Background

Spreading depolarization (SD) is a wave of neuronal and glial depolarization that spreads across the cortical surface of the brain. It is characterized by a transient, near-complete loss of membrane potential in affected neurons, leading to a marked change in electrical activity [1, 2]. SD typically propagates at a rate of approximately 3–5 mm per minute and is often associated with a cascade of biochemical and physiological changes, including alterations in ion concentrations, increased metabolic demand, and activation of inflammatory pathways [3]. SD is triggered by various factors, such as trauma, ischemia, or excitatory stimuli, and is most notably linked to migraine aura and other neurological conditions [4]. SD itself does not lead to tissue injury in the metabolically intact brain but recurrent spontaneous SDs triggered in the metabolically impaired brain (because of ischemia, hypoxia, hemorrhage or trauma) can impose additional metabolic burdens on cells and lead to cell damage and cell death [5, 6]. Increased energy expenditure to restore disrupted homeostasis but decreased energy production, because of mitochondrial dysfunction, can lead to a supply‒demand mismatch for energy in metabolically impaired tissue [7, 8]. Moreover the vasoconstrictive hemodynamic response to SDs in ischemic tissue can deepen hypoperfusion and hypoxia, and worsen supply‒demand mismatch [5, 9, 10]. When the inflammatory response triggered by SD is combined with these insults it can lead to cell death and lesion growth after primary injury [11]. However recent studies have failed to show a detrimental effect of SDs on peri-infarct tissue injury after ischemia when they are induced non-invasively via optogenetic approach, questioning the causal relationship [12]. Some studies suggest that SD may even serve protective functions in the brain such as ischemic preconditioning or preventing seizure generation [13,14,15]. (Fig. 1)

Fig. 1
figure 1

Effects, Interactions, and Consequences of Spreading Depolarization-Related Changes in the Brain. The metabolic effects and neuroinflammation triggered by SD do not lead to tissue injury in the metabolically intact brain. However, the activation of nociceptors and the trigeminovascular system (TGVS) resulting from SD can initiate the headache cascade [18, 79]. In contrast, in a metabolically impaired brain (due to ischemia, hypoxia, hemorrhage, trauma, etc.), SD-induced metabolic disturbances and oligemic hemodynamic responses can lead to worsened hypoperfusion and energy mismatch at the cellular level [5, 6]. The resulting neuroinflammation can exacerbate the insult, leading to further injury and cell death [11]. In this way, SD can contribute to secondary damage and lesion expansion in a metabolically compromised brain. However, some protective effects of SD, such as ischemic preconditioning and seizure suppression, may mitigate these processes and prevent tissue damage [13,14,15] (Abbreviations: MMA: Middle meningeal artery, PPE: Plasma protein extravasation, TGVS: Trigeminovascular system)

Full size image

SD is an energetically demanding event for cells in the central nervous system (CNS), characterized by massive ionic shifts and neurotransmitter release that require substantial energy to restore homeostasis [10]. In a healthy brain, where ischemia or hypoxia is absent, this cellular stress typically does not lead to cell death in neurons, astrocytes, or microglia [16, 17]. However, SD-induced cellular stress can trigger an inflammatory response which may be called as neuro-parainflammation [18]. Parainflammation is often seen as a physiological response that may contribute to the maintenance of tissue homeostasis and repair processes [19]. It occurs in response to various stressors, such as aging, metabolic changes, or tissue damage, without the occurrence of a full-blown inflammatory response. Hence, SD related parenchymal inflammatory changes can be considered neuro-parainflammation.

Parenchymal neuroinflammation

Initial evidence for SD induced parenchymal neuroinflammation was provided by Jander et al., who reported that SD increased the mRNA levels of pro-inflammatory cytokines interleukin-1β (IL-1β) and tumor necrosis factor α (TNF-α) in the rat brain cortex ipsilateral to the induction site [20]. The mRNA levels peaked at four hours and began to decline by 16 h post-SD. Notably, the increase in cytokine expression resembled that is seen in more pronounced inflammatory states such as ischemia. Immunostaining also revealed elevated protein levels of IL-1β in the ipsilateral cortex, predominantly within microglia [20]. In vitro studies with hippocampal organotypic cultures confirmed that SD increases the levels of both pro- and anti-inflammatory cytokines (IL-1α, IL-1β, IL-2, IL-4, IL-6, IL-10, interferon (IFN)-γ, and TNF-α), with IL-1β levels remaining significantly elevated for up to three days [21]. Further in vivo evidence from microarray analyses revealed increased mRNA levels of cytokines such as IL-1α, IL-1β, IL-6, and IL-13 as early as two hours after SD, which persisted for 48 h. In contrast, the levels of IL-2, IL-10, and IL-12 decreased at both time points. Additionally, expression of the inflammatory mediator cyclooxygenase-2 (COX-2) increased at three hours post-SD, with maximal protein levels detected one day after SD in the ipsilateral cortex [22]. Other inflammatory markers, including the cell adhesion molecules vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1) and chemokines such as macrophage chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein 1 (MIP-1), significantly increased at three hours after SD [22, 23].

The connection between SD and parenchymal inflammation was further elucidated by studies demonstrating that both single SD and multiple SDs open pannexin-1 (Panx1) megachannels in neurons, which are activated during excessive glutamatergic activity and cellular stress [18, 24]. This channel opening triggers caspase-1 activation and an increase in cleaved caspase-1 within neurons as early as five minutes post-SD. Activated caspase-1 promotes the release of IL-1β and high mobility group box 1 (HMGB1) from neurons [18]. HMGB1 acts as an alarmin molecule, signaling surrounding cells to initiate inflammatory pathways [25]. The release of HMGB1 from neurons subsequently activates nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) in astrocytes, leading to increased expression of inflammatory mediators such as COX-2 and inducible nitric oxide synthase (iNOS). The authors posited that this coordinated inflammatory signaling in neurons and astrocytes could underlie trigeminal activation and migraine headaches, as inhibiting Panx1 opening or the actions of HMGB1 and NF-kB alleviated increased blood flow in the middle meningeal artery, an indicator of trigeminal activation, and reversed headache behaviors in mice subjected to multiple SDs [18].

Further investigation of HMGB1 dynamics revealed that its expression significantly increased after multiple SDs but not after a single SD, peaking three hours post-SD and returning to baseline after 24 h [26]. HMGB1 was localized primarily in the nuclei of both neurons and astrocytes under basal conditions, but following SDs, its distribution shifted toward extranuclear staining, particularly in neurons. These findings suggest that HMGB1 is released from neurons after SD and may diffuse to the cerebrospinal fluid and meninges, highlighting its role in headache initiation, especially following multiple SDs [26].

The role of Panx1 in SD-induced neuroinflammation appears to be complemented by purinergic signaling. Panx1 forms a large pore complex with the purinergic receptor P2 × 7 upon activation, allowing the permeation of glutamate and ions—making it a relevant factor in SD [27]. Preventing the formation of the Panx1-P2 × 7 pore pharmacologically or genetically reduced susceptibility to SD and diminished the increases in IL-1β, COX-2, and iNOS expression following SD. This inhibition also decreased the expression of calcitonin gene-related peptide (CGRP) and neuronal activation in the trigeminal pain pathway, effects not observed with P2 × 7 inhibition alone [28]. The involvement of Src family kinase (SFK) signaling in this context was demonstrated, as inhibiting SFK or its interaction with the P2 × 7 pore reduced both SD susceptibility and inflammatory mediator expression [29]. Moreover, a recent study showed that P2 × 7R antagonism decreased neuroinflammation in the ipsilateral hemisphere where SD was induced but also in the contralateral hemisphere, and in the subcortical regions [30]. Historically, SD was thought to affect only a single hemisphere where the contralateral hemisphere was used as the control [31]. However, a recent study revealed that SD-induced neuroinflammation can be bilateral in the brains of wild-type and FHM1 mutant mice brains [32]. The contralateral neuroinflammatory responses could be explained by intense axonal volleys, which might propagate to contralateral cortex through the corpus callosum. Intense glutamatergic activity from excitatory firing, akin to epileptiform discharges, can activate Panx1 megachannels, thus triggering downstream pathways secondary to N-methyl-D-aspartate (NMDA) receptor overactivation [24, 30, 32, 33]. In support of this notion, the authors successfully attenuated the neuroinflammatory profile in the contralateral cortex and striatum by locally inhibiting NMDA receptors by MK801 in the contralateral cortex [32]. Human PET studies support these findings showing that neuroinflammation following SD is multiregional and can take place bilaterally. A recent PET-MRI study was conducted with patients who had at least 1 migraine attack with aura and healthy controls, used a (11 C)PBR28 radio ligand that binds to a glial marker 18 kDa translocator protein which increases during neuroinflammation, showed an increase in inflammatory signal in cortical and subcortical structures bilaterally in migraine patients and the signal intensity was found to be correlated with the number of migraine attacks [34].

In addition to SD, Panx1 channel activation may also result from an imbalance in synaptic energy usage. For example, the application of 1,4-dideoxy-1,4-imino-d-arabinitol (DAB) or sleep deprivation, both of which disrupt glycogen utilization in astrocytes, led to Panx1 opening in neurons. This activation is likely due to increased extracellular potassium (K⁺) concentrations caused by impaired K⁺ reuptake by astrocytes due to inadequate synaptic energy [35]. This channel opening also increased cleaved caspase-1 and facilitated the translocation of HMGB1 out of neuronal nuclei, suggesting that neurogenic inflammation can arise from various mechanisms. Indeed, both disrupted synaptic energy and elevated K⁺ levels (from DAB or sleep deprivation) reduce the SD threshold, a change reversible by glucose or lactate superfusion, underscoring the critical role of synaptic energy metabolism in maintaining synaptic activity and preventing SD generation [35].

Studies examining the link between SD and neuroinflammation have predominantly employed conventional methods of SD induction, such as craniotomy, pinprick, and intracortical injections. These invasive techniques can cause tissue injury, complicating the study of inflammation. To address this, Takizawa et al. utilized optogenetic SD induction to investigate the changes in the expression of inflammatory mediators [36]. Following six SD waves, they reported increased mRNA levels of COX-2, IL-1β, TNF-α, IL-6, ICAM-1, VCAM-1, and chemokine (C-C motif) ligand 2 (CCL2) in the ipsilateral cortex compared to the contralateral side. IL-1β levels peaked at one hour post-SD, while other mediators peaked around four hours and returned to baseline before 24 h. Importantly, a single SD also significantly elevated the expression of IL-1β, COX-2, TNF-α, and CCL2. Notably, IL-1β expression increased as early as ten minutes after a single SD, suggesting that it plays an upstream role in the inflammatory cascade. Experiments with IL-1 receptor knockout mice revealed diminished expression levels of CCL2 and TNF-α even after multiple SDs. Treatment with dexamethasone—administered before and after SDs—reduced this increase, whereas ibuprofen had no such effect. The rise in inflammatory mediators in the ipsilateral cortex was intrinsic to the brain, as it was not detectable in systemic circulation [36].

Emerging evidence suggests that inflammasome activation may link Panx1 channel opening to increased expression of inflammatory mediators following SD [18, 37]. Recent studies have indicated that nucleotide-binding domain (NOD)-like receptor family pyrin domain containing 3 (NLRP3) inflammasome expression, assembly, and cleaved caspase-1 levels increase in the brain cortex shortly after both multiple SDs and single SD [38]. Notably, this NLRP3 activation occurs predominantly in neurons, with no significant changes in glial cells. The activation of the NLRP3 inflammasome after SD leads to increased levels of IL-1β, COX-2, TNF-α, and CCL2, alongside the release of HMGB1 from neurons. Importantly, the increase in these inflammatory mediators was suppressed by pharmacological inhibition or genetic ablation of Panx1 and NLRP3, as well as IL-1β. Furthermore, inhibition of Panx1 and NLRP3 resulted in reduced trigeminovascular activation, as evidenced by diminished middle meningeal artery dilation and decreased CGRP and c-FOS expression typically observed after SD. Although microglia did not exhibit increased levels of the inflammatory mediators studied, they displayed a delayed morphology indicative of microglial activation following multiple SDs. Inhibiting microglial activation with minocycline reduced neuronal inflammation [38].

While the role of the NLRP3 inflammasome in SD-related parenchymal inflammation has been studied recently, the involvement of other inflammasome types remains less clear. The precise function of NLRP2 in immune responses is not fully understood; however, several studies suggest that it acts as an astrocytic inflammasome, contributing to inflammatory responses in various disease models, including cerebral ischemia, pain, and depression [39,40,41]. Additionally, NLRP2 has been implicated in anti-inflammatory processes through its regulation of NF-κB in different biological contexts [40]. A recent study examined the expression pattern and alterations of NLRP2 following optogenetically induced single SD in the mouse brain. Interestingly, NLRP2 immunoreactivity was observed predominantly in neurons, rather than in astrocytes or glial cells, in both sham and naive mouse brains. Notably, SD triggered a significant decrease in NLRP2 immunoreactivity in neurons at this early time point [42]. Hence, the specific expression of NLRP2 in neurons, along with its significant decrease following SD, highlights its potential involvement in neuronal responses to inflammatory challenges.

The interaction between HMGB1 and Toll-like receptors (TLR) 2/4 has been shown to be critical for microglial activation following SD [43]. The inhibition of TLR 2/4 effectively prevented microglial activation after SD, resulting in the downregulation of inflammatory mediators in neurons, thus emphasizing the interplay between neurons and microglia in SD-induced inflammation [38]. Studies have shown that HMGB1 is released from stressed neurons following SD, although the mechanism of its release was previously unclear. Recent findings clarified this by demonstrating that HMGB1 is released from neurons in extracellular vesicles after a single SD [44]. Upon release, these vesicles are primarily taken up by astrocytes rather than microglia, leading to NF-kB activation in astrocytes. This signaling appears to be localized to neurons and astrocytes, suggesting that it does not elicit a broader cellular immune response in the brain during mild inflammatory states, including SD [44].

Beyond the mediators already discussed, SD has also been shown to increase IFN-γ through T-cell activation in hippocampal slice cultures and in vivo, which is associated with myelin disruption in white matter and increased SD susceptibility [45]. These diverse effects of SD on neuroinflammation may be regulated by TLR3, a pattern recognition receptor crucial for the innate immune system. Application of the TLR3 agonist polyinosinic acid (Poly I) prior to SD attenuated the increases in TNF-α and IFN-γ levels in the brain [46]. Furthermore, Poly I treatment exhibited neuroprotective effects, increasing gamma aminobutyric acid (GABA)-A receptor levels in the brain and reducing the number of stressed neurons after multiple SDs. Notably, SD also influenced systemic inflammatory responses, increasing the production of IFN-γ, TNF-α, transforming growth factor-beta (TGF-β), and IL-4 in the spleen while promoting lymphocytic proliferation. Administration of Poly I diminished the production of TNF-α and IL-4 in the spleen following SD [46]. These findings indicate that these inflammatory processes may also have systemic implications, affecting immune responses beyond the CNS.

One of the ways SD can trigger inflammation can involve mitochondrial dysfunction. SD has been shown to cause calcium influx to the mitochondria and lead to mitochondrial membrane depolarization in neurons [47]. Importantly, SD can lead to changes in mitochondrial morphology that reflect mitochondrial dysfunction in neurons, even in healthy brain tissue [8]. Furthermore, it can also lead to opening of the mitochondrial permeability transition pore (mPTP) on the mitochondrial membrane [48]. mPTP opening can lead to the activation of innate immune system pathways in the cell and calcium release from the pore can also lead to apoptosis [49]. Conversely, mitochondrial stress or dysfunction may facilitate the generation of SD, as in the case of mitochondrial diseases and migraine coexistence, including myoclonic epilepsy with ragged red fibers (MERRF) and lactic acidosis and stroke-like episodes (MELAS) syndromes [50].

Cellular interplay between neurons, astrocytes, and microglia in SD-related inflammation

SD significantly disrupts the homeostasis of ions and neurotransmitters, leading to substantial transmembrane ionic shifts characterized by sodium (Na+) and calcium (Ca²⁺) influx, accompanied by water movement and K⁺ and glutamate efflux from affected brain cells [51]. Astrocytes are essential for clearing extracellular K⁺ and glutamate released during SD, and their Na⁺/K⁺ ATPase activity is critical for restoring homeostasis disrupted by SD [52]. These functions position astrocytes as key regulators of SD susceptibility [52, 53]. Moreover, SD itself profoundly affects astrocytes, which can become reactive in response to disturbances in tissue homeostasis, resulting in increased expression of glial fibrillary acidic protein (GFAP) and hypertrophy [54]. Following SD, increased GFAP staining has been observed in the affected cortex and chronic exposure to SD leads to a significant rise in astrocyte numbers with enhanced GFAP immunopositivity [55, 56]. Detailed analyses of astrocytes post-SD, both in ex vivo brain slices and in vitro, reveal an increase in hypertrophic astrocytes exhibiting elevated expression of markers associated with reactive astrocytosis, such as GFAP, vimentin, and S100β. Furthermore, astrocytes display significantly increased expression of inflammatory mediators, including IL-1β, TNF-α, and IL-6, and the pattern recognition receptors TLR3 and TLR4 after SD [57]. These findings indicate that astrocyte activation and their inflammatory signaling contribute to SD-related inflammatory changes.

The interaction between astrocytes and neurons is critical for initiating and propagating the inflammatory response associated with SD. The opening of Panx1 channels during SD leads to the release of the HMGB1 alarmin from neurons, which activates astrocytes through increased NF-κB signaling, resulting in increased production of inflammatory mediators such as NO, synthesized by iNOS, and COX-2. This astrocyte‒neuron interaction is implicated in the activation of the trigeminovascular system and headache onset following SD [18]. Beyond SD, astrocytes can also facilitate Panx1 channel opening in neurons during disruptions in energy utilization due to their close association with neuronal energy metabolism and ionic homeostasis [35]. Recent studies have shown that HMGB1 released from neurons in extracellular vesicles after SD is primarily taken up by nearby astrocytic processes, leading to their inflammatory activation [44]. This confined communication reinforces the notion that astrocyte‒neuron interactions are vital for the inflammatory signaling associated with SD.

As the primary resident immune cells of the brain, microglia perform diverse functions, from defending against pathogens to regulating neuronal and synaptic activity [58, 59]. They can secrete pro-inflammatory mediators such as IL-1β and TNF-α and promote a pro-inflammatory M1 polarization while also adopting an anti-inflammatory role to mitigate inflammation [60, 61]. Given their involvement in various inflammatory processes within the central nervous system, microglia are strongly implicated in SD-related neuroinflammation [61].

Early evidence of microglial activation in response to SD emerged from observations showing morphological changes in microglia—such as increased cell body size and de novo expression of major histocompatibility complex (MHC) class II antigens—after multiple SDs, particularly 24 h following induction [62]. Additionally, SD has been shown to enhance microglial motility, which inversely correlates with neuronal excitability [63]. In addition to morphological alterations, microglia have been implicated in the secretion of IL-1β and TNF-α following SD [20, 64]. Repeated SDs lead to a substantial increase in microglial proliferation, peaking two days after SD and persisting for up to four days [65]. These changes indicate robust microglial activation in response to multiple SDs. Moreover, microglial activity has been shown to increase in response to the increase in extracellular K⁺ concentrations caused by neuronal NMDA activity during SD [66]. It is recognized that a single SD typically causes less damage compared to clusters of SDs, which impose significant metabolic stress on brain tissue. Research by Takizawa et al. has shown that while a single SD does not elicit morphological changes in microglia, multiple SDs result in enlarged somas and thickened processes after 24 h [43]. These activated microglia also exhibit heightened immunopositivity for cathepsin D, which is indicative of increased lysosomal activity. Notably, HMGB1 released by neurons after SD plays a critical role in activating inflammatory pathways in astrocytes and enhances this release with repetitive SDs. Blocking HMGB1 can prevent microglial activation following multiple SDs, and its effects appear to be mediated through TLR2/TLR4 receptors, as knockout models for these receptors do not show microglial activation after SD [43]. These findings suggest that innate immune signaling triggered by SD is integral to microglial activation, particularly when HMGB1 levels are elevated following multiple SDs. Importantly, microglial activation following repeated SDs is crucial for the neuronal inflammatory response. Pretreatment with minocycline, a microglial activation inhibitor, or TLR2/4 inhibitors has been shown to attenuate neuronal NLRP3 inflammasome activity and reduce the levels of IL-1β and COX-2 [38].

The relationship between microglia and SD is bidirectional; microglia also regulate SD susceptibility. Microglial ablation has been shown to disrupt neuronal calcium responses and reduce SD susceptibility [67, 68]. Furthermore, multiple SDs appear to enhance microglia–neuron interactions, with increased numbers of microglial processes surrounding neurons observed after SD. These processes express the purinergic receptor P2Y12R, which plays a role in the chemotaxis of microglial processes, and the levels of P2Y12R on microglial processes increase following multiple SDs [69]. Notably, microglial ablation leads to a lower induction threshold for the first SD while raising thresholds for subsequent SDs. Moreover, both microglial ablation and P2Y12R knockout reduce the amplitudes of SDs. In microglia-ablated mice, extracellular K⁺ clearance is also accelerated [69]. Together, these findings suggest that microglia play critical roles in both the induction and resolution of SD.

Peripheral consequences of SD-related inflammation

Meningeal inflammation and trigeminovascular system activation

The inflammatory response associated with SD extends to peripheral structures such as the meninges and the trigeminovascular system, which has significant implications for nociceptor activation and headache initiation in migraine. Evidence of meningeal inflammation following SD includes increased diameter and blood flow in the middle meningeal artery (MMA), leakage of plasma proteins from dural vessels, and direct activation of meningeal immune cells [18, 70, 71]. Notably, increased blood flow in the MMA begins approximately five minutes after SD and can remain elevated for up to an hour [70]. This prolonged response is dependent on trigeminal and parasympathetic activation, as denervation of both systems abolishes the increase in blood flow. Plasma protein leakage occurs ipsilaterally to the SD and is associated with neurokinin-1 receptor activity and trigeminal activation, but it does not rely on parasympathetic input, as shown by chronic denervation’s lack of effect on leakage [70].

SD also activates meningeal nociceptors. Some Aδ and C fiber nociceptors in the meninges exhibit increased firing coinciding with the SD wave, whereas the majority show delayed activation starting 15–20 min after SD and last up to an hour. This timing aligns with the onset of aura and headache in migraine patients, underscoring the potential role of SD in headache initiation [72]. Subsequent studies revealed that the short-lived activation of nociceptors concurrent with SD can lead to prolonged activation in most nociceptors [73]. Both single and multiple SD events can trigger nociceptor activation [72]. However, it is unlikely that inflammatory mediators or metabolic disturbances associated with SD solely account for this nociceptor activation, as diminishing metabolic perturbations with naproxen reduced nociceptor sensitization but did not prevent nociceptor activation, suggesting that alternative mechanisms may be involved [74]. Importantly, mediators released in the brain parenchyma during SD—such as ATP, glutamate, and K+ can diffuse to the sub-arachnoid space where they can activate the trigeminal afferents around pial vessels [75,76,77]. Majority of trigeminal ganglion neurons that innervate pial vessels also have branches innervating dura mater so the activation of trigeminal afferents in leptomeninges (pia and arachnoid) can lead to the release of pro-inflammatory proteins in dura mater which is important for neurogenic meningeal inflammation [70, 77, 78]. Recent investigations into ATP and purinergic signaling have demonstrated that inhibiting P2 × 7 receptors (but not P2 × 2/3 receptors) reduced mechanical sensitization of meningeal nociceptors. Similar effects were observed when Panx1, which forms a pore complex with P2 × 7 and enhances ATP release, was inhibited. These findings suggest that meningeal nociceptor sensitization after SD may involve inflammatory mechanisms. However, neither P2 × 7 nor Panx1 inhibition affected the immediate or prolonged activation of meningeal nociceptors following SD, leading researchers to propose that the activation and sensitization of these nociceptors occur through different mechanisms [79].

The meninges contain a substantial number of mast cells, which are granulated innate immune cells that release various inflammatory mediators upon degranulation [80]. Their role in meningeal neurogenic inflammation has been recognized, particularly through the observation that neuropeptides such as CGRP can stimulate mast cells to release histamine, contributing to increased blood flow and permeability in the dura [81]. Moreover, SD has been shown to induce mast cell degranulation within 30 min after its onset [18]. Mast cell degranulation in the dura appears to create an inflammatory stimulus for nociceptor activation and sensitization, with serotonin identified as a principal mediator driving this process [82, 83]. The anti-migraine drug sumatriptan, a 5HT1B/D receptor agonist, has been shown to inhibit the activation of central neurons in trigeminovascular system without suppressing mast cell degranulation or the activation of meningeal nociceptors by serotonin [83]. This effect may be attributed to its ability to block synaptic transmission between meningeal nociceptors and central trigeminovascular neurons through the activation of pre-synaptic 5HT1B/D receptors in dorsal root ganglia, demonstrating a central effect on trigeminovascular system for mitigating migraine headache [84, 85].

There is also evidence suggesting that parenchymal inflammation can influence meningeal nociceptor activation. The delayed increase in MMA diameter observed after SD and mast cell degranulation in the meninges were blocked by the inhibition of Panx1 or HMGB1, highlighting their roles in trigeminal activation [18]. Additionally, the reduction in HMGB1 protein levels in the brain following SD—despite increased expression—implies that HMGB1 may be released into extracerebral regions, including the meninges, where it could participate in nociceptor activation [26]. Furthermore, inhibiting parenchymal NLRP3 inflammasome activation disrupted SD-mediated MMA dilation and the expression of CGRP and c-FOS in the trigeminovascular system [38].

Direct activation of meningeal immune cells by SD has been demonstrated in studies showing that SD activates meningeal macrophages and dendritic cells. Macrophages in the pia mater displayed a more circular morphology with fewer protrusions during the SD wave, whereas those in the dura exhibited similar changes with a delay of 20 min. These morphological alterations suggest macrophage activation. Dendritic cells, initially mobile, became stationary approximately six minutes after SD induction, further indicating their activation. To confirm that these changes were due to inflammatory effects, researchers injected lipopolysaccharide (LPS) into mice and observed analogous changes in macrophages and dendritic cells. Notably, these immune cells were found in close proximity to Trpv1 + nociceptors, suggesting their involvement in nociceptor activation following SD [71].

Prior research has demonstrated that SD events cause the release of small molecules into the cerebrospinal fluid (CSF) that activate and sensitize afferent trigeminal fibres in the meninges [72, 86]. It was previously assumed that the trigeminal ganglia, located outside the blood-brain barrier, were not directly exposed to CSF [87]. However, a recent study demonstrated that in an optogenetically triggered SD model, subarachnoid CSF carries signals from the cortex directly to trigeminal ganglion cell bodies, activating nociceptors through a pathway that bypasses meningeal trigeminal afferents [88]. This finding indicates that the trigeminal ganglia, at least proximal 1/3 part of it, are situated within the blood-brain barrier and highlights a previously unidentified mechanism connecting the aura and headache, potentially opening new avenues for migraine prevention and treatment. By employing a combination of proteomic, histological, imaging, and functional techniques, Rasmussen and colleagues demonstrated that SD results in alterations to the content of subarachnoid CSF, including an increase in CGRP levels [88]. It is noteworthy that CSF conditioned by SD has the capacity to activate trigeminal ganglion neurons, thereby establishing a connection between the central and peripheral nervous systems via an intracranial humoral pathway. This pathway may modulate neurochemical interactions among neurons, glia, and immune cells within the trigeminal ganglion [89]. Additionally, the rapid normalization of CSF composition changes induced by SD suggests that this pathway plays a crucial role in the very early phases of SD, whereas other processes, including parenchymal inflammatory signaling, may drive headache during later phases [88]. Furthermore, while the glymphatic system has been characterized primarily as the waste clearance mechanism of the brain, the study by Rasmussen et al. revealed a previously unrecognized role of glymphatic exchange in facilitating communication between distant brain regions [90]. Despite this, further research is required to fully understand the implications of the glymphatic system involvement in migraine, particularly with respect to fluid and solute transport in the brain. (Fig. 2)

Fig. 2
figure 2

Schematic illustration of SD associated inflammatory changes in the CNS. Panx1 channel opening and pore formation with P2 × 7 lead to the formation of the NLRP3 inflammasome in neurons, which activates caspase-1. Activated caspase-1 promotes the release of IL-1β and HMGB1 from neurons [18, 38]. In contrast, the level of the NLRP2 inflammasome has been shown to decrease in neurons after SD, which needs to be investigated further [42]. HMGB1 is released from neurons in extracellular vesicles, which are primarily taken up by astrocytes [44]. This triggers NF-κB activation in astrocytes, which can lead to the release of IL-1β, IL-6 and TNF-α from activated astrocytes and increased iNOS and COX2 levels in astrocyte end feets including the glia limitans [18]. Released HMGB1 from neurons can also trigger microglial activation through its interaction with TLR 2/4 receptors on microglia [43]. In the meninges, SD can lead to the activation of macrophages and dendritic cells and degranulation of mast cells which can lead to nociceptor activation in the meninges as well as vasodilation and plasma protein extravasation in the middle meningeal artery [18, 70, 71, 83]. Inflammatory activation in the brain parenchyma might also play a role in nociceptor activation as its inhibition prevents nociceptor activation [18, 38]. This could be due to the diffusion of inflammatory mediators directly or possibly through the release of HMGB1 to the meninges from the brain parenchyma [26]. Activation of meningeal trigeminal afferents leads to increased activity in the trigeminal ganglion and the trigeminal nucleus caudalis, as indicated by increased c-FOS expression [111]. Recent findings also suggest that increased mediators such as CGRP in the CSF after SD can travel directly to the trigeminal ganglion through subarachnoid CSF flow and activate trigeminal ganglion neurons independent from meningeal afferents [88]. One of the proposed ways that SD can activate inflammation in neurons is through mitochondrial dysfunction. SD has been shown to cause mitochondrial membrane depolarization and increased membrane permeability which can lead to the activation of innate immune system pathways such as the NLRP3 inflammasome [8, 47, 49]. This figure was prepared with the help of Servier Medical Art (https://smart.servier.com/) and the National Institutes of Health (NIH) BioArt Source (https://bioart.niaid.nih.gov/) (Abbreviations: SD: Spreading depolarization, Panx1: Pannexin-1, NLRP3: nucleotide-binding domain (NOD)-like receptor family pyrin domain containing 3, HMGB1: High mobility group box 1, IL-1β: Interleukin-1β, iNOS: İnducible nitric oxide synthase, COX2: cyclooxygenase-2, TNF-α: Tumor necrosis factor-α, TLR2/4: Toll like receptor 2/4, CGRP: Calcitonin gene related peptide, CSF: Cerebrospinal fluid, TNC: Trigeminal nucleus caudalis)

Full size image

SD-related inflammation in genetic migraine conditions

Migraine has a significant genetic component, encompassing both common susceptibility genes with smaller effects and rare monogenic forms that present additional features such as hemiparesis during aura, ataxia, and seizures alongside migraine headaches [91, 92]. Familial hemiplegic migraine (FHM) syndromes are frequently utilized in migraine research to develop transgenic mouse models due to their well-defined mutations that significantly affect protein functions and produce distinct phenotypes [93]. Four types of FHM have been established, each associated with different gene mutations that impact synaptic transmission and ion transport, resulting in increased brain excitability and heightened migraine susceptibility [52, 94, 95].

The CACNA1A gene linked to FHM type 1 (FHM1) encodes the voltage-gated Ca²⁺ channel Cav2.1. Gain-of-function mutations in this gene increase the probability of channel opening, leading to increased Ca²⁺ influx and heightened excitability in the brain [96]. Variations in mutations can result in a spectrum of phenotypes, from subtle changes in pain responses to severe attacks characterized by hemiparesis and cerebellar ataxia. Mice with the FHM1 mutation display signs of photophobia and heightened trigeminal nociception [93]. Notably, FHM1 mice show increased susceptibility to SD, with propagation rates that differ from those observed in wild-type mice, suggesting the presence of distinct mechanisms for SD in these models [97, 98].

FHM2, associated with mutations in the ATP1A2 gene, results in a loss of function in the α2 Na⁺/K⁺-ATPase. This impairment hampers K⁺ and glutamate clearance by astrocytes following neuronal activity, creating a hyperexcitable environment [52, 99]. Mice harboring the FHM2 mutation also exhibit increased SD susceptibility and speed of propagation, and they can experience tonic‒clonic seizures following SD induction [100]. The third FHM type involves gain-of-function mutations in the SCN1A gene, which encodes voltage-gated Nav1.1 sodium channels that are primarily found in cortical inhibitory interneurons, and are crucial for regulating excitability and firing patterns [101]. The FHM3 phenotype in humans is often characterized by prominent aura symptoms, although mice with FHM3 mutation tend to exhibit early mortality and abnormal behaviors, such as hindlimb jerks [102]. FHM3 mice also show increased susceptibility to SD and, notably, display spontaneous SDs that propagate from the visual cortex to the motor cortex, mirroring migraine aura phenomena [102]. Collectively, all FHM mutations predispose individuals to SD, with exaggerated effects observed in these models compared to wild-type mice [93]. FHM4 mutations occur in the PRRT2 gene, which is expressed in presynaptic terminals, where it interacts with proteins of the exocytosis complex, which inhibits voltage-gated sodium channels [103, 104]. In all four types of FHM mutations, there is an increase in K+ and glutamate in the synaptic cleft, which facilitates SD, the neurophysiological correlate of migraine aura.

Mice with FHM mutations demonstrate elevated neuroinflammatory activity even under basal conditions. Compared with those from wild-type mice, cultured trigeminal ganglia from FHM1 mice exhibit a greater number of activated macrophages, increased release of TNF-α, and greater P2 × 3 purinergic receptor currents in neurons [105, 106]. The heightened P2 × 3 receptor currents in FHM1 mice are already at peak activity in the absence of stimulation, indicating a predisposition for trigeminal nociception driven by enhanced neuroinflammatory activity [106]. Additional evidence of the neuroinflammatory environment in FHM mice comes from observations of astrogliosis and activated microglia in the brains of naive FHM1 mice, characterized by increased GFAP expression and increased branching in microglial processes, suggestive of heightened activation and surveillance [107]. In contrast, oligodendrocytes remain unchanged under basal conditions in FHM1 mice [107]. These findings establish that FHM mutations are correlated with intrinsic neuroinflammatory activity.

FHM brains not only facilitate the generation of SD but also experience more pronounced disruptive effects from it. Comparative gene expression analyses following multiple SD episodes revealed a stronger inflammatory response in FHM1 mice, particularly with elevated expression of genes associated with interferon-mediated inflammatory signaling [108]. These findings suggest distinct mechanisms of SD-associated inflammatory signaling in FHM1 mutants. Furthermore, compared with wild-type mice, FHM1 mice exhibited enhanced levels of pro-inflammatory mediators (IL-17 and IL-6) and chemokines (CCL2, CXCL2, and CXCL10), along with lower levels of anti-inflammatory mediators (IL-1 receptor antagonist, IL-4, IL-10, and IL-13) [109]. Collectively, these findings underscore the exaggerated and widespread inflammatory responses to SD in FHM models [32].

In FHM1 mice, the release of HMGB1 from neurons and the activation of NF-κB in astrocytes were observed in both the ipsilateral and contralateral hemispheres, as well as in subcortical regions such as the striatum and thalamus after SD. In contrast, such responses are typically confined to the ipsilateral cortex in wild-type mice. Notably, FHM1 mice exhibited HMGB1 release from neurons even in the absence of SD induction. This widespread inflammatory activation was mirrored by increased numbers of pERK-positive neurons [32]. The inflammatory signaling and neuronal activation in the contralateral hemisphere could be inhibited by local application of NMDA antagonists, confirming that the extensive inflammatory response in FHM1 brains is associated with excessive glutamatergic signaling, allowing SD to affect the contralateral hemisphere [32].

Not surprisingly, FHM mice also display sustained headache-related behaviors following SD. Optogenetically induced SDs in awake FHM1 and wild-type mice resulted in increased headache-related behaviors 30 min post-SD. While this behavior returned to baseline within 24 h in wild-type mice, FHM1 mice did not normalize until 48 h after SD [110]. Consistent with earlier findings, HMGB1 release after optogenetically induced SDs was more pronounced in both hemispheres of FHM1 mice compared to wild-type mice. Additionally, inhibiting Panx1 prior to SD induction diminished the increase in headache-related behaviors, highlighting the crucial role of Panx1 channels in mediating pain responses following SD [110].

Conclusion

The interplay between spreading depolarization and innate immunity in the CNS highlights a complex relationship that influences both local and systemic inflammatory responses. Understanding these mechanisms is essential for developing targeted therapeutic approaches to mitigate the harmful effects of neuroinflammation and improve outcomes in neurological disorders.