Keywords

Introduction

Potassium is an alkali metal and in water solutions it acts as a strong electrolyte if dissolved as KCl. As a consequence it is nearly 100% dissociated as a positive ion if present as KCl in a water solution. In animals potassium is a minor constituent of structures such as bone (Canas et al. 1969). Therefore it is mostly present as a fully dissociated cation in the body fluids. It is highly compartmentalized in that its concentration inside cells is much higher than in extracellular spaces. The concentration gradient is stabilized by various carriers, the most important is the Na/K-ATPase. The steep gradient of this cation and its high intracellular concentration make the potassium ion conductance useful as the most important contributor in setting a fairly negative resting membrane potential in many animal cells. Electrical signals like action and synaptic potentials very often involve changes in the conductance of the potassium ion. One has also to keep in mind that serum potassium shows fluctuations with dietary habits and has an effect on stroke risk. However, the potassium in this compartment is totally separate and in no way connected to the potassium dynamics of the brain syncytium unless a breach of the blood–brain barrier happened. The two factors are discussed separately below.

Requirement for Electroneutrality

Due to the large outwardly directed driving force, potassium ions exhibit a considerable net outward flux. If this flux is accomplished in an electroneutral way, net accumulation of potassium in the extracellular space will occur. Indeed, part of the potassium outflux occurs simultaneously as sodium influx, leading to sodium–potassium exchange. Part of the potassium efflux occurs together with chloride efflux, leading to KCl accumulation in the extracellular fluid spaces. Both, KCl outflow and Na:K exchange can occur simultaneously and can cause considerable potassium accumulation outside cells due to its electroneutrality In effect only during conditions of electroneutrality can a considerable shift of potassium from the inside to the outside of cells happen.

Dietary Potassium and Stroke Risk

Normal serum levels of potassium in humans are around 3.5–5.0 mM (Kratz et al. 2004). A recent meta-analysis (D’Elia et al. 2011) concluded that a 1.64-g-per-day higher potassium intake was associated with a lower stroke risk. Overall there was an inverse relationship between dietary potassium and stroke risk. As the permeability of the blood–brain barrier in adult rats and humans is relatively low for potassium with the only entrance mechanisms provided for by specific carriers and channels (Keep et al. 1995), this is a somewhat surprising finding. The cause for the protective effect of increased dietary potassium is therefore not obvious. The reason is that the blood pressure lowering effect of higher dietary potassium (Whelton et al. 1997) was adjusted for in the studies subject to the meta-analysis. Thus mechanisms other than the systemic effect of potassium on the blood pressure have to be taken into account. Several studies report the suppression of reactive oxygen species by higher dietary potassium, especially vascular lipid peroxides in stroke-prone spontaneously hypertensive rats (Ishimitsu et al. 1996; McCabe et al. 1994; Kido et al. 2008). Such interactions might not require an increased potassium flow into the brain parenchyma. In vitro experiments have shown that increased potassium concentrations inhibit the migration and proliferation of vascular smooth muscle cells (Ma et al. 2000a, b). In vivo potassium supplementation reduced salt-induced neointimal hyperplasia, adventitial macrophage infiltration, superoxide overproduction, and reduced nicotinamide-adenine dinucleotide phosphate oxidase activation and overexpression in rats (Kido et al. 2008). Thus it appears that the protective effect of high dietary potassium is exerted on the production of reactive oxygen species on the level of the artery wall and therefore counteracting the mechanisms leading to hypertension. This effect is not dependent on potassium crossing the blood–brain barrier.

Potassium Accumulation and Homeostasis During Normal Activity

The brain parenchyma is sheltered from the daily and long-term fluctuations in serum potassium concentrations by the blood–brain barrier. In the normal well-oxygenated and supplied brain parenchyma, extracellular potassium concentrations can still vary due to potassium outflow from neurons during action potential transmission in axons and synaptic potential propagation in synapses and dendrites as well as soma. between narrow confines due to regulatory mechanisms located in the neuronal, astrocytic, and oligodendrocytic cell membranes. Especially, the astrocytes seem to have developed a potassium homeostatic distribution system within their syncytium. Neuronal information processing is extremely sensitive to changes in the extracellular potassium concentration. The potassium concentration in the central nervous system has a resting level of 2.7–3.5 mM (Somjen 2002). Due to the high intracellular concentration and high resting permeability, small changes in these 3 mM translate into large changes in the equilibrium potential for potassium. This in turn affects membrane potential and therefore action potential frequency as well as synaptic potentials and their integration. Specifically, it was shown for example that increases of the external potassium from 3 to 5 mM change the action potential threshold in hippocampal neurons leading to hyperexcitability (Kreisman and Smith 1993). Increases which go beyond the 5 mM level change the efficacy of synaptic transmission in these neurons (Balestrino et al. 1986; Hablitz and Lundervold 1981). In these changes there is a suggestion that in addition the increased extracellular potassium concentration is due to a direct effect of the extracellular potassium ion on the gating mechanisms of participating ion channels (Leech and Stanfield 1981).

Pathological Potassium Accumulation: An Overview

What is the magnitude of the increase in the potassium concentration in the extracellular space in different situations? It is already known for some time that a single action potential will increase the potassium concentration by 1 mM above the 3 mM resting level (Frankenhaeuser and Hodgkin 1956; Adelman and Fitzhugh 1975; Clay 2005). The increase in extracellular potassium is correlated with the stimulation frequency and number of neuronal elements that contribute (Sykova 1991). Artificial, intensive stimulation of neuronal pathways causes an increase by nearly 5 mM above the resting level, but never exceeds this level. During seizures, the accumulation of extracellular potassium is further enhanced, but again there seems to be a ceiling level of 12 mM (Heinemann and Lux 1977). The reason for this is reuptake into neurons and various clearance mechanisms based in astrocytes and oligodendrocytes. These mechanisms are discussed elsewhere (Walz 2000).

In hypoxic/ischemic injury this ceiling is disrupted and the extracellular potassium concentration rises to extraordinarily high levels. The time course and extent of the extracellular potassium concentration is well established. It seems clear that there is a certain redistribution of intra- and extracellular potassium probably due to Na/K ATPase failure (D’Ambrosio et al. 2002). This is aggravated by swelling of the astrocytes and corresponding shrinkage of the extracellular space (Sykova and Nicholson 2008).

Stroke

During a stroke the extracellular potassium concentration is an almost accurate reflection of the health of the surrounding neurons. Anoxic depolarization is the first clear outward sign of energy failure. The subsequent events, however, differ in global and focal ischemia.

Global Ischemia

There is a difference between gray matter and white matter. Potassium increases in white matter have a slower time course and lower amplitude. This is probably due to the massive and uncontrolled release of excitatory transmitter substances in gray matter which contribute substantially to the potassium release (Ransom et al. 1992). The situation in gray matter was reviewed by Hansen (Hansen 1985). At the onset of a global ischemic, hypoxic, or anoxic event, there is a small transient potassium concentration increase beyond the normal 3 mM due to hyperexcitability. This is followed after cessation of hyperexcitability by a slow increase with the rate becoming increasingly faster and reaching about 10 mM (phase I). Then after about 2 min there is a fast, explosive increase to about 60 mM (phase II). Phase III thereafter increases the concentration much slower to 80 mM, which seems to represent total equilibrium between intra- and extracellular compartments. It has to be kept in mind that the volume of the extracellular space is shrinking considerably (Kraig and Nicholson 1978). It is assumed that at first hyperexcitability contributes to potassium, then energy failure with increased potassium conductances (Sick et al. 1982), followed by a sudden opening of unspecific pores in the cell membranes (Kraig and Nicholson 1978).

Focal Ischemia

Whereas the root cause of potassium and energy failure is similar in focal ischemia, the situation is far more complex. In a focal event, the necrotic core of an infarct is surrounded by a so-called penumbra, a peri-infarct tissue that is compromised, but still can be saved. In the infarct core the situation is not unlike the one in global ischemia (Nedergaard 1988). The core is heavily depolarized with a high extracellular potassium concentration. From there is a potassium gradient into the peri-infarct areas through penumbra to healthy tissue with normal concentrations. This increased potassium concentration in the peri-infarct area causes depolarizations of the neurons and worsens their energy state and therefore their health. The longer this lasts without recovery (reperfusion and reoxygenation) the worse the outcome for the survival of these peri-infarct neurons. Thus, such a necrotic infarct core will grow and expand at the cost of semi-healthy tissue in the penumbra. At worst the complete penumbra can be consumed by this state and consist of dying neurons (Balestrino 1995).

Spreading Depression Waves and Their Consequences

Things can get even more complicated. This is due to the regular spread of waves of depolarizations with high extracellular potassium from the core into surrounding penumbra and extending into healthy tissue. Spreading depression is a phenomenon which can be artificially elicited from rodent brain and sustained without detrimental effect through healthy tissue. For a long time it was questionable if spreading depression could be elicited and sustained in humans. This has now been shown to occur in humans and is a well-established fact (Dreier 2011).

The features and mechanisms of spreading depression have been reviewed elsewhere (Somjen et al. 1992). Here is a brief introduction into the phenomenon. A prerequisite for spreading depression is normal brain architecture with neurons and astrocytes. It cannot occur in white matter tracts or scar tissue devoid of neurons. It can be elicited by a massive depolarization irrespective of its origin, be it mechanical, electrical, or chemical stimuli. A long-lasting stimulus will trigger a spreading depression wave approximately every 10 min. Such a wave of a near-total membrane depolarization (representing near-total equilibration of the electrical gradients) migrates with a velocity of 1.5–7.5 mm/min through the parenchyma. It is a concentric wave which spreads from its origin outwards and is accompanied by a negative shift of the extracellular potential of up to 30 mV. The extracellular space contracts by 50% due to transient swelling of astrocytes and dendrites. The transient increase of the extracellular potassium concentration to values between 30 and 80 mM is the most distinct hallmark of such a wave, however. After 1–2 min, the potassium concentration will start to reverse to normal concentrations, as do other ionic changes including pH values. Neurons need several more minutes to regain their normal excitability and synaptic transmission. The wave encompasses dilated blood vessels and increased local blood flow. This is an indication that these waves, or better the re-establishment of normal gradients after a wave, require considerable amounts of energy. The mechanisms involved are not clear at all. It seems that astrocytes are not necessary and do not participate in the propagation (Largo et al. 1996). There is now good evidence that neuronal gap junctions play a role in the propagation (Somjen 2004).

In focal ischemia, spreading depression waves originate at the interphase between the core of the necrotic tissue and the penumbra at the infarct rim (Balestrino 1995). This is probably due to the increased extracellular potassium concentration of the core acting as a chemical stimulus. This stimulus will cause successive waves of spreading depression as soon as the tissue recovered enough and is ready to repeat the episode. This occurs approximately every 10 min at least in rats. The wave spreads through the penumbra into healthy tissue. The penumbra has a close-to-normal extracellular potassium concentration, but blood flow through it is reduced. Thus a wave of spreading depression is causing a mismatch between energy supply and demand. This is indicated by the fact that the duration of the spreading depression wave during its passage of a given location is approximately eight times longer in the penumbra than in healthy tissue (Nedergaard 1996). This results in considerable stress to the penumbra tissue. It will lead to increased damage and slow outward expansion of the necrotic core. This means the waves play an important role in extending the damage of a focal stroke. Indeed it was found that there is a linear relationship between the number of waves passing through the penumbra and infarct volume at least in rats (Mies et al. 1993). Blocking spreading depression waves which radiate out of a necrotic core into the penumbra and beyond has therefore a beneficial effect on reducing stroke volume and neuronal survival (Nedergaard and Diemer 1987). Also when rats are rendered hyperglycaemic before induction of focal ischemia, the number of waves and neuronal death are reduced presumably due to increased energy availability (Nedergaard et al. 1988).

The dynamics of the extracellular potassium concentration at a given location during the passing of a wave resembles the potassium increase in global ischemia. This is an indication that the underlying mechanisms are similar if unknown. It seems certain, however, that the sudden opening of pore-like structures in the cell membranes contribute heavily to the mechanism (Phillips and Nicholson 1979).

Blood–Brain Barrier and Potassium

There are indications that the clear separation of serum potassium and brain parenchymal potassium due to the function of the blood–brain barrier is not always accomplished. During normal conditions the permeability of the blood–brain barrier for potassium is low. A bolus of high potassium injected in the carotid artery of rats did not lead to a detectable rise of the potassium concentration in the brain as measured with ion-selective microelectrodes (Hansen et al. 1977). This was confirmed with permeability measurements of radiotracers (Smith and Rapoport 1986). Reduced blood flow in the rat brain, as low as a quarter of the normal rate and thus mimicking severe stroke, reduced the permeability for potassium of the blood–brain barrier (Hom et al. 2001). Recently there has been some evidence that a bumetanide-sensitive Na-K-Cl cotransporter in the blood–brain barrier and in astrocytes is activated in experimental stroke and contributing to brain edema. This has been confirmed with mice in which the NKCC1 transporter was knocked out and which have a considerable reduction in infarct volume and edema after transient ischemia (Kahle et al. 2009). Intravenous injection of bumetanide decreases edema formation after experimental stroke. As bumetanide is not crossing the blood–brain barrier and the NKCC1 transporter is located at the luminal side of endothelial cells, this is an indication that Na-K-Cl uptake from blood to parenchyma together with excessive water flow is the root cause of edema formation. Although this water net flow is a cause of edema, it is not clear that the amount of potassium transported across the barrier is of any significance other than as an osmotically active compound.

Reactive Astrocytes and Potassium Homeostasis

There is another more beneficial effect of spreading depression waves. If in an experimental situation repeated waves are created in an otherwise normal animal, at first reactive microglia and then reactive astrocytes are formed in this tissue (Gehrmann et al. 1993). Reactive astrocytes are in all likelihood beneficial in protecting ailing neurons, although not conducive to regeneration of axons which are sprouting from surviving neurons (Li et al. 2008). In fact it was shown that if a 2 h experimentally induced period of spreading depression waves in otherwise normal rats was followed 3 days later with a focal ischemic insult, these rats had a reduced infarct volume by 50% compared to controls (Matsushima et al. 1996). The difference was not caused by differences in blood flow rates. It is interesting that the one study which investigated potassium clearance properties of reactive astrocytes in situ found these more effective than in normal astrocytes (Walz and Wuttke 1999). Therefore a curious situation seems to exist: potassium supported spreading depression waves invade healthy tissue out of the penumbra of a focal stroke. This does not expand the penumbra into the healthy (but expands necrotic core into penumbra). However, it transforms normal astrocytes into reactive astrocytes within days. Another ischemic event is now much better managed due to the protective nature of these reactive astrocytes. This phenomenon is known as pre-ischemic conditioning. Among other features these astrocytes exhibit a vastly improved potassium clearance system. This will act against further potassium increases and therefore counteract new spreading depression waves.

Conclusions

Brain potassium may not be the root cause of events leading to stroke and causing damage. However, it is intimately involved in all the dynamics of expansion of the stoke volume and neuronal death. It is therefore a leading indicator of the state of health of the neurons. It is a vehicle for spreading depression and therefore acts as a double-edged sword: detrimental in the penumbra, protective in healthy tissue. Dietary potassium in the serum seems to be beneficial due to the strict separation of the two potassium pools in the serum and brain.