Investigating Migraine-Like Behavior Using Light Aversion in Mice

The light/dark assay is widely used to assess anxiety-like behavior¹². The assay relies on the innate aversion of mice to light and their drive to explore when placed into a novel environment (light zone). However, as we report here, this assay can also be used to assess light-aversive behavior as well.

It is critical to consider the number and necessity of pre-exposures prior to testing. This depends on the mouse strain or model. For example, in our light/dark assay protocol, naïve wildtype CD1 and C57BL/6J mice are pre-exposed to the light/dark chamber twice prior to undergoing the treatment test procedure, while optogenetic mice do not undergo pre-exposure. A recent publication reported that one pre-exposure is sufficient for CD1 mice to display light aversion after i.p. CGRP administration¹⁷. Consequently, the significance of the novelty parameter will have lessened upon arrival of the treatment day^10,16. Pre-exposures can unmask light-aversive phenotypes by reducing the exploratory drive and thus altering the balance between exploration and aversion. In some cases, pre-exposure is not necessary. For example, with genetically altered mice with increased CGRP receptors in the nervous system, pre-exposure was not necessary¹⁴. Likewise, with optogenetically manipulated mice, in which CaMKIIa-expressing neurons in the posterior thalamic nuclei were targeted for optical stimulation, pre-exposure was not necessary, presumably because the light-aversive response was so robust upon direct stimulation of the brain¹⁹. Thus, the number and necessity of pre-exposures to the chamber must be carefully considered when using different mouse strains or models. Indeed, overexposure of mice to the chamber may reduce exploratory behavior. This will lead to the mice preferentially occupying the dark zone, regardless of treatment, therein reducing the ability to observe a light-aversive response. Conversely, insufficient pre-exposure to the assay may lead to exploratory behavior masking potential light-aversive behavior.

A post-treatment exposure serves to identify whether a mouse has fully recovered from the CGRP injection administered 2 days prior. This is essential prior to running the open field assay or any other assay to confirm that no prolonged treatment effect is present that will affect future behavioral tests.

We opted for a 30-min protocol duration based on previous observations¹⁰. We have tested mice in the light/dark assay for 10 min¹⁵, 20 min¹⁶ and 30 min¹⁰ separately. CGRP decreased the amount of time mice spent in the light between 0-30 min, but past 30 min the control mice preferred to spend more time in the dark compared to 0-30 min, hence leading to the decision to test for 30 min. In a similar fashion, the testing duration can be adjusted with reference to the time-response curve for different mouse models. It should be noted that lengthening the exposure time to the light/dark chamber may reduce motivation to explore the light zone.

We analyzed many different parameters to assess the animal behavior. One essential feature of the light/dark assay is a measurement of the time a mouse spends in the light zone, directly reflecting light aversion. Percentage of time spent resting, the number of vertical beam breaks (to measure rearing activity) in light or dark zones, and the number of transitions between the two zones are used to assess motility. Resting time and vertical beam breaks are normalized to the time spent in each zone in order to avoid false conclusions regarding movement. We include all mice in the analyses except: mice that remain in the light zone for the entire 30 min of testing, mice that spend over 90% of time resting in total (both light and dark zones), and statistical outliers (>3 SDs from the mean). The number of mice that are excluded is generally less than 1%. For the open field assay, the percentage of time in center is the main measurement used to assess anxiety-like behavior.

In the modified light/dark assay, the positioning of the fiber-optic cannula at some brain regions can greatly restrict mouse movement and, in some instances, prevent the mouse from reaching the dark zone. Consequently, entry into the dark zone will be negatively reinforced and, after multiple attempts, the mouse may show a learnt preference for the light, even remaining in the light zone during the entire testing period. This can be rectified by modifying the size and shape of the opening in the dark insert. As an example, when fiber-optic cannulae were installed in the cerebellum of wildtype C57BL/6J mice, the mice had difficulty crossing the opening of the dark insert. After altering the width of the opening to 6.10 cm instead of 5.08 cm, the mice were able to traverse the opening freely.

A 30.5 cm fiber-optic patch cord is used in the modified light/dark assay, based on the size of the open field chamber, allowing the mouse to move freely. A shorter cord length will prevent a mouse from moving to the corners, while a longer cable may tangle and hinder movement. The length of the fiber-optic patch cord used for the modified open field assay is 50 cm. The length is not as strict as that in the light/dark assay since the height of the rotary joint can be adjusted according to the length of the fiber-optic patch cable, ensuring that the mouse is able to just reach the corners of the chamber.

Based on power analyses, 10-12 mice per group are needed for CD1 and C57BL/6J mice with i.p. CGRP, and for optogenetic C57BL/6J mice to detect significant light aversion. However, the C57BL/6J group size was considerably larger than the CD1 group size (Figure 2A,B) because the C57BL/6J mice were unresponsive to CGRP in a subset of the tests¹⁰, meaning multiple tests were conducted to account for this high variability in light-aversive behavior in these mice. Specifically, two experiments were combined for the CD1 mice and four experiments were combined for C57BL/6J mice with i.p. CGRP (Figure 2A,B)¹⁰. The reason for this variability is not known, but humans also show variability in their responses to CGRP and light. Intravenous (i.v.) injection of CGRP induced migraine attacks in around 63~75% of migraine patients, with 70~90% of patients who displayed migraine attacks exhibiting photophobia^22,23,24,25. Altogether, the assay has considerable variability and in addition to the number of mice, it is essential to do at least two and preferably three fully independent experiments with different cohorts of mice.

Bedding is not required in the light/dark chamber and the experimenter is not required to pre-handle or habituate the mice. As a precautionary measure the two pre-exposure procedures serve the purpose of acclimating the mice to the olfactory and physical cues of the experimenter; however, Ueno H. et al. demonstrated that there is no difference in time in light in the light/dark assay or time in center in the open field assay between mice after repeated handling and mice with no handling²⁶.

The open field assay can assess the contribution of anxiety towards a light-aversive phenotype. There are other well-validated anxiety-related assays, such as the elevated zero maze and the elevated plus maze²⁷; however, the open field assay is the most procedurally relevant control to the light/dark protocol since the same testing chamber is used for both assays. Even so, an assessment of anxiety can be strengthened by utilizing multiple assays or by measuring multiple parameters in a single test given that anxiety is a complicated and multifaceted behavior. Importantly, even if there is no anxiety phenotype in the open field assay, this does not rule out an anxiety component to the light-aversive phenotype. For example, light might be triggering an anxiety response. The open field assay only indicates that anxiety alone is not driving the response to light. While an anxiolytic drug, such as benzodiazepame, might be used in this assay, such an approach would have complications, e.g., anxiolytic drugs affect locomotion. Instead, we opted to use clinical anti-migraine medications, including sumatriptan, to validate the migraine-like status of the light-aversive phenotype. Sumatriptan successfully reversed CGRP-induced light aversion in both CD1 and C57BL/6J mice¹⁰.

Unlike the modified light/dark assay, the chamber on the pull-out drawer is outside of the cubicle with cubicle doors open in the modified open field assay due to the patch cord connecting to the mouse's head. Instead of 55 lux, the room light reaches the floor of the chamber at ~1000 lux. Even though the light intensity is different, the open field assay is a light-independent test. In detail, increasing the light intensity from 55 to 27,000 lux in the open field assay resulted in a trend of a decrease in time in the center in C57BL/6J mice, suggesting that the light intensity may influence mouse behavior²⁸. However, the difference between the control and experimental groups was not significant under neither 55 nor 27,000 lux²⁸. Additionally, the difference in light intensity between 55 and 1000 lux is far more subtle than between 55 and 27,000 lux. Wireless optogenetics can solve this problem as there would be no patch cord, allowing the open field chamber to be pushed inside of the sound-attenuating cubicle.

In addition, the patch cord still limits mouse movement despite selecting an optimal length. In the future, wireless optogenetics will offer a non-invasive alternative to cable-based optogenetic techniques.

It should be noted that we used acute injection of CGRP, which only replicates in part the prolonged CGRP release that accompanies migraine attacks. While we injected CGRP into mice to model migraine based on the premise that plasma CGRP levels were increased²⁹ and that i.v CGRP induced migraine attacks in migraine patients^{22,23,24,25,30}, this will not replicate the condition in the patient where CGRP is maintained at high levels for a relatively long time (patient measurements were taken at a median 3 hours after the migraine started²⁹), nor does it replicate chronic migraine where levels are reported to be elevated even between attacks³¹. Moreover, other pain-induced mediators have not been tested in our paradigm.

The Mogil group modified the elevated plus maze to measure light aversion in mice, with the closed arms being illuminated by bright light and the open arms remaining dark³². The standard elevated plus maze has often been used to detect anxiety-related behavior in animals. This assay is based on the conflict between a mouse's innate desire to explore a novel environment and being placed in a compromising position in the open unprotected maze arms. In the modified protocol, mice are forced to select between the closed arms, which are illuminated with bright light, and the open unprotected arms, which are dark. The preference to the former suggests anxiety overrules light aversion while the preference to the latter suggests light aversion takes precedence over anxiety. The Mogil group also conducted a standard elevated plus maze to evaluate anxiety-like behavior³². The purpose is the same as conducting the open field assay in our protocol. Cacna1a mutant mice, a familial hemiplegic migraine model, showed photophobia when the closed arms were bright. In contrast, anxiety-like behavior was not detected when the standard elevated plus maze was conducted³². In rats, by using both the modified elevated plus maze and the light/dark assay, it was demonstrated that nitroglycerin (NTG) was able to induce photophobia^33,34, which was rescued by sumatriptan³⁴. In the standard elevated plus maze setting where light is absent within the closed arms, NTG induced anxiety-like behavior in rats³⁴, suggesting that NTG-induced light aversion is accompanied with anxiety. To our knowledge, there are no publications using the light/dark assay and the modified elevated maze in the same mouse model. All in all, both the modified elevated plus maze and the light/dark assay proposed in this protocol have been demonstrated as effective measures of light-aversive behavior in mice.

We use the daylight LED panel with a daylight-balanced color (5600K), with a 60° flood beam spread, yielding no shadowing at a height of ~30 cm from the floor of chamber at either 55 lux or 27,000 lux. Other studies investigating light aversion have utilized the light/dark assay with varying modifications. For example, studies have used different light intensities for the light zone, ranging from hundreds to thousands of lux^35,36,37; used light at different wavelengths (e.g. blue and yellow)³⁸; or used different temperatures of light (cold and warm)³⁹. Caution should be taken for the heat produced by the light since it can affect the temperature of the dark and light zones and interfere with the mice's behavior, potentially causing a preference to a specific zone. Besides, it is also important to use the light with a good viewing angle to avoid shadow on the floor of the chamber. Light intensity is important for the test too. 25,000 -27,000 lux is approximately equivalent to bright daylight. By conducting the light/dark assay at such a high light intensity, it is possible to amplify the treatment effect; however, it is essential to consider the retinal damage⁴⁰ and the negative effect of such a high light intensity on a mouse's willingness to go into light. Some studies reported that mouse eyes exposed to direct light⁴¹ and mice exposed to bright light for several hours (e.g. 30,000 lux for 4 hrs⁴²) experienced retinal damage. In the light/dark assay, there is a dark zone for the mouse to escape from the bright light if the mouse desires. In addition, previous studies found that mice in the control group (C57BL/6J mice) spent a similar amount of time in the light zone under 55, 1000 and 27,000 lux²⁸. For CD1 mice, the control group spent about 1/3 of the time in the light under 27,000 lux¹⁰ and unpublished data had shown similar results at 55 lux. It suggests that 27,000 lux light on its own does not make CD1 and C57BL/6J mice distressed. Nonetheless, caution should be taken when opting for a higher light intensity.

Alongside differences in light setting, researchers have opted for a variety of approaches in analyzing the light/dark data. When assessing light aversion, the amount of time spent in the light zone with the light switched off (or with red light illumination of the light zone, given that mice eyes are less receptive to red light) are included in the calculation. For example, aversion index= (time in light_{0 lux}-time in light_{test lux})/ time in light_{0 lux} was used by the Gorin group to assess light aversion⁴³. Here, the 'light off' or 'red light' conditions are included to confirm that the avoidance of the light zone is conditional on light being present in opposed to simple place preference. We conducted this procedure with i.p. injection of CGRP and found that mice receiving CGRP did not have a place preference with light off in the light zone, confirming that CGRP-induced aversion is light-dependent¹⁶. Lastly, the Gorin group used the time mice spent in the periphery of the light zone in the light/dark assay as a measure of anxiety³⁶. We utilize a traditional test for anxiety, the open field assay. No matter which analysis method is chosen, it should be noted that the contribution of anxiety to light aversion cannot be ignored. This protocol attempts to partition out anxiety-like and light-aversive behavior by utilizing the light/dark and open field assays in tandem.

This protocol addresses the use of the light/dark and open field assays for the detection of light-aversive behavior in mice. This provides a useful tool to identify the mechanisms of neural circuits and brain regions driving photophobia. The test paradigm can be migraine-specific or can be expanded into other disorders involving photophobia. With respect to migraine, we have tested two other neuropeptides associated with migraine pathogenesis: pituitary adenylate cyclase-activating polypeptide (PACAP) and vasoactive intestinal peptide (VIP). PACAP and VIP were demonstrated to induce light aversion in CD1 mice^17,21. In addition to migraine, photophobia is also a symptom of many other disorders, including bradyopsia, acute ocular injury or inflammation, traumatic brain syndromes, Lyme disease, albinism and cone dystrophy³⁶. Thus, this test paradigm provides a tool to investigate mechanisms underlying photophobia-related disorders. Moreover, the pairing of optogenetic methods with conventional pharmacological approaches will undoubtedly assist in the development of novel therapeutics for photophobia-related disorders.