European Journal of Pharmacology 746 (2015) 78–88

Contents lists available at ScienceDirect

European Journal of Pharmacology journal homepage: www.elsevier.com/locate/ejphar

Neuropharmacology and analgesia

Bumetanide is not capable of terminating status epilepticus but enhances phenobarbital efficacy in different rat models Kathrin Töllner a,b,1, Claudia Brandt a,b,1, Thomas Erker c, Wolfgang Löscher a,b,n a

Department of Pharmacology, Toxicology, and Pharmacy, University of Veterinary Medicine Hannover, Germany Center for Systems Neuroscience, Hannover, Germany c Department of Medicinal Chemistry, University of Vienna, Austria b

art ic l e i nf o

a b s t r a c t

Article history: Received 8 July 2014 Received in revised form 21 October 2014 Accepted 27 October 2014 Available online 11 November 2014

In about 20–40% of patients, status epilepticus (SE) is refractory to standard treatment with benzodiazepines, necessitating second- and third-line treatments that are not always successful, resulting in increased mortality. Rat models of refractory SE are instrumental in studying the changes underlying refractoriness and to develop more effective treatments for this severe medical emergency. Failure of GABAergic inhibition is a likely cause of the development of benzodiazepine resistance during SE. In addition to changes in GABAA receptor expression, trafficking, and function, alterations in Cl
homeostasis with increased intraneuronal Cl
levels may be involved. Bumetanide, which reduces intraneuronal Cl
by inhibiting the Cl
intruding Na þ , K þ , Cl
cotransporter NKCC1, has been reported to interrupt SE induced by kainate in urethane-anesthetized rats, indicating that this diuretic drug may be an interesting candidate for treatment of refractory SE. In this study, we evaluated the effects of bumetanide in the kainate and lithium–pilocarpine models of SE as well as a model in which SE is induced by sustained electrical stimulation of the basolateral amygdala. Unexpectedly, bumetanide alone was ineffective to terminate SE in both conscious and anesthetized adult rats. However, it potentiated the anticonvulsant effect of low doses of phenobarbital, although this was only seen in part of the animals; higher doses of phenobarbital, particularly in combination with diazepam, were more effective to terminate SE than bumetanide/phenobarbital combinations. These data do not suggest that bumetanide, alone or in combination with phenobarbital, is a valuable option in the treatment of refractory SE in adult patients. & 2014 Elsevier B.V. All rights reserved.

Keywords: Kainate Pilocarpine Amygdala Cation cotransporters GABA receptors Chemical compounds studied in this article: Pilocarpine (PubChem CID: 5910)

1. Introduction Status epilepticus (SE), a life-threatening neurologic emergency requiring prompt treatment, has been shown to induce alterations in GABAA receptor function and neuronal chloride homeostasis that may underlie resistance to the anticonvulsant effect of benzodiazepines (Macdonald and Kapur, 1999; Chen and Wasterlain, 2006; Löscher, 2007; Deeb et al., 2012; Löscher et al., 2013). Benzodiazepines such as diazepam and lorazepam, which act by enhancing the inhibitory effect of GABA, are standard first-line treatments for SE; however, patients often become refractory to benzodiazepines when seizures are prolonged (Rossetti and Lowenstein, 2011). The regulation of intracellular chloride (Cl
) determines the polarity of GABAA-induced neuronal Cl
currents. In neurons, n Corresponding author at: Department of Pharmacology, Toxicology and Pharmacy, University of Veterinary Medicine, Bünteweg 17, D-30559 Hannover, Germany. Tel.: þ 49 511 856 8721; fax: þ49 511 953 8581. E-mail address: [email protected] (W. Löscher). 1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.ejphar.2014.10.056 0014-2999/& 2014 Elsevier B.V. All rights reserved.

Cl
concentrations depend on the activity of the Cl
extruding K þ , Cl
cotransporter KCC2 and the activity of the Cl
intruding Na þ , K þ , Cl
cotransporter NKCC1 (Blaesse et al., 2009). Alterations in the balance of NKCC1 and KCC2 activities may induce a switch from hyperpolarizing to depolarizing effects of GABA that may contribute to prolongation of SE and AED resistance (Deeb et al., 2012). In line with this hypothesis, Kapur and Coulter (1995), using the lithium/pilocarpine SE model in rats, reported a loss of GABA-mediated inhibition of hippocampal CA1 neurons after 45 min of SE, which resulted from a positive shift in EGABA in response to marked elevation in intracellular Cl
concentration. At about the same time following onset of SE, KCC2 expression was reported to markedly decrease in the hippocampus (Lee et al., 2010). We found that a pilocarpine-induced SE increases NKCC1 expression in the rat hippocampus, but the earliest time point studied was 24 h after onset of SE (Brandt et al., 2010). Assuming that both decreased KCC2 and increased NKCC1 contribute to the shift from hyperpolarizing to depolarizing GABA during SE, a selective NKCC blocker such as the loop diuretic bumetanide may be useful for treatment of intractable SE. The first

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

indication that blockade of Cl
cotransport interrupts SE was reported by Hochman et al. (1995), showing that the loop diuretic furosemide terminates a kainate-induced SE in urethane-anesthesized rats. Subsequently, they reported a similar effect for bumetanide (Schwartzkroin et al., 1998), which is more selective for NKCC1 than furosemide. Since both studies were performed in anesthetized rats, it was not clear whether the SE-blocking effect of the diuretics was due to an interaction with urethane, which is known to enhance GABA (Hara and Harris, 2002), or whether the diuretics alone exerted this effect. Holtkamp et al. (2003) reported that high doses of furosemide terminate an electrically induced SE in awake rats, but, to our knowledge, bumetanide has not yet been evaluated for such an effect in conscious animals. This prompted us to evaluate the effects of bumetanide in SE models in anesthetized and non-anesthetized rats. Our hypothesis was that the anticonvulsant effect reported by Schwartzkroin et al. (1998) was primarily due to an interaction between bumetanide and urethane, similar to the interaction between bumetanide and the GABApotentiating drug phenobarbital reported for other seizure models (Löscher et al., 2013). Thus, in addition to testing bumetanide's effects on SE in anesthetized and non-anesthetized rats, we also investigated whether bumetanide increases the anticonvulsant effect of phenobarbital in SE models.

79

2. Materials and methods

Since in most experiments, it was not possible to induce SE activity in the EEG with kainate in urethane-anesthetized rats (see Section 3), additional experiments with kainate were performed without anesthesia. In these experiments, SE was induced with 10–12 mg/kg kainate i.v.; in some rats an additional dose of 4 mg/ kg had to be administered to induce SE. Bumetanide (5 mg/kg i.v.) was administered repeatedly after SE onset (see Section 3). In some experiments, additional injections of bumetanide were combined with 10 mg/kg i.v. phenobarbital (Serva; Heidelberg, Germany). The dose of phenobarbital, which was dissolved as sodium salt in distilled water, was chosen from previous experiments with bumetanide and phenobarbital in the amygdalakindling model of epilepsy (Töllner et al., 2014), in which phenobarbital injected alone at a dose of 10 mg/kg did not exert significant anticonvulsant activity, but combined treatment with bumetanide resulted in anticonvulsant effects. Some experiments were also performed with female Sprague–Dawley rats, but the results did not differ from male rats. At the end of the experiment, we examined whether SE could be terminated by a higher dose (20–25 mg/kg i.p.) of phenobarbital injected alone or in combination with diazepam (10 mg/kg i.p.). For injection of diazepam, a commercial ethanol-containing aqueous solution (Faustans; Temmler Pharma, Marburg, Germany), which contains 5 mg diazepam per ml (the ethanol concentration is 18.6%), was used.

2.1. Animals

2.3. Experiments with lithium/pilocarpine

As in the experiments of Schwartkroin's group (Hochman et al., 1995; Schwartkroin et al., 1998), male Sprague–Dawley rats were used in all experiments unless otherwise indicated. They were obtained from Harlan (Horst, Netherlands) at a body weight of 200–224 g and were adapted to the laboratory for at least one week before starting the experiments. Animals were housed under controlled conditions (ambient temperature 22–24 1C, humidity 30–50%, lights on from 6:00 a.m. to 6:00 p.m.). Food (Altromin 1324 standard diet) and water were freely available. All experiments were done in compliance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and had institutional approval (LAVES, Oldenburg, Germany). All efforts were made to minimize both the suffering and the number of animals.

As with kainate, lithium/pilocarpine was used to induce SE in conscious and anesthetized rats. Since in our hand urethane did not induce surgical anesthesia (see Section 3), chloral hydrate (400 mg/kg i.p.) was used in these experiments. EEG was recorded from cortical electrodes as in the kainate experiments (for electrode locations see above). Lithium chloride (127 mg/kg p.o.) was administered 12–18 h and methyl-scopolamine (1 mg/kg i.p.) 30 min before pilocarpine. In both conscious and anesthetized rats, pilocarpine was administered as an i.p. bolus of 30 mg/kg, followed by repeated injections of 10 mg/kg at intervals of 30 min until onset of SE. In conscious rats, one additional injection of pilocarpine was sufficient to induce SE, where following chloral hydrate, 5 additional injections were needed. Bumetanide (5 mg/ kg i.v.) was injected 30, 60, 90, and 150 min following SE onset, either alone or in combination with phenobarbital (10 mg/kg i.v.).

2.2. Experiments with kainate In order to replicate the experiments of Schwartkroin's group in the kainate model as closely as possible, we asked Daryl Hochman to provide details of the experiments that were published by Hochman et al. (1995) and Schwartzkroin et al. (1998) with furosemide and bumetanide in urethane-anesthetized rats. 3–4 weeks before the experiment, rats were implanted with electrodes under anesthesia with chloral hydrate as described previously (Rattka et al., 2012) to record electroencephalographic (EEG) activity in the fronto–parietal cortex (mm from bregma: AP,
2.2; L,71.5; Paxinos and Watson, 2007) as in the experiments of Hochman et al. (1995). Some rats were also implanted with electrodes into the dentate gyrus (AP,
3.9; L,
1.7; and V,
3.5). For the experiments with urethane anesthesia, rats received 1.25– 1.4 g/kg urethane (Sigma-Aldrich; Taufkirchen, Germany) i.p. and, about 60 min later, they received an i.v. injection of 10–12 mg/kg kainate [(
)-(α)-kainic acid; Cayman Chemical; Ann Arbor, MI, USA ]. If no SE was induced, additional kainate doses (4–8 mg/kg) were administered. Bumetanide (Sigma-Aldrich; dissolved in distilled water by means of dilute NaOH) was then injected at two boluses of 5 mg/kg i.v. each, at 80 and 115 min after the last kainate injection as described by Schwartzkroin et al. (1998). All i.v. injections were done in a tail vein.

2.4. Experiments with electrical SE induction In view of the resistance of chemically-induced SE to bumetanide (see Section 3), further experiments were performed with electrical SE induction, using a model previously described and characterized in detail by us (Brandt et al., 2003). In this model, a self-sustained SE is induced by electrical stimulation of the basolateral amygdala (BLA). For this purpose, electrodes were stereotactically implanted into the right anterior BLA under anesthesia as described in detail previously (Brandt et al., 2003) and served for electrical stimulation and recording of the EEG. About 2 weeks after electrode implantation, 28 rats were electrically stimulated via the BLA electrode for induction of a self-sustained SE as described previously (Brandt et al., 2003; Bethmann et al., 2007). The following stimulus parameters were chosen: stimulus duration 25 min; stimulus consisting of 100 ms trains of 1 ms alternating positive and negative square wave pulses. The trains were given at a frequency of 2/s and the intra-train pulse frequency was 50/s. Peak pulse intensity was 700 mA. For this pulsed-train stimulation, an Accupulser A310C stimulator connected with a Stimulus Isolator A365 (World Precision Instruments, Berlin, Germany) was used. In all rats, the EEG was recorded via the BLA electrode during self-sustained SE, which typically lasts up to 8 h in

80

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

male Sprague–Dawley rats (Brandt et al., 2003). As described in detail previously (Brandt et al., 2003), sustained BLA stimulation results in three different SE types, i.e., type 1, partial SE with nonconvulsive seizure activity; type 2, partial SE with secondarily generalized seizures; and type 3, generalized convulsive SE, which differ in their EEG characteristics. As in the experiments with kainate and pilocarpine, bumetanide was repeatedly injected at doses of 5 mg/kg i.v. at different times after onset of SE, either alone or in combination with phenobarbital (10 mg/kg i.v.). Drug administration was started 3–4 h after SE onset, i.e., an SE duration which is sufficient to induce development of epilepsy in this model (Brandt et al., 2003). For control, phenobarbital alone or vehicle was injected. Furthermore, we also investigated whether a lipophilic prodrug of bumetanide, BUM5 (the N,N-dimethylaminoethylester of bumetanide), which recently has been shown to produce significantly higher bumetanide levels in the brain of rats than obtained with administration of bumetanide (Töllner et al., 2014), exerts anticonvulsant activity in this model. BUM5 was synthesized and dissolved as described recently (Töllner et al., 2014) and injected at an equimolar dose (6.5 mg/ kg i.v.) compared to bumetanide except one experiment, in which a lower dose (1.3 mg/kg i.v.) was injected to examine any dose-effect response relationship. All experiments with electrically induced SE

were performed in conscious rats. At the end of the experiment, we examined in several rats whether SE could be terminated by a higher dose (15–25 mg/kg i.p.) of phenobarbital or diazepam (20 mg/kg i.p.). Furthermore, in an additional group of 8 rats (not included in the bumetanide experiments), we examined whether SE can be more effectively terminated by a combination of diazepam (10 mg/kg i.p.) and phenobarbital (25 mg/kg i.p.), because we previously found that this drug combination is more effective than either drug alone to terminate SE in the lithium–pilocarpine model (Bankstahl and Löscher, 2008). 2.5. EEG analysis In all experiments in the different SE models, the EEG was continuously monitored during SE, and effects of the diverse treatments on paroxysmal EEG alterations were evaluated as in the studies of Hochman et al. (1995) and Schwartzkroin et al. (1998) with furosemide and bumetanide in the kainate model. For EEG-monitoring, the system consisted of one-channel amplifiers (ADInstruments Ltd, Sydney, Australia) and analogue-digital converters (PowerLab/16 SP1396 and 8/30 ML870, ADInstruments). The data were recorded and analyzed with the Chart 4 or LabChart 6 for Windows software

Table 1 Details of the experiments with bumetanide and phenobarbital in SE models in rats. Only the 22 rats that could be used for final analysis of drug effects are shown. Drugs that were used for final SE termination (high doses of phenobarbital or diazepam alone or in combination; see text) are not shown. Rat #

Illustrated in Fig.

Anesthesia

SE model

Bumetanide (or BUM5) alone Phenobarbital (PB; alone or in Vehicle combination with bumetanide [BUM]) Dose (mg/kg i.v.)

Time after SE onset

Dose (mg/kg i.v.)

Time after SE onset

Time after SE onset

– 120 and 300 min –

– –

ne nce

– – 90 and 120 min

Aggravation of EEG activity ne nce

– –

ac (after PBþ BUM) ne

– 210 and 240 min – 210 and 240 min –

ac (after PBþ BUM) nce

BUM 104 BUM 107

1 2

Urethane –

Kainate Kainate

BUM (5) BUM (5)

80 min – 90 and 270 min PB (10) þBUM (5)

CK 93

–

–

Kainate

BUM (5)

BUM 101 BUM 110

– –

– –

Kainate Kainate

BUM (5) –

75, 105, – 135 min 80 and 115 min – – PB (10)

BUM 103 BUM 113

3 4

Pilocarpine BUM (5) Pilocarpine BUM (5)

30 min PB (10) þBUM (5) 60, 90, 150 min –

BUM115 BUM 116

– –

– Cloral hydrate – –

BLA BLA

BUM (5) –

210 min –

PB (10) þ5 BUM (5) 240 min PB (10) 240 min

BUM 117 BUM120

5 7

– –

BLA BLA

BUM (5) –

210 min –

PB (10) þBUM (5) PB (10)

240 min 240 min

BUM 136

–

–

BLA

BUM5 (6.5) 180 min

210 min

BUM 137

–

–

BLA

–

PB (10) þBUM5 (6.5) PB (10)

BUM 138

–

–

BLA

BUM5 (6.5) 210 min

BUM 139

6

–

BLA

BUM5 (6.5) 210 min

BUM 144

–

–

BLA

–

–

BUM 145

–

–

BLA

BUM5 (1.3)

180 min

BUM148 BUM149

– –

– –

BLA BLA

BUM (5) –

BUM 153

–

–

BLA

–

BUM 154

–

–

BLA

BUM5 (6.5) 210 min

BUM 155

–

–

BLA

BUM5 (6.5) 210 min

–

PB (10) þBUM5 (6.5) PB (10) þBUM5 (6.5) PB (10)

Effect on SE

– 120 and 300 min 60 and 120 min –

210

ac (after PBþ BUM) nce ac (after PBþ BUM5)

240

180 and 210 min –

nce

240 min

–

nce

210 min

210 and 240 min 210 and 240 min –

210 min –

PB (10) þBUM5 (1.3) PB (10) þBUM (5) PB (10)

240 min 240 min

–

PB (10)

210 min

PB (10) þBUM5 (6.5) PB (10) þBUM5 (6.5)

nce

nce ne

240 min

– 210 and 240 min 180 and 210 min –

nce

240 min

–

nce

nce nce nce

Abbreviations: ac, anticonvulsant effect; BLA, basolateral amygdala; BUM, bumetanide; BUM5, N,N-dimethylaminoethylester of bumetanide; CK, cocktail; ne, no effect; nce, no clear effect.

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

(ADInstruments; sampling rate 200 Hz; time constant 0.1 s; low pass filter 60 Hz). 2.6. Statistics Fisher's exact test was used to calculate whether the frequency of anticonvulsant responses to bumetanide, BUM5 or phenobarbital alone differed significantly from that of combined treatment with bumetanide or BUM5 and phenobarbital. Furthermore, it was tested whether the frequency of anticonvulsant responses to combined treatment with bumetanide or BUM5 and phenobarbital differed significantly from that of treatment with high doses of phenobarbital and/or diazepam. A Po0.05 was considered significant.

3. Results Overall, 44 rats were used for the present experiment with bumetanide and SE induction by different methods: 12 rats for kainate, 4 rats for lithium–pilocarpine, and 28 rats for the BLA model. Inclusion criteria for final analysis of drug effects were induction of obvious and sustained SE activity in the EEG for studying the effects of bumetanide, BUM5 or phenobarbital on SE, and no loss of electrode assembly or death during the recordings. Twenty-two of the 44 rats fulfilled these criteria and could thus be used for final analysis of data. Details of the experiments in these 22 rats are shown in Table 1. 3.1. Experiments with bumetanide in the kainate model At the doses of urethane (1.25–1.4 g/kg i.v.) used, only shallow anesthesia was obtained, characterized by sedation

81

with loss of righting reflexes. However, urethane inhibited the convulsant activity of kainate. In 2 of 4 rats in which kainate (10–12 mg/kg i.v.) was injected 60 min after onset of urethane anesthesia, no SE-like paroxysmal activity was observed in the EEG, even though one of the rats received an additional injection of 8 mg/kg kainate. One urethane-anesthetized rat that exhibited weak paroxysmal activity following kainate is illustrated in Fig. 1. As previously described for intrahippocampal kainate injection (Bragin et al., 2009; Rattka et al., 2013), systemic administration of kainate in this rat resulted in an increase in the amplitude and frequency of the EEG signal; periods of high-frequency spiking with higher amplitude were only occasionally observed (Fig. 1B). Bumetanide did not alter this activity (Fig. 1C and D). Since of the problems in inducing typical SE activity in the EEG following urethane anesthesia, additional experiments were performed in unanesthetized (conscious) rats. A typical SE with limbic and convulsive motor seizures accompanied by high-amplitude, high-frequent bursting in the EEG (Fig. 2) was induced in 7/8 rats used for these experiments. However, 3 rats lost their EEG electrode assembly during the SE (because of intense motor seizures), so that only 4 rats could be finally evaluated. When bumetanide (5 mg/kg i.v.) was repeatedly injected in 3 of these rats starting 80 min after kainate, seizure-like activity was not abolished in the EEG (Table 1). In the rat illustrated in Fig. 2, the frequency of the bursting seemed to be reduced by bumetanide (5 mg/kg) and the second bumetanide injection also reduced the amplitude of the spiking, but this became more marked when combining bumetanide with phenobarbital (10 mg/kg). However, none of these treatments terminated the SE. Repeated administration of saline or phenobarbital (10 mg/kg) alone did not alter SE severity (Table 1).

Baseline (60 min after urethane) mV

0.2 0.0 -0.2

Status epilepticus (120 min after kainate) 0.4

mV

0.2 0.0 -0.2 -0.4

20 min after bumetanide (160 min after kainate) mV

0.2 0.0 -0.2

110 min after bumetanide (250 min after kainate) 0.6

mV

0.4 0.2 0.0 -0.2 -0.4

10 s Fig. 1. Representative example of EEG recordings from a rat (# BUM104) in which kainate was injected under anesthesia with urethane. EEG was recorded via cortical electrodes above the fronto–parietal cortex. Kainate, 12 mg/kg, was administered i.v. 60 min after urethane (1.25 g/kg i.p.). Since no clear SE was observed, a 2nd i.v. dose of kainate, 4 mg/kg, was administered 60 min after the 1st injection. (A) Baseline EEG 60 min after urethane. (B) EEG 120 min after kainate, showing increased EEG amplitude and frequency but only occasional spike complexes. (C) EEG 20 min after bumetanide, 5 mg/kg i.v.; no effect of bumetanide was observed. (D) EEG 110 min after bumetanide with high-amplitude, high-frequency spike complexes.

82

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

Baseline mV

0.2 0.0 -0.2

Status epilepticus (120 min after kainate) 0.4

mV

0.2 0.0 -0.2 -0.4

20 min after bumetanide (170 min after kainate) 0.4

mV

0.2 0.0 -0.2 -0.4

20 min after bumetanide plus phenobarbital (200 min after kainate) 0.4

mV

0.2 0.0 -0.2 -0.4 -0.6

20 min after 2nd bumetanide (350 min after kainate) mV

0.2 0.0 -0.2

20 min after 2nd bumetanide plus phenobarbital (380 min after kainate) mV

0.2 0.0 -0.2

60 min after high-dose phenobarbital (540 min after kainate) mV

0.2 0.0 -0.2

10 s Fig. 2. Representative example of EEG recordings from a rat (# BUM107) in which kainate was injected without anesthesia. EEG was recorded via cortical electrodes above the fronto–parietal cortex. Kainate, 12 mg/kg, was administered i.v.; Since no clear SE was observed, a 2nd i.v. dose of kainate, 4 mg/kg, was administered 60 min after the 1st injection. (A) Baseline EEG before kainate. (B) Status epilepticus (SE) induced by kainate; note the difference to the EEG shown in Fig. 1B in the anesthetized rat. (C) Bumetanide, 5 mg/kg i.v., seems to reduce frequency of the paroxysmal EEG activity. (D) A similar effect is seen when administering a combination of bumetanide, 5 mg/ kg i.v., with phenobarbital, 10 mg/kg i.v., 30 min after the 1st bumetanide bolus. (E) Following a second bumetanide bolus of 5 mg/kg i.v. (150 min after the drug injection in D), the amplitude of the SE EEG activity is reduced. (F) This becomes more marked when administering again the bumetanide (5 mg/kg i.v.)/phenobarbital (10 mg/kg i.v.) combination (30 min after the drug injection in E), but SE still continues. (G) A higher dose of phenobarbital (20 mg/kg i.v.) almost abolishes SE.

When the dose of phenobarbital was increased to 20 mg/kg, SE activity was almost abolished (Fig. 2G). In another rat, SE could be completely blocked by combining phenobarbital (25 mg/kg) with diazepam (10 mg/kg) (not shown). 3.2. Experiments with bumetanide in the lithium/pilocarpine model In both conscious and anesthetized rats, pilocarpine induced typical high-amplitude, high-frequency discharges in the EEG (Figs. 3 and 4) that were accompanied by limbic and generalized convulsive motor seizures in conscious animals. Only few

experiments with lithium/pilocarpine were performed, because no clear anticonvulsant effects of bumetanide were observed (Table 1), independently of whether the experiments were performed in conscious (Fig. 3) or anesthetized rats (Fig. 4). In the experiment illustrated in Fig. 3, combined administration of bumetanide and phenobarbital seemed to reduce both amplitude and frequency of the paroxysmal activity in the cortical EEG. In the experiment under chloral hydrate anesthesia illustrated in Fig. 4, the third bumetanide injection seemed to reduce frequency of the paroxysmal activity, but the amplitude was increased.

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

83

mV

Baseline 0.2 0.0 -0.2

Status epilepticus (20 min after onset) 1.0

0.5

mV

0.0

-0.5

-1.0

-1.5

20 min after bumetanide (50 min after onset of SE) 1.0

mV

0.5

0.0

-0.5

-1.0

60 min after bumetanide plus phenobarbital (120 min after onset of SE) 1.0

mV

0.5

0.0

-0.5

-1.0

60 min after 2nd bumetanide plus phenobarbital (180 min after onset of SE) 1.0

mV

0.5

0.0

-0.5

-1.0

10 s

Fig. 3. Representative example of EEG recordings from a rat (# BUM103) in which pilocarpine was injected without anesthesia. EEG was recorded via cortical electrodes above the fronto–parietal cortex. Pilocarpine was injected i.p. following pretreatment with lithium as described in Section 2. (A) Baseline EEG before pilocarpine. (B) Status epilepticus (SE) induced by pilocarpine. (C) Bumetanide (5 mg/kg i.v.) alone does not affect SE activity. (D) Similarly, a combination of bumetanide, 5 mg/kg i.v., with phenobarbital, 10 mg/kg i.v., does not affect SE, when injected 30 min after the 1st bumetanide injection. (E) However, when administering this combination a 2nd time (60 min after the 1st time), SE activity is reduced. Typically, SE activity as shown in (B) lasts for several hours in this model if not terminated by anticonvulsant drugs.

84

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

Baseline

mV

0.4 0.0 -0.4

Status epilepticus (20 min after onset) 1.5

mV

1.0

0.5

0.0

-0.5

20 min after bumetanide (80 min after onset of SE) 1.5

mV

1.0

0.5

0.0

-0.5

20 min after 2nd bumetanide (110 min after onset of SE) 1.5

mV

1.0

0.5

0.0

-0.5

20 min after 3rd bumetanide (170 min after onset of SE) 2.0

1.5

mV

1.0

0.5

0.0

-0.5

-1.0

10 s Fig. 4. Representative example of EEG recordings from a rat (# BUM113) in which pilocarpine was injected under anesthesia with chloralhydrate. EEG was recorded via cortical electrodes above the fronto–parietal cortex. Pilocarpine was injected i.p. following pretreatment with lithium as described in Section 2. (A) Baseline EEG before pilocarpine. (B) Status epilepticus (SE) induced by pilocarpine 90 min after chloral hydrate. (C) Bumetanide (5 mg/kg i.v.) alone does not affect SE activity. (D) A 2nd bumetanide bolus (5 mg/kg i.v.; 30 min after 1st bolus) also exerts no clear effect on SE. (E) A 3rd bumetanide bolus (5 mg/kg i.v.; 60 min after 2nd bolus) seems to reduce the frequency of paroxysmal activity, but not its amplitude, and SE continues.

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

3.3. Experiments with bumetanide in the electrical BLA stimulation model All experiments with electrically induced SE were performed in conscious rats. As previously reported (Brandt et al., 2003), prolonged (25 min) electrical stimulation of the BLA resulted in a self-sustained partial (limbic) or generalized convulsive SE in the majority of rats. In 9 rats, in which either bumetanide (5 mg/kg) or an equimolar dose of its prodrug BUM5 were injected alone, no obvious effect on SE was observed, irrespective of SE type (Table 1). Representative examples are illustrated in Figs. 5 and 6. Similarly, a low dose of phenobarbital (10 mg/kg i.v.) did not block the SE when given alone (Fig. 7), which was tested in 6 rats (Table 1). However, when bumetanide or BUM5 was combined with this dose of phenobarbital, a marked anticonvulsant effect was observed in some rats (Table 1). A representative example is illustrated in Fig. 5D. Overall, such an anticonvulsant effect of bumetanide (or BUM5) and phenobarbital was observed in 3/9 rats, which was not significantly different from the 0/9 rats with anticonvulsant effect after bumetanide or BUM5 alone (P¼ 0.2059; Fisher's

85

exact test) or the 0/6 rats with anticonvulsant effect after phenobarbital alone (P¼0.2286). However, statistical comparison of the 3/9 rats with an anticonvulsant effect of combined administration with the 15 rats in which either bumetanide/BUM5 or phenobarbital alone did not exert an anticonvulsant effect resulted in a significant difference (P¼ 0.0415), indicating that the combined administration of bumetanide (or BUM5) and phenobarbital resulted in a significant anticonvulsant effect. In 6 rats, we evaluated whether SE could be terminated by increasing the dose of phenobarbital to 15–25 mg/kg or injection of a high dose of diazepam (20 mg/kg); in 4 of these rats (3 with diazepam, 1 with phenobarbital), a rapid anticonvulsant effect was observed (see Fig. 7E), while in the other 2 rats (both with phenobarbital) the anticonvulsant effect was more retarded (not illustrated). Thus, all 6 rats showed SE suppression with high doses of diazepam or phenobarbital, which was significantly different from the efficacy obtained by combining bumetanide with a lower dose of phenobarbital (6/6 vs. 3/9; P¼0.0278). In an additional group of 8 rats (not included in Table 1, because these rats were not treated with bumetanide or low doses of phenobarbital), we evaluated the effect of combined administration of diazepam (10 mg/kg i.p.) and

Baseline mV

0.4 0.2 0.0

Status epilepticus (180 min after beginning of stimulation) 1.0

mV

0.5

0.0

-0.5

20 min after bumetanide (230 min after beginning of stimulation) 1.5

mV

1.0

0.5

0.0

-0.5

60 min after bumetanide plus phenobarbital (300 min after beginning of stimulation) mV

0.4 0.2 0.0 -0.2

10 s Fig. 5. Representative example of EEG recordings from a rat (# BUM117) in which a generalized convulsive (type 3) status epilepticus (SE) was induced by sustained (25 min) electrical stimulation of the basolateral amygdala (BLA) in the absence of anesthesia. EEG was recorded via the BLA electrode. (A) Baseline before onset of BLA stimulation. (B) Self-sustained SE activity. (C) Bumetanide, 5 mg/kg i.v., does not affect SE activity. (D) A combination of bumetanide, 5 mg/kg i.v., with phenobarbital, 10 mg/kg i.v. (injected 30 min after the 1st bumetanide injection), interrupts SE; only few isolated spikes and waves are observed. However, this anticonvulsant effect is not long-lasting, but SE recurs in subsequent hours (not illustrated).

86

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

Baseline mV

0.4 0.0

Status epilepticus (180 min after beginning of stimulation)

mV

0.4 0.0

20 min after BUM5 (230 min after beginning of stimulation) mV

0.4 0.0

60 min after BUM5 plus phenobarbital (300 min after beginning of stimulation) mV

0.4 0.0

10 s Fig. 6. Representative example of EEG recordings from a rat (# BUM139) in which a partial (limbic) and generalized convulsive (type 2) status epilepticus (SE) was induced by sustained (25 min) electrical stimulation of the basolateral amygdala (BLA) in the absence of anesthesia. EEG was recorded via the BLA electrode. (A) Baseline before onset of BLA stimulation. (B) Self-sustained SE activity. (C) The bumetanide prodrug, BUM5, 6.5 mg/kg i.v., does not affect SE activity. (D) Similarly, a combination of BUM5, 6.5 mg/ kg i.v., with phenobarbital, 10 mg/kg i.v. (injected 30 min after the 1st BUM5 injection), does not interrupt SE. A similar lack of any anticonvulsant activity in this model was also seen with bumetanide in the majority of rats (see text).

phenobarbital (25 mg/kg i.p.) on SE induced by BLA stimulation. The drugs were administered after 4 h of SE, resulting in an immediate SE termination within 1–12 min after injection in all rats (not shown). Thus, treatment with this drug combination was more effective to terminate SE than combined treatment with a low dose of phenobarbital and bumetanide or BUM5 (8/8 vs. 3/9; P¼0.009).

4. Discussion Unexpectedly, we could not reproduce the marked anticonvulsant effect of bumetanide on kainate-induced SE in urethaneanesthetized rats reported by Schwartzkroin et al. (1998). In the latter study, the authors extended previous observations with furosemide in the same model (Hochman et al., 1995). Male Sprague–Dawley rats were anesthetized with 1.25 g/kg urethane i.p. and, following termination of surgical procedures (cannulation of jugular vein and stereotactic insertion of cortical EEG electrodes) animals were allowed to recover for 30 min before injection of kainate (10–12 mg/kg i.v.). Anesthesia was maintained by additional urethane injections. Once stable EEG seizure activity was evident, the rats were injected with bumetanide (two boluses of 5 mg/kg, i.v. each, at 80 and 115 min after kainate), resulting in complete blockade of the seizure-like EEG activity within 20– 30 min (Schwartzkroin et al., 1998). However, only one rat was illustrated in the report by Schwartzkroin et al. (1998), and the authors did not describe in how many rats they evaluated this treatment and how robust the finding was. In our hands, using the same rat strain and gender (male Sprague–Dawley rats) as in the experiments of Schwartzkroin et al. (1998), the first obvious difference to the previous reports with furosemide and bumetanide was that by 1.25 g/kg urethane no surgical anesthesia was obtained. Even increasing the urethane dose to 1.4 g/kg did not result in surgical anesthesia, so that it was not possible to perform any surgical preparation under this anesthesia,

but EEG electrodes had to be implanted before the experiments with a more effective anesthesia (chloral hydrate). Second, even though the urethane anesthesia was only shallow, it almost completely blocked SE induction by kainate. Third, even the weak seizure activity observed under these conditions could not be blocked by bumetanide. Therefore, we performed additional kainate experiments without anesthesia, again resulting in no obvious anticonvulsant effects of bumetanide, using the same dosing scheme as in the experiments described by Schwartzkroin et al. (1998). One possible explanation for these apparent discrepancies between the present findings and the report by Schwartzkroin et al. (1998) is intra-strain differences between the Sprague–Dawley rats used by Schwartzkroin's group and the Sprague–Dawley rats used for the present study. We have recently reported marked differences in SE induction in Sprague– Dawley rats purchased from different breeders (Harlan, Charles River, Janvier, Taconic)(Langer et al., 2011). Sprague–Dawley rats are randomly outbred, hence allelic variations can occur across separate colonies. Sprague–Dawley rats from different sources may have little in common with each other besides their names and similarities in pelage, since many of the commercially available animals are outbred and have heterogeneous genetic backgrounds (Festing, 1993; Kacew and Festing, 1996). Genetic divergence between outbred subpopulations may arise from a number of processes, including mutation, natural selection, unconscious selection, and random genetic drift (White and Lee, 1998). However, even if such factors may partly explain the striking differences between our study and the study of Schwartzkroin et al. (1998), the lack of any clear anticonvulsant effect of bumetanide in the kainate model in our study argues against any robust efficacy of this compound to terminate SE. The unexpected lack of such an effect of bumetanide in either anesthetized or conscious rats prompted us to examine whether bumetanide is capable of blocking SE in other rat models. Using the same dosing scheme of bumetanide as in the study of Schwartzkroin et al. (1998) in the kainate model, bumetanide did not exert any obvious anticonvulsant effect in the lithium/pilocarpine model or

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

87

mV

Baseline 0.4 0.2

Status epilepticus (180 min after beginning of stimulation)

mV

0.4 0.0 -0.4

20 min after NaCl (230 min after beginning of stimulation)

mV

0.4 0.0 -0.4

20 min after NaCl plus phenobarbital (260 min after beginning of stimulation)

mV

0.4 0.0 -0.4

20 min after high-dose diazepam (390 min after beginning of stimulation) mV

0.4 0.2 0.0

10 s Fig. 7. Representative example of EEG recordings from a rat (# BUM120) in which a partial (limbic) and generalized convulsive (type 2) status epilepticus (SE) was induced by sustained (25 min) electrical stimulation of the basolateral amygdala (BLA) in the absence of anesthesia. EEG was recorded via the BLA electrode. (A) Baseline before onset of BLA stimulation. (B) Self-sustained SE activity. (C) I.v. injection of saline does not affect SE activity. (D) Similarly, a combination of saline and phenobarbital (10 mg/kg i.v.) does not alter the paroxysmal activity. (E) However, a higher dose of diazepam (20 mg/kg i.v.) terminates SE.

an electrical SE model, in which SE is induced by sustained stimulation of the BLA. Furthermore, the bumetanide prodrug BUM5 was without any detectable anticonvulsant effect. Bumetanide has previously been reported to enhance the anticonvulsant efficacy of phenobarbital in neonatal seizure models (Dzhala et al., 2008; Cleary et al., 2013) and adult amygdalakindled rats (Töllner et al., 2014), which prompted us to evaluate this combination also in SE models. In all three SE models used in this study, combined administration of bumetanide and a relatively low dose (10 mg/kg) of phenobarbital was more effective than either treatment alone, but complete blockade of SE was only observed in the electrical SE model. However, we observed such an effect in the BLA model only in 3 out of 9 rats treated with this drug combination, indicating that about 60% of the rats in this model are resistant to this combined drug activity. If responders and nonresponders to bumetanide or bumetanide/phenobarbital combinations also exist in other SE models, this may explain the striking differences between our and previous studies (Schwartzkroin et al., 1998) in the kainate model of SE. In the present experiments, phenobarbital was administered at a dose (10 mg/kg i.v.), which we recently demonstrated to be ineffective against fully kindled seizures (Töllner et al., 2014), so that it was likely that this dose would not be capable of blocking SE when administered alone. This was proven in the kainate and BLA SE models. However, increasing the dose of phenobarbital to 20–25 mg/kg blocked SE when administered alone or in combination with diazepam, substantiating

previous experiments with these drugs in the lithium–pilocarpine and BLA SE models (Bankstahl and Löscher, 2008). Recent electrophysiological experiments showed that a combination of phenobarbital (100 mM) and bumetanide (10 mM) largely reversed seizure-induced changes in EGABA in CA1 hippocampal neurons from a neonatal seizure model, whereas phenobarbital alone had no effect on GABA reversal potential (Cleary et al., 2013). It is not known whether the same is true in adult seizure models, but our data with bumetanide do not seem to support that perturbed chloride homeostasis is critically involved in maintenance of SE in rat models. Interestingly, similar to the present data with bumetanide in SE models, Cleary et al. (2013) reported that this drug alone did not exert anticonvulsant effects, whereas Dzhala et al. (2005) reported an anti-seizure effect of bumetanide in the kainate model of neonatal seizures in rats. Most previous studies with systemic administration of bumetanide in adult rodent models of seizures or epilepsy have been negative or equivocal (Löscher et al., 2013). In this respect, it is important to note that bumetanide, because of high degree of protein-binding and ionization in the blood, only poorly penetrates into the brain after systemic administration, and is rapidly metabolized by adult rodents, so that high doses of the drug have to be administered to reach brain concentrations in the range known to inhibit NKCC1 (Brandt et al., 2010; Töllner et al., 2014; Töpfer et al., 2014). Even with such high doses of Z5 mg/kg as used in the present study, maximal brain levels are only in the range of 100–300 nM in adult rats, and rapidly decline with a

88

K. Töllner et al. / European Journal of Pharmacology 746 (2015) 78–88

half-life of about 20 min (Töllner et al., 2014; Töpfer et al., 2014). With the bumetanide prodrug BUM5, bumetanide brain levels could be increased to about 0.8 mM in rats (Töllner et al., 2014). Although bumetanide has been reported to inhibit neuronal NKCC1 with an IC50 of 200–300 nM in certain preparations (O’Grady et al., 1987; Russell, 2000; Hannaert et al., 2002), the bumetanide concentrations typically used in in vitro studies for effective CNS inhibition of NKCC1 are in the range of 2–10 mM (Puskarjov et al., 2014), which is not reached by systemic administration of bumetanide or BUM5 in rats up to doses of 10–15 mg/kg (Brandt et al., 2010; Töllner et al., 2014; Töpfer et al., 2014). In light of bumetanide's poor BBB penetration, its effects on the CNS have often been a priori attributed in part to enhanced brain accumulation as a result of a seizureinduced breakdown of the blood–brain barrier (Puskarjov et al., 2014). However, the only study to investigate this in neonatal rats did not show any significant increase in brain accumulation of bumetanide (given at 0.2–0.3 mg/kg i.p.) following hypoxia-induced seizures (Cleary et al., 2013). In line with this, after a relatively high dose of bumetanide (10 mg/kg i.v.), no difference in total hippocampal bumetanide levels were reported between nonkindled and fully kindled adult rats (Töpfer et al., 2014), or in total brain bumetanide levels between adult mice 24 h after pilocarpine SE or sham treatment (Töllner et al., 2014). Thus, the lack of any robust anticonvulsant effect of bumetanide or BUM5 in different SE models in adult rats seen in the present study may be a consequence of the low bumetanide brain levels achieved with such treatment.

5. Conclusions Our experiments did not reproduce previous findings of a marked anticonvulsant effect of bumetanide in the kainate SE model in adult rats (Schwartzkroin et al., 1998). Furthermore, bumetanide failed to interrupt SE in other SE models, but potentiated the anticonvulsant efficacy of phenobarbital in part of the animals. However, higher doses of phenobarbital, particularly in combination with diazepam, were more effective to terminate SE than combinations of bumetanide and phenobarbital. Based on preclinical data in neonatal seizure models, two large clinical trials currently investigate bumetanide as an add-on therapy to phenobarbital in newborns with refractory seizures (Kahle and Staley, 2012). The present data do not suggest that bumetanide, alone or in combination with phenobarbital, is a valuable option in the treatment of refractory SE in adult patients.

Acknowledgments We thank Daryl W. Hochman for detailed informations about his previous experiments with furosemide and bumetanide in the rat kainate model (Hochman et al., 1995; Schwartzkroin et al., 1998). The experiments were supported by a Grant (Lo 274/11-2) from the Deutsche Forschungsgemeinschaft (Bonn, Germany). We appreciate the skilful technical assistance of M. Weißing and J. Hinz. The authors have no conflict of interest to declare. References Bankstahl, J.P., Löscher, W., 2008. Resistance to antiepileptic drugs and expression of P-glycoprotein in two rat models of status epilepticus. Epilepsy Res. 82, 70–85. Bethmann, K., Brandt, C., Löscher, W., 2007. Resistance to phenobarbital extends to phenytoin in a rat model of temporal lobe epilepsy. Epilepsia 48, 816–826. Blaesse, P., Airaksinen, M.S., Rivera, C., Kaila, K., 2009. Cation–chloride cotransporters and neuronal function. Neuron 61, 820–838. Bragin, A., Azizyan, A., Almajano, J., Engel Jr., J., 2009. The cause of the imbalance in the neuronal network leading to seizure activity can be predicted by the electrographic pattern of the seizure onset. J. Neurosci. 29, 3660–3671.

Brandt, C., Glien, M., Potschka, H., Volk, H., Löscher, W., 2003. Epileptogenesis and neuropathology after different types of status epilepticus induced by prolonged electrical stimulation of the basolateral amygdala in rats. Epilepsy Res. 55, 83–103. Brandt, C., Nozadze, M., Heuchert, N., Rattka, M., Löscher, W., 2010. Diseasemodifying effects of phenobarbital and the NKCC1 inhibitor bumetanide in the pilocarpine model of temporal lobe epilepsy. J. Neurosci. 30, 8602–8612. Chen, J.W., Wasterlain, C.G., 2006. Status epilepticus: pathophysiology and management in adults. Lancet Neurol. 5, 246–256. Cleary, R.T., Sun, H., Huynh, T., Manning, S.M., Li, Y., Rotenberg, A., Talos, D.M., Kahle, K.T., Jackson, M., Rakhade, S.N., Berry, G., Jensen, F.E., 2013. Bumetanide enhances phenobarbital efficacy in a rat model of hypoxic neonatal seizures. PLoS One 8, e57148. Deeb, T.Z., Maguire, J., Moss, S.J., 2012. Possible alterations in GABAA receptor signaling that underlie benzodiazepine-resistant seizures. Epilepsia 53 (Suppl. 9), S79–S88. Dzhala, V.I., Talos, D.M., Sdrulla, D.A., Brumback, A.C., Mathews, G.C., Benke, T.A., Delpire, E., Jensen, F.E., Staley, K.J., 2005. NKCC1 transporter facilitates seizures in the developing brain. Nat. Med. 11, 1205–1213. Dzhala, V.I., Brumback, A.C., Staley, K.J., 2008. Bumetanide enhances phenobarbital efficacy in a neonatal seizure model. Ann. Neurol. 63, 222–235. Festing, M.F., 1993. Genetic variation in outbred rats and mice and its implications for toxicological screening. J. Exp. Anim. Sci. 35, 210–220. Hannaert, P., Alvarez-Guerra, M., Pirot, D., Nazaret, C., Garay, R.P., 2002. Rat NKCC2/ NKCC1 cotransporter selectivity for loop diuretic drugs. Naunyn-Schmiedeberg's Arch. Pharmacol. 365, 193–199. Hara, K., Harris, R.A., 2002. The anesthetic mechanism of urethane: the effects on neurotransmitter-gated ion channels. Anesth. Analg. 94, 313–318. Hochman, D.W., Baraban, S.C., Owens, J.W., Schwartzkroin, P.A., 1995. Dissociation of synchronization and excitability in furosemide blockade of epileptiform activity. Science 270, 99–102. Holtkamp, M., Matzen, J., Buchheim, K., Walker, M.C., Meierkord, H., 2003. Furosemide terminates limbic status epilepticus in freely moving rats. Epilepsia 44, 1141–1144. Kacew, S., Festing, M.F., 1996. Role of rat strain in the differential sensitivity to pharmaceutical agents and naturally occurring substances. J. Toxicol. Environ. Health 47, 1–30. Kahle, K.T., Staley, K.J., 2012. Neonatal seizures and neuronal transmembrane ion transport. In: Noebels, J.L., Avoli, M., Rogawski, M.A., Olsen, R.W., DelgadoEscueta, A.V. (Eds.), Jasper's Basic Mechanisms of the Epilepsies, 4th edition Oxford University Press, New York, pp. 1066–1076. Kapur, J., Coulter, D.A., 1995. Experimental status epilepticus alters gammaaminobutyric acid type A receptor function in CA1 pyramidal neurons. Ann. Neurol. 38, 893–900. Langer, M., Brandt, C., Löscher, W., 2011. Marked strain and substrain differences in induction of status epilepticus and subsequent development of neurodegeneration, epilepsy, and behavioral alterations in rats. Epilepsy Res. 96, 207–224. Lee, H.H., Jurd, R., Moss, S.J., 2010. Tyrosine phosphorylation regulates the membrane trafficking of the potassium chloride co-transporter KCC2. Mol. Cell. Neurosci. 45, 173–179. Löscher, W., 2007. Mechanisms of drug resistance in status epilepticus. Epilepsia 48 (Suppl. 8), S74–S77. Löscher, W., Puskarjov, M., Kaila, K., 2013. Cation–chloride cotransporters NKCC1 and KCC2 as potential targets for novel antiepileptic and antiepileptogenic treatments. Neuropharmacology 69, 62–74. Macdonald, R.L., Kapur, J., 1999. Acute cellular alterations in the hippocampus after status epilepticus. Epilepsia 40, S9–S20. O’Grady, S.M., Palfrey, H.C., Field, M., 1987. Characteristics and functions of Na–K–Cl cotransport in epithelial tissues. Am. J. Physiol. 253, C177–C192. Paxinos, G., Watson, C., 2007. The Rat Brain in Stereotaxic Coordinates, 6th ed. Elsevier, Amsterdam. Puskarjov, M., Kahle, K.T., Ruusuvuori, E., Kaila, K., 2014. Pharmacotherapeutic targeting of cation–chloride cotransporters in neonatal seizures. Epilepsia 55, 806–818. Rattka, M., Brandt, C., Löscher, W., 2012. Do proconvulsants modify or halt epileptogenesis? Pentylenetetrazole is ineffective in two rat models of temporal lobe epilepsy. Eur. J. Neurosci. 36, 2505–2520. Rattka, M., Brandt, C., Löscher, W., 2013. The intrahippocampal kainate model of temporal lobe epilepsy revisited: epileptogenesis, behavioral and cognitive alterations, pharmacological response, and hippoccampal damage in epileptic rats. Epilepsy Res. 103, 135–152. Rossetti, A.O., Lowenstein, D.H., 2011. Management of refractory status epilepticus in adults: still more questions than answers. Lancet Neurol. 10, 922–930. Russell, J.M., 2000. Sodium–potassium–chloride cotransport. Physiol. Rev. 80, 211–276. Schwartzkroin, P.A., Baraban, S.C., Hochman, D.W., 1998. Osmolarity, ionic flux, and changes in brain excitability. Epilepsy Res. 32, 275–285. Töllner, K., Brandt, C., Töpfer, M., Brunhofer, G., Erker, T., Gabriel, M., Feit, P.W., Lindfors, J., Kaila, K., Löscher, W., 2014. A novel prodrug-based strategy to increase effects of bumetanide in epilepsy. Ann. Neurol. 75, 550–562. Töpfer, M., Töllner, K., Brandt, C., Twele, F., Bröer, S., Löscher, W., 2014. Consequences of inhibition of bumetanide metabolism in rodents on brain penetration and effects of bumetanide in chronic models of epilepsy. Eur. J. Neurosci. 39, 673–687. White, W.J., Lee, C.S., 1998. The development and maintenance of the Crl:CD(SD)IGS BR rat breeding system. In: CD(SD)IGS Study Group (Ed.), Biological Reference Data on CD(SD)IGS Rats. Best Printing Co. Ltd., Tokyo, pp. 8–14.