Preservation of neuronal [Ca²⁺]_i and [Na⁺]_i homeostasis is dependent on ATPases, ion exchangers and cotransporters, and disruption of these during ischemia has a major impact on cell survival. For example, the Na⁺-K⁺-Cl^- cotransporter 1 (NKCC1) (X. Chen et al., 2008; Luo et al., 2008) promotes neuronal [Na⁺]_i overload upon reoxygenation, triggering reverse-mode operation of the Na⁺/Ca²⁺ exchanger, NCX1, Ca²⁺ influx, and cell death (Pignataro et al., 2004; Shono et al., 2010). In contrast, upregulation of the endoplasmic reticulum Ca(2+)-ATPases (SERCA2b subtype) in hippocampal neurons decreases Ca²⁺ store dysfunction and is neuroprotective during oxygen-glucose deprivation (Kopach et al., 2016).

Ionic imbalances are also a contributing factor to the acidotoxicity observed following cerebral ischemia (Siesjo, 1988). Our laboratory has shown that acidosis synergistically potentiates the intracellular Ca²⁺ dysregulation evoked by ischemia in cortical neurons, enhancing neuronal death (Mari et al., 2010). The acid-sensing ion channel, ASIC1a, contributes to this potentiation, but these channels alone cannot account for the long-lived synergy (Mari et al., 2010), since ASIC1a rapidly inactivates and release of Ca²⁺ from intracellular stores was also observed following ASIC1a activation (Herrera et al., 2008; Mari, Katnik, and Cuevas, 2010). It is of significant interest to identify other contributors to this synergistic potentiation of [Ca²⁺]_i dysregulation during ischemia and acidosis since these may, in part, account for the expansion of the ischemic lesion following stroke.

One possible molecular mechanism for this long-lived ionic imbalance during ischemia-acidosis is the prolonged activation of NKCC1. The role of NKCC1 during stroke and ischemia has been previously studied using the NKCC1-selective inhibitor, bumetanide. Bumetanide was shown to reduce reperfusion injury following focal cerebral ischemia in rats, presumably due to the inhibition of NKCC1 (Wang, Huang, He, Ruan, and Huang, 2014). Chronic bumetanide treatment following stroke has been shown to enhance neurogenesis and behavioral recovery in rats, although this effect was not unequivocally linked to NKCC1 (Xu et al., 2016). In vitro experiments have shown bumetanide can inhibit ischemia-induced [Na⁺]_i and [Ca²⁺]_i dysregulation in both neurons and astrocytes (X. Chen et al., 2008; Kintner et al., 2007; Lenart, Kintner, Shull, and Sun, 2004). These effects of bumetanide were suggested to be due to block of NKCC1.

Experiments were carried out to determine if inhibition of NKCC1 with bumetanide or ethacrynic acid alters ischemia-acidosis evoked [Na⁺]_i and [Ca²⁺]_i overload in neurons. Fluorometric Na⁺ and Ca²⁺ measurements showed that bumetanide and ethacrynic acid inhibited ischemia-acidosis induced [Ca²⁺]_i, but not [Na⁺]_i, overload. The effects of bumetanide and ethacrynic acid were sufficiently different to suggest that these compounds were acting on distinct off-target sites. Whole-cell membrane current measurements indicated that both bumetanide and ethacrynic acid inhibit voltage-gated calcium and voltage-gated sodium channels, but that neither affects glutamatergic ionotropic receptors or acid-sensing ion channels. Ethacrynic acid was found to be a more efficacious inhibitor of VGCC than bumetanide, which may explain why it has a greater effect on ischemia-acidosis evoked [Ca²⁺]_i overload. The effects of bumetanide on ischemia-acidosis evoked [Ca²⁺]_i overload may in part explain the benefits of this compound following stroke. However, results presented here suggest that ethacrynic acid may be a superior compound for reducing stroke injury.

Our laboratory has shown that the concurrence of ischemia and acidosis, which occurs during stroke, results in a synergistic [Ca²⁺]_i overload in neurons and concomitant cell death (Mari et al., 2010). Given that NKCC1 has been implicated in [Ca²⁺]_i overload during reperfusion following oxygen-glucose deprivation (X. Chen et al., 2008), we examined the effects of the NKCC1 inhibitors, bumetanide and ethacrynic acid, on the [Ca²⁺]_i burden produced by simultaneous ischemia and acidosis in neurons. Consistent with previous studies, in vitro ischemia-acidosis (Isc. pH 6) induced a rapid, transient increase in [Ca²⁺]_i in cortical neurons followed by a slow steady rise in the continued presence of the Isc. pH 6 solution. Upon washout of the solution, the [Ca²⁺]_i rebounded to a second, slower decaying peak, before returning to near baseline levels (Figure 1A). Application of 100 μM bumetanide (Figure 1B, blue trace) or 100 μM ethacrynic acid (Figure 1C, red trace) resulted in a reduction in [Ca²⁺]_i elevations produced by ischemia-acidosis. Additionally, [Ca²⁺]_i recovered more slowly in the presence of these inhibitors. Combined data from identical experiments showed both bumetanide and ethacrynic acid reduced the initial peak increases and the sustained levels of [Ca²⁺]_i triggered by ischemia-acidosis. However, ethacrynic acid produced a statistically greater reduction in both of these [Ca²⁺]_i increases relative to bumetanide. Furthermore, ethacrynic acid, but not bumetanide, reduced the peak rebound increase in [Ca²⁺]_i observed upon washout of ischemia-acidosis (Figures 1D,E). While the loop diuretics inhibited the transient increases, they also decreased the rate of recovery from the elevated [Ca²⁺]_i, resulting in an elevated [Ca²⁺]_i at the end of the recording period (9 min) relative to Control (Figures 1D,E). The total increase in [Ca²⁺]_i induced by ischemia-acidosis, as measured by integrating under the traces (Total [Ca²⁺]_i), is dependent on the transient increases and as well as the degree of recovery during the recording period. For cells incubated in bumetanide, the decreased elevations in [Ca²⁺]_i occurring during ischemia-acidosis were offset by the reduced recovery during washout. Thus, bumetanide failed to reduce net [Ca²⁺]_i increases relative to Control. In contrast, ethacrynic acid, which produced a similar degree of inhibition of recovery as bumetanide, but a greater block of initial peak, sustained and rebound peak [Ca²⁺]_i than bumetanide, reduced the net [Ca²⁺]_i increase produced by ischemia-acidosis (Figures 1D,E).

In addition to affecting [Ca²⁺]_i, NKCC1 has been implicated in elevations in [Na⁺]_i during post-ischemia reperfusion (X. Chen et al., 2008). This [Na⁺]_i overload causes cellular edema and promotes cell death. Thus, it was of interest to determine if NKCC1 plays a significant role in intracellular Na⁺ homeostasis during ischemia-acidosis and during reperfusion. SBFI loaded cultured cortical neurons were imaged to determine the effects of ischemia-acidosis on intracellular Na⁺ by monitoring R_SBFI. Ischemia-acidosis was found to produce an immediate rapid rise in [Na⁺]_i followed by a slow steady increase that persisted for approximately 1 min following washout of the ischemia-acidosis solution before [Na⁺]_i returned towards baseline (Figure 2A). Application of either 100 μM bumetanide or 100 μM ethacrynic acid failed to significantly alter increases in [Na⁺]_i produced by ischemia-acidosis, but both effected recovery from these [Na⁺]_i elevations (Figures 2B,C). Compiled data showed that basal [Na⁺]_i was not affected by either bumetanide or ethacrynic acid (Figures 2D,E). Similarly, neither the initial peak [Na⁺]_i nor the maximum peak [Na⁺]_i were significantly affected by the NKCC1 inhibitors (Figures 2D,E). However, opposite effects on the recovery from these induced [Na⁺]_i increases were noted for the two NKCC1 inhibitors, with bumetanide slowing down recovery and ethacrynic acid accelerating return to baseline [Na⁺]_i (Figure 2D).

The fact the two NKCC1 antagonists, bumetanide and ethacrynic acid, failed to block elevations in [Na⁺]_i observed following ischemia-acidosis suggests that NKCC1 is not the site of action responsible for the mitigation of [Ca²⁺]_i overload by these drugs under these conditions. Our laboratory has previously shown the initial, transient rise in [Ca²⁺]_i following ischemia-acidosis is dependent on activation of acid-sensing ion channels (ASIC) (Mari et al., 2010). Since this rise was BMN- and EA-sensitive, we examined the effects of these loop diuretics on ASIC-mediated whole-cell currents in isolated cortical neurons. Application of acidic PSS (pH 6.0) evoked transient inward currents in neurons voltage-clamped at −70 mV, consistent with ASIC activation (Figure 3A). Bumetanide (100 μM) was found to increase peak current amplitudes by 13% and decay times by 25% which resulted in no significant difference in the net inward current induced by 10 s acidosis compared to control (Figures 3B,D,E). In contrast, while ethacrynic acid (100 μM) did not statistically alter the peak inward current amplitude or the time constant of its decay, the small apparent decrease in both resulted in a 10% block of the net inward current produced by 10 s acidosis (Figure 3C,D,E). These results, however, are not consistent with the 10 and 20% blocks of initial [Ca²⁺]_i increases by BMN and EA, respectively (see Figure 1), which suggests that these drugs are modulating Ca²⁺ influx pathways distinct from ASIC.

Activation of NMDA receptors has also been shown to occur during ischemia-acidosis and produces increases in neuronal [Ca²⁺]_i (Mari et al., 2010). To determine if bumetanide and/or ethacrynic acid inhibit ionotropic glutamatergic currents, whole cell currents were recorded in neurons voltage clamped at -70 mV and perfused with 20 μM glutamate in the absence and presence of the two drugs (Figures 4A–C). The NCCK1 inhibitors did not produce appreciable changes in ionotropic glutamatergic responses (Figures 4B,C). Figure 4D shows bar graph of mean peak, steady-state and net inward currents (Charge Influx) recorded from multiple neurons during 10 s glutamate applications. Neither 100 μM bumetanide nor 100 μM ethacrynic acid significantly inhibited any of the components of the glutamate-evoked currents (Figures 4D).

The inability of bumetanide to inhibit acidosis- and glutamate-evoked inward currents in cultured cortical neurons raises the possibility that the reduction of ischemia-acidosis induced [Ca²⁺]_i increases produced by this NCCK1 inhibitor might be due to block of Ca²⁺ influx through voltage-gated Ca²⁺ channels (VGCC). Previous studies in our laboratory have shown VGCC are activated downstream of ASIC following ischemia-acidosis (Mari et al., 2010). Neurons were voltage-clamped at -70 mV and depolarizing membrane pulses applied to -10 mV, in the presence of TTX and TEA, to isolate currents through VGCC. Figure 5A shows representative membrane currents recorded from a single neuron in response to membrane depolarizations in the absence (Control) and presence of 10 μM bumetanide (BMN). At this concentration, bumetanide decreased the peak current amplitude by approximately 40%. Figure 5B shows representative membrane currents recorded from a single neuron in response to membrane depolarizations in the absence (Control) and presence of 56 μM ethacrynic acid (EA). In this cell, 56 μM ethacrynic acid produced an approximately 60% reduction in the VGCC-mediated current. Using identical methods, concentration-response relationships were constructed for bumetanide and ethacrynic acid inhibition of VGCC-mediated currents (Figure 5C). Data points were best fit with a single-site Langmuir-Hill equation (Cuevas and Adams, 1997) with half-maximal inhibitions (IC₅₀), Hill coefficients and non-reducible current (asymptotic minimum) values of 8 μM, 0.41 and 0.37, respectively for BMN (blue circles and line) and 36 μM, 0.75 and 0.0 for EA (red circles and line) (Figure 5C). Thus, ethacrynic acid has lower affinity for the channel, but greater efficacy than bumetanide for blocking VGCCs in cortical neurons.

Our laboratory has shown that elevations of [Ca²⁺]_i triggered by acidosis can also be reduced via the inhibition of voltage-gate Na⁺ channels (VGSC). Furthermore, blocking of VGSC has been shown to lessen neuronal [Ca²⁺]_i increases observed after oxygen-glucose deprivation (LoPachin, Gaughan, Lehning, Weber, and Taylor, 2001; Herrera et al., 2008). Thus, we examined if the loop diuretics, bumetanide and ethacrynic acid, also effected VGSC in cortical neurons. Inward VGSC currents were activated by stepping voltage clamped neurons from −70 mV to −30 mV in the presence of TEA, Ba²⁺ and Cd²⁺. Representative VGSC currents recorded from a single neuron are shown in Figure 6A and demonstrate BMN inhibit VGSC. Similarly, VGSC currents were reduced by micromolar concentrations of EA (Figure 6B). Concentration-response relationships were constructed using measurements from 86 neurons and concentrations of bumetanide ranging from 0.3 to 1,000 μM and on 46 neurons using concentrations of ethacrynic acid ranging from 3 to 1,000 μM. Figure 6C shows the data points were best fit with a single-site Langmuir-Hill equation with IC₅₀, Hill coefficient and non-reducible current (asymptotic minimum) values of 19 μM, 0.46 and 0.21, respectively for BMN (blue circles and line) and 36 μM, 0.72 and 0.22, respectively for EA (red circles and line). These results indicate nearly 20% of the VGSC current is BMN and EA-insensitive.

It should also be noted that the blocks of these voltage-gated currents by the loop diuretics were not voltage dependent. There were no observable shifts in the voltages for maximal current amplitude (Supplementary Figure S1) in the presence of half-maximal concentrations of BMN or EA compared to control.

The 70% block of VGSC currents by maximal concentrations of bumetanide only occurred in a fraction of the neurons tested. Of the 103 cells used to measure bumetanide inhibition of VGSC currents, 31 were found to be inhibited by ∼20% by millimolar concentrations of bumetanide. A concentration-response relationship for VGSC currents in bumetanide-resistant neurons failed to exhibit any concentration-dependence of block (solid line), in contrast to the dose-response relationship observed in the bumetanide-sensitive neurons (dashed line) (Figure 7A). Figure 7B shows current traces from two different voltage clamped neurons stepped from a holding potential of −70 to −30 mV in the absence and presence of 1 mM bumetanide, 100 nM TTX and 1 mM BMN + 100 nM TTX. The current traces on the left are from a representative neuron resistant to bumetanide block, whereas the traces on the right are from a neuron sensitive to bumetanide block. The data from 10 cells exposed to this protocol were grouped according to bumetanide sensitivity and each group analyzed with a two-way ANOVA to determine if there were interactions between TTX and bumetanide (Figure 7C). For the bumetanide resistant currents, there was no interaction between the two compounds (p = 0.336) (Figure 7C), while, in bumetanide sensitive neurons there was a statistically significant interaction between the drugs, such that the response to bumetanide was dependent on the presence of TTX (p < 0.02) (Figure 7C). While both BMN and TTX reduced the VGSC current amplitude relative to control alone (p < 0.001 and p = 0.002, respectively), the combination BMN + TTX was not significantly different from either BMN (p = 0.981) or TTX (p = 0.49) alone (Figure 7C). To further facilitate comparison, we determined the percent block observed for each condition, BMN, TTX and BMN + TTX. Figure 7D shows the percent inhibition (mean ± SEM) observed for both bumetanide-resistant (left panel) and bumetanide-sensitive (right panel) neurons. In bumetanide-resistant neurons, 1 mM BMN inhibited VGSC currents by only ∼26%, while TTX and BMN + TTX inhibited the responses by over 70% (Figure 7D). In these neurons, there were statistically significant blocks produced by TTX and BMN + TTX. The block by BMN alone was significantly different from inhibition with TTX present (p < 0.001). In contrast, VGSC currents from the bumetanide-sensitive cells were all blocked by >70% by BMN, TTX and BMN + TTX (Figure 7C) and these inhibitions were not significantly different from each other.

The primary finding of this study is the loop diuretics bumetanide and ethacrynic acid inhibit voltage-gated Ca²⁺ and Na⁺ ion channels that contribute to intracellular Ca²⁺ dysregulation evoked by ischemia-acidosis in neurons. The IC₅₀ values of BMN and EA for these channels are consistent with the concentrations of the loop diuretics required to inhibit the ischemia-acidosis evoked [Ca²⁺]_i overload. Neither loop diuretic significantly blocked acidosis- or glutamate-activated whole cell inward currents, suggesting that the inhibition of ischemia-acidosis induced [Ca²⁺]_i overload was not due to block of ASIC or glutamatergic ion channels. In addition, neither BMN nor EA significantly altered ischemia-acidosis induced increases in [Na⁺]_i. Thus, the increase in [Na⁺]_i induced by ischemia-acidosis is not mediated by Na⁺ influx through either VGSCs or NKCC. These observations further suggest that loop diuretic suppression of [Ca²⁺]_i overload during ischemia-acidosis is not a downstream effect of mechanisms activated to preserve [Na⁺]_i homeostasis.

Our laboratory first showed ischemia and acidosis interact to produce a synergistic increase in neuronal [Ca²⁺]_i overload (Mari et al., 2010). While several ion channels, including acid-sensing ion channels, VGCC and NMDA receptors were all found to contribute to the increases in [Ca²⁺]_i, use of specific blockers of these channels suggested additional molecular mechanisms were likely involved in the ischemia-acidosis evoked [Ca²⁺]_i dysregulation. Both bumetanide and ethacrynic acid inhibited multiple components of the ischemia-acidosis evoked [Ca²⁺]_i overload, including the initial transient increase, sustained levels and rebound peak increase in [Ca²⁺]_i. This suggested bumetanide and ethacrynic acid are either inhibiting multiple proteins that cause this [Ca²⁺]_i overload or are affecting mechanisms that directly or indirectly influence [Ca²⁺]_i handling throughout the ischemia-acidosis event. Activation of ASIC, NMDA and VGCC all produce rapid increases in neuronal [Ca²⁺]_i and impact the initial rapid transient increase in [Ca²⁺]_i (Mari et al., 2010; Mari, Katnik, and Cuevas, 2015). However, neither bumetanide nor ethacrynic acid were found to appreciably block ASIC- or NMDA-mediated currents at concentrations as high as 100 μM. Thus, it is unlikely inhibition of these channels by the loop diuretics produces the reduction in ischemia-acidosis evoked synergistic increases in neuronal [Ca²⁺]_i overload. In contrast, both bumetanide and ethacrynic acid were found to inhibit VGCC.

These loop diuretics are inhibitors of the renal Na-K-Cl cotransporters, NKCC1 and NKCC2. Bumetanide has been shown to inhibit rat NKCC1 and NKCC2 with IC₅₀ values of ∼6 μM (Hannaert, Alvarez-Guerra, Pirot, Nazaret, and Garay, 2002). The IC₅₀ observed in the current study for bumetanide inhibition of voltage-gated Ca²⁺ channels, 8 μM, is nearly identical to that for NKCC1/NKCC2. However, while bumetanide at the highest concentrations (>50 μM) blocked nearly all of the activity of both NKCC1 and NKCC2 (Hannaert et al., 2002), over 40% of the VGCC current was resistant to 300 μM bumetanide. It remains to be determined if this bumetanide-resistant component represents a specific VGCC subtype or if the compound has low efficacy for VGCC in general. While the effects of loop diuretics on neuronal VGCC were previously unknown, bumetanide can inhibit VGCC in cardiac myocytes at low μM concentrations (Shimoni, 1991). Inhibition of VGCC in cardiomyocytes by bumetanide was found to be as high as 80% of the peak VGCC current and varied significantly from cell to cell (Shimoni, 1991).

Ethacrynic acid is structurally dissimilar to bumetanide and higher concentrations are required to inhibit both transporter subtypes. EA inhibits NKCC1 in avian erythrocytes and NKCC2 in canine renal epithelial cells with IC₅₀ values of 180 and 20 μM, respectively, (Rugg, Simmons, and Tivey, 1986; Palfrey and Leung, 1993). Given the IC₅₀ value for EA inhibition of VGCC is 36 μM, EA is a more potent inhibitor of VGCC than NKCC. While EA was less potent than bumetanide at inhibiting VGCC, EA exhibited greater efficacy for these channels, with 178 μM EA blocking ∼80% of the VGCC currents in the neurons. Unlike bumetanide, ethacrynic acid, which is structurally dissimilar and does not contain a sulfonamide substituent, has not been previously shown to modulate VGCC in any cell type.

Bumetanide blocked voltage-gated Na⁺ channels with an IC₅₀ (19 μM) similar to its IC₅₀ for NKCC1 and NKCC2. In contrast to observations of VGCC block by bumetanide in the current study, a population of cortical neurons expressed VGSC that were resistant to bumetanide block, even at the highest concentrations (1 mM). In both bumetanide-sensitive and bumetanide-insensitive cells, VGSC currents were blocked by >70% by TTX (100 nM). Bumetanide application did not have an additive effect with TTX, suggesting bumetanide specifically inhibits one of the TTX-sensitive channel subtypes expressed in cortical neurons. Cortical neurons express a variety of TTX-sensitive VGSC, including NaV1.1, NaV1.2, NaV1.3, NaV1.6 and NaV1.7 and the TTX-insensitive NaV1.9 (Jeong et al., 2000; de Lera Ruiz and Kraus, 2015; Rubinstein et al., 2016). Thus, bumetanide does not appear to affect NaV1.9. The specific TTX-sensitive NaV subtype affected by bumetanide remains to be determined.

The ethacrynic acid block of VGSC (IC₅₀ = 36 μM) was more potent than its block of NKCC1 but comparable to its block of NKCC2. Like bumetanide, EA failed to completely block VGSC currents, with approximately 20% of the current remaining at 1 mM EA.

Inhibition of voltage-gated sodium channels by BMN and EA did not alter [Na⁺]_i accumulation caused by ischemia-acidosis. Neither the initial rapid rise in [Na⁺]_i due to ischemia-acidosis, nor the slow rise in [Na⁺]_i during the ischemic event were reduced by either compound at concentrations that significantly inhibit VGSC (100 μM). Both ischemia and acidosis are known to evoke increases in [Na⁺]_i resulting in reverse-mode activity of the Na⁺/Ca²⁺ exchanger (Lenart et al., 2004; Kintner et al., 2007; Luo et al., 2008). Extrusion of Na⁺ by NCX produces Ca²⁺ influx and [Ca²⁺]_i elevations. Given [Na⁺]_i is not affected by the loop diuretics, it does not appear a reduction in [Na⁺]_i produced by inhibition of VGSC and concomitant lessening of Ca²⁺ influx via NCX can explain the ability of these loop diuretics to mitigate ischemia-acidosis evoked [Ca²⁺]_i overload. NKCC activity has been implicated in the [Na⁺]_i accumulation that precedes NCX activity and [Ca²⁺]_i overload in astrocytes (Lenart et al., 2004; Kintner et al., 2007; Luo et al., 2008). Neither bumetanide nor ethacrynic acid reduced the [Na⁺]_i elevation in neurons suggesting NKCC is not a major conduit for ischemia-acidosis induced Na⁺ influx which leads to [Ca²⁺]_i overload. The lack of NKCC contribution to [Na⁺]_i increases may be due to reduced activity of the cotransporter during ischemia-acidosis. It was previously reported the activity of NKCC1 is reduced by ∼80% when extracellular pH approaches 6.0 (Hegde and Palfrey, 1992).

Stroke-induced gray and white matter injury in mice was shown to be reduced in NKCC knockout animals (H. Chen, Luo, Kintner, Shull, and Sun, 2005). Similarly, bumetanide was shown to reduce infarct volume in a rat middle cerebral artery occlusion stroke model (O'Donnell, Tran, Lam, Liu, and Anderson, 2004). NKCC1 effects on [Na⁺]_i in neurons and astrocytes appeared during re-oxygenation rather than during the ischemic event (H. Chen et al., 2005), consistent with our observation that bumetanide does not reduce [Na⁺]_i accumulation during ischemia-acidosis. While reduced edema associated with NKCC1 inhibition by bumetanide may lessen stroke injury (O'Donnell et al., 2004), the blunting of [Ca²⁺]_i overload in response to ischemia-acidosis by BMN would also improve outcomes in stroke. Similarly, inhibition of voltage-gated channels may explain how bumetanide reduces glutamate-mediated excitotoxicity (Beck, Lenart, Kintner, and Sun, 2003).

Bumetanide has been shown to decrease seizure activity in humans (Kahle and Staley, 2008). Results from the current study suggest this effect may be due to the inhibition of voltage-gated channels. The inhibition of voltage-gated Ca²⁺ channels is known to contribute to the actions of antiepileptic drugs, such as levetiracetam (Niespodziany, Klitgaard, and Margineanu, 2001; Yan et al., 2013). Similarly, inhibiting TTX-sensitive sodium channels, such as NaV1.6 has been shown to blunt seizure activity in rat epilepsy models (Hargus, Nigam, Bertram, and Patel, 2013; Shao et al., 2017). Finally, loop diuretics, including bumetanide, have been shown to promote direct vasorelaxation in the concentration range shown here to be effective for VGCC and VGSC block (Pickkers, Russel, Thien, Hughes, and Smits, 2003). Therefore, direct inhibition of these channels which contribute to vascular tone may explain these effects.

Clinically, bumetanide and ethacrynic acid, are used in edematous states, such as heart failure, to promote diuresis and natriuresis (Somberg and Molnar, 2009). This effect is due to the ability of these compounds to block the NKCC2 cotransporter, primarily in the Loop of Henle, and prevent the reuptake of Na⁺, K⁺, and Cl⁻ from the tubular fluid. Loop diuretics have additional effects, such as the lowering of blood pressure, which is often observed even prior to diuresis (Gabriel, 1983). The effective concentrations of bumetanide and ethacrynic acid reported here are consistent with clinically relevant doses (Ward and Heel, 1984; Lacreta et al., 1994; van der Heijden et al., 1998), and thus may contribute to the systemic effects of these compounds. It will be important to determine if other off-target effects of these loop diuretics contribute to the decrease in [Ca²⁺]_i overload reported here.

In conclusion, the loop diuretics, bumetanide and ethacrynic acid effectively suppress ischemia-acidosis induced [Ca²⁺]_i overload in neurons at concentrations near their respective IC₅₀ values for NKCC inhibition. These effects appear to be in part mediated via the inhibition of voltage-gated Ca²⁺ and Na⁺ channels and are not due to any direct effects on ionotropic glutamatergic receptors or acid-sensing ion channels. Furthermore, the inability of the loop diuretics to reduce ischemia-acidosis evoked [Na⁺]_i elevations suggest NKCC cotransporters are not involved in neuronal ionic imbalances during ischemia-acidosis; and inhibition of these transporters does not account for the beneficial effects of bumetanide and ethacrynic acid under these conditions. However, the ability of these loop diuretics to lessen ischemia-acidosis induced [Ca²⁺]_i overload suggests that they may be useful for reducing stroke injury and that their effect on Ca²⁺ and Na⁺ channels may in part explain observations made in previous studies.