM. Statistics are described in Table 1.

We then tested whether preventing CGRP-induced drop in blood pressure would prevent CGRP-induced light aversion. Mice were treated with either vehicle, CGRP, phenylephrine or CGRP plus phenylephrine at the doses described in Fig. 2A. Co-administration of phenylephrine with CGRP did not inhibit the CGRP-induced behavior (Fig. 2B). The degree of light aversion was indistinguishable between mice treated with CGRP alone compared to CGRP plus phenylephrine, spending an average of 51 s and 46 s in the light, respectively (Fig 2B). Consistent with the inability of phenylephrine to attenuate light aversion, it also did not block the CGRP-evoked resting in the dark (Fig. 2C) and decreased rearing (Fig. 2D). Unexpectedly, administration of phenylephrine alone also caused significant light aversion (Fig. 2B). Likewise, phenylephrine also increased resting time in the dark and decreased rearing, comparable to CGRP (Fig. 2C, D).

Inhibition of CGRP-induced vasodilation with either endothelin-1 or caffeine partially attenuates light aversion

While normalization of CGRP-induced drop in mean arterial pressure with phenylephrine failed to block light aversion, since phenylephrine alone caused light aversion, we tested two other vasoconstrictors with different mechanisms of action: endothelin-1 and caffeine. Endothelin-1 is the most potent vasoconstrictor in the body (42). It is released by endothelial cells and induces sustained vasoconstriction through activation of endothelin A/B receptors on smooth muscle cells (42). Co-administration of endothelin-1 (0.1 mg/kg) with CGRP abrogated CGRP-induced drop in blood pressure (Fig. 3A). The mean arterial pressure in CGRP plus endothelin-1 treated mice (119 mm Hg) was statistically indistinguishable from the vehicle-treated group (107 mm Hg) during the 30 min testing period. Attempts to use a lower dose of endothelin-1 (0.01 mg/kg), which had been reported to induce hypertension in rats when administered intravenously (43), was not sufficient to normalize CGRP-induced drop in mean arterial pressure during the first 25 min after injection and did not consistently normalize the pressure during the 30 min testing period compared to vehicle (data not shown). Once again, the diastolic blood pressure results showed the same time course (Suppl. Fig. 1). Hence, the 0.1 mg/kg dose was chosen for testing in the light aversion assay.

Figure 3. Partial attenuation of CGRP-induced light aversion by co-administration of endothelin-1.

A. Mean arterial pressure (MAP) was measured by telemetry from CD1 mice at baseline and following i.p. injection of vehicle (n=11), CGRP (0.1 mg/kg, n=8), endothelin-1 (ET-1) (0.1 mg/kg, n=6), and CGRP+ET-1 (0.1 mg/kg, n=6); 2 experiments. Data are mean ± SEM. B. Light aversive behavior after CGRP and endothelin-1 co-administration. Mice were treated with vehicle (n=10), CGRP (0.1 mg/kg, n=12), CGRP+ET-1 (0.1 mg/kg, n=18), or ET-1 (1 mg/kg, n=10), and light aversion was tested 30 min after i.p. injection; 2 experiments. The mean ± SEM time spent in the light compartment every 5 min over a 30 min period is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. C. Resting in the dark and light zones. Percentage time spent resting in the dark (upper panel) and the light (lower panel) per 5 min intervals during the same experiment as in panel B. Data are the mean ± SEM. D. Rearing in the dark and light zones. Average time spent rearing in the dark (upper panel) and the light (lower panel) per 5 min intervals following treatments. Data are the mean ± SEM. Statistics are described in Table 1.

We then tested whether endothelin-1 could attenuate CGRP-induced light aversion. In the presence of endothelin-1 plus CGRP, the mice spent an intermediate time in the light (63 s), which falls in between the levels observed with vehicle (107 s) and CGRP alone (24 s), without being significantly different from either (Fig. 3B). This suggests that endothelin-1 partially diminishes the light aversion induced by CGRP. Interestingly, co-administration of CGRP plus endothelin-1 failed to attenuate the change in resting (Fig. 3C) and rearing behavior (Fig. 3D) induced by CGRP, which is similar to what was observed with phenylephrine. As with phenylephrine, endothelin-1 alone caused significant light aversion (Fig. 3B). Likewise, endothelin-1 alone also increased resting time in the dark and decreased rearing, comparable to CGRP (Fig. 3C, D).

We then tested a third mechanism of vasoconstriction. Caffeine acts directly on the vasculature via inhibition of adenosine receptors (44). Furthermore, caffeine is used along with nonsteroidal anti-inflammatory agents to alleviate migraine pain (45). As a starting point, a dose of 5 mg/kg caffeine was chosen since it is comparable to the average caffeine consumption per day in humans (46) and causes a significant decrease in cerebral blood flow (47), although it is recognized that the pharmokinetics of oral vs i.p. delivery in human vs mouse preclude a direct comparison. The CGRP-induced decrease in mean arterial pressure (63 mm Hg) was attenuated by co-administration of caffeine with CGRP (88 mm Hg) during the light/dark testing period, which was statistically indistinguishable from vehicle (100 mm Hg) and caffeine (98 mm Hg) (Fig. 4A). Dissimilar to the other vasoconstrictors used in this study, this dose of caffeine alone did not induce a hypertensive effect.

Figure 4. Partial attenuation of CGRP-induced light aversion by co-administration of caffeine.

A. Mean arterial pressure (MAP) was measured by telemetry from CD1 mice at baseline and following i.p. injection of vehicle (n=5), CGRP (0.1 mg/kg, n=5), caffeine (5 mg/kg, n=5), and CGRP+caffeine (n=5); 1 experiment. Data are mean ± SEM. B. Light aversive behavior after CGRP and caffeine co-administration. Mice were treated with vehicle (n=21), CGRP (0.1 mg/kg, n=20), CGRP+caffeine (n=21), or caffeine (5 mg/kg, n=20), and light aversion was tested 30 min after i.p. injection; 2 experiments. The mean ± SEM time spent in the light compartment every 5 min over a 30 min run is shown on the left panel. The right panel shows the mean ± SEM in light per 5 min interval for individual mice. C. Resting in the dark and light zones. Percentage time spent resting in the dark (upper panel) and the light (lower panel) per 5 min intervals during the same experiment as in panel B. Data are the mean ± SEM. D. Rearing in the dark and light zones. Average time spent rearing in the dark (upper panel) and the light (lower panel) per 5 min intervals following treatments. Data are the mean ± SEM. Statistics are described in Table 1.

The efficacy of caffeine was then tested in the light aversion assay. As with endothelin-1, in the presence of caffeine plus CGRP, the mice spent an intermediate time in the light (54 s), which falls in-between the levels observed with vehicle (73 s) and CGRP alone (20 s), without being significantly different from either (p=0.052) (Fig. 4B; Table 1). This suggests that caffeine partially diminishes the light aversion induced by CGRP. Moreover, co-administration of caffeine with CGRP completely abrogated both the increase in resting time in the dark (Fig. 4C) and restored rearing behavior to levels similar to the vehicle-treated mice (Fig. 4D). Unlike phenylephrine and endothelin-1, caffeine alone did not cause significant light aversion (76 s) (Fig. 4B). Likewise, caffeine on its own did not affect resting or rearing behaviors (Fig 4C, D).

Neither phenylephrine, endothelin-1 or caffeine induce anxiety in an open-field assay

To determine whether the light aversive phenotype induced by phenylephrine and endothelin-1 was driven by an increased anxiety state, mice that had been previously tested in the light/dark assay were placed in an open field test. The mice were tested 30 min following injection of either vehicle, CGRP, phenylephrine, endothelin-1, or caffeine at the doses and light intensity used for the light/dark assay. There was no significant difference in time spent in the center of the chamber for any of the treatments compared to vehicle. The mean +/− SEM time spent in the center by mice treated with vehicle was 36% +/− 3%, CGRP was 29% +/− 4% (p=0.65), phenylephrine was 32% +/− 5% (p=0.88), endothelin-1 was 26% +/− 4% (p=0.28), and caffeine was 44% +/− 4% (p=0.66) (Table 1).

Discussion

In this study, we have investigated the functional significance of vasodilation on a migraine-like symptom in mice. The causal contribution of vasodilation in migraine pathophysiology is still intensely debated and new evidence often contradicts previous conclusions (1–4). To address this controversy, we first established that there is an apparent association between light aversion and the timing of decreased mean arterial pressure in response to two vasodilators, CGRP and VIP. The CGRP results are consistent with clinical studies reporting that infusion of CGRP causes migraine (14, 48) and vascular changes that can last up to ~80–120 min (12, 14), including dilation of the middle meningeal artery (10, 12). Indeed, there is an immediate onset of vascular changes upon CGRP infusion in subjects, as well as a rapid development of photophobia (median onset ~20 min) and mild headache (median onset ~20 min) (48). An association between intradural middle meningeal artery dilation and headache has recently been proposed (12). However, the delayed migraine-like headache (median onset ~3 h) (14, 48) and associated symptoms can last for many hours to days, far beyond the vascular changes. We do not see this prolonged phenotype in the mice. The reason is not known but may reflect that prolonged effects of CGRP infusion are only seen in migraine patients, not in control subjects (49), and we have only tested wild-type mice in this study. Furthermore, we have only examined aversion to bright light in wildtype mice and while the similar phenotypes of aversion to bright and dim light suggests similar mechanisms (35, 36), we cannot rule out the possibility of different mechanisms with different time courses. It will be interesting to see if genetically engineered “migraine mice”, such as hRAMP1 or other migraine models (50) have more prolonged responses to CGRP.

In contrast to CGRP, VIP causes dilation of cranial arteries without causing migraine symptoms in patients (23). However, in mice VIP did cause light aversion that correlated with a drop in mean arterial pressure. The duration of decreased pressure and light aversion was relatively short compared to CGRP, which is consistent with the shorter half-life of VIP (48 – 60 s (51, 52)) compared to CGRP (6.9 min (53)). Based on the transient response in mice, we suggest that the inability of VIP to induce migraine in patients may be due to pharmacokinetics and that VIP receptors (VPAC1 and VPAC2) should also be considered as targets for migraine therapeutics. While speculative, if PACAP acts on both VPAC and PAC1 receptors in migraine, this might help explain the recent failure of a PAC1-blocking antibody in a migraine clinical trial (https://clinicaltrials.gov/ct2/show/NCT03238781).

The main take home from this study is that CGRP has actions on both vascular tone and non-vasomotor activities that contribute to light aversive behavior, suggesting that vasodilation can in some cases contribute to CGRP-induced light aversion, but is not necessary. The role of vascular tone is suggested by the finding that two constrictors (endothelin-1 and caffeine) prevented vasodilation and partially attenuated the magnitude of light aversion. However, in the case of caffeine, we cannot exclude an antinociceptive contribution from inhibition of central adenosine receptors. While an analgesic effect generally requires higher doses (35–100 mg/kg), effects can also be seen at doses used in this study (54). Moreover, it is possible that the delayed effects of endothelin-1 and caffeine on blunting the CGRP-induced drop in blood pressure may have allowed early vasodilation signals to initiate processes leading to later light aversion.

Mechanical activation of trigeminal perivascular afferents can occur by intracranial vasodilation. Recently, it has been reported that activation of Piezo mechanoreceptors in the trigeminal ganglion can cause nociceptive firing of trigeminal nerve branches in the meninges and promote release of CGRP (55). A similar role may be played by TRP mechanosensitive channels (56). Therefore, it is possible that CGRP-induced vasodilation activates mechanoreceptors on nearby neurons, causing further release of CGRP, and vasodilation of cerebral blood vessels in a feedback loop (2). While this hypothesis does not agree with a reported lack of dural CGRP action on mechanoreceptors (57), it should be noted that study only used male rats. While the sex dependence of CGRP is not resolved, the recent observation that dural application of CGRP causes tactile hypersensitivity in only females (58), suggests that CGRP can activate some trigeminal mechanoceptors, possibly by changing vascular tone. Receptors for CGRP and the constrictors used in this study are present in the vasculature of the meninges. Specifically, meningeal vessels express the canonical CGRP receptor, Amy1 receptor, endothelin-A receptor, adenosine receptors, and alpha 1 adrenergic receptor (59–62). Interactions between these receptors, such as CGRP increasing dissociation of endothelin-1 from the endothelin-A receptor (63, 64), could further fine-tune the vascular tone.

Two other possible sites of CGRP action beyond the meninges are the eye and trigeminal ganglia. With respect to the eye, a bright light stimulus has previously been shown to evoke neural activity in central trigeminal neurons in rats, which could be inhibited by intravitreal injection of phenylephrine (65). The possibility that CGRP-evoked light aversion might be mediated by vascular events within the eye, Indeed, there are CGRP-immunoreactive sensory fibers and vascular CGRP receptors in the eye and intravitreal injections of CGRP can change intraocular pressure (66). A role for the ganglion is suggested by a recent report that intraganglionic injection of CGRP induces both photic sensitivity and facial allodynia in male and female rodents (67). We also did not detect a significant difference between sexes, consistent with our previous studies (35, 68). This is in contrast to a recent report that dural CGRP causes tactile hypersensitivity in female, but not male rodents (58). We believe a key difference may be that we used systemic administrations that would reach multiple target sites in addition to the dura, such as the eye and ganglia. Hence, a robust sex effect may be observed only when the drug is applied locally to the dura. It will be interesting to see whether there is cross-talk between CGRP actions on trigeminal cell bodies in the ganglia and fibers in the dura and eye.

Nonetheless, CGRP must also have non-vasomotor activities because some light aversive behavior was still observed in the presence of three vasoconstrictors that all normalized mean arterial pressure at the time of behavioral testing. In particular, the constrictor phenylephrine prevented CGRP-induced drop in pressure and the resultant light aversion was indistinguishable from CGRP alone. Similarly, light aversion was also still seen after normalization of mean arterial pressure by the constrictors endothelin-1 and caffeine, although, as mentioned above, there was a partial rescue.

This leaves the question of how might perivascular CGRP be acting beyond vasodilation? At the vasculature, CGRP actions on smooth muscle and endothelium can potentially release paracrine substances, such as cytokines, that sensitize nociceptors (1). Alternatively, the dura mater is heavily populated with an array of immune cells (69, 70). Mast cells have been implicated in the sensitization of sensory neurons in migraine and CGRP receptors have been reported on murine mast cells (71). Other immune cells, such as T cells, dendritic cells, and macrophages also express functional CGRP receptors (72). It is possible that a rich network of immune cells in the meninges may play a role in sensitizing the trigeminal nerve. Additionally, within the trigeminal ganglion, neurons, satellite glia, and Schwann cells express functional CGRP receptors and may play a role in peripheral sensitization (73, 74). Future studies are needed to dissect the role of CGRP at these vascular and perivascular targets in migraine.

In addition to light aversion, the associated symptoms of resting in the dark and rearing were also monitored. We have previously compared increased resting in the dark in mice to human behavior when a migraineur goes to lay in a dark room during an attack (35). Of note, only caffeine could reverse the motility phenotypes. The apparent dissociation of light aversion and motility responses following normalization of mean arterial pressure may reflect the multifactorial complexity of migraine.

There are two major caveats of this study. First, we have relied on a systemic measure of mean arterial pressure as a surrogate for vasodilation in the meninges. However, meningeal vascular tone can vary independently of systemic arterial pressure. The middle meningeal artery has its own intrinsic independent autoregulation and only minimally (or differentially) responded to drugs able to induce larger changes in the other vascular beds (75–78). Likewise, a study from Ashina and colleagues showed that while there were no differences in extracranial blood vessels during spontaneous migraines in humans, intracranial arteries had a slight dilation during the attacks (13). Moreover, it is now known that even subtle changes in the middle meningeal artery could be sufficient to activate migraine/pain pathways (1). A recent magnetic resonance angiography study showed a unilateral increase in the middle meningeal artery on the pain side during unilateral migraine compared to baseline (79), further supporting the importance of monitoring local changes. Future studies are needed to measure meningeal dilation in conjunction with migraine-like behaviors in mice.

A second caveat is that both phenylephrine and endothelin-1 unexpectedly induced light aversion when administered alone, which may account for the light aversion still seen following treatment with the combination of CGRP and these agents. However, even if so, it still remains that systemic pressure was normalized, so light aversion does not apparently require concurrent vasodilation. Phenylephrine is a sympathomimetic agent that is used in clinic to dilate the iris (80) and there are other effects associated with the eye, including lacrimation, eye pain, and visual sensitivity to light (81, 82). Hence, it is possible that phenylephrine could promote an affinity towards the dark zone by directly affecting the eye. Alternatively, Dux et al. have recently reported that phenylephrine causes trigeminal nociception and increases meningeal blood flow through TRPV1 receptor activation and release of CGRP (83). Likewise, endothelin-1 on its own can induce pain-like behaviors in rodents (84) and can sensitize nociceptors and enhance release of glutamate in sensory neurons (85). Moreover, reports show that endothelin-1 can amplify basal CGRP release approximately 6-fold in cultured neurons and that CGRP₍8_–37₎ fully attenuates endothelin-1 induced tactile allodynia at the later stages (86). Conversely, a recent clinical study showed endothelin-1 did not induce aura or headache in migraine patients (87), which was surprising since endothelin-1 is a potent inducer of CSD in animals (88). Therefore, it is plausible that the actions of phenylephrine and endothelin-1 in the light aversion assay could be influenced by an additive effect of endogenous CGRP. In addition, both phenylephrine and endothelin-1 on their own increased resting in the dark and decreased rearing. Therefore, while not negating the conclusion regarding CGRP actions, it is intriguing to speculate that vessels might contribute to migraine through activation of mechanoreceptors by contraction and this change from basal tone may be sufficient to sensitize trigeminal neurons.

Conclusion

The present study shows that, in mice, CGRP and VIP can induce a migraine-like symptom at the same time that they decreased mean arterial pressure. While we were not able to definitely elucidate the role of vasodilation, we can conclude that perivascular CGRP acts by multiple mechanisms involving both vascular tone and non-vasomotor actions to cause photophobia-like behavior in mice.

Article Highlights.

• CGRP and VIP can induce a migraine-like symptom in mice that is associated with decreased mean arterial pressure.

• Co-administration of vasoconstrictors that normalized CGRP-induced drop in systemic blood pressure did not prevent light aversion, although two of the constrictors reduced the magnitude of the response.

• CGRP acts by multiple mechanisms involving both vasomotor and non-vasomotor actions to cause photophobia-like behavior in mice.