Is there a role of calcitonin gene-related peptide in cortical spreading depression mechanisms?– Argument pro

Migraine ranks as one of the most debilitating medical conditions worldwide, characterized by debilitating headaches associated with a plethora of accompanying symptoms [1, 2]. Approximately one-third of the patients with migraine experience aura, a reversible neurological phenomenon that manifests with visual, sensory, speech, and motor neurologic symptoms, usually lasting 5–60 min, either preceding or accompanying headache pain [3]. Cortical spreading depression (CSD), originally described by Leão in 1944 in a rabbit model, is a wave of depolarization of neuronal and glial cells spreading across the cerebral cortex at a rate of 2–5 mm/min, followed by a long period of hyperpolarization, has been proposed as the pathophysiological mechanism of migraine aura [4].

The depolarization typical of CSD is associated with local ionic shifts and neurotransmitter and metabolite release [5]. This includes the massive increases in extracellular potassium, intracellular sodium and calcium, and glutamate release [5]. Among the neurotransmitters, glutamate is prominent in initiating and propagating CSD by activating N-methyl-D-aspartic acid (NMDA) receptors [6, 7]. Vascular reactivity and blood flow modifications follow the change in neuronal excitability, with an increase in regional cerebral blood flow for 3–5 min, followed by hypoperfusion (spreading oligemia) lasting for approximately 1–2 h [8, 9].

Experimental evidence shows that CSD results in the activation of trigeminovascular neurons in a temporal way that resembles the onset of headache after aura in patients [10]. This suggests that activation of the trigeminovascular system (TGVS) is crucial for initiating migraine attacks [10]. The activation leads to the release of neuropeptides, such as calcitonin gene-related peptide (CGRP) and pituitary adenylate cyclase-activating polypeptide (PACAP). The resultant arteriolar vasodilation, neurogenic inflammation, and mast cell degranulation may contribute to the activation or sensitization of dural nociceptors [11,12,13]. Evidence of CSD, as the underlying mechanism of aura, was shown in humans by functional magnetic resonance imaging (fMRI) [14, 15]. However, the existence of a relationship between CSD mechanisms and the subsequent head pain remains debated [13].

Calcitonin gene-related peptide, discovered in 1982, is a 37-amino acid vasodilator peptide that belongs to the calcitonin peptide family and is produced in peripheral sensory neurons and multiple sites throughout the CNS [16], that has been regarded as a target for the development of acute and prophylactic treatment for migraine [17]. It is the most abundant neuropeptide within the TGVS, where it is expressed in 50% of neurons in the human trigeminal ganglia [18].

The first evidence supporting the involvement of CGRP in the pathogenesis of migraine dates back to 1990, when elevated CGRP levels were observed during spontaneous migraine attacks in the external jugular blood, into which extracerebral tissues drain [19]. In 1993, it was discovered that stimulation of the trigeminal ganglion in cats resulted in the release of CGRP [20]. Both human and animal studies have shown that CGRP levels were significantly reduced by sumatriptan, in parallel with the relief of migraine symptoms in humans [20]. Additional evidence that CGRP has a role in migraine derives from clinical studies in which the infusion of exogenous CGRP could trigger migraine attacks in patients prone to migraine, suggesting a causative role of this peptide [21, 22].

More recently, several studies have confirmed elevated levels of CGRP in plasma, serum, saliva, tear fluid, and cerebrospinal fluid (CSF) in ictal and interictal phases in migraine patients compared to healthy controls [23, 24].

Lastly, the strongest ex adiuvantibus implication of CGRP in migraine comes from the development of treatments targeting the CGRP pathway, including monoclonal antibodies against the CGRP receptor or ligand (anti-CGRP mAbs) and small molecule antagonists of the CGRP receptor (i.e., gepants), that have shown safety and effectiveness as prophylactic and acute therapy for of migraine [25]. The role of CGRP in CSD mechanisms has been a matter of debate.

Relation between CSD and CGRP in animal models

The hypothesis that CGRP is only involved in the pain phase of migraine is reductive. There are several preclinical lines of evidence to support the involvement of CGRP in CSD processes and migraine aura, as well as the coupling of CSD with the activation of the TGVS in migraine pathophysiology with the subsequent release of CGRP from peripheral terminals [26, 27]. These hypotheses are supported by the loss of CSD-dependent neurogenic inflammatory response observed after sensory denervation of the meninges [28].

CGRP could also be released centrally in the cerebral cortex, influencing neural activity and vascular tone [29,30,31]. Tozzi and colleagues demonstrated that, in rat brain, endogenous CGRP was released in a calcium-dependent manner during potassium-induced CSD [6]. Three CGRP receptor antagonists (MK-8825, olcegepant, and CGRP 8–37) determined a dose-dependent inhibitory effect on CSD in neocortical slices, suggesting a pivotal role of CGRP in this event [6]. Interestingly, in this study, CSD was also blocked by NMDA but not AMPA receptor antagonists. Moreover, this phenomenon was also inhibited by topiramate but not carbamazepine [6]. Similarly, preliminary evidence showed that the systemic administration of olcegepant to mice in vivo inhibited repetitive CSD and altered the vascular response to CSD [32]. Fremanezumab, given intravenously in anesthetized rats, prevented the activation and sensitization of high-threshold trigeminovascular neurons by CSD [33]. The intraperitoneal application of MK-8825 in awake rats with intact blood-brain barrier (BBB), instead, failed to block CSD waves in the cortex and the CSD-associated hemodynamic changes but reversed the CSD-induced behaviors associated with pain and blocked the neuronal activation in the spinal trigeminal nucleus [34]. Lastly, exogenous CGRP in mice brain slices reversed the effects of prolonged CSD latency caused by anti-CGRP antibodies [35].

This data supports the hypothesis that CGRP antagonists could act directly or indirectly at a central level [6].

CSD also seems to influence CGRP expression and synthesis. Repeated experimental-induced CSD events increased CGRP mRNA and peptide levels at 24 h post-CSD in the rat cortex ipsilaterally [29]. In agreement, induced CSD in rats significantly enhanced CGRP expression in both protein and mRNA levels in the trigeminal ganglion [36, 37]. In a mouse model of migraine, CSD was also able to induce substantial changes in CSF composition, with 11% of the CSF proteome being altered in concentration, with up-regulation of proteins involved in the activation of receptors in the trigeminal ganglion [38]. Among these ligands, CGRP doubled in concentration [38]. The authors proposed that the CSF transport of CGRP to the trigeminal ganglion could be directly involved in the development of migraine headache [38].

An elevation of CGRP synthesis in association with CSD might contribute to an increased susceptibility to migraine through a positive feedback, where CSD triggers CGRP release and synthesis, thereby increasing the probability of subsequent CSD and migraine events.

CGRP may also mediate the coupling between neuronal activity and cerebral hyperemia observed during CSD [30, 31]. In a rabbit and cat model of CSD, the topical administration of an inhibitor of the CGRP receptor, [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37], reduced CSD-induced pial dilation [30, 31]. In accordance with these findings, brain topical application of a CGRP receptor antagonist also decreased hyperperfusion associated with CSD [39]. These results suggest that CGRP may act as a vascular modulator of CSD, and the local release of CGRP in the meninges could contribute to CSD-induced dilation. In this view, there could be a bidirectional relationship where modifications in vascular tone can modulate neuronal activity, namely vascular-neuro coupling [40].

Human migraine models

Provocation experiments have demonstrated that intravenous infusion of CGRP can induce migraine-like attacks in patients with migraine, with and without aura [41]. Even to a lesser extent, CGRP provocation experiments have also been conducted in patients with aura [41]. In patients who experienced only migraine attacks with typical aura, CGRP infusion triggered migraine-like attacks without aura in 57% of patients, with four (28%) reporting typical aura symptoms [42].

In an open-label, single-arm, non-randomized trial, CGRP was administered intravenously to 34 patients with migraine with aura. Thirteen (38%) of 34 participants developed an attack of migraine with aura after CGRP infusion [43]. The authors hypothesized that CGRP acts on meningeal arteries, initiating the transmission of nociceptive information that reaches the somatosensory cortex and other cortical/subcortical areas implicated in the perception of migraine pain. In susceptible individuals, excitatory stimuli reaching regions like the visual cortex may reach the threshold for triggering CSD [43].

Altogether, these findings demonstrated that the majority of patients with migraine, with or without aura, developed delayed migraine-like attacks after the infusion of CGRP, including migraine aura episodes, indicative of a CGRP-related common cascade originating cephalic pain and aura in both phenotypes of patients.

Clinical evidence of the effectiveness of anti-CGRP mAbs in migraine aura

An indirect evidence of the role of CGRP in CSD is the effect of anti-CGRP mAbs in reducing the frequency and intensity of aura episodes. Anti-CGRP drugs show similar effectiveness in patients both with and without aura because they reduce the number of migraine attacks with aura in parallel with the reduction of migraine attacks without aura in both real-world studies and post-hoc analyses of clinical trials [44,45,46,47,48].

In post hoc analyses of clinical trials, the effectiveness of galcanezumab, eptinezumab, and erenumab was similar in patients with and without aura [46,47,48].

In other real-world prospective and retrospective studies or case series of patients with a diagnosis of migraine with aura treated with anti-CGRP mAbs, a reduced incidence, intensity, and duration of aura was observed regardless of the anti-CGRP mAb used or the type of migraine (episodic or chronic migraine) [44, 49, 50], with some studies showing a reduction in the number of auras regardless of responder status [51] or the complete disappearance of aura [52, 53].

A recent case series described the effectiveness of galcanezumab in patients with sporadic and familial hemiplegic migraine, with improvement in weakness symptoms in all patients except for two [54]. In line with these results, mice expressing a mutation in SCN1A, which is associated with familial hemiplegic type three in humans, exhibited spontaneous CSDs that propagate from the visual to the motor cortex [55]. CGRP and CGRP receptors are located throughout the CNS. Due to their dimension and peptidic nature, anti-CGRP mAbs and gepants poorly cross the BBB due to their molecular size [16, 56]. In rats, 0.1–0.3% of the plasma concentration of galcanezumab was found in the CNS [57].

Their action is thought to be through a peripheral site of action, and whether the action of CGRP blockade in migraine is mediated via central or peripheral mechanisms remains unclear [58]. The trigeminal ganglion has always been considered to lack BBB; therefore, it was thought it could represent the targets of gepants and anti-CGRP mAbs [16].

However, a recent paradigm shift study demonstrated that CSF carries solutes from the cortex to the trigeminal ganglion, facilitating nonsynaptic communication between the CNS and the periphery [38]. Furthermore, these observations are still preliminary, and it has not been elucidated whether the BBB remains intact during the course of spontaneous migraine attacks with aura or if it is uniformly tight throughout the brain, as specific regions may be less protected by the BBB than others [59].

Thus, if aura originates in the cortex, a direct action of anti-CGRP mAbs at this level remains to be demonstrated; alternatively, the inhibition of CGRP or CGRP receptor in the periphery could indirectly influence brain functioning, including CSD.

Response to Melo-Carrillo A. The Journal of Headache and Pain 2025 [60]

We appreciate the author’s impressive and thorough review of the literature on this topic and the effort in supporting the “cons” hypothesis. We acknowledge the limitations in reproducing CSD in animal models that the author raised.

However, it remains challenging to dismiss the role of CGRP in CSD entirely, as evidence from electrophysiological, in vivo, and in vitro models suggests that CGRP may have a dual role, both centrally and peripherally [6]. This is further supported by the fact that CGRP antagonism may disrupt neurovascular coupling in CSD to the degree that it could reduce or even prevent the occurrence of neuronal and/or vascular events [30, 31].

In summary, in vitro studies demonstrate that endogenous CGRP was released during potassium-induced CSD and CGRP receptor antagonists inhibited CSD initiation [6], and in vivo studies revealed that topical administration of CGRP receptor antagonists reduced CSD-induced pial dilation and hyperperfusion associated with CSD [30, 31].

CGRP could also have an excitatory effect itself. Gimeno-Ferrer and colleagues demonstrated that topical application of CGRP in rats in vivo triggered episodes of local ictal discharge activity related to CSD that the CGRP receptor antagonist BIBN4096BS prevented. Similarly, in vitro recordings from slices of mouse cortex showed that the application of CGRP evoked periods of synchronized activity [61]. In the same study, topical application of CGRP to rat cortex induced plasma extravasation [61].

Such excitatory effects of CGRP, producing repetitive action potentials, were also observed in other studies. Ryu and colleagues, in the immature rat in vitro spinal cord slice-dorsal root ganglion preparation, showed that application of CGRP produced a slow reversible depolarization in about one-third of the cells. The CGRP-evoked depolarization was associated with enhanced excitability in most neurons tested [62]. Similarly, in rat brain slices, Han and colleagues showed that CGRP increased excitatory postsynaptic currents in the amygdala and increased neuronal excitability. An NMDA receptor antagonist reversed CGRP-induced synaptic facilitation [63]. In vitro and in vivo models of non-migraine pain showed that CGRP receptor blockade in the amygdala directly inhibited NMDA-evoked, but not AMPA-evoked, membrane currents [64].

Altogether, these data suggest an important role of CGRP both in the regulation of neuronal excitability, possibly favoring CSD, and in the regulation of brain blood flow [61].

Furthermore, these findings reinforce the connection between central CGRP and the modulation of NMDA receptors and glutamate, as also demonstrated by Tozzi and colleagues, who showed that NMDA receptor antagonism inhibits CSD [6]. These mechanisms collectively enhance neural signaling in the cortex and mutually potentiate their effects [64]. Notably, glutamate, a key excitatory neurotransmitter, plays a fundamental role in CSD susceptibility, initiation, and propagation [65]. CGRP, functioning among other roles as a glutamatergic co-transmitter, contributes to sustaining dysfunctional neuronal excitation within the CNS, an event that is likely to occur in the brains of individuals with migraine [64].

A second key strength for CGRP implication in migraine CSD/aura is still the clinical evidence supporting the effectiveness of anti-CGRP mAbs in migraine aura. Indeed, anti-CGRP mAbs can be effective in reducing migraine with aura in parallel with the reduction of the number of migraine attacks [44,45,46,47,48]. However, increasing evidence reports a reduction in the number of auras regardless of the reduction of migraine days [51], or the complete disappearance of aura [52, 53], suggesting an independent and direct effect on the aura, which may not be mediated necessarily by a peripheral action.

Similarly, along with aura, findings on a positive effect of anti-CGRP mAbs on prodromal and accompanying symptoms have been reported in the literature, alongside aura frequency reduction [50, 66]. Some accompanying symptoms, such as photophobia or phonophobia, seem to be associated with higher levels of CGRP during migraine attacks [67].

In addition, intracerebroventricular administration of CGRP in transgenic mice (nestin/hRAMP1) caused a significant increase in light aversion prevented by CGRP receptor antagonist olcegepant [68].

These symptoms clearly originate within central structures, including the cortex, thalamus, brainstem, and hypothalamus [69]. However, as these regions lie beyond the BBB, the mechanism by which anti-CGRP mAbs, acting hypothetically at meningeal trigeminal afferents, prevent the development of migraine symptoms arising from the CNS, including aura, remains unclear. Furthermore, these symptoms frequently precede the activation of nociceptive trigeminovascular fibers and the subsequent migraine pain phase.

Comparably, other studies have shown neuroimaging and neurophysiological parameters changes with anti-CGRP mAbs [70, 71]. For example, Ziegeler and colleagues demonstrated through fMRI that erenumab led to a decrease in the activation of a specific network following trigeminal nociceptive input, including the secondary somatosensory cortex, the thalamus, and the insular cortex [70]. While a significant reduction of hypothalamic activation was found only in patients responders to erenumab [70].

These findings further support the notion that anti-CGRP mAbs mitigate “central symptoms” by inhibiting CGRP-mediated neurotransmission either directly or indirectly within the brain. Notably, radiolabeled galcanezumab has been shown to accumulate in the brain parenchyma and CSF of male rats, even if at low concentrations [57]. These findings cannot allow us to exclude that a small amount of active medication that reaches the CNS could be sufficient for a central effect. It is important to highlight that only 21 mg of erenumab significantly inhibits capsaicin-induced dermal blood flow (a model that assesses the target engagement of CGRP blocking agents), while the clinically approved dosages for migraine prophylaxis are 70 and 140 mg [72].

Furthermore, the status of the BBB during spontaneous migraine attacks with aura remains uncertain. Most neuroimaging studies have found no evidence of increased BBB permeability during aura, whereas animal models of CSD suggest a potential BBB alteration induced by CSD [73, 74]. However, it is important to highlight that certain structures within the circumventricular organs, which may lack BBB, such as the area postrema, have been shown in rat models to be functionally connected to key regions involved in migraine pathophysiology, including the hypothalamic nuclei and the spinal trigeminal nucleus. These regions may represent potential targets for anti-CGRP mAbs [75].

In conclusion, the available clinical and preclinical evidence does not allow us to rule out a significant role for CGRP in the mechanisms of CSD and its central function. Targeting the CGRP pathway could also impact aura, the clinical manifestation of CSD.