B OTH resting and action potentials of neurons are

J Neurosurg 62:698-703, 1985 The role of calcium and cellular membrane dysfunction in experimental trauma and subarachnoid hemorrhage OTAKAR R. HUBSCHMANN, M.D., AND DOUGLAS C. NATHANSON, M.A. Department of Neurosurgery, Veterans Administration Medical Center, East Orange, and University of Medicine and Dentistry, Newark, New Jersey v- Acute subarachnoid hemorrhage (SAH) and intraparenchymal hematoma (IPH) in cats are accompanied by massive cellular depolarization. This depolarization, characterized by potassium (K +) effiux and calcium (Ca ++) influx, results in membrane destabilization, osmotic imbalance, and a decrease in electrical conduction. The Ca ++ influx appears to initiate a chain reaction that, in some instances, may result in delayed cell destruction. The ionic dysequilibrium probably contributes to both brain engorgement and spasm in large vessels. The cellular depolarization and calcium-induced cell membrane injury at the moment of impact may play a greater role in the pathophysiology of head trauma than previously thought. KEY WORDS ~ intraparenchymal hematoma 9 subarachnoid hemorrhage 9 head injury 9 potassium 9 calcium 9 neuronal membrane B OTH resting and action potentials of neurons are dependent on the selective permeability of cell membranes and the energy-dependent function of electrochemical pumps. The propagation of action potentials is produced by changes in ionic permeability to various ions in a given area of the neuronal membrane, while restoration of resting potentials depends on the function of electrochemical pumps. In general terms, the neuronal potential is a function of the ratio of the concentration of ions outside the cell to the ions inside the cell. Normal equilibrium of ions inside and outside the cell membrane is essential for normal neuronal function. Although equilibrium of all the ions is important, potassium plays a particularly important role in determining the resting potential. The low concentration of potassium in the extracellular space is kept remarkably constant at around 3.0 mm/liter. 24 Any significant increase in the K + concentration in the extracellular space has a profound effect on the resting membrane potential and the electrical properties of the axons. ~-~3 A comparable situation exists at synapses, where calcium plays a similar role. 17 At the arrival of an action potential, increased permeability of the presynaptic membrane to calcium ions causes calcium ion cellular entry from extracellular spaces, where it is in much higher concentration, and the release of a neurotransmitter. Low extracellular calcium levels will impair synaptic function, and under certain conditions may prevent it altogether. Intracellular ions and neurotransmitters liberated during mechanical displacement of the membrane upon impact in head injury also have the capacity to inflict further damage upon cells in the vicinity. These two factors could therefore, at least theoretically, significantly enlarge the extent of the original injury and account for some of the signs of clinical deterioration. Experimental data from our laboratory lj-~3 indicate that subarachnoid hemorrhage (SAH) and experimental injury are accompanied by movements of potassium (K +) and calcium (Ca ++ ) at the moment of trauma. In addition to the effects an intracellular movement of Ca ++ may have on the resting and action potentials, recent data indicate that it may induce a chain reaction, eventually leading to cell destruction. 9'1~ In this investigation, we have studied the movements of K + and Ca ++ using ion-specific electrodes in experimental SAH and intraparenchymal hemorrhage (IPH), and the effect of hemorrhage on the electrochemical membrane pumps. We have also tested the hypothesis that uncontrolled fluxes of K + and Ca ++ not only result from membrane dysfunction caused by the trauma, but may also increase the extent of damage subsequent to the initial injury. Materials and Methods Two groups consisting of 10 adult cats each were used. In addition, four animals that received either 698 J. Neurosurg. / Volume 62/May, 1985

Calcium and membrane dysfunction in trauma and SAH intraparenchymal or subarachnoid injections of saline served as controls. The animals were anesthetized with sodium pentobarbital (Nembutal, 35 mg/kg) and maintained on regular maintenance intravenous doses, using the animals' reaction to pain, pupillary size, and electrocorticogram (ECoG) pattern as indicators of the level of anesthesia. A tracheostomy was performed, and the femoral vein and artery were cannulated. Blood pressure was monitored. Blood gases were controlled to keep the pco2 between 27 and 32 mm Hg by allowing animals to breathe unassisted or on a respirator subsequent to a paralyzing dose of Flaxedil (gallamine triethiodide, 3 mg/kg). The cerebral cortex was exposed through a left frontal craniectomy and the dura was opened under an operating microscope. Recordings were made from the posterior sigmoid gyms. The cortical surface was kept moist with constant irrigation of mock cerebrospinal fluid (CSF), and cortical and rectal temperatures were maintained between 36* and 380C by means of a heating lamp and pad connected to a temperature monitor and controller. Electrical activity of the cortical cells was monitored by ECoG, with a surface Ag-AgC1 electrode referenced against cervical muscles in a monopolar fashion. Potassium (K +) and calcium (Ca ++) ion-selective microelectrodes were manufactured applying our modifications to techniques previously described. ~ Glass micropipettes were prepared from borosilicate glass tubing using a Narashige puller. Pipettes with tips between 5 and 10 um in diameter were siliconized with 1% Siliclad in water by drawing a small column of the solution into the tip and baking at 200~ for 2 hours. The tips of the electrodes were then filled with either a K selective ion-exchanger (Corning 477317 resin) or a Ca++-selective ion-exchanger (WPI IE-200 resin). Potassium and calcium electrodes were then backfilled with 100 mm KCI and 100 mm CaCI2, respectively. Reference electrodes were prepared by filling microelectrodes with 5 to 10 um tip diameters with 150 mm NaC1. The reference and ion-selective electrode were then cemented together so that their tips were within 50 um of each other. The activity from the electrode pairs was fed directly into matched high-impedance probes (1015) and differentially amplified with a common-mode rejection ratio of 10,000:1. The K + and Ca + activity, cortical field potential (recorded from the reference electrodes), and ECoG were displayed on a polygraph and oscilloscope. The electrode pairs were calibrated in solutions of 150 mm NaC1 containing amounts of KC1 (between 3 and 30 mm) or CaC12 (between 1 and 10 mm) just prior to and immediately following use in the animals. The electrodes exhibited a response time of less than 200 msec and high selectivity against interfering cations. Only electrodes with slopes greater than 50 mv per decade change in K + concentration and 26 mv per decade change in Ca ++ concentration were used. The IPH or SAH was produced by 0.5 to 1.0 cc of autologous blood obtained from the femoral vein being injected into the frontal subcortical white matter or subarachnoid space, respectively. These models of IPH and SAH have been studied previously in our laboratory. Injection of saline in four cats was used as the control technique. Following termination of the experiments, the animals were sacrificed with an overdose of Nembutal. The K release and Ca ++ depletion from the extracellular space that followed the cardiac standstill were used both as a model of global ischemia and as confirmation that the electrodes were functional. The animals were perfused with 10% formaldehyde and the brains were removed. The brains were stored in 10% formaldehyde solution for 4 weeks, after which they were cut serially in sections 1.0 cm thick. The extent of the IPH or SAH, location of the IPH, configuration and size of the ventricles, presence of herniation, and presence of macroscopic changes in the brain stem were evaluated. Sections 50 ~m thick were taken from the recording areas, from the vicinity of the hemorrhage, and from the brain stem at the midcollicular level. The sections were stained with cresyl violet and examined with x 40 magnification. Results Intraparenchymat Hemorrhage Following the injection of blood into the brain tissue, there was a rapid increase in the extracellular K + concentration from the preinjection level of 3.20 + 0.41 mm to peak levels ranging between 14 and 32 ram. This increase in extracellular K + was accompanied by a depletion of extracellular Ca ++ levels from a preinjection baseline of 1.11 + 0.07 mm to levels between 0.8 and 0.24 mm. Although the actual values varied from experiment to experiment, this combination of K + release and Ca ++ depletion was present in every instance. Along with these ionic changes, there was an attenuation of the ECoG, characterized by a depression in voltage, the disappearance of spindle activity, and a cortical depolarization of 15 to 23 mv. In five animals, the K + concentration returned to preinjection levels within 5 minutes, while in three other animals K + remained elevated 0.25 to 1.5 mm above baseline for the duration of the experiment (2 to 3 hours). In two animals with herniated cerebral cortex, the K + levels remained elevated at 9.3 and 13.4 mm for the duration of the experiment. In all of the animals with IPH there was no return of the Ca ++ to baseline levels. In eight animals, the Ca ++ remained depressed at levels ranging between 0.68 and 0.90 mm, and in the two animals with ruptured cerebral cortex it was between 0.28 and 0.33 mm. No improvement was seen in the ECoG patterns for any of the animals (Fig. 1). The pathological examination confirmed the presence of an IPH. While there was some subarachnoid blood present at the injection site, there was none at the recording site. The major portion of the hemorrhage J. Neurosurg. / Volume 62/May, 1985 699

O. R. Hubschmann and D. C. Nathanson FIG. 1. A permanent decrease in Ca ++ in the extracellular space seen in a model of intraparenchymal hemorrhage (IPH). Ca ++ = calcium ion-specific electrode recording; K = potassium ion-specific electrode recording; fpk = field potential K+; fpca = field potential Ca and ECoG = electrocorticogram. The injection of blood into the brain is indicated by arrow (IPH). was confined to the subcortical white matter and varied from 0.5 to 1.0 cm in diameter (Fig. 2). There was no involvement of the thalamus or midline structures, nor was there penetration into the ventricles. Ventricular size and configuration were equal on both sides. The structures of the cells at the recording area, examined at 40, showed no gross evidence of destruction in any of the experiments, with the exception of the two animals with ruptured cortex. Subarachnoid Hemorrhage Injection of blood into the subarachnoid space resulted in a cellular response similar but not identical to that seen following IPH. Subsequent to SAH there was FIG. 2. Brain section showing the location and extent of intraparenchymal hemorrhage. The recordings were made from the gyrus overlying the hematoma. The anatomically intact cells of the area from which the recordings were made represent the marginal zone in the traumatic penumbra according to our hypothesis. rapid extracellular accumulation of K +, extracellular Ca ++ depletion, profound cellular depolarization, and, in four cases, attenuation of the ECoG. Cortical extracellular baseline levels of Ca ++ and K + averaged 1.21 + 0.06 mm and 3.18 _+ 0.43 mm, respectively. The production of SAH resulted in a rapid increase in the extracellular K + concentration to a peak level of 16 to 28 mm, and a reduction of the extracellular Ca ++ concentration to a level of 0.4 to 0.7 mm. Within 5 minutes, the extracellular K + and field potentials returned to normal levels. Two basic patterns of Ca ++ changes occurred following the initial decrease. The first Ca ++ pattern, seen in four animals, showed a return to normal values following SAH. In two of these instances a period of transient extracellular hypercalcemia was seen, where the Ca ++ pattern was dissociated from the K + levels, and the return of Ca ++ to normal preceded the return of K + (Fig. 3). In the second pattern, seen in the remaining six animals, the Ca ++ did not return to normal for the duration of the experiment, but remained depressed at levels between 0.6 and 0.9 mm. Histological examination confirmed the presence of SAH over the cortex. No gross trauma was identifiable at x 40 magnification at the injection or recording sites in any of the animals. Control Studies Injection of saline into the brain or subarachnoid space resulted in little or no effect on the parameters being measured. There was a small transient elevation of K and a drop in Ca ++ which returned to baseline levels within 1 minute. There was no effect on the ECoG and no consistent change in field potential recordings following control injections of saline. The reasons for these changes may be related to the mechanical manipulation of the cortex produced by the injection of saline. 700 J. Neurosurg. / Volume 62/May, 1985

Calcium and membrane dysfunction in trauma and SAH FIG. 3. Recordings showing the intracellular entry of calcium and its extrusion as seen in some instances of experimental subarachnoid hemorrhage (SAH). Transient extracellular Ca depression is followed by its elevation, most likely representing increased activity of the membrane ionic pumps. Ca ++ = calcium ion-specific electrode recording; K + = potassium ion-specific electrode recording; fpk = field potential K+; fpca = field potential Ca+ and ECoG = electrocorticogram. The injection of blood into the brain is indicated by arrow (SAH). Discussion Extravasation of blood into the subarachnoid space or brain parenchyma, similar to that created in our model, appears in all forms of head trauma or spontaneous SAH in humans. Our experiments indicate that both instances are accompanied by a massive cell membrane dysfunction. This dysfunction initiates the movement of K out of and Ca ++ into the cells. As reported previously, this does not appear to be caused by a decrease in blood supply. ~-~3 In our model of IPH, ionic movements also take place in cells that have not been directly damaged at the time of injury and are located at some distance away from the lesion. This is of particular importance because it indicates that there is a population of cells of which the membranes have not been destroyed by the impact but, judging from the ionic fluxes, which have developed an early membrane instability. This observation suggests that, similar to ischemic damage, a traumatic penumbra of cells with varying degrees of membrane injury develops shortly after head trauma (Fig. 4). During the peak of the K release, levels in the extracellular space are high enough to suppress axonal and synaptic transmission and lead to an increase in glial metabolism and swelling. 7'~'12 The effect on the microvasculature is less clear, but varying degrees of vasodilation, vasoconstriction, and an increase in vascular resistance have been described. 2'5"~5,22,25 In addition, it is known that K + elevations are accompanied by simultaneous changes in Na +, CI-, and extracellular osmolality. 23 Changes in the reactivity and diameter of FIG. 4. Diagram illustrating the traumatic penumbra. Traumatic injuries result in development of zones of cell injury of different intensity. The cells that are exposed to the full impact, or those that are more sensitive to injury because of their anatomical or biochemical properties, are destroyed (black circle, dead cells). A large population of cells have their membranes destabilized but not destroyed, and have the potential to recover (marginal zone). Among the factors capable of enlarging the traumatic penumbra, converting a subcritical injury (marginal zone) to a critical one (black circle, dead cells), are the K+/Ca ++ fluxes, extracellular edema, ischemia, hypoxia, and increased intracranial pressure, alone or in combination. small vessels following the K + release may have particular importance in changing the brain-blood volume ratio, due to pooling of blood in the cerebral vessels, resulting in the development of an increase in intracranial pressure after head trauma.16'~8 It is generally accepted that Ca ++ plays an essential role in maintenance of membrane stability, transmitter release, and transmembrane electrical charges. 3,4,6'17 In all our experiments, the K + release was accompanied by a simultaneous disappearance of Ca ++ from the extracellular space, indicating its movement into the cells. Two basic patterns of Ca ++ uptake are seen. The first pattern, where Ca ++ intake is followed immediately by its extrusion, suggests that the cells were not permanently damaged (Fig. 3). The transient Ca ++ elevation most likely represents an increased activity of the Ca ++ adenosine triphosphatase (ATPase) pumps aimed at restoration of the Ca ++ homeostasis. The inward Ca ++ movement, without its extrusion, that was seen in most animals with SAH and in all animals with IPH probably indicates more serious damage to the membrane and its ionic pumps (Fig. 1). Movement of Ca ++ into the cells and an increase in its cytoplasmic concentration have been described in several other pathological states. Most investigators consider the influx of Ca ++ to be a pivotal factor in determining the survival of the cell. Schanne and his colleagues 2~ reported that the viability of cells in tissue culture, when exposed to 10 different toxins, was directly related to the presence of J. Neurosurg. / Volume 62/May, 1985 701

O. R. Hubschmann and D. C. Nathanson FIG. 5. Diagram illustrating the calcium theory of traumatic cell membrane disruption. Mechanical trauma results not only in anatomical destruction of some parenchymal and vascular elements but also in dynamic changes in ionic fluxes and neurotransmitters. Chemical substances present in extravasated blood, in combination with the K + and neurotransmitters released from the destroyed and destabilized cells, adversely affect all elements of the normal neuropil. The most susceptible elements are the destabilized cells. Although not destroyed by the initial trauma, they may undergo a process of membrane autolysis started by the movement of calcium into the cytoplasm at the moment of impact. The hypothetical sequence of events at the cell membrane that may lead to "self-destruction" are depicted on the fight side of the diagram. Normal neuropil: N = neuron, G = glia, V = microvasculature. Ca ++ in the media. Absence of calcium in the media prevented the toxins from causing cell death. Farber, et al., 9 implicated the intracellular influx of Ca** during postischemic reperfusion or partial ischemia in causing an irreversible cell injury. Siesj6 zl similarly implicated Ca *+ in ischemic cell death in the brain; and Schanne, et al.,2~ considered the calcium entry into the cell to be the final common pathway in cell death from a variety of causes. In general, it appears that movement of Ca ++ from the extracellular space into the cytoplasm sets off a cascade of events leading to membrane dysfunction and possibly to complete cell destruction/9 Elevated cell Ca ++ levels lead to the activation of the phospholipases A and C that attack the phospholipid component of the cell membrane. This, in turn, results in the liberation of arachidonic acid which, in the presence of some blood flow and oxygen, will be metabolized by normally present enzymatic systems within the cell. 9'19 The metabolic degradation of arachidonic acid leads to the liberation of prostaglandins, free oxygen radicals, thromboxane, and leukotrienes. All of these substances can further damage the cell membrane and some can affect the macro- and microvasculature as well. 9' 10,14,19,26 At the level of large vessels, the movement of Ca ++ into the cells of the muscularis layer may result in the development of spasm and secondary ischemia not uncommon in head trauma and SAH.I'8 At the cell level, the eventual result is the development of a membrane that is no longer selectively permeable, but grossly "leaky," allowing uncontrollable expansion of the cell volume. This inability to control its volume is the ultimate sign of cell dysfunction and leads, in the most severe instances, to its destruction (Fig. 5). Even without mechanical destruction, however, the changes in ionic homeostasis within the cell cause the cell to swell and lead to a temporary cessation of cell function. Some of these cells will recover their normal function, while others will be permanently destroyed. Which factors determine the recovery or destruction are not known, but ischemia, hypoxia, or increased intracranial pressure are thought to be able to convert a subcritical to a critical injury (Fig. 4). It appears, however, that the initial influx of Ca ++ alone may have the capability to start an independent process of membrane autolysis, and may under certain circumstances lead to delayed destruction of cells sometime after the initial injury (Fig. 5). 702 J. Neurosurg. / Volume 62/May, 1985

Calcium and membrane dysfunction in trauma and SAH Summary At the cell level there are at least two important processes in the immediate response of the brain to spontaneous SAH and trauma. The first, well described in the past, results from disruption of the blood-brain barrier and extravasation of fluid into the extracellular space. The second, not fully appreciated in the past, which may precede or take place simultaneously with the first, is the development of massive cellular depolarization. Movements of K + out of injured cells are coupled with movements of Ca ++ into the cells, both in the immediate vicinity and at some distance from the region of the impact. The combination of these ionic changes results in a significant depression of neuronal function and an increase in glial metabolism. It most likely plays a role in the development of changes in vascular caliber and cerebral swelling. Delayed cell death may be caused by the calcium influx, via activation of enzymatic processes attacking the cell's own membrane. This may eventually lead to membrane destruction in cells only subcritically injured at the time of impact. Our results suggest that cell membrane dysfunction leading to delayed cell destruction may be addressed therapeutically. Prevention of Ca ++ influx, early Ca ++ extrusion from the intracellular space, early interruption of the arachidonic acid degradation process, or fortification of the membranes against the cell's own enzymes are only a few of the theoretical therapeutic possibilities requiting investigation. References I. Allen GS, Gross CJ, Henderson LM, et al: Cerebral arterial spasm. Part 4: In vitro effects of temperature, serotonin analogues, large non-physiological concentrations of serotonin, and extracellular calcium and magnesium on serotonin-induced contractions of the canine basilar artery. J Neurosurg 44:585-593, 1976 2. Auen EL, Bourke RS, Barron KD, et al: Alteration in cat cerebrocortical capillary morphometrical parameters following K+-induced cerebrocortical swelling. Acta Neuropathol 47:175-181, 1979 3. Baker PF: Transport and metabolism of calcium ions in nerve. Prog Biophys Mol Biol 24:177-223, 1972 4. Baker PF, Reute H: Calcium Movement in Excitable Cells. Oxford: Pergamon Press, 1976 5. Betz E, Enzenross HG, Vlahov V: Interaction of H + and Ca ++ in the regulation of local pial vascular resistance. Pfluegers Arch 343:79-88, 1973 6. Blaustein MP: The interrelationship between sodium and calcium fluxes across cell membranes. Rev Physiol Biochem Pharmacol 70:33-82, 1974 7. Bourke RS, Nelson KM: Further studies on the K +- dependent swelling of primate cerebral cortex in vitro: the enzymatic basis of the K+-dependent transport of chloride. J Neurochem 19:663-685, 1972 8. Columella F, Delzanno GB, Gaist G, et al: Angiography in traumatic cerebral lacerations with special regard to some less common aspects. Acta Radiol (Diagn) 1: 239-247, 1963 9. Farber JL, Chien KR, Mittnacht S Jr: The pathogenesis of irreversible cell injury in ischemia. Am J Pathol 102:271-281, 1981 10. Hass WK: Beyond cerebral blood flow, metabolism and ischemic thresholds: an examination of the role of calcium in the initiation of cerebral infarction, in Meyer JS, Lechner H, Reivich M, et al (eds): Cerebral Vascular Disease 3. Amsterdam: Excerpta Medica, 1981, pp 3-17 11. Hubschmann OR, Kornhauser D: Cortical cellular response in acute subarachnoid hemorrhage. J Neurosurg 52:456-462, 1980 12. Hubschmann OR, Kornhauser D: Effect of subarachnoid hemorrhage on the extracellular microenvironment. J Neurosurg 56:216-221, 1982 13. Hubschmann OR, Kornhauser D: Effects of intraparenchymal hemorrhage on extracellular cortical potassium in experimental head trauma. J Neurosurg 59:289-293, 1983 14. Kontos HA, Wei EP, Povlishock JT, et al: Cerebral arteriolar damage by arachidonic acid and prostaglandin G2. Science 209:1242-1245, 1980 15. Kuschinsky W, Wahl M, Bosse O, et al: The dependency of the pial arterial and arteriolar resistance on the perivascular H + and K + concentrations. A micropuncture study. Eur Neurol 6:92-95, 1971/72 16. Langfitt TW, Weinstein JD, Kassell NF: Cerebral vasomotor paralysis produced by intracranial hypertension. Neurology 15:622-641, 1965 17. Llinas R: The role of calcium in neuronal function, in Schmitt FO, Worden FG (eds): The Neurosciences: Fourth Study Program. Cambridge, Mass: MIT Press, 1979, pp 555-571 18. Marshall LF: Treatment of brain swelling and brain edema in man, in Cerv6s-Navarro J, Ferszt R (eds): Brain Edema. Advances in Neurology, Vo128. New York: Raven Press, 1980, pp 459-470 19. Raichle ME: The pathophysiology of brain ischemia and infarction. Clin Neurosurg 29:379-389, 1982 20. Schanne FAX, Kane AB, Young EE, et al: Calcium dependence of toxic cell death: a final common pathway. Science 206:700-702, 1979 21. Siesj6 BK: Cell damage in the brain: a speculative synthesis. J Cereb Blood Flow Metab 1:155-185, 1981 22. Somlyo AV, Haeusler G, Somlyo AP: Cyclic adenosine monophosphate: potassium-dependent action on vascular smooth muscle membrane potential. Science 169: 490-491, 1970 23. Van Harreveld A: Brain Tissue Electrolytes. London: Butterworths, 1966 24. Vyskocil F, Kriz N, Bures J: Potassium-selective microelectrodes used for measuring the extracellular brain potassium during spreading depression and anoxic depolarization in rats. Brain Res 39:255-259, 1972 25. Wade JG, Amtorp O, Sorensen SC: No-flow state following cerebral ischemia. Role of increase in potassium concentration in brain interstitial fluid. Arch Neurol 32: 381-384, 1975 26. Wolfe LS: Eicosanoids: prostaglandins, thromboxanes, leukotrienes, and other derivatives of carbon-20 unsaturated fatty acids. J Neurochem 38: I- 14, 1982 Manuscript received August 24, 1984. Address reprint requests to." Otakar R. Hubschmann, M.D., University of Medicine and Dentistry, 100 Bergen Street, Newark, New Jersey 07103. J. Neurosurg. / Volume 62/May, 1985 703