Spreading Depolarization Waves in Neurological Diseases: A Short Review about its Pathophysiology and Clinical Relevance

1.1. Mechanisms of Spreading Depolarization and Its Importance as a Pathological Mechanism in Human Neurological Conditions

For many years, the difficulty and inadequacy of detec-ting SD in routine scalp EEG recordings led scientists to think that SD was just an artifact, which occurs in the rodent brains and during experimental conditions. Hence, its significance and translation to human pathological conditions were obscure and there were opponent arguments about its presence and implication for neurological diseases especially for migraine. Unfortunately, the routine scalp EEG was not sensitive enough due to the high impedance of dura and skull to record SD waves from a relatively narrow cortical area. With the development of new imaging technologies, first clues about its presence and the role in disease pathophysiology, especially for migraine aura, have emerged. A wave of reduced blood flow propagation across the brain during an attack of a patient with migraine aura has been elegantly demonstrated by Hadjikhani et al. via blood flow changes detected with bold magnetic resonance imaging and travel velocity of blood flow changes over cortex was similar to the SD waves observed in experimental animals [7]. Later on, studies of several clinical and preclinical groups clearly demonstrated that SD waves can occur in several pathological conditions in humans like migraine, stroke, subarachnoid hemorrhage, trauma, etc [2, 3, 8-12]. These researchers recorded SD waves through cranial windows or with direct electro-corticographic (ECoG) recordings and laser speckle imaging. Additionally, ECoG recordings of operative patients with acute brain injury obtained during intensive care suggest that SD waves play a role in the deterioration of cognitive functions and promote lesion progression [13].

SD waves can occur in both physiological and pathological conditions. They usually start from a point on the cortex and travel through with a velocity of 2 to 5 mm/min and the length of travel depends on the severity or amplitude of depolarization waves as well as origin [1]. In healthy brain tissue, as the SD propagates, both the spontaneous and the evoked synaptic activity is silenced for 5-15 minutes, then it spontaneously returns to normal; whereas in pathological conditions (such as head trauma, hypoxia-ischemia, hypoglycemia) depolarization waves start spontaneously and the recovery period is prolonged [14].

Spreading depolarizations of neurons and glial cells on the cortical surface are preceded by propagated field oscillations that create a brief moment of hyperexcitability that covers distances up to 1 mm [15]. There is a complete silence of electrical activity after these oscillations and this electrical silence turns back to normal in 5 to 10 minutes [1]. The silencing of the neurons is accompanied by perturbations in ionic homeostasis and increased release of excitatory amino acids from neurons [16].

Events leading to initiation of SD waves are not clearly understood yet, but increased extracellular levels of K⁺ [17] and excitatory amino acids especially glutamate are proposed as triggers for SD occurrence [18]. Moreover, depolarization of neurons removes the voltage-sensitive Mg2+ block of the N-methyl-D-aspartate (NMDA) receptors, and makes the receptor more sensitive to fluctuations of interstitial glutamate levels [19]. Interactions between glutamate and NMDA receptors trigger further K⁺ and glutamate release hence increase the excitability of brain tissue and make the neuronal depolarization pass through to neighboring regions [17].

During the passage of spreading depolarization waves, extracellular K⁺ levels increase, whereas Ca⁺², Cl^- and Na⁺ levels significantly decrease [14, 20]. At the same time, pH declines from 7.3 to 6.9 and extracellular space shrinks due to increased water uptake into neurons [21]. This event leads to reversible neuronal swelling while the volume of astrocytes remains stable as they express significant amounts of volume sensitive channels that allow rapid efflux of taurine, aspartate and glutamate [22-25]. There is also a massive efflux of glutamate and aspartate from the neurons during the depolarization wave [26, 27]. Recently, neuronal swelling was found to be associated with a novel chloride channel SLC26A11 mutation leading to Na⁺ and Cl^- influx, but it was independent of Ca⁺² influx leading to cytotoxic neuronal edema [28].

SD is also associated with an increased synthesis of matrix metalloproteinase-9 (MMP-9) [29] Fig. (2). Increased MMP-9 activity causes opening of the blood-brain barrier and results with vasogenic edema formation [30]. Other pro-inflammatory cytokine pathways are also found to be upregulated [31]. There are also signs of immediate early gene expression in response to SD [32].

1.2. Effects on Blood Flow and Energy Metabolism: Accompanying Conditions

Inside the depolarized tissue, cerebral blood flow briefly decreases just before the onset of depolarization [33]. As the ionic changes are restored, cerebral blood flow (CBF) increases by approximately 100% to 200% in anesthetized animals [34], which lasts for 1 to 2 minutes. This period is called hyperemic phase of SD. It is followed by a persistent blood flow reduction of 20% to 30% of baseline value that lasts 1 to 2 hours, termed as oligemic phase. However, this pattern of brief hyperperfusion followed by hypoperfusion is mainly observed in healthy and energetically stable brain tissue. Recovery from SD and normalization of membrane potentials and extracellular as well as intracellular ionic hemostasis requires optimum amount of energy. Hence, during pathological conditions accompanied by energy deprivation like stroke and trauma, SD is observed to trigger a cortical spreading ischemia [35, 36], and in stroke patients the cerebrovascular changes triggered by SD vary greatly in the penumbra depending on the distance to the infarct core [37]. The regional CBF changes might also create a mismatch between oxygen demand and supply for the already energetically compromised tissue, which may lead to tissue hypoxia [23, 38]. Recovery of the ionic homeostasis after SD is an energy demanding process and metabolic rate rises significantly during this process [39]. For this reason, continuity of normal blood flow supply is critical to guarantee the brain’s needs of glucose and oxygen during this high-energy requiring period.

Main energy consuming enzyme during SD is Na⁺/K⁺-ATPase. This pump restores extracellular K⁺ levels using ATP [40]. During an SD wave, extracellular K⁺ levels rises up to 60-100 mmol/L and Na⁺/K⁺-ATPase tries to drop it to around 3 mmol/L. Re-establishment of ionic hemostasis enables neurons to repolarize. The rise in cytosolic Ca⁺² during SD depolarizes neuronal mitochondria and increases the oxygen use through the Ca⁺² uniporter that is located in the inner mitochondrial membrane [24]. A recent study in mice demonstrated that the disruption of mitochondria in ischemic and traumatic injuries is reversible and can be a possible therapeutic target [41]. Additionally, presynaptic NMDAR activation creates a positive feedback loop of glutamate induced glutamate release via both Ca⁺² influx from extracellular space and mitochondria by mitochondrial Na⁺ / Ca⁺² exchanger [42]. Takano et al. observed that even when there are no significant changes in cerebral perfusion, or at the hyperemic phase, local tissue hypoxia occurred during SD due to the increased metabolic rate and energy demand [23]. This prolonged mismatch between oxygen demand and supply has been suggested to contribute to the pathogenesis of some other neurodegenerative diseases [43]. Takano et al. also pointed out that there are transient morphological changes in dendritic structures accompanying SD such as spine loss, that are comparable to the ones observed during anoxic depolarization [23]. A recent in vivo experiment also showed that there is dendritic beading during SD and this beading was strictly dependent on the presence of chloride ions, and the blockage of all chloride coupled transporters

significantly reduced dendritic beading but did not have any effect on the SD waves [44]. SD also activates glycolytic pathways leading to an increase in brain lactate levels and reduction in brain glycogen. On the other hand, the concentration of energy-rich phosphate compounds remains the same in normal brains [45, 46]. There are also reports suggesting that depolarizations are associated with progressive reduction in brain tissue glucose via monitoring of extracellular brain glucose concentration with microdialysis in patients with traumatic brain injury [46, 47]. Whether this state of energy depletion delays the recovery of cortical function or not, depends on the persistence of oxygen and glucose supply. Acute CBF responses during SD waves are also found to be varying between species [48].

All of these listed evidences for SD induced blood flow, metabolic and enzymatic changes point out the importance of spreading depolarization in nervous tissue and SDs may have a significant role in several neurological diseases as a main mechanism or as secondary conditions leading to detrimental effects in tissue. Some examples of aforementioned neurological conditions are discussed below in regard to pathophysiological effects of spreading depolarizations.

1.3. Previously Tested Therapeutic Approaches

As SD clusters are proven to play a significant role in the pathogenesis of aforementioned neurological diseases, prevention of recurring SD waves in acutely injured brain tissue has been an intriguing and challenging subject. Studies made in gyrencephalic swine brain have shown that NMDAR antagonist ketamine decreases speed, duration, spread of SDs in low-dose infusion (2 mg/kg/h), and inhibits induction and expansion of SDs in high-dose infusion (4 mg/kg/h) [125]. Similar results were obtained in rats [126], swine [127] and human patients [128, 129]. There are studies suggesting that low dose of ketamine (2 mg/kg/h) in swine brain was not effective at preventing the hemodynamic response in the oligemic phase, but reduced the initial hyperemic response to SDs [127]. Sedatives that are frequently used in intensive care units such as isoflurane, sevoflurane and propofol have been studied as potential therapeutic agents. Takagaki et al. has shown that isoflurane is more effective than propofol in preventing SDs and suggested isoflurane to be preferred as the sedative agent in acute ischemic stroke and trauma patients [130]. Animal experiments also show that isoflurane reduces the frequency of SDs during focal cerebral ischemia in rats and therefore have neuroprotective effects [131]. Another study of isoflurane shows that it prevents acquired epilepsy in rat models of temporal lobe epilepsy [132]. Experiments on pentobarbital have conflicting results [131, 133]. Chronic administration of migraine prophylactic drugs such as topiramate, valproate, amitriptyline, and methysergide has also shown to suppress SDs in a dose dependent manner, but chronic D-propranolol treatment was found to be ineffective at suppressing SDs [134]. Tonabersat is another promising SD inhibitor drug with potential uses in migraine prevention. It has been shown to reduce the median frequency of aura attacks by 71%, but has no effect on non-aura attacks [135]. Pregabalin, a GABA analogue, has also been shown to be effective at suppressing the initiation, wave speed and subcortical propagation of SDs in vitro [136]. Although there are many different experimental therapeutic agents (Fig. 3), there is no consensus on a particular agent in humans to be used.

1.4. Non-pharmacological Management Options of SDs at Intensive Care Units

In addition to pharmacologic agents, there are several clinically relevant applications that might inhibit the occurrence of SD waves and stabilize the perilesional tissue [13]. Normobaric hyperoxia [137], temperature control [138] and glucose management [139-141] have been shown to decrease the rates of SD waves and they must be considered as important management points for the patients followed in intensive care units.

SD stimulates fast glucose consumption leading to glucose depletion within minutes [39, 46, 142-147] and hypoglycemia prolongs the recovery period from SD, suggesting that glucose is essential in restoring the transmembrane ionic gradients after SD [39, 144, 148]. It has been shown that peri-infarct SDs can be suppressed in hyperglycemic patients [140, 149, 150]. Hoffmann et al. demonstrated hyperglycemia in rats elevated electrical thresholds for triggering SD, and reduced frequency of KCl-induced SD [139].

SDs are shown to increase brain temperature in intra-cerebral hemorrhage and TBI patients [119, 138, 151]. Hartings et al. also discussed that high core temperatures were associated with higher risk of depolarization, suggesting temperature control as a potential preventative measure for secondary insults [119].

Several studies have also shown that personalized management of cerebral perfusion pressure (CPP) is beneficial to patients with TBI [152, 153], and low CPP values are known to increase depolarization risk [154, 155] and reduce the neurovascular responses to depolarization [119, 156, 157].

Altogether, if the factors listed above may be followed and/or managed closely at intensive care units, morbidity and mortality of patients with stroke, SAH and traumatic brain injury may decrease significantly and may help to limit secondary tissue damage.

1.5. Future Directions and Research Limitations

In the last 15 years, monitorization of SDs in patients after stroke and brain trauma at intensive care units has been established in multiple centers under the framework of Co-Operative Studies on Brain Injury Depolarizations (COSBID). More than 500 patients have been monitored in multiple centers in various countries, and multiple trials are currently in process, such as DISCHARGE-1, Invasive and Non-Invasive Monitoring of Spreading Depolarization by Electrocorticography in Trauma and Stroke (INSPECT), NEWTON trial (Nimodipine microparticles to Enhance recovery While reducing TOxicity after subarachNoid hemorrhage) and MHS trials [158]. Although the monitorization of the actual patients pointed out some pathological pathways, there have also been some limitations of human trials of SDs. As ECoG is an invasive technique that requires electrode strips placed directly on top of the cortical surface, it can only be done in patients who are in need of neurosurgery, and as the level of consciousness alters in these patients, it is difficult to assess clinical symptoms of SD waves. More importantly, there is no clinical trial that demonstrates a benefit on the patient outcome. Secondly, initial lesions differ between patients and the location of the electrode and its distance to the lesion are also factors that might alter the results. The question that arises from these limitations is, how can SDs be identified in the absence of cortical strips. Marching pattern and transient nature of neurological deficits might make a physician assume that there are SDs, but forming a noninvasive diagnostic method is still an intriguing research subject. There have been some attempts of ima-ging SDs noninvasively. Previously, voxel based morphometry in high-resolution, three dimensional magnetic-resonance imaging was found to be effective in measuring learning induced brain plasticity [159]. A recently published article also suggested that structural MRI might also be used in monitoring functional changes like blood volume, blood flow and tissue oxygenation [160]. bold-fMRI has also been proposed as an alternative to gadolinium contrast agent-based perfusion assessment in acute stroke as it can significantly identify microvascular hypoperfusion in severely hypoperfused tissue in acute stroke [161].

While SDs are present as aggravating factors in many pathological conditions such as migraine, ischemic stroke, trauma and subarachnoid hemorrhage, there is also evidence suggesting that they might have neuroprotective effects [162]. It has been discussed that these neuroprotective properties may be due to uncoupling protein-5 (UCP-5) trigger [162], which is known to have a long-term effect upon neuron protection [163]. Also, brief acidosis in ischemic conditions are considered neuroprotective, which is called “pH paradox” [164-170]. This brief acidosis period was also reported to reduce stroke infarct size [164, 169]. Additionally, studies show that pre-conditioning cortex with SDs provides neuroprotection against subsequent ischemia which is characterized by decrease in final cell death and infarct size [171-175], and is thought to act likely by upregulation of cytokines and growth hormones [171, 176, 177], as well as down regulation of metabolism [178] and altered neurotransmission [179]. Also SDs potentially have a role in induction of neurogenesis and plasticity [180-182]. Therefore, it is still a topic of debate whether SDs should be considered completely pathological, or sometimes beneficial, and whether these two properties are mutually exclusive and if so how to distinguish them.

Another important issue is the translation of the know-ledge gathered through experimental animals to humans. Humans, swine and cats have a highly folded and more complex cortex, termed as ‘gyrencephalic’ compared to small animals that are widely used in research, such as mice and rats (lissencephalic brains). They have much less cortical foldings. A recent study compared SDs in swine, rat and human brains and observed that in addition to the differences in the propagation patterns due to anatomical differences, both lissencephalic and gyrencephalic brains exhibit heterogeneous propagation of SD waves [183]. Furthermore, Santos et al. described these heterogeneous SD propagation patterns in gyrencephalic brains [184], and suggested using movement-compensated intrinsic optical signal imaging (IOS) for the analysis of haemodynamic responses to SDs during secondary brain damage [185]. In the light of these findings, researchers must be cautious when translating research findings and observations on lissencephalic brains to human.