Is there a role of calcitonin gene-related peptide in cortical spreading depression mechanisms?– argument con.

Abstract

Cortical spreading depression (CSD) is a wave of neuronal and glial depolarization followed by suppressed neural activity, thought to underlie migraine aura. While Calcitonin Gene-Related Peptide (CGRP) is well established in migraine pathophysiology, its role in CSD remains uncertain. This comment evaluates evidence suggesting that CGRP is not directly involved in CSD initiation or propagation but may contribute to nociceptive activation associated with migraine. While some studies report CGRP-related effects on CSD susceptibility, methodological limitations raise concerns about their interpretation. Electrophysiological data indicate that CGRP does not influence the ionic mechanisms driving CSD. However, CGRP plays a key role in sensitizing nociceptive neurons, and CGRP-targeting drugs effectively modulate migraine pain without altering CSD dynamics. Clinical findings further suggest that peripheral CGRP inhibition reduces headache burden, potentially allowing the brain to recover from chronic pain states. In conclusion, while CGRP is integral to migraine pain modulation, its direct involvement in CSD appears minimal, highlighting distinct pathways for aura and headache pathophysiology.

Cortical spreading depression (CSD) is a wave of neuronal and glial depolarization followed by a period of suppressed neural activity, and is thought to be the phenomenon that underlies the migraine aura. While Calcitonin Gene-Related Peptide (CGRP) has a well-recognized role in migraine pathophysiology, its role in CSD remains controversial.

As reviewed in detail below, CSD produces activation of the nociceptive sensory pathway that is thought to be responsible for the initiation of the migraine headache, and this CSD-induced nociceptive activation is dependent on CGRP. There are two potential sites of action for CGRP in this process: it could affect the sensitivity of the nociceptive neurons, or it could affect the CSD wave itself. Below, we present the argument that the role of CGRP in CSD-induced nociceptive activation is through an action on the nociceptive neurons and not on the CSD wave itself.

To date, there is limited evidence to support the involvement of CGRP in CSD. However, some observations suggest its potential role. For example, Tozzi et al. [1] reported that blocking CGRP receptors inhibited CSD initiation, while Wang et al. [2, 3] found that intraventricular perfusion of anti-CGRP antibodies in rats reduced susceptibility to CSD. Tozzi et al. [1] also observed the release of cortical CGRP when the cortex was exposed to elevated potassium. Similarly, Wang et al. [2, 3] demonstrated that repeated CSD events upregulated CGRP gene expression and increased peptide levels in multiple brain regions. While these findings suggest a possible role for CGRP in CSD, limitations inherent to the methodologies used in animal studies may amplify the reported effects. For instance, in slice preparations, repeated CSD waves are often induced. This poses a challenge, as CSD recovery requires several minutes to hours, and the repetition of CSDs may artificially amplify CGRP-related effects. In contrast, in humans, a single wave of CSD is sufficient to trigger aura symptoms [4, 5]. Additionally, some experimental protocols expose brain tissue to vehicles like DMSO (Dimethyl sulfoxide), used to dilute CGRP antagonists, which can be toxic to neurons. Missing appropriate controls for these vehicles may confound results. Despite these limitations, the evidence suggests CGRP could play a minor role in CSD, and so might have some influence on the perception of CSD (e.g., aura symptoms) in patients.

Electrophysiological studies further support the notion that CGRP is not directly involved in CSD. CSD initiation and propagation are driven by ionic fluxes, particularly sodium and calcium influxes and potassium effluxes, resulting in a massive depolarization wave. To the best of our knowledge, CGRP does not directly modulate these ionic movements.

While the evidence supporting CGRP’s involvement in CSD is relatively sparse, there is a large body of evidence that CGRP exert strong actions on the sensory neurons that have been implicated in the initiation of the migraine headache. CSD is widely accepted as the physiological basis of migraine aura and can activate peripheral and central trigeminovascular neurons [6–8], leading to sensitization of second-order neurons in the spinal cord. In our research, we used CSD to activate the trigeminovascular system and test migraine treatments such as atogepant and fremanezumab (a CGRP receptor antagonist and a monoclonal antibody against CGRP, respectively) [9–12]. Across these studies, CSD was reliably induced regardless of drug administration. However, peripheral CGRP appears to play an important role in migraine mechanisms. Atogepant and fremanezumab blocked CSD-induced activation of Aδ neurons in the periphery and High Threshold (HT) neurons centrally. Interestingly, atogepant also affected Wide Dynamic Range (WDR) neurons, possibly due to its broader effects on C fibers (potentially mediated via amylin receptors, as CGRP antagonists are not entirely selective). Importantly, systemic CGRP antagonists and monoclonal antibodies act peripherally, as demonstrated by our findings and supported by other groups [13–15]. It is important to mention that all the cited studies used peripheral (intravenous or intraperitoneal) injections, meaning the observed effects are driven by peripheral mechanisms.

After these series of results, it was important to test the effect of fremanezumab on the CSD itself, so in a subsequent study we measured CSD characteristics: amplitude, spreading velocity and recovery period [16]. For this study, consistent with our previous studies involving fremanezumab, we included appropriate controls to account for the specificity of the monoclonal antibody. Thus, we tested saline, fremanezumab and an isotype control antibody (an IgG antibody targeting a protein not present in rats). Surprisingly, fremanezumab reduced the recovery period and spreading velocity to the same extent as the isotype control, suggesting a non-CGRP-dependent effect. In a following set of experiments [17], we tag fluorescently the fremanezumab in an effort to figure out if it would cross the blood brain barrier. Given the methodology used in our experiments (electrocorticograms) it is possible that the Blod Brain Barrier (BBB) might be broken at the area where we implant the electrodes. The results showed that fremanezumab actually reaches cortical areas surrounding the electrode. If CGRP were involved in the effects we observed earlier it should have only happened with the fremanezumab and not with the isotype. In my opinion this give us the possibility that either the CGRP does not play any role in CSD or the possibility that its role might be too small to actually create an effect that is measurable by this methodology. It is important to note that all of our experiments induced a single wave of CSD, mimicking human aura conditions, where one wave is sufficient to trigger symptoms. Models using repeated CSD waves may observe amplified effects due to methodological differences.

We also need to take into consideration that for all our animal studies we are using naïve animals (non migraineurs rats), which I think is very important if we consider the following: Clinical data shows that migraine actually evolves over the years that the patients suffer from it. So, it is possible that the continuous and prolonged nociceptive events that the patients experience are actually modifying the brain and the pathways related to such nociception, creating what we know as the “migraine brain”. Now it is true that this migraine brain might be something that patients are born with, thus predisposing some patients to experience migraine in their lives, but for other patients, it might be something that develops through the years. In a subset of these patients the auras can also evolve, becoming longer, and becoming accompanied by other sensory symptoms, not only visual. In some cases, patients who initially experience migraine without aura may later develop aura symptoms. When we tested galcanezumab for all the triggers and prodromal symptoms that patients experience, we included the aura symptoms and to our surprise, those aura symptoms changed. Either the patients experienced the aura but did not get a headache, or the aura was shorter or the scotoma was smaller in some cases. But what was common to all the patients in which their aura was modified was the fact that the headache also diminished during the 3, 6 and 12 months of treatment. The patients that experienced these changes in aura were all responders to the treatment (exhibited reduction in headache), and we concluded that it is very likely that by modifying the periphery with the anti-CGRP drug, the patients experienced fewer attacks, allowing the brain to actually recover from years of pain, and when that happened not only the aura changed but also many prodromes and triggers [18, 19]. Related to this is the work by our group where we used high-resolution magnetic resonance imaging before and after treatment with galcanezumab and showed that those who responded to the treatment had a decreased cortical thickness (compared to pre-treatment baseline), in regions of the somatosensory cortex, anterior cingulate cortex, medial frontal cortex superior frontal cortex and supramarginal gyrus. We interpret the cortical thinning seen in the responder group as suggesting that reduction in head pain could lead to changes in neural swelling and dendritic complexity and that such changes reflect the recovery process from maladaptive neural activity, which is supported by the previous studies mentioned before in which the incidence of premonitory symptoms and prodromes decreased [20].

Since the antibodies do not enter the brain, these findings represent a peripheral action, and show that we might not need to modify CGRP centrally to produce an effect on the CSD (aura), as this effect could occur secondarily to the reduction in headache.

In summary, the available evidence from electrophysiological, pharmacological and clinical studies collectively suggests that CGRP might not play a significant role in the mechanisms of cortical spreading depression. While CGRP is undoubtedly involved in the modulation of migraine-related pain and vascular changes, its involvement in the core processes of CSD initiation and propagation might be minimal. Therefore, targeting CGRP may alleviate migraine symptoms without directly affecting the occurrence or characteristics of CSD, underscoring the distinct pathways involved in these phenomena.

Response to Romozzi. M & Calabresi P. The Journal of Headache and Pain 2025 [21].

I would like to begin by expressing my appreciation for the extensive review and thoughtfulness with which Professors Romozzi and Calabresi have presented their perspective on the role of CGRP in the mechanisms underlying cortical spreading depression (CSD) and migraine aura. Their discussion of the experimental and clinical evidence, ranging from animal models to human provocation studies, provides a robust backdrop for understanding the complex interplay between neuronal, vascular, and nociceptive systems in migraine.

In addressing their arguments, I would like to clarify several points where our interpretations diverge, particularly regarding the direct involvement of CGRP in CSD propagation.

Experimental evidence and methodological considerations

The authors present data from in vitro and in vivo models where CGRP release is observed during CSD and where CGRP receptor antagonists appear to modulate certain electrophysiological or vascular responses [1–3]. While these findings are compelling, it is important to recognize that many of these studies involve experimental conditions—such as repeated CSD events in brain slices or conditions of potentially compromised blood–brain barrier integrity—that may not accurately mirror the single, isolated CSD wave associated with the human migraine aura. The work from our lab emphasizes that in models mimicking the human condition (i.e., a single CSD wave), the influence of CGRP on the core ionic events that drive CSD propagation appears minimal. Repeated or exaggerated CSD induction in some animal studies may amplify CGRP-related effects, thereby leading to interpretations that might not be directly transferable to clinical phenomena.

Peripheral versus central actions of CGRP

The evidence presented by Drs. Romozzi and Romozzi also suggests that CGRP, when released in the cortex, might influence neural activity and vascular tone in a manner that could feed back into CSD mechanisms. However, our data [7–10, 14, 15]—particularly those derived from studies using peripheral administration of CGRP antagonists or monoclonal antibodies—consistently indicate that while these agents effectively attenuate migraine-related nociceptive activation, they do not significantly alter the electrophysiological parameters of CSD itself (e.g., amplitude, propagation velocity, recovery period). Given that the anti-CGRP drugs under discussion have limited blood–brain barrier penetration, it is plausible that their primary site of action is peripheral. This supports the view that the observed clinical benefits, including changes in aura characteristics, are more likely secondary to reductions in peripheral nociceptive input and subsequent central modulation, rather than a direct modulation of the CSD phenomenon by CGRP.

Clinical correlations and the “migraine brain”

Clinical observations further complicate the interpretation of CGRP’s role. Although anti-CGRP monoclonal antibodies and gepants have shown efficacy in reducing both headache and aura frequency, it is crucial to distinguish between direct effects on CSD and indirect effects mediated through a decrease in peripheral nociceptive signaling. In our clinical studies, modifications in aura characteristics were temporally associated with reduced headache severity over months of treatment. This suggests that the therapeutic modulation of CGRP may allow the “migraine brain” to recover from chronic nociceptive bombardment rather than directly preventing the initiation or propagation of CSD. Moreover, our findings that changes in cortical thickness correlate with headache reduction further support the hypothesis that the beneficial effects of CGRP modulation are mediated by peripheral actions leading to central recovery, rather than by an intrinsic suppression of CSD mechanisms [16–18].

Synthesis and future directions

While the evidence reviewed by Drs. Romozzi and Calabresi provides a strong argument for a role of CGRP in the context of migraine pathophysiology—including a potential contribution to CSD-related vascular changes—the preponderance of data from studies that mimic human migraine more closely suggests that CGRP’s primary impact is exerted on the trigeminovascular system and peripheral nociceptive pathways. It is my opinion that the influence of CGRP on the initiation and propagation of the CSD wave, if present, is likely too subtle to be detected by the methodologies currently employed in human-relevant models.

Nevertheless, these differences in interpretation underscore the need for further nuanced research. Future studies employing models that faithfully replicate human migraine conditions, alongside advanced imaging and electrophysiological techniques, may help reconcile these observations and more clearly delineate the distinct yet interconnected roles of CGRP in both central and peripheral mechanisms.

In summary, while we acknowledge the evidence that supports CGRP release during CSD and its modulatory effects on vascular and nociceptive pathways, our data and interpretation lead us to conclude that CGRP does not play a significant, direct role in the core electrophysiological mechanisms of CSD. Rather, its predominant effects appear to be peripheral, influencing the subsequent activation of nociceptive circuits that underlie migraine headache. I look forward to continued collegial dialogue on this important topic as I definitely believe that is a clear path for our field to move forward.

Author contributions

AMC wrote the comment.

Funding

There was no funding for this comment. Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics declaration

Not applicable.

Consent for publication

not applicable.

Competing interests

The authors declare no competing interests.