Faculty Bibliography

Exposure to hypoxic, acidic, ion-shifted Ringer (HAIR) for 15-40 min has been shown to cause rapid astrocyte death upon reperfusion with normal media. The ion shifts of the HAIR solution included a rise in extracellular K(+) (e.g., [K(+)](o)) and a fall in [Na(+)](o), [Cl(-)](o), and [Ca(2+)](o), characteristic of ischemic-traumatic brain insults. We investigated the ionic basis of the HAIR-induced injury. After HAIR exposure, reperfusion in 0 Ca(2+)/EGTA media completely protected astrocytes. Preincubation of cells in BAPTA-AM ester was also protective, indicating that the injury was triggered by Ca(2+) influx during reperfusion. Neither nimodipine, CNQX, APV, nor TTX reduced injury. Astrocyte death could be blocked by 100 microM Ni(2+) or 100 microM benzamil, suggesting involvement of Na(+)-Ca(2+) exchange. KB-R7943, which preferentially inhibits reverse Na(+)-Ca(2+) exchange, also protected astrocytes. Elevation of [K(+)](o) was not necessary for astrocyte death. However, when [Na(+)](o) was maintained at 151 mM throughout the HAIR protocol, cell death was markedly reduced. We postulate that [Na(+)](o) shifts aid reversal of Na(+)-Ca(2+) exchange by favoring cytosolic Na(+) loading. Possible means of astrocytic Na(+) accumulation are discussed

Death of astrocytes requires hours to days in injury models that use hypoxia, acidosis, or calcium paradox protocols. These methods do not incorporate the shifts in extracellular K(+), Na(+), Cl(-), and Ca(2+) that accompany acute brain insults. We studied astrocyte survival after exposure to hypoxic, acidic, ion-shifted Ringer (HAIR), with respective [Ca(2+)], [K(+)], [Na(+)], [Cl(-)], and [HCO(-)(3)] of 0.13, 65, 51, 75, and 13 mM (15% CO(2)/85% N(2), pH 6.6). Intracellular pH (pH(i)) was monitored with the fluorescent dye BCECF. Cell death was indicated by a steep fall in the pH-insensitive, 440-nm-induced fluorescence (F440) and was confirmed by propidium iodide staining. After 15-40-min HAIR exposure, reperfusion with standard Ringer caused death of most cultured (and acutely dissociated) astrocytes within 20 min. Cell death was not prevented if low Ca(2+) was maintained during reperfusion. Survival fell with increased HAIR duration, elevated temperature, or absence of external glucose. Comparable durations of hypoxia, acidosis, or ion shifts alone did not lead to acute cell death, while modest loss was noted when acidosis was paired with either hypoxia or ion shifts. Severe cell loss required the triad of hypoxia, acidosis, and ion shifts. Intracellular pH was significantly higher in HAIR media, compared with solutions of low pH alone or with low pH plus hypoxia. These results indicate that astrocytes can be killed rapidly by changes in the extracellular microenvironment that occur in settings of traumatic and ischemic brain injury

We tested the hypothesis that extracellular membrane-bound carbonic anhydrase (CA) type IV is responsible for the regulation of interstitial pH (pH(o)) transients in brain. Rat hippocampal slices were incubated in phosphatidylinositol-specific phospholipase C (PI-PLC), which cleaves the link of CA IV to the external face of plasma membranes. Then evoked alkaline pH(o) shifts were studied in a recording chamber, using pH microelectrodes. Incubation fluid was saved for later analysis. The ability to buffer a rapid alkaline load was reduced markedly in PI-PLC-treated tissue as compared with adjacent, paired control slices. The effect of benzolamide (a poorly permeant CA inhibitor) on evoked pH(o) shifts was diminished greatly in the PI-PLC-treated tissue, consistent with the washout of interstitial CA. Treatment of the incubation fluid with SDS abolished nearly all of the CA activity in fluid from controls, whereas an SDS-insensitive component remained in the fluid from PI-PLC-treated slices. These data suggested that CA type II (which is blocked by SDS) leaked from injured glial cells in both slice preparations, whereas CA type IV (which is insensitive to SDS) was liberated selectively into the fluid from PI-PLC-treated tissue. Western blot analysis was consistent with this interpretation, demonstrating a predominance of CA IV in the incubation fluid from PI-PLC-treated tissue and variable amounts of CA II in fluid from PI-PLC-treated and control slices. These results demonstrate that interstitial CA activity brain is attributable principally to membrane-bound CA IV

Spreading depression (SD) and related phenomena have been implicated in hypoxic-ischemic injury. In such settings, SD occurs in the presence of marked extracellular acidosis. SD itself can also generate changes in extracellular pH (pH(o)), including a pronounced early alkaline shift. In a hippocampal slice model, we investigated the effect of interstitial acidosis on the generation and propagation of SD in the CA1 stratum radiatum. In addition, a carbonic anhydrase inhibitor (benzolamide) was used to decrease buffering of the alkaline shift to investigate its role in the modulation of SD. pH(o) was lowered by a decrease in saline HCO(3)(-) (from 26 to 13 to 6.5 mM at 5% CO(2)), or by an increase in the CO(2) content (from 5 to 15% in 26 mM HCO(3)(-)). Recordings with pH microelectrodes revealed respective pHo values of 7.23 +/- 0. 13, 6.95 +/- 0.10, 6.67 +/- 0.09, and 6.97 +/- 0.12. The overall effect of acidosis was an increase in the threshold for SD induction, a decrease in velocity, and a shortened SD duration. This inhibition was most pronounced at the lowest pH(o) (in 6.5 mM HCO(3)(-)) where SD was often blocked. The effects of acidosis were reversible on return to control saline. Benzolamide (10 microM) caused an approximate doubling of the early alkaline shift to an amplitude of 0.3-0.4 U pH. The amplified alkalosis was associated with an increased duration and/or increased velocity of the wave. These effects were most pronounced in acidic media (13 mM HCO(3)(-)/5% CO(2)) where benzolamide increased the SD duration by 55 +/- 32%. The initial velocity (including time for induction) and propagation velocity (measured between distal electrodes) were enhanced by 35 +/- 25 and 26 +/- 16%, respectively. Measurements of [Ca(2+)](o) demonstrated an increase in duration of the Ca(2+) transient when the alkaline shift was amplified by benzolamide. The augmentation of SD caused by benzolamide was blocked in media containing the N-methyl-D-aspartate (NMDA) receptor antagonist DL-2-amino-5-phosphonovaleric acid. These data indicate that the induction and propagation of SD is inhibited by a fall in baseline pH characteristic of ischemic conditions and that the early alkaline shift can remove this inhibition by relieving the proton block on NMDA receptors. Under ischemic conditions, the intrinsic alkalosis may therefore enable SD and thereby contribute to NMDA receptor-mediated injury

The type II isoform of carbonic anhydrase is abundant in astrocytes and oligodendroglia. To explore whether the expression of the type II isoform is required for interstitial carbonic anhydrase activity, we studied extracellular pH transients in hippocampal slices from mutant mice devoid of carbonic anhydrase type II and from wild-type littermates. Stimulation of the Schaffer collateral afferents evoked similar extracellular pH transients in the CA1 stratum pyramidale, consisting of a predominant alkaline shift and little or no subsequent acidosis. After 5-s stimulus trains at 10 Hz, alkaline shifts were not significantly different in carbonic anhydrase II-deficient and wild-type preparations, averaging 0.09 +/- 0.04 and 0.08 +/- 0.04 unit pH, respectively. Addition of 1.5 microM benzolamide amplified the alkaline shifts by 385 +/- 146 and 345 +/- 75% in the mutant and wild-type preparations, respectively. Dose response studies with benzolamide displayed similar sensitivity to this carbonic anhydrase inhibitor over a concentration range of 0. 03-10 microM. These data indicate that interstitial carbonic anhydrase activity is effectively unaltered in brains devoid of carbonic anhydrase type II. The results are consistent with the interpretation that a distinct extracellular isoform of carbonic anhydrase exists in brain.

Interstitial ionic shifts that accompany ouabain-induced spreading depression (SD) were studied in rat hippocampal and cortical slices in the presence and absence of extracellular Ca(2+). A double-barreled ion-selective microelectrode specific for H(+), K(+), Na(+), or Ca(2+) was placed in the CA1 stratum radiatum or midcortical layer. Superfusion of 100 microM ouabain caused a rapid, negative, interstitial voltage shift (2-10 mV) after 3-5 min. The negativity was accompanied by a rapid alkaline transient followed by prolonged acidosis. In media containing 3 mM Ca(2+), the alkalosis induced by ouabain averaged 0.07 +/- 0.01 unit pH. In media with no added Ca(2+) and 2 mM EGTA, the alkaline shift was not significantly different (0.09 +/- 0.02 unit pH). The alkaline transient was unaffected by inhibiting Na(+)-H(+) exchange with ethylisopropylamiloride (EIPA) or by blocking endoplasmic reticulum Ca(2+) uptake with thapsigargin or cyclopiazonic acid. Alkaline transients were also observed in Ca(2+)-free media when SD was induced by microinjecting high K(+). The late acidification accompanying ouabain-induced SD was significantly reduced in Ca(2+)-free media and in solutions containing EIPA. The ouabain-induced SD was associated with a rapid but relatively modest increase in [K(+)](o). In the presence of 3 mM external Ca(2+), the mean peak elevation of [K(+)](o) was 12 +/- 0.62 mM. In Ca(2+)-free media, the elevation of [K(+)](o) had a more gradual onset and reached a significantly larger peak value, which averaged 22 +/- 1.1 mM. The decrease in [Na(+)](o) that accompanied ouabain-induced SD was somewhat greater. The [Na(+)](o) decreased by averages of 40 +/- 7 and 33 +/- 3 mM in Ca(2+) and Ca(2+)-free media, respectively. In media containing 1.2 mM Ca(2+), ouabain-induced SD was associated with a substantial decrease in [Ca(2+)](o) that averaged 0.73 +/- 0. 07 mM. These data demonstrate that in comparison with conventional SD, ouabain-induced SD exhibits ion shifts that are qualitatively similar but quantitatively diminished. The presence of external Ca(2+) can modulate the phenomenon but is irrelevant to the generation of the SD and its accompanying alkaline pH transient. Significance of these results is discussed in reference to the propagation of SD and the generation of interstitial pH changes

The generation of activity-evoked extracellular alkaline shifts has been linked to the presence of external Ca(2+) or Ba(2+). We further investigated this dependence using pH- and Ca(2+)-selective microelectrodes in the CA1 area of juvenile, rat hippocampal slices. In HEPES-buffered media, alkaline transients evoked by pressure ejection of RS-alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) averaged approximately 0.07 unit pH and were calculated to arise from the equivalent net addition of approximately 1 mM strong base to the interstitial space. These alkaline responses were correlated with a mean decrease in [Ca(2+)](o) of approximately 300 microM. The alkalinizations were abolished reversibly in zero-Ca(2+) media, becoming indiscernible at a [Ca(2+)](o) of 117+/-29 microM. Addition of as little as 30-50 microM Ba(2+) caused the reappearance of an alkaline response. In approximately one-fourth of slices, a persistent alkaline shift of approximately 0.03 unit pH was observed in zero-Ca(2+) saline containing EGTA. In HEPES media, addition of 300 microM Cd(2+), 100 microM Ni(2+), or 100 microM nimodipine inhibited the alkaline shifts by roughly one-half, one-third, and one-third, respectively, whereas Cd(+) and Ni(2+) in combination fully blocked the response. In bicarbonate media, by contrast, Cd(+) and Ni(2+) blocked only two-thirds of the response. In the presence of bicarbonate, Ni(2+) caused an unexpected enhancement of the alkalinization by approximately 150%. However, when the extracellular carbonic anhydrase was blocked by benzolamide, addition of Ni(2+) reduced the alkaline shift. These results suggested that Ni(2+) partially inhibited the interstitial carbonic anhydrase and thereby increased the alkaline responses. These data indicate that an activity-dependent alkaline shift is largely dependent on the entry of Ca(2+) or Ba(2+) via voltage-gated calcium channels. However, sizable alkaline transients still can be generated with little or no external presence of these ions. Implications for the mechanism of the activity-dependent alkaline shift are discussed

Rapid extracellular alkalinizations accompany normal neuronal activity and have been implicated in the modulation of N-methyl-D-aspartate (NMDA) receptors. Particularly large alkaline transients also occur at the onset of spreading depression (SD). To test whether these endogenous pH shifts can modulate SD, the alkaline shift was amplified using benzolamide, a poorly permeant inhibitor of interstitial carbonic anhydrase. SD was evoked by microinjection of 1.2 M KCl into the CA1 stratum radiatum of rat hippocampal slices and recorded by a proximal double-barreled pH microelectrode and a distal potential electrode. In Ringer solution of pH 7.1 containing picrotoxin (but not at a bath pH of 7.4), addition of 10 microM benzolamide increased the SD alkaline shift from 0.20 +/- 0.07 to 0.38 +/- 0.17 unit pH (means +/- SE). This was correlated with a significant shortening of the latency and an increase in the conduction velocity by 26 +/- 16%. In the presence of the NMDA receptor antagonist DL-2-amino-5-phosphonovaleric acid (APV), benzolamide still amplified the alkaline transient, however, its effect on the SD latency and propagation velocity was abolished. The intrinsic modulation of SD by its alkaline transient may play an important role under focal ischemic conditions by removing the proton block of NMDA receptors where interstitial acidosis would otherwise limit NMDA receptor activity

Activity-dependent extracellular pH shifts were studied in slices of the rat dorsal lateral geniculate nucleus (dLGN) using double-barreled pH-sensitive microelectrodes. In 26 mM HCO3--buffered media, afferent activation (10 Hz, 5 s) elicited an early alkaline shift of 0.04+/-0.02 pH units associated with a later, slow acid shift of 0.05+/-0.03 pH units. Extracellular pH shifts in the ventral lateral geniculate nucleus were rare, and limited to acidifications of approximately 0.02 pH units. The alkaline shift in the dLGN increased in the presence of benzolamide (1-2 microM), an extracellular carbonic anhydrase inhibitor. The mean alkaline shift in benzolamide was 0.10+/-0.05 pH units. In 26 mM HEPES-buffered saline, the alkaline response averaged 0.09+/-0.03 pH units. The alkaline shifts persisted in 100 microM picrotoxin (PiTX) but were blocked by 25 microM CNQX/50 microM APV. If stimulation intensity was raised in the presence of CNQX/APV, a second alkalinization arose, presumably due to direct activation of dLGN neurons. The direct responses were amplified by benzolamide, and blocked by either 0 Ca2+/EGTA, Cd2+ or TTX. In 0 Ca2+, addition of 500 microM-5 mM Ba2+ restored the alkalosis. Alkaline shifts evoked with extracellular Ba2+ were larger and faster than those elicited by equimolar Ca2+. In summary, synchronous activation in the dLGN results in an extracellular H+ sink, via a Ca2+-dependent mechanism, similar to activity-dependent alkaline shifts in hippocampus