Involvement of the somatostatin-2 receptor in the anti-convulsant effect of angiotensin IV against pilocarpine-induced limbic seizures in rats

Authors

  • Bart Stragier,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Ralph Clinckers,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Alfred Meurs,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Dimitri De Bundel,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Sophie Sarre,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Guy Ebinger,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Yvette Michotte,

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Ilse Smolders

    1. Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author

Address correspondence and reprint requests to Yvette Michotte, Department of Pharmaceutical Chemistry, Drug Analysis and Drug Information, Research Group Experimental Pharmacology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium.
E-mail: ymichot@vub.ac.be

Abstract

The anti-convulsant properties of angiotensin IV (Ang IV), an inhibitor of insulin-regulated aminopeptidase (IRAP) and somatostatin-14, a substrate of IRAP, were evaluated in the acute pilocarpine rat seizure model. Simultaneously, the neurochemical changes in the hippocampus were monitored using in vivo microdialysis. Intracerebroventricularly (i.c.v.) administered Ang IV or somatostatin-14 caused a significant increase in the hippocampal extracellular dopamine and serotonin levels and protected rats against pilocarpine-induced seizures. These effects of Ang IV were both blocked by concomitant i.c.v. administration of the somatostatin receptor-2 antagonist cyanamid 154806. These results reveal a possible role for dopamine and serotonin in the anti-convulsant effect of Ang IV and somatostatin-14. Our study suggests that the ability of Ang IV to inhibit pilocarpine-induced convulsions is dependent on somatostatin receptor-2 activation, and is possibly mediated via the inhibition of IRAP resulting in an elevated concentration of somatostatin-14 in the brain.

Abbreviations used
Ang

angiotensin

cyanamid 154806

Ac-(4-Nitro)Phe-D-Cys-Tyr-D-Trp-Lys-Thr-Cys-D-Tyr-NH2

DA

dopamine

5-HT

serotonin

i.c.v.

intracerebroventricularly

i.p.

intraperitoneally

IRAP

insulin-regulated aminopeptidase

RAS

renin-angiotensin system

SSS

seizure severity score

sst2

somatostatin-2 receptor

TSSS

total seizure severity score

The renin-angiotensin system (RAS) consists of a number of angiotensin (Ang) peptides. Among them, Ang II is well known for its hypertensive effect in which angiotensin 1 (AT1) receptors play a major role (de Gasparo et al. 2000). Besides Ang II, the central effects of the Ang II (3–8) fragment, termed Ang IV, have gained a lot of interest (for a review see Mustafa et al. 2001). In this context it has already been shown that Ang IV considerably improves the learning of conditioned avoidance responses and facilitates the recall of passive avoidance behaviour in rats (Braszko et al. 1988). In subsequent reports intracerebroventricularly (i.c.v.)-injected Ang IV has also been found to prevent learning and memory deficits caused by scopolamine (Pederson et al. 1998) and bilateral knife cuts of the perforant path (Wright et al. 1999). On the other hand, it has been shown that Ang IV protects against the clonic convulsions, increases the seizure threshold and alleviates the seizure intensity in kindled mice (Tchekalarova et al. 2001, 2004, 2005). The underlying mechanisms of this anti-convulsant effect of Ang IV are not yet known. Roberts et al. (1995) demonstrated that Ang IV induces c-fos expression in seizure-prone brain areas. A close relationship between Ang IV and the dopamine (DA) neurotransmission in the brain has been proposed (Tchekalarova and Georgiev 2005). In this context we already showed that Ang IV causes a dose-dependent increase of extracellular DA levels in the striatum, a brain region linked to the control of movement (Stragier et al. 2004). As it is known that DA can prevent epileptogenesis by curbing neuronal hyperexcitability in the hippocampus (Starr 1996), the present in vivo microdialysis study was conceived to evaluate the effect of i.c.v.-administered Ang IV on the hippocampal DA concentration. Moreover, these effects were linked to the anti-convulsant effect of Ang IV in the acute pilocarpine rat epilepsy model. Interactions between Ang IV and other neurotransmitters that are related to seizure occurence such as γ-aminobutyric acid (GABA), glutamate (Glu) (Bradford 1995) and serotonin (5-HT) (Lu and Gean 1998) were also examined.

Initially, the ‘AT4 receptor’ was postulated as the pharmacodynamic target mediating the effects elicited by Ang IV (de Gasparo et al. 2000). Later, it was shown that the Ang-IV binding site corresponds to insulin-regulated aminopeptidase (IRAP) (Albiston et al. 2001), a membrane-associated aminopeptidase. IRAP is abundantly present in the hippocampus (Fernando et al. 2005). Lew et al. (2003) showed that Ang IV is a competitive inhibitor of IRAP with an IC50 value of 0.2 µm. IRAP mediates the degradation of different neuropeptides such as vasopressin, substance P and somatostatin in vitro (Herbst et al. 1997; Albiston et al. 2001). Among these, only somatostatin has anti-convulsant properties (Vezzani and Hoyer 1999; Cammalleri et al. 2004) via the somatostatin-2 receptor (sst2) (Cammalleri et al. 2004). In the current work, the effect of i.c.v.-administered Ang IV on hippocampal DA, 5-HT, GABA and Glu was compared with that of somatostatin-14. Also, the effect of somatostatin-14 against pilocarpine-induced limbic seizures was evaluated. The involvement of the sst2 receptor in the anti-convulsant properties of Ang IV was examined by co-administering the selective sst2 receptor antagonist Ac-(4-Nitro)Phe-D-Cys-Tyr-D-Trp-Lys-Thr-Cys-D-Tyr-NH2 (cyanamid 154806) (Feniuk et al. 2000).

Materials and methods

Animals

All experiments were carried out on freely moving rats, according to the national guidelines on animal experimentation and were approved by the Ethical Committee for Animal Experiments of the Faculty of Medicine and Pharmacy of the Vrije Universiteit Brussel. All efforts were made to minimize pain and discomfort of the animals.

The experiments were performed on male Wistar rats (Iffa Credo, Brussels, Belgium), weighing 250–300 g. The rats were given free access to water and standard rat chow.

Surgery

The rats were anaesthetized with a mixture of ketamine : diazepam (60 : 4.5 mg/kg intraperitoneally, i.p.) and placed on a stereotaxic frame. The skull was exposed and two burr holes were drilled to implant two cannulas with replaceable inner guide (CMA Microdialysis, Solna, Sweden) according to the atlas of Paxinos and Watson (1986). The first guide cannula was positioned 3 mm above the right hippocampus (the CA1–CA3 region); coordinates relative to bregma were L, + 4.6; A, −5.6; V, + 4.6. A second guide cannula was implanted into the right lateral ventricle; coordinates relative to bregma were L, + 1.4; A, −0.9; V, + 3.5.

After surgery, the rats received 4 mg/kg ketoprofen i.p. as an analgesic. Immediately after surgery, a microdialysis probe with a membrane length of 3 mm was introduced into the hippocampus via the cannula (CMA 12; CMA Microdialysis). The probe was perfused with modified Ringer's solution containing 147 mm NaCl, 4 mm KCl and 2.3 mm CaCl2 at a constant flow-rate of 2 µL/min using a CMA 100 microdialysis pump (CMA Microdialysis).

Animals were allowed to recover from surgery overnight and dialysate collection was started the day after surgery.

In vivo microdialysis experiments

Samples were collected every 20 min, yielding 40 µL dialysates. A 15-µL sample was used to measure either the GABA or the Glu dialysate concentration; 6.25-µL of filtered antioxidant mixture (0.1 m acetic acid, 3.3 mm l-cystein, 0.27 mm Na2EDTA, 12.5 µm ascorbic acid) was added to the remaining volume to prevent oxidation of DA and 5-HT prior to their determination.

Before any pharmacological manipulation was performed, six basal dialysate samples were collected. The mean of these neurotransmitter dialysate concentrations was taken as baseline value at time zero. Pharmacological manipulations were carried out according to the respective protocols. In all experiments, the drugs were dissolved in the modified Ringer's solution. A dose–response relationship was determined for the i.c.v.-administered peptides. A rate of 1 nmol per hour was chosen as the lowest dose. However, as 1 nmol per hour had no effect, further experiments were performed with a dose of 10 nmol per hour. A dose–response relationship (2.5–50 pmol per hour) was also determined for cyanamid 154806. Candesartan was infused in a dose previously used by Konishi et al. (2002) (2.25 nmol per hour). At the end of the experiment, the rats were killed with an overdose of pentobarbital (Nembutal®).

Experimental groups designed to examine the anti-convulsant effect of Ang IV

Group 1: effect of i.c.v.-administered modified Ringer's solution, Ang IV, somatostatin-14 and cyanamid 154806. After the measurement of six baseline samples, an infusion probe (a microdialysis probe without membrane, CMA 12; CMA Microdialysis) was introduced in the lateral ventricle via the cannula (CMA 12; CMA Microdialysis) and modified Ringer's solution (n = 11), Ang IV (n = 22), somatostatin-14 (n = 21) or cyanamid 154806 (n = 6) were administered i.c.v. via continuous infusion (0.2 µL/min) for 220 min. Ang IV was administered in a dose of 1 (n = 8) and 10 (n = 14) nmol per hour. Somatostatin-14 was administered in a dose of 1 (n = 7) and 10 nmol (n = 14) per hour. Cyanamid 154806 was administered in a dose of 50 pmol per hour. After this, the experiment was continued for 60 min to determine the reversibility of the effect.

Group 2: pilocarpine control group. After the measurement of six baseline samples with modified Ringer's solution, convulsions were evoked by hippocampal perfusion with pilocarpine in a concentration of 10 mm (n = 15) for 40 min. Subsequently, sample collection was continued for 100 min.

Group 3: effect of i.c.v.-administered Ang IV, somatostatin-14 or cyanamid 154806 on seizures induced by the hippocampal perfusion of pilocarpine (10 mm). After the measurement of six baseline samples, an infusion probe was introduced in the lateral ventricle via the cannula. Ang IV (10 nmol per hour; n = 15), somatostatin-14 (10 nmol per hour; n = 11) or cyanamid 154806 (50 pmol per hour; n = 8) were administered i.c.v. via continuous infusion (0.2 µL/min) for 220 min. Pilocarpine (10 mm) was perfused intrahippocampally for 40 min starting 80 min after commencing the i.c.v. infusion.

Group 4: effect of the co-infusion with either cyanamid 154806 or candesartan on the anti-convulsant effect of Ang IV and the effect of the co-infusion with cyanamid 154806 on the anti-convulsant effect of somatostatin-14. The same protocol was followed as described for group 3, but with the co-infusion of the sst2 receptor antagonist cyanamid 154806 in a dose–response relationship (2.5, 25 and 50 pmol per hour) and Ang IV (10 nmol per hour) for 220 min (n = 21). The AT1 receptor is the best characterized receptor of the RAS, and Ang IV in a concentration above 1 µm is able to stimulate this receptor (Le et al. 2002). In order to exclude possible AT1 effects, additional experiments were performed in which Ang IV (10 nmol per hour for 220 min) was co-administered with the AT1 antagonist candesartan (2.25 nmol per hour; Konishi et al. 2002) (n = 9) according to the same protocol as described for group 3.

The sst2 receptor antagonist cyanamid 154806 (50 pmol per hour) was co-infused with somatostatin-14 (10 nmol per hour; n = 10) for 220 min, again according to the same protocol as described for group 3.

Experimental group with a non-convulsant dose of pilocarpine

Group 5: effect of i.c.v.-administered somatostatin-14 on hippocampal perfusion of pilocarpine (100 μm). The same protocol was followed as described for group 3, but with hippocampal perfusion of pilocarpine in a subtreshold concentration (100 µm; n = 8).

Control experiments were also performed with pilocarpine in a concentration of 100 µm, according to the protocol of group 2 (n = 5; data not shown).

Seizure severity score

During the experiments in which pilocarpine was administered, seizure severity was assessed. This was based on the observation of behavioural manifestations for each collection period starting with the first sample in which pilocarpine was perfused. The observation period was 140 min because most of the epileptic phenomena occurred within this time range (as observed in our laboratory). After 140 min the seizure severity diminishes. The seizure severity score (SSS) was adapted from Racine's scale (Racine 1972) to take into account the typical behavioural changes associated with pilocarpine-induced motor seizures. This adapted scale consists of six scores that correspond to the successive developmental stages of motor seizures: (0) normal non-epileptic activity; (1) mouth and facial movements, hyperactivity, grooming, sniffing, scratching, wet-dog shakes; (2) head nodding, staring, tremor; (3) forelimb clonus, forelimb extension; (4) rearing, salivating; (5) falling, status epilepticus. Seizure severity was then determined by the summation of the highest SSS obtained during each collection period, resulting in a total seizure severity score (TSSS) for each individual animal.

Chromatographic assay

For the determination of the DA and 5-HT concentration in the dialysates, an isocratic microbore liquid chromatography (LC) assay was used (C8, 5 µm; 100 × 1 mm; Unijet, Bioanalytical Systems, West Lafayette, IN, USA) coupled to amperometric detection (Decade, Antec, Leiden, the Netherlands) as previously described in detail (Sarre et al. 1997). The detection limit was 0.05 nm for both monoamines.

Chromatographic conditions and pre-column derivatization procedures for GABA and Glu analyses have been previously described (Smolders et al. 1995; Van Hemelrijck et al. 2005). Reversed-phase isocratic microbore LC (C8, 5 µm; 100 × 1 mm; Unijet, Bioanalytical Systems) with amperometric detection was used for the analysis of GABA after pre-column derivatization with o-phtalaldehyde/2-methyl-2-propanethiol and iodoacetamide. The limit of detection was 1 nm. The Glu content of the dialysates was determined by reversed-phase LC (C18, 5 µm; 250 × 2 mm; Shiseido, CAPCELL PAK MG, Analis, Suarlée, Belgium) with gradient elution and fluorescence detection (Shimadzu, Duisburg, Germany). o-Phtalaldehyde/β-mercaptoethanol was added to the sample to yield fluorescent derivatives. The limit of detection was 0.03 µm.

Chemicals

Angiotensin IV and somatostatin-14 were supplied by NeoMPS (Strasbourg, France). Pilocarpine. HCl and cyanamid 154806 were purchased from Sigma (St Louis, MO, USA). All other chemicals were either analytical grade or better and supplied by Merck (Darmstadt, Germany). Aqueous solutions were prepared in fresh water purified by a Seralpur Pro 90 CN system (Merck Belgolabo, Overijse, Belgium) and filtered through a membrane filter with a pore size of 0.2 µm.

Statistical analysis

The extracellular concentrations of DA, 5-HT, GABA and Glu in baseline conditions were calculated in nm (mean ± SEM). Extracellular neurotransmitter levels after the pharmacological manipulation were expressed as percentages (mean ± SEM) of the mean baseline value, expressed as 100%. No corrections were made for probe recovery across the dialysis membrane. Therefore, the reported extracellular concentrations of the transmitters are actually dialysate concentrations. The acquired SSSs and TSSSs are represented as the mean ± SEM. For testing statistical significance of differences in neurotransmitter levels and mean SSSs after the administration of drugs compared with baseline values, a one-way analysis of variance (anova) for repeated measures was used. When the anova was significant, the Wilcoxon test was carried out to compare the extracellular neurotransmitter levels and the mean SSSs after the pharmacological manipulation at the different time points with the baseline value. A Mann–Whitney test was used for comparison of either mean neurotransmitter concentrations or SSSs at a certain time point and of mean TSSSs between groups. For all statistical analyses α = 0.05.

Results

Mean basal output of DA, 5-HT, GABA and Glu in the hippocampus of the freely moving Wistar rat

The mean (± SEM) basal output in the hippocampus of the freely moving Wistar rat was 0.21 ± 0.03 nm (n = 54) for DA, 0.27 ± 0.05 nm (n = 54) for 5-HT, 658 ± 94 nm (n = 55) for Glu and 33.23 ± 4.30 nm (n = 54) for GABA.

Experimental groups designed to examine the anti-convulsant effect of Ang IV

Group 1: effect of i.c.v.-administered modified Ringer's solution, Ang IV, somatostatin-14 and cyanamid 154806. The i.c.v. administration of modified Ringer's solution (Fig. 1a) and cyanamid 154806 (50 pmol per hour) (Fig. 1d) had no effect on hippocampal neurotransmitter release.

Figure 1.

 Effect of the i. c.v. infusion of (a) modified Ringer's solution, (b) Ang IV (10 nmol per hour), (c) somatostatin-14 (10 nmol per hour) and (d) cyanamid 154806 (50 pmol per hour) on the extracellular DA, 5-HT, Glu and GABA levels in the hippocampus of the rat. Baseline levels are set to 100% (mean of six values). The values at all other points are expressed as a percentage of the baseline value. Each value is the mean ± SEM. The modified Ringer's solution, Ang IV and somatostatin-14 are administered for 220 min (shown by the horizontal bar). Data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values significantly different from the corresponding baseline value (*p < 0.05).

The i.c.v. administration of Ang IV via a continuous infusion in a dose of 1 and 10 nmol per hour caused drinking behaviour in all rats, which could not be blocked by the co-administration of the AT1 antagonist candesartan (2.25 nmol per hour). Figure 1(b) only depicts the effect on hippocampal neurotransmitter release of Ang IV at a dose of 10 nmol because the administration of 1 nmol Ang IV per hour had no effect. Ang IV (10 nmol per hour) provoked a significant increase in the extracellular hippocampal DA and 5-HT levels (Fig. 1b; time 20–220 min). For DA, the increase was maximal after 160 min of administration of Ang IV and was 245% of baseline (p < 0.001). The 5-HT level was significantly elevated with a maximal increase of 215% of baseline after 200 min (p < 0.001). The hippocampal DA and 5-HT levels remained elevated during Ang IV administration and returned subsequently to baseline (Fig. 1b; time 240–280 min). I.c.v.-administered Ang IV had no effect on the hippocampal Glu release (Fig. 1b; 20–220 min) and caused a significant decrease in the hippocampal GABA levels (Fig. 1b; 20–220 min). The decrease was maximal after 80 min of Ang IV administration and was 50% of baseline (p < 0.001). The GABA levels remained lowered during the infusion of Ang IV and returned to baseline afterwards (Fig. 1b; 240–280 min).

Somatostatin-14 was i.c.v.-administered in a dose of 1 and 10 nmol per hour. Figure 1(c) only depicts the effect of somatostatin-14 on hippocampal neurotransmitter release at a dose of 10 nmol becase the administration of the lower dose caused no significant effect. Somatostatin-14 provoked a significant increase of the hippocampal DA and 5-HT concentration (Fig. 1c; 20–220 min). For DA, the increase was maximal following 160 min of administration and was 325% of baseline (p < 0.001). The effect on 5-HT was maximal after 220 min of administration with a maximal increase of 250% of baseline (p < 0.001). The hippocampal DA and 5-HT levels remained elevated during somatostatin-14 administration and returned to baseline afterwards (Fig. 1c; 240–280 min). I.c.v.-administered somatostatin-14 had no effect on the hippocampal Glu and GABA overflow (Fig. 1c; 20–220 min).

Group 2: pilocarpine control group. Pilocarpine was perfused in the hippocampus in a concentration of 10 mm (for 40 min), and acutely caused pre-convulsive symptoms such as hyperactivity, wet-dog shakes and tremors. After approximately 40 min, all rats developed full-blown limbic seizures characterized by salivating, rearing and falling. The mean TSSS was 14.7 ± 0.5 (n = 7; Fig. 2a). The maximal seizure severity (SSS) was obtained 120 min after commencing the pilocarpine perfusion and was 3.4 ± 0.3 (Fig. 2a). Local administration of pilocarpine caused an increase of the extracellular hippocampal DA, 5-HT, GABA and Glu concentrations (Fig. 3). DA and 5-HT levels were significantly elevated with maximal increases 280% of (p < 0.001) and 230% (p < 0.001) of baseline, respectively, after 40 min of perfusion. After the administration of pilocarpine, DA and 5-HT concentrations returned to baseline levels. The Glu concentration was significantly increased to a maximum of 250% after 20 min of perfusion (p = 0.032) and returned to baseline after the administration of pilocarpine. The pilocarpine-induced increase of the hippocampal GABA concentration was maximal 40 min after the cessation of the pilocarpine administration and was 220% of baseline (p < 0.001). The GABA concentration remained elevated until the end of the experiment.

Figure 2.

  (a) The effect of Ang IV (10 nmol per hour), somatostatin-14 (10 nmol per hour) and of the sst2 receptor antagonist cyanamid 154806 (50 pmol per hour) on pilocarpine-induced seizures. (b) The effect of Ang IV (10 nmol per hour), Ang IV (10 nmol per hour) in combination with cyanamid 154806 (2.5, 25 and 50 pmol per hour) and with candesartan (2.25 nmol per hour) on pilocarpine-induced seizures. (c) The effect of somatostatin-14 (10 nmol per hour) and somatostatin-14 (10 nmol per hour) in combination with cyanamid 154806 (50 pmol per hour) on pilocarpine-induced convulsions. The seizure severity of each set of experiments is expressed as SSSs during each microdialysis collection interval and as a TSSS. For the SSSs, data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values significantly different from the corresponding baseline value (*p < 0.05). A Mann–Whitney test was used to compare the TSSS of the different experiments with the TSSS obtained from the pilocarpine control experiments (a), the pilocarpine-Ang IV experiments (b) and the pilocarpine-somatostatin-14 experiments (c). Statistical significance (§p < 0.05).

Figure 3.

 Effect of the hippocampal perfusion of pilocarpine (10 mm) on the extracellular DA, 5-HT, Glu and GABA concentration in the hippocampus of the rat. Baseline levels are set to 100% (mean of six values). The values at all other points are expressed as a percentage of the baseline value. Each value is the mean ± SEM. Pilocarpine is perfused for 40 min (shown by the horizontal bar). Data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values significantly different from the corresponding baseline value (*p < 0.05).

Group 3: effect of i.c.v.-administered Ang IV, somatostatin-14 or cyanamid 154806 on pilocarpine-induced seizures. The i.c.v. administration of Ang IV in a dose of 10 nmol per hour was able to protect the rats from pilocarpine-induced seizures. The mean TSSS (4.2 ± 1.1, n = 15) was significantly decreased (p < 0.05) when compared with the pilocarpine control group (Fig. 2a). The maximum SSS was obtained after 80 min and was 1.0 ± 0.4, which was significantly lower than the SSS of the pilocarpine control group at the same time point (p < 0.05) (Fig. 2a). The animals only occasionally developed mild pre-convulsive phenomena like wet-dog shakes and hyperactivity. During the pilocarpine administration, the Ang IV-induced increased DA and 5-HT levels (Fig. 4a; 20–80 min) were further increased up to 400% and 335% of baseline, respectively (Fig. 4a; 100–120 min), which was significantly higher (p < 0.05) compared with the pilocarpine control group (Fig. 3; 20–40 min). Cessation of the pilocarpine administration resulted in a restoration of the baseline DA and 5-HT levels, even while the infusion of Ang IV continued (Fig. 4a; 140–220 min). Compared with the pilocapine control experiments (Fig. 3; 20–140 min), Ang IV did not attenuate the pilocarpine-induced GABA increase (p > 0.05) (Fig. 4a; 140–220 min) and abolished the pilocarpine-induced Glu increase (p < 0.05) (Fig. 4a; 100–120 min).

Figure 4.

 Effect of an i. c.v. infusion of (a) Ang IV (10 nmol per hour), (b) somatostatin-14 (10 nmol per hour) and (c) cyanamid 154806 (50 pmol per hour) on pilocarpine (10 mm)-induced changes in the DA, 5-HT, Glu and GABA release in the hippocampus of the rat. Baseline levels are set to 100% (mean of six values). The values at all other points are expressed as a percentage of the baseline value. Each value is the mean ± SEM. Ang IV, somatostatin-14 and cyanamid 154806 are administered for 220 min and pilocarpine (10 mm) for 40 min (shown by the horizontal bars). Data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values that are significantly different compared with the corresponding baseline value (*p < 0.05).

The i.c.v. administration of somatostatin-14 in a dose of 10 nmol per hour was also able to protect the rats from pilocarpine-induced seizures. The mean TSSS (2.6 ± 0.8, n = 11) was significantly decreased (p < 0.05) when compared with the pilocarpine control group (Fig. 3a). The maximum SSS was obtained after 60 min and was 1.4 ± 0.4, which was significantly lower than the SSS of the pilocarpine control group at the same time point (p < 0.05) (Fig. 2a). After the somatostatin-14-induced DA and 5-HT increase (Fig. 4b; 20–80 min), the administration of pilocarpine caused a further increase of the DA and 5-HT concentration up to 400% of baseline (Fig. 4b; 100–120 min), which is again significantly higher (p < 0.05) compared with the pilocarpine control group (Fig. 3; 20–40 min). Cessation of the pilocarpine administration resulted in a restoration of the baseline DA and 5-HT levels (Fig. 4b; 140–220 min). Compared with the pilocarpine control experiments (Fig. 3; 20–140 min), somatostatin-14 had no significant effect on the pilocarpine-induced GABA increase (p > 0.05) (Fig. 4b; 140–220 min), and abolished the pilocarpine-induced Glu increase (p < 0.05) (Fig. 4b; 100–120 min).

The i.c.v. administration of cyanamid 154806 in a dose of 50 pmol per hour had no effect on pilocarpine-induced seizures. The mean TSSS (14.5 ± 1.4, n = 4) was not significantly different from the pilocarpine control group (p > 0.05) (Fig. 2a). The maximum SSS was obtained after 80 min and was 3.3 ± 0.4, which was not significantly different from the SSS of the pilocarpine control group at the same time point (p > 0.05) (Fig. 2a). Compared with the pilocarpine control group (Fig. 3; 20–40 min), co-administration of pilocarpine and cyanamid 154806 caused a significantly higher (p < 0.05) increase (of up to 400% of baseline) of the extracellular hippocampal DA and 5-HT levels (Fig. 4c; 100–120 min). Cyanamid 154806 was not able to block the pilocarpine-induced Glu increase (Fig. 4c; 100–120 min). Compared with the pilocarpine control experiments (Fig. 3; 20–140 min), cyanamid 154806 had no significant effect (p > 0.05) on the pilocarpine-induced GABA increase (Fig. 4c; 140–220 min).

Group 4: effect of cyanamid 154806 and candesartan on the anti-convulsant effect of Ang IV and the effect of cyanamid 154806 on the anti-convulsant effect of somatostatin-14. Co-infusion with Ang IV at a rate of either 2.5 or 25 pmol per hour of the sst2 receptor antagonist cyanamid 154806, and of candesartan (2.25 nmol per hour), did not abolish the anti-convulsant effect of Ang IV. The mean TSSS remained significantly (p < 0.05) lower compared with the pilocarpine control group, namely 3.0 ± 1.7 (n = 3), 7.0 ± 0.9 (n = 4) and 2.9 ± 0.5 (n = 9), respectively (Fig. 2b).

In contrast, cyanamid 154806 in a dose of 50 pmol per hour was able to abolish (p < 0.05), the anti-convulsant activity of both Ang IV (Fig. 2b) and somatostatin-14 (Fig. 2c), resulting in a mean TSSS of, respectively, 13.1 ± 1.3 (n = 14) and 12.2 ± 1.0 (n = 10), which are not significantly different from the pilocarpine control group (p > 0.05) (Fig. 2b).

In comparison to the Ang IV–pilocarpine and somatostatin-14–pilocarpine co-administration experiments (Figs 4a and b; 20–80 min), the pre-pilocarpine DA and 5-HT increase was blocked (p < 0.05) by co-infusion with cyanamid 154806 in a dose of 50 pmol per hour (Figs 5a and b; 20–80 min). However, cyanamid 154806 did not alter the effect on the DA and 5-HT levels during pilocarpine administration (Figs 5a and b; 100–120 min) compared with the Ang IV–pilocarpine and somatostatin-14–pilocarpine co-administration experiments (p > 0.05) (Figs 4a and b; 100–120 min).

Figure 5.

 Effect of an i. c.v. infusion of (a) Ang IV (10 nmol per hour) in combination with the somatostatin receptor 2 antagonist cyanamid 154806 (50 pmol per hour), (b) of somatostatin-14 (10 nmol per hour) in combination with cyanamid 154806 (50 pmol per hour) and of (c) Ang IV (10 nmol per hour) in combination with candesartan (2.25 nmol per hour) on pilocarpine-induced changes in DA, 5-HT, Glu and GABA release in the hippocampus of the rat. Baseline levels are set to 100% (mean of six values). The values at all other points are expressed as a percentage of the baseline value. Each value is the mean ± SEM. The drugs were administered for 220 min and pilocarpine for 40 min (shown by the horizontal bars). Data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values significantly different from the corresponding baseline value (*p < 0.05).

Cyanamid 154806 (50 pmol per hour) neither changed the effect of Ang IV (p > 0.05) (Fig. 4a; 20–80 min) nor the effect of the co-administration of either Ang IV or somatostatin-14 and pilocarpine (p > 0.05) (Figs 4a and b; 140–220 min) on the hippocampal GABA concentration (Figs 5a and b; 20–220 min).

The i.c.v. co-administration of either Ang IV or somatostatin-14 and cyanamid 154806 caused no effect on the hippocampal Glu concentration (Figs 5a and b; 20–80 min). Similar to the pilocarpine control experiments (Fig. 3; 20–40 min), but in contrast to the effect of either Ang IV or somatostatin-14 combined with pilocarpine (Figs 4a and b; 100–120 min), co-infusion of either Ang IV or somatostatin-14 and cyanamid 154806 combined with intrahippocampal pilocarpine administration caused a significant (p < 0.001) increase of the hippocampal Glu concentration (Figs 5a and b; 100–120 min).

Co-infusion with candesartan (2.25 nmol per hour) neither changed the effect of Ang IV (Fig. 5c; 20–80 min) (p > 0.05) nor the effect of co-administration of Ang IV and pilocarpine (Fig. 5c; 100–120 min) (p > 0.05) (Fig. 4a) on the hippocampal DA, 5-HT, GABA and Glu concentration (Fig. 5c).

Experimental group with a non-convulsant dose of pilocarpine

Group 5: effect of i.c.v.-administered somatostatin-14 on the hippocampal perfusion of pilocarpine (100 µm). In the experiments in which either Ang IV or somatostatin-14 were combined with pilocarpine, the hippocampal DA and 5-HT levels returned to baseline level after the cessation of the pilocarpine perfusion, although the peptide was further administered. Possibly this is caused by the elevation of the GABA concentration in the hippocampus, as a result of the perfusion of pilocarpine in a high concentration (10 mm). Therefore, we combined pilocarpine in a lower concentration (100 µm) with somatostatin-14. This combination had a similar effect on the extracellular hippocampal DA, 5-HT, Glu and GABA levels (Fig. 6; 20–220 min) as compared with the experiments administering somatostatin-14 alone (Fig. 1c; 20–220 min) (p > 0.05). In contrast to the somatostatin-14-pilocarpine (10 mm) co-administration experiments (Fig. 1c; 100–120 min), pilocarpine in a concentration of 100 µm did not cause an increase of the hippocampal Glu and GABA levels during pilocarpine administration (Fig. 6; 100–120 min) (p < 0.05).

Figure 6.

 Effect of the co-administration of somatostatin-14 (10 nmol per hour) and pilocarpine (100 µm) on the DA, 5-HT, Glu and GABA release in the hippocampus of the rat. Baseline levels are set to 100% (mean of six values). The values at all other points are expressed as a percentage of the baseline value. Each value is the mean ± SEM. Somatostatin-14 is administered for 220 min and pilocarpine (100 µm) for 40 min (shown by the horizontal bars). Data were analysed by anova for repeated measures followed by a Wilcoxon test. Asterisks denote the values significantly different from the corresponding baseline value (*p < 0.05).

Hippocampal perfusion with pilocarpine in a subthreshold concentration (100 µm) did not evoke seizures and caused no significant effect on hippocampal neurotransmitter levels (data not shown).

Discussion

The inhibition of pilocarpine-induced seizures by Ang IV: involvement of the somatostatin receptor 2

This study shows that i.c.v.-administered Ang IV in a dose of 10 nmol per hour protects rats against pilocarpine-induced seizures. The sst2 receptor antagonist cyanamid 154806 in a dose of 25 pmol per hour partially inhibited this effect of Ang IV. The anti-convulsant properties of Ang IV were completely blocked by cyanamid 154806 in a dose of 50 pmol per hour. Lew et al. (2003) showed that Ang IV is a potent and competitive inhibitor of IRAP, which was shown to cleave somatostatin-14 in vitro into the inactive de-[Ala-Gly]-somatostatin-14 fragment (Matsumoto et al. 2000). Moreover, the IRAP protein is abundantly present in the central nervous system (Albiston et al. 2001; Fernando et al. 2005). Taken together, this could suggest that an i.c.v. administration of Ang IV leads to an increase of somatostatin-14 in the brain, which could then be responsible for the inhibition of pilocarpine-induced seizures. Indeed, the present study shows that, similar to Ang IV, somatostatin-14 protects rats against pilocarpine-induced seizures. Previously, it has already been shown that intracerebral injections of somatostatin-14 counteracts seizures in other animal models of epilepsy (Mazarati and Teledgy 1992; Monno et al. 1993; Vezzani and Hoyer 1999), probably via the activation of postsynaptic K+ currents in hippocampal pyramidal neurones, which then become hyperpolarized (Schweitzer et al. 1998). In rats, this appears to be mediated by sst2 receptors (Tallent and Siggins 1999), which is in agreement with our data. Indeed, the sst2 receptor antagonist cyanamid 154806 blocked the ability of somatostatin-14 to inhibit pilocarpine-induced seizures. To our knowledge, this is the first in vivo study in which an effect of Ang IV can be blocked with an sst2 antagonist.

Next to the degradation of somatostatin-14, IRAP also cleaves vasopressin, substance P, oxytocin and Met-enkephalin in vitro (Herbst et al. 1997; Matsumoto et al. 2000, 2001; Albiston et al. 2001). However, it is not likely that Ang IV exerts its anti-convulsant properties via an increase of one of these peptides. Indeed, vasopressin (Croiset and De Wied 1997), substance P (Liu et al. 1999), Met-enkephalin (Tanaka et al. 1989) and oxytocin (Leventhal and Reid 1968; Abouleish 1976) have proconvulsive properties.

The in vivo substrates of IRAP are not yet known. It can therefore not be excluded that other neuropeptides, which can be cleaved by IRAP, are also involved in the ability of Ang IV to block pilocarpine-induced seizures. Nevertheless, it seems clear that somatostatin-14 is involved in the anti-convulsant effect of Ang IV, as this can be reversed with the sst2 antagonist cyanamid 154806.

In our study, i.c.v. administration of Ang IV but not of somatostatin-14 caused characteristic drinking behaviour, which is associated with AT1 receptor activation in the lamina terminalis, the anteroventral third ventricle region and the periventricular preoptic nuclei in the brain (Song et al. 1992; Lenkei et al. 1997; Allen et al. 2000). It has been observed that pre-activated human AT1 receptors display a high affinity for Ang IV and can be fully activated by this peptide (Le et al. 2002). However, the occurrence of such a ‘pre-activated’ receptor conformation could not be revealed in in vivo situations. Moreover, in the hippocampus, there is only a low level of expression of AT1 receptors (Song et al. 1992; Lenkei et al. 1997). We excluded possible effects caused by the binding of Ang IV to AT1 receptors by co-administering Ang IV and the AT1 receptor antagonist candesartan in a dose of 2.25 nmol per hour (Ki = 50 pm) (Fierens et al. 2001; Konishi et al. 2002). Candesartan was unable to block the anti-convulsant effect of Ang IV, the Ang IV-induced neurotransmitter changes in hippocampus nor the Ang IV-induced drinking behaviour. This suggests that AT1 receptors do not play a role in the anti-convulsant effect of Ang IV, and that the drinking behaviour is caused via another mechanism than the activation of AT1 receptors. Indeed, a high density of IRAP has been shown in brain regions involved in drinking behaviour (Fernando et al. 2005). As previously mentioned, inhibition of IRAP can cause an increase of the vasopressin concentration (Herbst et al. 1997; Matsumoto et al. 2000, 2001; Albiston et al. 2001). Taken together with the dipsogenic effect caused by i.c.v.-administered vasopressin (Szczepanska-Sadowska et al. 1982), we hypothesize that centrally administered Ang IV causes drinking behaviour via the inhibition of IRAP in brain regions that control water intake.

Role of DA and 5-HT in the anti-convulsant effect of Ang IV

In order to correlate the neurochemical changes to the anti-convulsant properties of Ang IV, we monitored neurotransmitter alterations in the hippocampus before and during seizures via in vivo microdialysis. The hippocampus is one of the most vulnerable brain areas for epilepsy related brain damage and plays a major role in the development and maintenance of limbic seizures. The i.c.v.-administered Ang IV (10 nmol per hour) caused an increase in the extracellular DA and 5-HT concentration in the hippocampus. Although the dopaminergic innervation of the hippocampus is sparse, it has been found that DA is a functionally important neurotransmitter in the hippocampus and that DA receptors are widely distributed (Starr 1996). Behavioural and electrophysiological studies showed that the endogenous hippocampal DA system is involved in curbing neuronal hyperexcitability. Hippocampal DA exerts an inhibitory effect on epileptogenesis, mediated via DA D2 receptors (Suppes et al. 1985; Alam and Starr 1993). It has been suggested that DA prevents epileptic discharges in the hippocampus and limits propagation through the basal ganglia at the level of the limbicostriatal interface (Starr 1996). 5-HT can elicit an inhibitory response in the hippocampus, mediated via 5-HT1A receptors (Colino and Halliwel 1987; Salgado and Alkadhi 1995). 5-HT is able to limit the development of seizures and inhibit the propagation throughout the brain (Zhang et al. 1994; Lu and Gean 1998; Schmitz et al. 1998; Hernandez et al. 2002). Many anti-epileptic drugs cause significant increases in DA and 5-HT in the hippocampus, which are considered to be important contributors to the overall anti-convulsant effect (Clinckers et al. 2005b).

We previously showed that increased hippocampal DA and 5-HT can protect rats from pilocarpine-induced seizures as long as the hippocampal DA and 5-HT levels are elevated by 70–400% for DA and 80–350% for 5-HT (Clinckers et al. 2004). After the i.c.v. administration of Ang IV (10 nmol per hour) the hippocampal monoamine levels were elevated within that range. Well-known anti-epileptic drugs like sodium valproate, carbamazepine, lamotrigine and zonisamide induce similar monoamine increases (Biggs et al. 1992; Yan et al. 1992; Baf et al. 1994a,b; Dailey et al. 1996, 1997a,b; Smolders et al. 1997a; Graumlich et al. 1999; Murakami et al. 2001; Clinckers et al. 2005b). Surprisingly, in the control experiments with pilocarpine, the hippocampal DA and 5-HT concentrations also increased in the anti-convulsant range. However, after the administration of pilocarpine, the monoamine concentrations returned to baseline level and the animals suffered from full-blown convulsions. We showed that only before the administration of pilocarpine, increased levels of hippocampal DA and 5-HT can effectively protect against seizures (Clinckers et al. 2004, 2005a). Possibly, this increase triggers an inhibitory response via D2 and 5-HT1A receptors, which makes the brain less prone to the development of convulsions.

The i.c.v. co-administration of the sst2 receptor antagonist cyanamid 154806 (50 pmol per hour) blocked the Ang IV-induced DA and 5-HT increase. Moreover, i.c.v.-administered somatostatin-14 (10 nmol per hour) also caused an increase in the hippocampal DA and 5-HT levels, which was also sst2-receptor dependent. These results provide further support for our hypothesis that the effect of Ang IV is mediated via the inhibition of IRAP and a consequent elevation of somatostatin-14 in the brain, which then causes this effect on the hippocampal monoamine release by acting on sst2 receptors. It is noticeable that the effect of somatostatin-14 on hippocampal DA and 5-HT release is somewhat more pronounced than the effect of Ang IV. This could be explained by the fact that somatostatin-14 is the mediator of this effect, whereas Ang IV only causes an increase of somatostatin-14, which is a more indirect and thus possibly a less potent way to cause this effect. On the other hand, the effect caused by Ang IV is longer lasting than the effect of somatostatin-14. This is possibly caused by the fact that Ang IV was bound to IRAP even after the infusion had already stopped, thereby inhibiting the enzyme and producing a quantity of somatostatin-14, which then caused its prolonged effect.

The i.c.v.-administered Ang IV and somatostatin-14 caused an increase of the hippocampal DA and 5-HT concentration, which at least remained until the cessation of the peptide administration. In contrast, in the Ang IV-pilocarpine and somatostatin-14-pilocarpine co-administration experiments, the DA and 5-HT levels returned to baseline after hippocampal perfusion of pilocarpine, even while the i.c.v. infusion of either Ang IV or somatostatin-14 continued. This discrepancy is possibly the consequence of an inhibition of the hippocampal DA and 5-HT release by an elevated GABA concentration in the hippocampus. Indeed, it is already known that intrahippocampal perfusion with vigabatrin, which causes an increase of the hippocampal GABA levels, leads to a decrease of the hippocampal DA concentration (Smolders et al. 1997a). In our study, the hippocampal perfusion of pilocarpine in a concentration of 10 mm caused an increase of the hippocampal GABA levels, probably as a consequence of seizures. In order to clarify this, we performed experiments with pilocarpine in a non-convulsant concentration (100 µm). The co-administration of pilocarpine in a concentration of 100 µm and somatostatin-14 caused no effect on the hippocampal GABA concentration. Moreover, the somatostatin-14-induced DA and 5-HT increases remained until the end of the somatostatin-14 infusion. These results seem to confirm our hypothesis that pilocarpine in a concentration of 10 mm abolished the effect of somatostatin-14 and Ang IV on DA and 5-HT release, possibly via an increase of the hippocampal GABA concentration.

Hippocampal perfusion of pilocarpine in a concentration of 10 mm caused an increase of the DA and 5-HT concentration up to about 250%, whereas the combination of pilocarpine (10 mm) and either Ang IV or somatostatin-14 led to a significantly higher increase of the DA and 5-HT levels in the hippocampus. These data point to an interaction between the somatostatinergic and cholinergic system, resulting in a potentiation of the effect on the release of monoamines in the hippocampus. It has already been shown that somatostatin-14 can enhance acetylcholine-induced excitations in the hippocampus (Mancillas et al. 1986). In our study, this interaction seems independent from sst2 receptor activation, as the co-infusion with cyanamid 154806 did not abolish the effect of Ang IV and somatostatin-14 on the pilocarpine-induced DA and 5-HT increase. Cyanamid 154806 alone combined with pilocarpine (10 mm) also caused a higher increase of the hippocampal DA and 5-HT levels as compared with the pilocarpine (10 mm) control experiments. The reason for this discrepancy remains unclear.

Despite the fact that cyanamid 154806 did not block the increase in DA and 5-HT release induced by Ang IV and somatostatin-14 during pilocarpine administration, it is able to abolish the anti-convulsant effect of these peptides. This shows that the anti-convulsant effect of Ang IV and somatostatin-14 is not dependent on the increase of the pilocarpine-induced DA and 5-HT release. Instead, the Ang IV- and somatostatin-14-induced increase of DA and 5-HT before the administration of pilocarpine seems more important as this was blocked by cyanamid 154806. It has already been shown that only before pilocarpine administration, increased DA and 5-HT concentrations in the hippocampus can trigger an anti-convulsant effect (Clinckers et al. 2004, 2005a), probably by the initiation of inhibitory processes, which eventually curtail the excessive excitatory process caused by muscarinic stimulation.

Involvement of Glu and GABA in the anti-convulsant properties of Ang IV

Next to the monoamines, GABA and Glu are also implied to be involved in the occurrence of seizures (Bradford 1995). Increased GABA and Glu concentrations in the hippocampus are often used as markers for pilocarpine-induced convulsions (Smolders et al. 1997a). Pilocarpine (10 mm), locally administered in the hippocampus, caused an increase of the extracellular hippocampal Glu (250% of baseline) and GABA (220% of baseline) concentrations. Well-known anti-epileptic drugs such as carbamazepine, phenobarbital and lamotrigine are able to suppress these pilocarpine-induced increases of GABA and Glu (Smolders et al. 1997a,b,c). Ang IV and somatostatin-14 (10 nmol per hour i.c.v.) are able to completely abolish the pilocarpine-induced increase of Glu. This finding points to an interaction between somatostatin-14 and the hippocampal glutamate release. There is indeed compelling evidence for a close relationship between the somatostatinergic and glutamatergic system in the hippocampus, which might be crucial for the anti-convulsant effect of somatostatin-14 (Binaschi et al. 2003) and thus of Ang IV. On the other hand, because the i.c.v. administration of Ang IV and somatostatin-14 caused no effect on the baseline hippocampal Glu levels, it is also possible that the effect on the pilocarpine-induced glutamate release is caused by preventing seizures via other mechanisms rather than a direct interference with the baseline hippocampal glutamate release. The blocking of seizures indeed inhibits the neurochemical processes that accompany seizures and that possibly cause an elevation of the hippocampal Glu concentration. The abolishment of the pilocarpine-induced Glu increase by the co-administration of either Ang IV or somatostatin-14 is sst2-receptor dependent. Indeed, co-administration of Ang IV and cyanamid 154806 together with pilocarpine caused an increase of the hippocampal Glu levels.

As Ang IV inhibits pilocarpine-induced convulsions, and as a pilocarpine-induced increase in GABA concentration is used as a marker for convulsions, it seems surprising that Ang IV was unable to attenuate the elevation of the hippocampal GABA concentration following pilocarpine. The reason for this discrepancy is unknown and remains puzzling as Ang IV per se caused a decrease (50% of baseline) of the hippocampal GABA levels. On the other hand, it is clear that the Ang IV-induced GABA decrease in the hippocampus is not dependent on an interaction with the sst2 receptor. Indeed, i.c.v.-administered somatostatin-14 (1–10 nmol per hour) was unable to cause an effect on the hippocampal GABA levels. Moreover, i.c.v.-administered cyanamid 154806 (50 pmol per hour) could not block the Ang IV-induced GABA decrease. This argues that apparently not all Ang IV-induced neurochemical effects are a consequence of the release of somatostatin-14. Until now, the mechanism by which this effect on GABA is mediated was unkown.

Conclusion

The present in vivo microdialysis study shows the ability of i.c.v.-administered Ang IV to protect rats from limbic seizures in the pilocarpine model for focal epilepsy. Inhibition of IRAP, resulting in an elevated concentration of somatostatin-14 in the brain, is proposed as a possible anti-convulsant mechanism of action of Ang IV. Indeed, the anti-convulsant effect and the neurochemical changes that may contribute to the anti-convulsant action of Ang IV could be blocked with the somatostatin receptor-2 antagonist, cyanamid 154806. Moreover, somatostatin-14 itself also protects rats from pilocarpine-induced seizures. The ability of Ang IV and somatostatin-14 to inhibit pilocarpine-induced convulsions is partly attributed to an increase of the hippocampal DA and 5-HT concentration. In the current experimental set-up, the role of GABA and Glu in the anti-convulsant effect of Ang IV could not be revealed.

Taken together, this study underlines the importance of the development of non-peptidergic inhibitors of IRAP, which could penetrate the blood–brain barrier and could be important in the treatment of epilepsy.

Acknowledgements

BS is a research fellow (Nr. 13317) from the ‘Instituut voor de aanmoediging van innovatie door Wetenschap en Technologie in Vlaanderen’ (IWT). RC and IS are postdoctoral fellows of the ‘Fund for Scientific Research-Flanders’ (FWO-Vlaanderen), Belgium. AM and DDeB are both research assistants of the FWO-Vlaanderen. This work was also supported by research grants from the VUB and the FWO-Vlaanderen. The authors wish to acknowledge the excellent assistance of S. Van Lint, R. Berckmans, G. De Smet, C. De Rijck and R.-M. Geens.

Ancillary