Solidifying the Link Between Nitric Oxide and Migraine

Two recent studies – one in people, one in animals – build the case for aberrant nitric oxide signaling in the pathophysiology of migraine. Experiments in mice implicate sex differences in mitochondrial function.

Nitric oxide (NO) signaling, when it goes astray, is thought to contribute to migraine, but investigators are still working to understand the extent of its involvement and to uncover the pathophysiological mechanisms at play. A recent human study further implicates aberrant NO signaling in people with migraine, particularly as an initiator of a migraine attack. Meanwhile, new research in mice points the finger at a particular molecule in the NO pathway as a culprit.

The human study, from Michel Ferrari and colleagues at Leiden University Medical Center in the Netherlands, reports decreased interictal levels of L-arginine in the cerebrospinal fluid, but not plasma, of patients who had migraine with or without aura, compared to healthy volunteers. L-arginine is an amino acid whose oxidation produces NO.

That this decrease of L-arginine occurred interictally, that is, outside of a migraine attack, suggests that permanent abnormalities of NO signaling may make people with migraine susceptible to recurrent attacks, according to the investigators.

“[T]his work is significant. It strengthens the hypothesis that nitric oxide metabolism is involved in migraine generation, at least in the CNS (the spinal cord or the spinal trigeminal nucleus),” wrote Karl Messlinger, Universität Erlangen-Nürnberg, Germany, in an email to MSC.

“The greatest strengths [of the research],” Messlinger added, “are the comparably high number of participants and the careful and strict statistical analysis, which is superior to all comparable studies. It is challenging (in terms of logistics and medical routine) to recruit so many patients and healthy volunteers for an invasive method like lumbar puncture,” according to Messlinger, who was not part of the current study.

The mouse study, from Greg Dussor and colleagues at the University of Texas at Dallas, US, shows that neutralizing peroxynitrite (PN), a molecule in the NO pathway, decreased periorbital hypersensitivity in a restraint stress-induced model of migraine, and also weakened the response to sodium nitroprusside, a compound that releases NO.

Accompanying those beneficial changes were improvements in the neuronal hyperexcitability and molecular alterations resulting from the generation of PN. Interestingly, the study found that PN regulated mitochondrial function, but only in stressed male animals.

“I think this is a beautiful piece of experimental research,” said Jon Borkum, a pain and headache expert at Health Psych Maine, Waterville, US, and the University of Maine, Orono, US, who has written about the role of molecules, like NO, that affect mitochondria during migraine but who was not involved with the new research. “I loved the precision of the methods and the precision of the thinking, so certainly this research will advance the field,” said Borkum.

The clinical work appeared in the April 2023 issue of Annals of Neurology, whereas the mouse research was published in the March 1, 2023, issue of the Journal of Neuroscience.

A human study
In their clinical study, Ferrari and colleagues were interested in why people experience migraine attacks in the first place.

“There are all kinds of studies that really don’t address the issue of the initiation of an attack,” Ferrari told MSC.

The investigators were particularly focused on metabolites in the cerebrospinal fluid (CSF). The idea was that if migraine originates in the brain, measuring those metabolites in the plasma would not give the investigators meaningful information about how migraines start. They focused on measuring amines, a chemical class that includes amino acids, neurotransmitters, and hormones, since previous work had linked amine metabolism to migraine.

So the researchers recruited study participants from the LUMINA (Leiden University Medical Center Migraine Neuro Analysis) project. LUMINA includes people with migraine, as well as healthy controls, from the general Dutch population who had agreed to take part in migraine-related research. Some study participants were recruited through a public advertisement or via the Leiden University Medical Center Headache Clinic.

The team would ultimately obtain CSF, collected via lumbar puncture, from nearly 300 study participants, age 18 to 69 years, over a 10-year period beginning in April of 2008. The study sample included CSF from 99 people with migraine with aura, and 98 who had migraine without aura. Notably, the study also included 96 healthy volunteers, overcoming a limitation of previous studies where “healthy” control CSF was not truly healthy. That’s because those samples usually came from individuals with neurological symptoms who had undergone lumbar puncture so that doctors could rule out neurological disease.

Importantly, considering their interest in understanding attack initiation, the group obtained the CSF samples during the interictal period, from patients who had been free of an attack for at least three days. In addition, in order to avoid another hurdle of prior research, they collected the samples between 8:30 in the morning and 1:00 in the afternoon, in random and interchangeable fashion. This was to avoid any influence of diurnal or seasonal influences on amine levels.

In terms of outcomes, the investigators examined levels of individual amines, global amine profiles, and specific amine pathways. To measure amine levels, they used ultraperformance liquid chromatography mass spectrometry, which is a chemistry technique that allows for measurement of metabolite levels in plasma and CSF.

The results are in
The team detected 30 amines in the CSF, and 31 in the plasma, overall. Comparisons of amine levels between the different study groups revealed 10.4% lower levels of one amine, L-arginine, in the CSF of people with migraine with aura, and 5% lower levels of L-arginine in the CSF of those with migraine without aura, compared to healthy controls.

The authors say such differences may seem small but are actually quite meaningful, considering that the amine measurements were not only taken interictally, a time when there are no symptoms, but also taken from the whole CSF.

“If migraine is a focal disorder – it starts somewhere in the brain, presumably the hypothalamus, though we don’t know for sure – and there is a biochemical dysregulation only there in the brain, then that dysregulation will be diluted in the CSF. So to us, that we were able to measure this [at all] is very supportive of a real abnormality” in those with migraine, according to Ferrari.

There was no significant difference in CSF L-arginine levels between those with migraine with aura and those who had migraine without aura. Nor were there any differences between migraine and control groups when measuring amine levels in plasma.

Since L-arginine may be related to cardiovascular status, the researchers examined their results after exclusion of 12 people who had possible cardiovascular comorbidities, and 20 people who were taking blood pressure medication as a migraine preventative treatment. But the differences in CSF L-arginine levels between migraine and control groups remained significant even after this adjustment.

Next, rather than looking at levels of individual amines, the investigators took a more global approach where they measured all amines in the CSF, which resulted in a metabolic profile that they could compare between groups. Here, consistent with their earlier findings, they found significant differences in global metabolic amine profiles between both migraine groups versus healthy controls. And, as with the individual amines, there was no difference in global profiles between the two migraine groups. All of these results held true with regard to global amine levels in the plasma.

Finally, the researchers used a computational tool, called genome-scale metabolic modeling, which provides a mathematical simulation of an organism’s metabolism, including metabolites and related genes and proteins. Using this tool, the researchers could test for associations in their CSF samples between migraine and specific metabolic pathways. Here, the most notable result was an association with pathways related to L-arginine metabolism.

It’s all about NO
Ferrari said that he and his colleagues’ new findings, when considered along with research showing that NO or NO donors (substances that release NO) cause migraine attacks only in people with migraine but not in healthy controls, make a convincing case for aberrant NO signaling in migraine.

“The two [lines of research] combined are very strongly suggestive that nitric oxide dysregulation is primary in migraine,” he said.

One surprise to the investigators was that there was no association between migraine and metabolites of glutamate, an excitatory neurotransmitter that has previously been linked to migraine.

“We ourselves have done all kinds of studies suggesting that glutamate is important in migraine, but we didn’t find that here. We don’t know whether that is because glutamate metabolites are diluted in the CSF, since glutamate is a very rapidly metabolized amine. So it could be that levels are still abnormal, but we couldn’t measure it.”

So what’s next in this line of research?

“I think the obvious next step is a focused, hypothesis-driven study on more metabolites involved in the nitric oxide pathway. One of the fair criticisms of the paper is that we did not measure all amines involved in the NO pathway. That was because we couldn’t do that and at the same time also measure glutamine and other compounds,” according to Ferrari.

Ferrari told MSC that he and his colleagues are now redesigning their biochemical detection platform to focus more squarely on the NO pathway and all of its metabolites.

Noting that the authors wrote in their paper that “sex was found to be a highly significant factor determining L-arginine levels,” Messlinger said he would have liked to see more information about whether male or female CSF has more L-arginine, considering all the other research in the field trying to understand why migraine is more prevalent in females. Further examination of sex differences in amines appears as an interesting research avenue to pursue.

Finally, one of the fruits of the researchers’ efforts is their large collection of CSF samples taken interictally from individuals with migraine. Interestingly, Ferrari said he is unaware of any other episodic brain disorder, such as epilepsy, where biochemical abnormalities have been identified during the interictal period, which makes their collection of CSF samples during the interictal period of migraine quite unique. Having CSF samples from healthy controls is also an exceptional resource.

“We now have a real goldmine of CSF samples, and we are happy to offer other people in the world a collaboration to use the many CSF samples we have for other biochemical studies,” he said.

A preclinical study
When asked about how the animal study on PN came about, Dussor told MSC that he actually traces the research all the way back to the 19th century.

“When I give talks on this research, I start with the real background where this project came from, which is the Italian chemist at the University of Turin who originally made nitroglycerin over 150 years ago. He was the first person to ever synthesize it.”

Dussor was referring to Ascanio Sobrero, who found that tasting nitroglycerin, which is an NO donor, caused severe headaches. “And the joke is always, of course, that any good chemist has to taste everything that is made.”

Since then, many animal and human studies have implicated NO signaling in migraine, but exactly how it contributes to this condition has remained unclear. “There’s a strong link between nitric oxide and migraine, but we just don’t know what the actual mechanism is,” Dussor said.

The researchers, including first author Jacob Lackovic, hypothesized that focusing on parts of the NO pathway downstream from NO itself – NO first reacts with a compound called superoxide to form PN – might ultimately be a better way to pharmacologically target the pathway for migraine treatment. PN, which has also been linked to pain conditions in animal studies, fit the bill, especially since NO and NO donors, which ultimately give rise to PN, appear to cause headache specifically.

Pointing the finger at peroxynitrite
The researchers began with a preclinical model of migraine in which mice are restrained in their cages, which causes stress – the most common trigger of migraine in people – followed by injection of a normally harmless dose of the NO donor sodium nitroprusside (SNP). The idea is that the stressed animals are “primed” to respond to the NO donor, even though the dose used would not otherwise cause problems.

The team found that animals receiving, before SNP administration, a compound that modulated PN activity showed less facial mechanical hypersensitivity, and less facial grimacing, in response to SNP, compared to control animals receiving a saline solution. The researchers administered the PN-modulating compound 14 days after stressing the animals, but they reported similar results when giving the compound at multiple earlier time points ranging from one hour to three days after stress. They also saw that a single dose of the compound one hour after stress reduced facial hypersensitivity and grimacing in response to SNP.

The findings suggested that the effects of the PN-modulating compound depended on when the animals received it, and that the formation of PN immediately after stress exposure played a critical role.

In contrast, modulating PN had no effects in another migraine model, where mice that received interleukin-6 are primed to respond to application of a synthetic solution, at a pH of 7.0, to the dura. This is a neutral solution that does not normally cause hypersensitivity, except in mice primed to respond to it.

“Essentially, the areas where we saw the effects were only in response to stress itself and in response to an NO donor. And that just tells us that those conditions are more likely to be generating peroxynitrite, and the behavioral outputs are more likely to depend on peroxynitrite,” Dussor said.

Peter Grace, a pain researcher at MD Anderson Cancer Center, Houston, US, told MSC that a role for PN only in the restraint stress model is a finding that stands out to him.

“That’s an interesting observation. Maybe these models are modeling different mechanisms that might be present to different extents in human patients. It’s possible that peroxynitrite might play a role in a subgroup of patients, if we were to extrapolate from this data,” according to Grace, who was not involved with the study.

Where is it? What is it doing?
Next, the team asked where in the nervous system PN played an active role. The results showed an increase of a marker of PN activity in both the trigeminal (TG) ganglia and dura of stressed mice, compared to controls. This was also the case in response to SNP in the stressed animals. Further, a PN-modulating compound administered after stress lessened these increases.

Considering that PN changes the activity of proteins and essential enzymes, which could lead to changes in the electrical excitability of neurons, the team also examined the electrical properties of TG neurons in cell culture.

Here, they saw increases in the excitability of the TG neurons in response to an NO donor. But treatment with a PN-modulating compound prevented the increased excitability. Whether PN would affect the electrical excitability of human neurons is an open question.

A sex difference in the functioning of mitochondria
Since NO and molecules derived from it can change the functioning of mitochondria, the researchers then tested whether this held true in their restraint stress model of migraine, and whether PN could affect the workings of the “powerhouses of the cell.”

To do so, the group harvested and cultured TG neurons from stressed male and female animals. Twenty-four hours after three days of restraint stress, the TG cell cultures from male and female mice showed significant increases in maximal respiration levels, a parameter of mitochondrial function.

But, interestingly, 14 days after stress, there were increases in spare respiratory capacity – a measure of mitochondrial function that reflects how the cell adapts to environmental stressors – but only in stressed female mice. This, the researchers argue, suggests that female TG mitochondria are better able to adapt to stress than males. It also may explain another finding: SNP given two weeks after stress only affected mitochondrial function in male, but not female, TG mitochondria.

“If female mitochondria, specifically in the trigeminal ganglia, are better able to adapt to changes in the environment around them, then that would make sense of why SNP did not have as robust of an effect on them,” said Lackovic.

PN appeared to play a role in the sex-specific changes in the mitochondria. That’s because a PN-modulating compound administered before the mice received SNP alleviated the changes in mitochondria caused by SNP in stressed male animals.

The findings about the mitochondria were surprising at first to the researchers, but not so surprising upon a second glance.

“We were surprised about the sex differences, until we actually went and looked in the literature. There’s a fairly robust literature about how mitochondria have different properties in males and females,” Dussor said.

The sex differences in mitochondrial function were of note to Grace.

“The increase in bioenergetic function in males – they are more sensitive to peroxynitrite decomposition – jumped out at me because it went in the opposite direction from what I might have initially predicted; certainly in a lot of the chronic pain models, mitochondrial energetics is impaired, and here they show that it’s increased. We don’t yet know what the functional significance of that is, and I think that is something that needs to be explored.”

A promising target?
There is good reason to hypothesize that PN might make an attractive drug target for migraine therapy. Perhaps the most promising thing about PN is that, unlike NO, it seems to be involved primary in pathophysiological changes rather than in the normal physiological functioning of cells.

“If we were able to go at something downstream of nitric oxide that was really contributing to a more specific pain-like response or a neuronal-sensitization response, but didn’t impact the normal functions of nitric oxide, then that would clearly be an advantage over preventing the production of nitric oxide in the first place,” Dussor said.

Combining a PN-modulating drug with other drugs that target other components of the NO pathway, such as an inhibitor of the NO receptor, might be the most promising way to go, Lackovic said.

“Certainly, blocking peroxynitrite is not going to be an all or nothing thing. I think we’ll have to have multiple targets” from a drug development perspective, Lackovic said.

Targeting PN would be a different way to treat migraine, since current drugs like calcitonin gene-related peptide (CGRP) inhibitors and triptans affect different pathophysiological mechanisms. Potentially, some of these approaches could be combined, according to Dussor.

But first things first: Researchers will need to understand if the current study results apply to people, and whether targeting PN would be a safe pharmacological approach to treat them.

There are also important questions for animal researchers to address, Dussor said.

“What part of the migraine pathology is nitric oxide contributing to? We still don’t really know. And is nitric oxide an equal contributor in all forms of migraine?”

Neil Andrews is a science journalist and executive editor of the Migraine Science Collaborative. Follow him on Twitter @NeilAndrews

References
Cerebrospinal fluid and plasma amine profiles in interictal migraine.
Onderwater et al.
Ann Neurol. 2023 Apr;93(4):715-728.

Peroxynitrite contributes to behavioral responses, increased trigeminal excitability, and changes in mitochondrial function in a preclinical model of migraine.
Lackovic et al.
J Neurosci. 2023 Mar 1;43(9):1627-42.