Mechanisms of GTN-induced migraine: Role of NOS isoforms,sGC and peroxynitrite in a migraine relevant mouse model

Abstract

Background

Migraine research has highlighted the pivotal role of nitric oxide (NO) in migraine pathophysiology. Nitric oxide donors such as glyceryl trinitrate (GTN) induce migraine attacks in humans, whereas spontaneous migraine attacks can be aborted by inhibiting NO production. The present study aimed to investigate how GTN triggers migraine through its three nitric oxide synthase (NOS) isoforms (neuronal NOS (nNOS), endothelial NOS (eNOS) and inducible NOS (iNOS)) via a suspected feed-forward phenomenon.

Methods

Migraine-relevant hypersensitivity was induced by repeated injection of GTN in an in vivo mouse model. Cutaneous tactile sensitivity was assessed using von Frey filaments. Signaling pathways involved in this model were dissected using non-selective and selective NOS inhibitors, knockout mice lacking eNOS or nNOS and their wild-type control mice. Also, we tested a soluble guanylate cyclase inhibitor and a peroxynitrite decomposition catalyst (N_total= 312).

Results

Non-selective NOS inhibition blocked GTN-induced hypersensitivity. This response was partially associated with iNOS, and potentially nNOS and eNOS conjointly. Furthermore, we found that the GTN response was largely dependent on the generation of peroxynitrite and partly soluble guanylate cyclase.

Conclusions

Migraine-relevant hypersensitivity induced by GTN is mediated by a possible feed-forward phenomenon of NO driven mainly by iNOS but with contributions from other isoforms. The involvement of peroxynitrite adds to the notion that oxidative stress reactions are also involved.

Graphical abstract

This is a visual representation of the abstract.

Keywords

GTN migraine signaling mouse model nitric oxide NOS peroxynitrate von Frey

Introduction

In recent decades, there has been significant progress in our understanding of the signaling mechanisms of migraine, a condition that affects one billion people worldwide (1). These advances were achieved using human and animal provocation models (2). It was shown that nitric oxide (NO) is a crucial signaling molecule in migraine, as evidenced by the induction of migraine-like headaches in humans and animals using NO donors such as glyceryl trinitrate (GTN) and the therapeutic efficacy of inhibiting NO production in animals (3,4). Spontaneous migraine attacks in humans can be aborted by l-N^G methylarginine hydrochloride (l-NMMA) (5), a non-selective inhibitor of all three nitric oxide synthases (NOS) that are responsible for the endogenous production of NO. Also, various components of the NO signaling cascade such as NO precursor l-arginine and NO receptor soluble guanylate cyclase (sGC) are suggested to be involved in migraine (6,7). Interestingly, NO can also trigger tension-type headache and cluster headache in adults with these disorders (8,9). These findings strengthen the idea that altered NO pathways play a crucial role in migraine and potentially other headache disorders.

In human provocation models, GTN infusion for 20 minutes caused an immediate headache during and shortly after infusion, as well as a delayed migraine with a peak after 5.5 hours that fulfilled the international diagnostic criteria for a migraine attack (3). This model was translated into a mouse model using either single or repeated injections of GTN. The model has been validated by its response to receptor-specific migraine drugs i.e., sumatriptan (5-HT_1B/1D receptor agonists) and olcegepant (calcitonin gene-related peptide (CGRP) receptor antagonist) (4,10,11). These studies posed a crucial question: What happens in the hours long lag phase between immediate headache and delayed migraine? So far, this has not been clarified. NO mechanisms may also be active throughout spontaneous migraine attacks because they respond to treatment with l-NMMA (5). A feed-forward phenomenon induced by the NO donor sodium nitroprusside has previously been shown in rat medullary brainstem (12). Yet such a feed-forward mechanism is contrary to normal agonist biology which associates increased ligand concentration with decreased receptor sensitivity/density. We decided to study this phenomenon in the GTN mouse model. In addition to a non-selective NOS inhibitor, we used genetically modified mice not expressing endothelial NOS (eNOS) or neuronal NOS (nNOS), as well as chemical antagonists selective for inducible NOS (iNOS). Our data suggest the presence of a feed-forward phenomenon induced by GTN and studied its downstream effects by blocking sGC and stimulating the break-down of peroxynitrite.

Methods

Experimental animals

We used a total of 216 wild-type (WT) C57Bl/6J BomTac mice (Taconic, Denmark) along with 96 transgenic knockout (KO) mice aged 7–12 weeks (20–29 g) with an equal distribution of both sexes. Mice not born in-house had a minimum acclimatization period of one week before the initiation of experiments. Age-related variations were not observed. Consistent with previously published protocols (13), mice received uniform conditions of care. They were housed in a room with controlled temperature and lighting (lights on at 7:00 am, 12 h light/dark photocycle) and had ad libitum access to food and water. Cages were adequately enriched with shelters and nesting material. All mice were regularly weighed on test days as part of a general health assessment. The experiments adhered to ARRIVE guidelines (https://arriveguidelines.org) and complied with the European Community guide for the care and use of animals (2010/63/UE). Ethical approval was obtained from the Danish Animal Experiments Inspectorate (ethical approval numbers 2017-15-0201-01358 and 2022-15-0201-01347).

Transgenic mice and genotyping

Adult nNOS KO mice (JAX strain 002986) for experiments were delivered from the Jackson Laboratory (Bar Harbor, ME, USA). eNOS KO and WT controls were bred in-house by mating of homozygous eNOS KO (JAX strain 002684) to WT C57Bl6/J BomTac. Subsequent breeding for homozygous eNOS KO and WT animals was continuously monitored through genotyping. The PCR protocol and primers used was based on a standard protocol from JAX (https://www.jax.org/Protocol?stockNumber = 002684&protocolID = 23416) (14).

Experimental design and protocols

We employed a well-established mechanistic mouse model of migraine using GTN as migraine trigger (4,10) and assessed effects by quantifying cutaneous sensitivity, as described previously (4,15). This model uses GTN injection followed by assessment by von Frey filaments as a proxy for migraine-associated pain (4). Group sizes were twelve mice in all experiments, except for eNOS, where the group size was 14. Each experiment was conducted independently in separate cohorts of mice. The study protocol is outlined in Figure 1, whereas Table 1 summarizes the experiments and compounds administered. Mice of both sexes were tested every other day over a nine-day period, encompassing five test days. Inhibitory compounds were given as pre-treatment 15 minutes before GTN injection on each test day. Baseline cutaneous sensitivity was measured prior to injections on every test day (i.e. 48 hours after the last injection), and acute responses assessed two hours post-GTN injection. All assessments were conducted in low-light conditions (20–30 lux) between 8:00 am and 3:00 pm by an experimenter blinded to the experimental conditions.

Figure 1.

Design and experimental timeline of test paradigm for glyceryl trinitrate (GTN) injection. Following 1 day of habituation (day 0), five tests were performed every other day over the course of nine days. All compounds were injected intraperitoneally (i.p.). Pre-treatment of N^G-nitro-l-arginine methyl ester hydrochloride (l-NAME), S-methyl-l-thiocitrulline (SMTC), 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), Fe(III)5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato chloride (FeTPPS) or S-methylisothiourea hemisulfate salt (SMT) was given on every test day 15 minutes prior to the injection of GTN. On every test day, the basal threshold of cutaneous sensitivity was measured using von Frey filaments prior to injections and the acute response after injections was measured two hours after GTN.

Table 1.

Overview of study compounds used in vivo (alphabetical order).

Compound	Provider	Mechanism of action	Dose, route of administration	Time of injection	Vehicle
GTN (glyceryl trinitrate)	Cambrex (Germany) via RegionH Pharmacy (Denmark)	Nitric oxide (NO) donor	10 mg/kg, i.p.	Two hours prior to test	12.2% alcohol in saline
l-NAME (N^G-nitro-l-arginine methyl ester hydrochloride)	Tocris/Bio-Techne (UK)	Non-selective NOS inhibitor	100 mg/kg, i.p.	15 minutes prior to GTN injection	Saline
SMTC (S-methyl-l-thiocitrulline)	Cayman/BioNordika (Denmark)	Semi-selective nNOS inhibitor	60 mg/kg, i.p.	15 minutes prior to GTN injection	PBS
ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one)	Tocris/Bio-Techne (UK)	Selective soluble guanylate cyclase (sGC) inhibitor	10 mg/kg, i.p.	15 minutes prior to GTN injection	5% DMSO + 5% tween80 in saline
FeTPPS (Fe(III)5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato chloride)	Cayman/BioNordika (Denmark)	Peroxynitrite decomposite catalyst	30 mg/kg, i.p.	15 minutes prior to GTN injection	Saline
SMT (S-methylisothiourea hemisulfate salt)	Sigma-Aldrich/Merck (Belgium)	Selective iNOS inhibitor	10 mg/kg, i.p.	15 minutes prior to GTN injection	Saline
Midazolam	Hameln Pharma GMBH (Germany) via RegionH Pharmacy (Denmark)	Benzodiazepine	2 mg/kg, i.p.	10 minutes prior to test	Saline

Behavioral tests

Cutaneous sensitivity to tactile stimulation

Cutaneous sensitivity was evaluated on the left plantar region of the hind paw starting at 0.16 g using von Frey filaments, with a range of 0.008–2.0 g (excluding 1.4 g; sourced from Ugo Basile, Gemonio, Italy). For our assessment, we employed the up-down method (16). The 50% withdrawal threshold in grams was calculated using a freely accessible online tool (https://bioapps.shinyapps.io/von_frey_app) (17) and subsequently square root transformed (SQRT) for improved normal distribution for statistical analyses. For the test, mice were placed in a transparent plexiglas chamber with a mesh net bottom (IITC Life Science, Woodland Hills, CA, USA). To familiarize them with the testing environment, mice were allowed to acclimate in these chambers for 45 minutes before the first test (day 0 and a 30-minute habituation period for subsequent tests (18).

Motor function (Rotarod)

We evaluated general motor function using a rotarod (Rotarods Advanced; IITC Life Science) to ensure that potential side effects of the study drugs, such as impaired motor function or sedation, did not bias the cutaneous sensitivity test with von Frey filaments. The rotarod assessment was conducted before the initial drug injection and immediately following the last von Frey test on test day 9. Each mouse was given a single attempt on the rotarod, starting from 0 rpm and gradually accelerating to 30 rpm within 45 seconds, with a maximum duration of 150 seconds. The duration spent on the rotarod was recorded. Midazolam (2 mg/kg, intraperitoneally (i.p.)) or saline (i.p.) was administered as a positive or negative control, respectively, 10 minutes prior to testing. Mice were randomized into treatment groups when receiving midazolam.

Test compounds

Compounds were administered i.p. in the lower right quadrant of the abdomen. They were diluted in saline (Fresenius Kabi, Germany) and injected at a volume of 10 ml/kg unless otherwise specified. Details regarding compounds, vehicles, concentrations and doses can be found in Table 1.

The NO donor GTN (7.89 mg/ml in 96% ethanol) was administered to mice at a dose of 10 mg/kg after dilution in saline to a concentration of 1 mg/ml (4,15). For vehicle treatment in the GTN model, an equivalent volume of ethanol was dissolved in saline.

Statistical analysis

Mice were randomized balanced according to 50% withdrawal thresholds measured at baseline, sex and home cage. Treatment groups and sex were equally divided throughout the test day (19). Group sizes were based on our previous work with these models (10,13,20) where 12 animals per group provided sufficient power to detect intermediate effects (18,21). Withdrawal thresholds of 50% were SQRT for improved normal distribution and data were analyzed using two-way repeated measures analysis of variance (ANOVA). A subsequent Tukey’s post-hoc test was performed comparing all groups on each of the five test days. p < 0.05 was considered statistically significant. Data are shown as the mean ± SEM. In In the figures, significance levels are reposted shown as: *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. In the figures, baselines are shown for visualization but were not included in the statistical analyses. Rotarod data are analyzed by a Kruskal–Wallis test (non-parametric one-way ANOVA on ranks) with Dunn's post-hoc comparison. All statistical analyses and graphs were conducted in Prism, version 9 (GraphPad Software Inc., San Diego, CA, USA).

Results

N^G-nitro-l-arginine methyl ester hydrochloride (l-NAME) blocks GTN-induced hypersensitivity

We tested whether l-NAME (100 mg/kg) could block GTN-induced hypersensitivity. The dose of l-NAME was selected based on a dose response pilot study, we conducted prior the main experiment using three different doses of l-NAME (25, 50 and 100 mg/kg) in WT mice (data not shown). In this pilot study, 100 mg/kg was shown to be the optimal dose for complete blockade of the GTN response. We found a significant difference between the test group (l-NAME + GTN) and the positive control group (VEH + GTN) on all test days at the two-hour time point (Figure 2) and at basal levels on day 5, 7 and 9 (p < 0.01 to 0.001) (data not shown). The test group (l-NAME + GTN) showed no statistical difference from the negative control group (VEH + VEH) on all test days. The positive control group (vehicle + GTN) was significantly different from the negative control groups (vehicle or l-NAME + vehicle) (p < 0.05 to 0.0001). On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for the test group, positive control group and negative control group were 1.07 ± 0.09, 0.45 ± 0.04 and 0.88 ± 0.06, respectively. Thus, l-NAME blocked GTN-induced hypersensitivity.

Figure 2.

N^G-nitro-l-arginine methyl ester hydrochloride (l-NAME) blocks glyceryl trinitrate (GTN)-induced hypersensitivity. (A) Responses two hours after GTN (10 mg/kg, i.p.) or vehicle (VEH) injection (i.p.) in wild-type mice (n = 12 per group) on five test days. Mice were pre-treated with l-NAME (100 mg/kg, i.p.) or VEH on every test day. A black asterisk (*) indicates a difference between test group (l-NAME + GTN, red) and positive control group (VEH + GTN, blue), whereas a grey asterisl (*)indicates a difference between positive (VEH + GTN, blue) and negative control groups (VEH + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive visualization of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

nNOS may be involved in GTN-induced hypersensitivity

The effects of semi-selective nNOS inhibitor S-methyl-l-thiocitrulline (SMTC) on GTN provocation was studied in WT mice. We found a significant difference between the test group (SMTC + GTN) and the positive control group (VEH + GTN) on almost all test days at the two-hour time point (Figure 3) (p < 0.01 to 0.001). Both the test group and positive control group receiving GTN were significantly different from the negative control (VEH + VEH) on all test days (p < 0.05 to 0.0001) and at basal levels on day 5, 7 and 9 (p < 0.001) (data not shown). On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for the test group, positive control group and negative control group were 0.63 ± 0.06, 0.38 ± 0.08 and 1.00 ± 0.08, respectively.

Figure 3.

Glyceryl trinitrate (GTN)-induced hypersensitivity was partially diminished by neuronal nitric oxide synthase (nNOS) inhibitor S-methyl-l-thiocitrulline (SMTC). (A) Responses two hours after GTN (10 mg/kg, i.p.) or vehicle (VEH) injection (i.p.) in wild-type mice (n = 12 per group) on five test days. SMTC (60 mg/kg, i.p.) or VEH was given on day 0. A black asterisk (*) indicates difference between test group (SMTC + GTN, red) and positive control group (VEH + GTN, blue), whereas a grey asterisk (*) indicates difference between positive and negative control groups (VEH + GTN, blue vs. VEH + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive visualization of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

To rule out speculative effects of eNOS inhibition by SMTC, we investigated the role of nNOS on GTN-induced hypersensitivity using genetically modified mice that lack the nNOS enzyme. In contrast to our experiment using the semi-selective chemical nNOS inhibitor (SMTC), nNOS KO mice showed no difference in GTN response compared to WT (p > 0.05) (Figure 4). Both nNOS KO and WT mice displayed hypersensitivity after GTN injection compared to vehicle controls on all test days at the two-hour time point (p < 0.0001). On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for nNOS KO mice and WT mice that received GTN were 0.49 ± 0.05 and 0.41 ± 0.07, respectively. The result of nNOS inhibition suggests that nNOS alone does not mediate the effect of GTN, but that nNOS and eNOS together have a partial role.

Figure 4.

Glyceryl trinitrate (GTN)-induced hypersensitivity in both wild-type (WT) and neuronal nitric oxide synthase (nNOS) knockout (KO) mice, but we found no difference in GTN response between WT and nNOS KO mice. (A) Responses two hours after administration of GTN (10 mg/kg, i.p.) or vehicle (VEH) (i.p.) to nNOS KO mice (n = 12 per group) on five test days. A grey asterisk (*) indicates difference between positive and negative control groups (WT + GTN, blue vs. WT + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive representation of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

eNOS is not involved in GTN-induced hypersensitivity

We also investigated the role of eNOS on GTN-induced hypersensitivity using genetically modified mice that do not express the eNOS enzyme. eNOS KO mice showed no difference in GTN response compared to WT on all test days (p > 0.05) (Figure 5). Both eNOS KO and WT mice displayed migraine-like hypersensitivity after GTN injection compared to vehicle controls on day 5, 7 and 9 (p < 0.05). In addition, GTN response showed no difference between the two vehicle groups. On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for eNOS KO mice and WT mice that received GTN were 0.24 ± 0.05 and 0.35 ± 0.07, respectively. As for nNOS, eNOS alone could not be identified as the mediator of GTN induced hypersensitivity.

Figure 5.

Glyceryl trinitrate (GTN) significantly induced hypersensitivity in both wild-type (WT) and endothelial nitric oxide synthase (eNOS) knockout (KO) mice on day 5, 7 and 9 (P < 0.05), but we found no significant difference in GTN response between WT and eNOS KO mice (P > 0.05). (A) Responses two hours after administration of GTN (10 mg/kg, i.p.) or vehicle (VEH) (i.p.) to eNOS KO mice (n = 14 per group) on five test days. A grey asterisk (*) indicates difference between positive and negative control groups (WT + GTN, blue vs. WT + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive representation of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

iNOS is partially involved in GTN-induced hypersensitivity

The effects of selective iNOS inhibitor S-methylisothiourea hemisulfate salt (SMT) on GTN provocation was studied in WT mice. We found a significant difference between the test group (SMT + GTN) and the positive control group (VEH + GTN) on almost all test days at the two-hour time point (Figure 6) (p < 0.05 to 0.01). Both the test group and positive control group receiving GTN were significantly different from the negative control (SMT + VEH) on all test days (p < 0.0001) and at basal levels (p < 0.01) (data not shown). On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for the test group, positive control group and negative control group were 0.71 ± 0.06, 0.43 ± 0.04 and 1.04 ± 0.07, respectively.

Figure 6.

Glyceryl trinitrate (GTN)-induced hypersensitivity is partially dependent on inducible nitric oxide synthase (iNOS). (A) Responses two hours after GTN (10 mg/kg, i.p.) or vehicle (VEH) injection (i.p.) in wild-type mice (n = 12 per group) on five test days. S-methylisothiourea hemisulfate salt (SMT) (10 mg/kg, i.p.) or VEH was given on day 0. A black asterisk (*) indicates difference between test group (SMT + GTN, red) and positive control group (VEH + GTN, blue), whereas a grey asterisk (*) indicates difference between positive and negative control groups (VEH + GTN, blue vs. SMT + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive visualization of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

GTN-induced hypersensitivity is partially dependent on sgc

The effects of sGC inhibitor 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) on GTN response was studied in WT mice. We found a significant difference between the test group (ODQ + GTN) and the positive control group (VEH + GTN) on test days 7 and 9 (Figure 7) (p < 0.01). The positive control group receiving GTN were significantly different from the negative control (VEH + VEH) on all test days (p < 0.01 to 0.0001) (Figure 7) and at basal levels on day 5, 7 and 9 (p < 0.001). On day 9, the mean ± SEM SQRT 50% withdrawal thresholds for the test group, positive control group and negative control group were 0.75 ± 0.09, 0.29 ± 0.07 and 0.98 ± 0.06, respectively.

Figure 7.

Soluble guanylate cyclase inhibitor 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) diminishes glyceryl trinitrate (GTN)-induced hypersensitivity on later test days. (A) Responses two hours after GTN (10 mg/kg, i.p.) or vehicle (VEH) injection (i.p.) in wild-type mice (n = 12 per group) on five test days. ODQ (10 mg/kg, i.p.) or VEH was given on day 0. A black asterisk (*) indicates difference between test group (ODQ + GTN, red) and positive control group (VEH + GTN, blue), whereas a grey asterisk (*) indicates difference between positive and negative control groups (VEH + GTN, blue vs. VEH + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive visualization of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

GTN-induced hypersensitivity is partly mediated via generation of peroxynitrite

The effects of Fe(III)5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato chloride (FeTPPS), a peroxynitrite decomposite catalyst, on GTN response was studied in WT mice. We found a significant difference between the test group (FeTPPS + GTN) and the positive control group (VEH + GTN) on all test days (p < 0.01 to 0.001) (Figure 8). The positive control group receiving GTN was significantly different from the negative control (VEH + VEH) on all test days (p < 0.05 to 0.0001) (Figure 8) and at basal levels an all test days (p < 0.01). On day 9, the mean ± SEM 50% withdrawal thresholds for the test group, positive control group and negative control group were 0.84 ± 0.07, 0.38 ± 0.07 and 0.80 ± 0.08, respectively.

Figure 8.

Fe(III)5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrinato chloride (FeTPPS), a peroxynitrite decomposite catalyst, diminishes glyceryl trinitrate (GTN)-induced hypersensitivity. (A) Responses two hours after GTN (10 mg/kg, i.p.) or vehicle (VEH) injection (i.p.) in wild-type mice (n = 12 per group) on five test days. FeTPPS (30 mg/kg, i.p.) or VEH was given on day 0. A black asterisk (*) indicates difference between test group (FeTPPS + GTN, red) and positive control group (VEH + GTN, blue), whereas a grey asterisk (*) indicates difference between positive and negative control groups (VEH + GTN, blue vs. VEH + VEH, black). Repeated measure two-way ANOVA with Tukey's post-hoc comparison between groups. (B) Descriptive visualization of individual data points for baseline and day 9 of (A) with the mean ± SEM (○ = males, ● = females).

Motor function was unaffected by all combinations of test compounds

To assure that the von Frey tests were not biased by impaired motor coordination, this was examined in mice before and after injection of study compounds using the Rotarod test. We found that motor function was unaffected in any of the KO mice or by any of the compounds used in the study. We found no differences between the negative control groups (vehicle) and the test groups (p > 0.67). As a validity of the test, we observed a significant difference between the positive control groups, which were injected with midazolam, and the tested groups (p < 0.0001 in all experiments). Thus, the compounds used in the experiments did not affect motor coordination, nor did they have sedative effects.

No adverse effects were observed following treatment with GTN, SMTC, ODQ or FeTPPS in any of the experimental mice. All mice appeared healthy and had normal stools and stable weight throughout the experiment, as well as normal motor function.

Discussion

We investigated the possible feed-forward phenomenon of GTN-induced hypersensitivity in a mouse model of migraine. GTN-induced hypersensitivity was fully dependent on NOS enzymes but neither of the isoforms alone had potent inhibitory effects. Furthermore, both sGC and peroxynitrite generation contributed to the effect. Overall, our results demonstrate the importance of the NO feed-forward phenomenon and suggest that iNOS may be the most important mediator of this response. Downstream pathways involving sGC and peroxynitrite generation both appear to be important, suggesting that GTN-induced hypersensitivity in mice is mediated via multiple pathways rather than one. The GTN induced feed-forward phenomenon and the complex interaction of multiple pathways may explain the gap of several hours between GTN infusion and induction of migraine-like headache in humans.

Feed-forward phenomenon of NO

GTN can liberate NO in tissues with and without active NOS enzymes. Both nNOS and eNOS are constitutively active (22), whereas iNOS is not constitutively active but is induced by bacterial infection and pro-inflammatory cytokines (23). We suggest that an amplifying wave of iNOS and/or nNOS/eNOS induction (possibly uncoupled) may cause the delayed provoked migraine via subsequent high endogenous NO production leading to oxidative stress (Figure 9). In spontaneous migraine, non-selective NOS inhibition effectively aborted attacks (5). It has previously been suggested that GTN administration in rats may cause a delayed upregulation of iNOS mRNA level within dural meningeal macrophages after two hours, followed by a corresponding increase in protein expression after four to 10 hours (24).

Figure 9.

Proposed nitric oxide (NO) feed-forward mechanisms based on present study and cited literature. Additionally, TRPA1 is closely linked to these pathways.

NO and NOS

Over the past two decades, inhibiting NO production through NOS inhibition has been investigated as a therapeutic strategy. Two proof of concept clinical trials showed that a non-selective NOS inhibitor was effective in acute migraine attacks and chronic tension-type headache (5,25). Several preclinical studies have also validated the efficacy of non-selective NOS inhibition in animal models of migraine (26). The non-selective NOS inhibitors have poor oral absorption, a short duration of action and cause significant side effects including increased blood pressure (27). Thus, the development of non-vascular NOS inhibitors may be a more promising strategy for migraine.

Based on present and previous preclinical studies, targeting iNOS may appear most effective because of its involvement in dural inflammation that may contribute to trigeminovascular activation (28). However, two phase II trials using a highly potent and selective iNOS inhibitor, GW274150, were negative for acute and preventive migraine treatment (29,30). Perhaps NOS isoform involvement is different in spontaneous vs. GTN induced migraine. On the other hand, iNOS inhibition as a therapeutic target may not need to be abandoned solely because of the lack of effect from a single chemical entity (GW274150). Targeting eNOS for migraine treatment has always been problematic because it seems unsafe as a result of its primary presence in vascular endothelium, essential for healthy cardiovascular function (31). Furthermore, our results do not support eNOS as a viable drug target because eNOS KO did not prevent hypersensitivity by GTN. Targeting nNOS is a promising option as a result of its expression in both the central and peripheral neuronal pain pathways of migraine (22,32,33). Two clinical trials using a combined nNOS inhibitor and 5HT-1B/1D receptor agonist, NXN-188, suggested that this drug may be effective in migraine treatment (34,35). Yet, a follow-up trial showed that NXN-188 did not have a substantial effect on migraine headache in adults with migraine with aura (36). This study was limited by a high drop-out rate in a relatively small sample. Thus, the effectiveness of NXN-188 in migraine remains uncertain. In summary, specific NOS targeting has had limited success, reinforcing our conclusion that NO acts via synergistic effects of the NOS isoforms to induce migraine. This is supported by a prior study investigating the effects of each NOS isoform on neurogenic and CGRP-induced meningeal vasodilation, suggesting a synergistic effect of these isoforms (37).

sGC

NO activates the sGC pathway that catalyzes the conversion of GTP to second messenger cGMP (31). The sGC pathway is a relevant signaling mechanism in migraine pathophysiology because the phosphodiesterase 5 inhibitor, sildenafil, induces migraine attacks in adults with migraine (7). Preclinical experiments using different pharmacological agents have likewise identified the sGC pathway as a key signaling mechanism regulating migraine pathophysiology (26). The phosphodiesterase 5 inhibitor sildenafil induced both acute and chronic hyperalgesia in mice (4) and a selective sGC stimulator, VL-102, induced cephalic and hindpaw hyperalgesia that was sensitive to both acute and preventive migraine treatments (38). Furthermore, the sGC inhibitor ODQ semi-effectively blocked migraine-associated acute and chronic hypersensitivity induced by GTN, even without exogenous NO stimulation. This suggest a potential link between endogenous NOS activity and chronic migraine maintenance (38) or the fact that GTN is acting directly on the sGC-cAMP pathway without NO. Despite the need for a ten-fold higher dose and less robust results, our findings partly reproduce the findings that ODQ inhibits GTN-induced hypersensitivity.

Peroxynitrite

NO can induce oxidative stress through cGMP-independent pathways, potentially contributing to migraine pathophysiology. One such pathway involves the formation of peroxynitrite, a downstream product of NO (39). Clinical studies have revealed disturbances in platelet levels of peroxynitrite-related proteins among migraine patients (40,41). Furthermore, inhibition of peroxynitrite production has been demonstrated to both prevent and terminate nociceptive responses, supporting its significance as a target for pain management (39,42). A recent preclinical study, focusing on peroxynitrite in a migraine model, demonstrated its involvement in modulating ongoing and evoked responses in dural-trigeminocervical neurons, as well as their response to CGRP (43). Another recent study showed that compounds modulating peroxynitrite alleviate the behavioral, neuronal, and molecular alterations induced by repeated stress and nitric oxide donors (44). The induction of oxidative stress pathways by GTN, is in line with previous studies in rats and mice showing that GTN-induced responses are dependent on TRPA1 channel activation (18,45,46). GTN and NO do not activate TRPA1 directly, but likely via the endogenous reactions leading to the formation of electrophilic agents like oxygen species or nitroxyl (47,48). Our findings strongly support that the peroxynitrite pathway is important in migraine. This opens a possible role for oxidative stress and mitochondrial function (44,49,50) in migraine susceptibility. Thus, peroxynitrite and other reactive oxygen species may be targets for migraine and migraine-like headache disorders.

Strengths and limitations

We used a well-validated mouse model to investigate signaling pathways of GTN (4,15,18). The compounds administered in our experiments had no impact on motor coordination, assuring the reliability of von Frey measurements.

We acknowledge some limitations of our study. Although GTN can release NO in tissues (51), it can also directly act on the sGC-cAMP pathway without involving NO (52). Therefore, it is possible that GTN may not be a strong NO liberator, and its effect could be through direct activation of sGC. The modest effect of blocking sGC seen in this study suggests a more important role of NO liberation in this particular model. However, other more genuine NO donors might function differently from GTN and more accurately mimic NO generation through the three NOS isoforms.

Our contrasting results with pharmacological nNOS inhibition and nNOS KO may be a result of compensatory mechanisms in nNOS KO mice. It is known that in some genetically modified mice, backup and compensatory mechanisms can maintain functions (53). Ideally, we should have investigated all three NOS both chemical inhibitors and KO for better comparison. This was not possible because we were not able to acquire a selective chemical inhibitor for eNOS or a KO mouse strain for iNOS. Also, we did not explore the combination of chemical inhibitors and KO mice, which would be interesting as future experiments.

Regarding eNOS, we cannot say with full certainty that it does not have a role in GTN-induced hypersensitivity. The given dose of the semi-selective nNOS inhibitor SMTC was based on literature, where 10–200 mg/kg has been shown to be effective in various mouse models (54,55). SMTC is a potent NOS inhibitor with selectivity toward nNOS compared to eNOS and iNOS with K_i values of 1.2, 11, and 40 nM for nNOS, eNOS and iNOS, respectively (56). When nNOS selectivity ceases in vivo is unfortunately unknown. However, the lack of effect on GTN response in eNOS KO mice indicates that eNOS involvement alone is unlikely.

We used plantar tactile sensitivity as primary outcome, although some suggest periorbital sensitivity is more relevant (18). Both anatomical areas show increased mechanical sensitivity to various migraine-inducing compounds (4,18,38,57). To our knowledge, there have been no studies demonstrating exclusive hypersensitivity in the cephalic region of mice without affecting the plantar area. Migraine-specific drugs without general analgesic effects (58,59). In our experience, periorbital sensitivity is more variable and has a narrower effect window (60,61), necessitating larger experimental groups thus reducing feasibility (18,38). Likewise, other potential endpoints such as light sensitivity or grimacing, also require larger mouse cohorts because of increased variability (62).

Conclusions

The induction of migraine-relevant hypersensitivity by GTN involves a NOS mediated feed-forward phenomenon in which subsequent high NO-levels are maintained and engaged in both cGMP-dependent and cGMP-independent pathways via sGC and peroxynitrite generation. Our results suggest that inhibition of one NOS enzyme is not a likely migraine drug target, whereas antagonizing soluble guanylyl cyclase or peroxynitrite may be more promising. Clinical implications

Migraine-relevant hypersensitivity induced by GTN is mediated by a feed-forward phenomenon of NO driven mainly by iNOS but with contributions from other isoforms.

Peroxynitrite and oxidative stress reactions are also involved in migraine-relevant hypersensitivity induced by GTN.

Our results suggest that inhibition of one NOS enzyme is not a likely migraine drug target, whereas antagonizing soluble guanylyl cyclase or peroxynitrite may be more promising.

Footnotes

Author contributions

CE, JO, SLC, DMK and SG contributed to the conception, design and planning of the experiments. SG, CE, KOR and SLC contributed to data acquisition by performing the in vivo experiments and wrote the appropriate manuscript sections. Data were analyzed by SG, SLC and CE. The manuscript was written by SG with assistance from SLC and JO. All authors read, reviewed and approved the final version of the manuscript submitted for publication .

Data availability

The datasets used or analyzed during the current study are available from the corresponding author upon reasonable request.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Song Guo is supported by the BRIDGE – Translational Excellence Programme (bridge.ku.dk) at the Faculty of Health and Medical Sciences, University of Copenhagen, funded by the Novo Nordisk Foundation. Grant agreement no. NNF20SA0064340 (2021 fellows). Also, this work was supported by Candys Foundation, Migrænefonden, Torben og Alice Frimodts Fond and Ragna Rask-Nielsens Grundforskningsfond.

Ethical statement

Ethical approval was obtained from the Danish Animal Experiments Inspectorate.

ORCID iD

Song Guo

References

Stovner

Nichols

Steiner

, et al. Global, regional, and national burden of migraine and tension-type headache, 1990–2016: a systematic analysis for the global burden of disease study 2016. Lancet Neurol 2018; 17: 954–976.

Ashina

Hansen

Á Dunga

, et al. Human models of migraine - short-term pain for long-term gain. Nat Rev Neurol 2017; 13: 713–724.

Thomsen

Kruuse

Iversen

. A nitric oxide donor triggers genuine migraine attacks. Eur J Neurol 1994; 1: 73–80.

Pradhan

Smith

McGuire

, et al. Characterization of a novel model of chronic migraine. Pain 2014; 155: 269–274.

Lassen

Ashina

Christiansen

, et al. Nitric oxide synthase inhibition in migraine. Lancet 1997; 349: 401–402.

Onderwater

GLJ

van Dongen

Harms

, et al. Cerebrospinal fluid and plasma amine profiles in interictal migraine. Ann Neurol 2023; 93: 715–728.

Kruuse

Thomsen

Birk

, et al. Migraine can be induced by sildenafil without changes in middle cerebral artery diameter. Brain 2003; 126: 241–247.

Olesen

Iversen

Thomsen

. Nitric oxide supersensitivity: a possible molecular mechanism of migraine pain. Neuroreport 1993; 4: 1027–1030.

Fanciullacci

Alessandri

Figini

, et al. Increase in plasma calcitonin gene-related peptide from the extracerebral circulation during nitroglycerin-induced cluster headache attack. Pain 1995; 60: 119–123.

10.

Christensen

Petersen

Kristensen

, et al. Targeting CGRP via receptor antagonism and antibody neutralisation in two distinct rodent models of migraine-like pain. Cephalalgia 2019; 39: 1827–1837.

11.

Bates

Nikai

Brennan

, et al. Sumatriptan alleviates nitroglycerin-induced mechanical and thermal allodynia in mice. Cephalalgia 2010; 30: 170–178.

12.

Schlechtweg

Röder

Fischer

MJM

, et al. Increase in NADPH-diaphorase-positive and neuronal NO synthase immunoreactive neurons in the rat spinal trigeminal nucleus following infusion of a NO donor–evidence for a feed-forward process in NO production involved in trigeminal nociception. Cephalalgia 2009; 29: 566–579.

13.

Ernstsen

Christensen

Olesen

, et al. No additive effect of combining sumatriptan and olcegepant in the GTN mouse model of migraine. Cephalalgia 2021; 41: 329–339.

14.

Liu

Xiang

, et al. Nitric oxide synthase-3 deficiency results in hypoplastic coronary arteries and postnatal myocardial infarction. Eur Heart J 2014; 35: 920–931.

15.

Tipton

Tarash

McGuire

, et al. The effects of acute and preventive migraine therapies in a mouse model of chronic migraine. Cephalalgia 2016; 36: 1048–1056.

16.

Chaplan

Bach

Pogrel

, et al. Quantitative assessment of tactile allodynia evoked by unilateral ligation of the fifth and sixth lumbar nerves in the rat. J Neurosci 1994; 53: 55–63.

17.

Christensen

Hansen

Storm

, et al. Von Frey testing revisited: provision of an online algorithm for improved accuracy of 50% thresholds. Eur J Pain 2020; 24: 783–790.

18.

Christensen

Rasmussen

Ernstsen

, et al. CGRP-dependent signalling pathways involved in mouse models of GTN- cilostazol- and levcromakalim-induced migraine. Cephalalgia 2021; 41: 1413–1426.

19.

Minett

Eijkelkamp

Wood

. Significant determinants of mouse pain behaviour. PLoS One 2014; 9: e104458.

20.

Christensen

Ernstsen

Olesen

, et al. No central action of CGRP antagonising drugs in the GTN mouse model of migraine. Cephalalgia 2020 Aug; 40: 924–934.

21.

Potvin

Schutz

. Statistical power for the two-factor repeated measures ANOVA. Behav Res Methods Instrum Comput 2000; 32: 347–356.

22.

Moncada

Higgs

. The discovery of nitric oxide and its role in vascular biology. Br J Pharmacol 2006; 147 Suppl 1:: S193–S201 .

23.

Bryan

Bian

Murad

. Discovery of the nitric oxide signaling pathway and targets for drug development. Front Biosci 2009; 14: 1–18.

24.

Reuter

Bolay

Jansen-Olesen

, et al. Delayed inflammation in rat meninges: implications for migraine pathophysiology. Brain 2001; 124: 2490–2502.

25.

Ashina

Lassen

Bendtsen

, et al. Effect of inhibition of nitric oxide synthase on chronic tension-type headache: a randomised crossover trial. Lancet 1999; 353: 287–289.

26.

Pradhan

Bertels

Akerman

. Targeted nitric oxide synthase inhibitors for migraine. Neurotherapeutics 2018; 15: 391–401.

27.

Olesen

. Nitric oxide-related drug targets in headache. Neurotherapeutics 2010; 7: 183–190.

28.

Messlinger

Suzuki

Pawlak

, et al. Involvement of nitric oxide in the modulation of dural arterial blood flow in the rat. Br J Pharmacol 2000; 129: 1397–1404.

29.

Palmer

. A randomised, single-blind, placebo-controlled, adaptive clinical trial of GW274150, a selective iNOS inhibitor, in the treatment of acute migraine. Cephalalgia 2009; 29: 124–124.

30.

Høivik

Laurijssens

Harnisch

, et al. Lack of efficacy of the selective iNOS inhibitor GW274150 in prophylaxis of migraine headache. Cephalalgia 2010; 30: 1458–1467.

31.

Moncada

Palmer

Higgs

. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 1991; 43: 109–142.

32.

Ramachandran

Ploug

Hay-Schmidt

, et al. Nitric oxide synthase (NOS) in the trigeminal vascular system and other brain structures related to pain in rats. Neurosci Lett 2010; 484: 192–196.

33.

Csati

Tajti

Kuris

, et al. Distribution of vasoactive intestinal peptide, pituitary adenylate cyclase-activating peptide, nitric oxide synthase, and their receptors in human and rat sphenopalatine ganglion. Neuroscience 2012; 202: 158–168.

34.

Medve

. Effects of fixed dose combination of nNOS inhibition and 5HT agonism on progrression of migraine with and without aura. Cephalalgia 2009; 29: 126–126.

35.

Medve

. A phase 2 multicenter, randomized, double-blind, parallel-group, placebo-controlled study of NXN-188 dihydrochloride in acute migraine without aura. J of Headache and Pain 2010; 11: S38–S3S.

36.

Hougaard

Hauge

Guo

, et al. The nitric oxide synthase inhibitor and serotonin-receptor agonist NXN-188 during the aura phase of migraine with aura: a randomized, double-blind, placebo-controlled cross-over study. Scand J Pain 2013; 4: 48–52.

37.

Akerman

Williamson

Kaube

, et al. Nitric oxide synthase inhibitors can antagonize neurogenic and calcitonin gene-related peptide induced dilation of dural meningeal vessels. Br J Pharmacol 2002; 137: 62–68.

38.

Ben Aissa

Tipton

Bertels

, et al. Soluble guanylyl cyclase is a critical regulator of migraine-associated pain. Cephalalgia 2018; 38: 1471–1484.

39.

Little

Doyle

Salvemini

. Reactive nitroxidative species and nociceptive processing: determining the roles for nitric oxide, superoxide, and peroxynitrite in pain. Amino Acids 2012; 42: 75–94.

40.

Shimomura

Kowa

Nakano

, et al. Platelet superoxide dismutase in migraine and tension-type headache. Cephalalgia 1994; 14: 215–218.

41.

Taffi

Vignini

Lanciotti

, et al. Platelet membrane fluidity and peroxynitrite levels in migraine patients during headache-free periods. Cephalalgia 2005; 25: 353–358.

42.

Salvemini

Little

Doyle

, et al. Roles of reactive oxygen and nitrogen species in pain. Free Radic Biol Med 2011; 51: 951–966.

43.

Akerman

Salvemini

Romero-Reyes

. Targeting reactive nitroxidative species in preclinical models of migraine. Cephalalgia 2021; 41: 1187–1200.

44.

Lackovic

Jeevakumar

Burton

, et al. Peroxynitrite contributes to behavioral responses, increased trigeminal excitability, and changes in mitochondrial function in a preclinical model of migraine. The J of Neuroscience 2023; 43: 1627–1642.

45.

Marone

De Logu

Nassini

, et al. TRPA1/NOX In the soma of trigeminal ganglion neurons mediates migraine-related pain of glyceryl trinitrate in mice. Brain 2018; 141: 2312–2328.

46.

Demartini

Tassorelli

Zanaboni

, et al. The role of the transient receptor potential ankyrin type-1 (TRPA1) channel in migraine pain: evaluation in an animal model. J Headache Pain 2017; 18: doi.org/10.1186/s10194-017-0804-4.

47.

Taylor-Clark

. Role of reactive oxygen species and TRP channels in the cough reflex. Cell Calcium 2016; 60: 155–162.

48.

Dux

Will

Vogler

, et al. Meningeal blood flow is controlled by H2 S-NO crosstalk activating a HNO-TRPA1-CGRP signalling pathway. Br J Pharmacol 2016; 173: 431–445.

49.

Borkum

. Migraine triggers and oxidative stress: a narrative review and synthesis. Headache 2016; 56: 12–35.

50.

Gross

Putananickal

Orsini

, et al. Mitochondrial function and oxidative stress markers in higher-frequency episodic migraine. Sci Rep 2021; 11: 4543.

51.

Kozlov

Dietrich

Nohl

. Various intracellular compartments cooperate in the release of nitric oxide from glycerol trinitrate in liver. Br J Pharmacol 2003; 139: 989–997.

52.

Kleschyov

Oelze

Daiber

, et al.

Does nitric oxide mediate the vasodilator activity of nitroglycerin?

Circ Res 2003; 93: e104–e112.

53.

Ivic

Fulop

Juhasz

, et al. Backup mechanisms maintain PACAP/VIP-induced arterial relaxations in pituitary adenylate cyclase-activating polypeptide-deficient mice. J Vasc Res 2017; 54: 180–192.

54.

Kelley

Anderson

Altmann

, et al. Long-term memory of visually cued fear conditioning: roles of the neuronal nitric oxide synthase gene and cyclic AMP response element-binding protein. Neuroscience 2011; 174: 91–103.

55.

Kelley

Anderson

Itzhak

. Pharmacological modulators of nitric oxide signaling and contextual fear conditioning in mice. Psychopharmacology (Berl) 2010; 210: 65–74.

56.

Narayanan

Spack

McMillan

, et al. S-alkyl-L-thiocitrullines. Potent stereoselective inhibitors of nitric oxide synthase with strong pressor activity in vivo. J Biol Chem 1995; 270: 11103–11110.

57.

Guo

Ernstsen

Hay-Schmidt

, et al. PACAP Signaling is not involved in GTN- and levcromakalim-induced hypersensitivity in mouse models of migraine. J Headache Pain 2022; 23: 155.

58.

Brennum

Brinck

Schriver

, et al. Sumatriptan has no clinically relevant effect in the treatment of episodic tension-type headache. Eur J Neurol 1996; 3: 23–28.

59.

Moreno-Ajona

Pérez-Rodríguez

Goadsby

. Gepants, calcitonin-gene-related peptide receptor antagonists. Curr Opin Neurol 2020; 33: 309–315.

60.

Ernstsen

Christensen

Rasmussen

, et al.

The PACAP pathway is independent of CGRP in mouse models of migraine: possible new drug target?

Brain 2022; 145: 2450–2460.

61.

Moye

Tipton

Dripps

, et al. Delta opioid receptor agonists are effective for multiple types of headache disorders. Neuropharmacology 2019; 148: 77–86.

62.

Christensen

SLT

Petersen

Sørensen

, et al. Infusion of low dose glyceryl trinitrate has no consistent effect on burrowing behavior, running wheel activity and light sensitivity in female rats. J Pharmacol Toxicol Methods 2016; 80: 43–50.

Mechanisms of GTN-induced migraine: Role of NOS isoforms,sGC and peroxynitrite in a migraine relevant mouse model

Abstract

Background

Methods

Results

Conclusions

Keywords

Introduction

Methods

Experimental animals

Transgenic mice and genotyping

Experimental design and protocols

Behavioral tests

Cutaneous sensitivity to tactile stimulation

Motor function (Rotarod)

Test compounds

Statistical analysis

Results

NG-nitro-l-arginine methyl ester hydrochloride (l-NAME) blocks GTN-induced hypersensitivity

nNOS may be involved in GTN-induced hypersensitivity

eNOS is not involved in GTN-induced hypersensitivity

iNOS is partially involved in GTN-induced hypersensitivity

GTN-induced hypersensitivity is partially dependent on sgc

GTN-induced hypersensitivity is partly mediated via generation of peroxynitrite

Motor function was unaffected by all combinations of test compounds

Discussion

Feed-forward phenomenon of NO

NO and NOS

sGC

Peroxynitrite

Strengths and limitations

Conclusions

Footnotes

Author contributions

Data availability

Declaration of conflicting interests

Funding

Ethical statement

ORCID iD

References

N^G-nitro-l-arginine methyl ester hydrochloride (l-NAME) blocks GTN-induced hypersensitivity