The role of muscarinic acetylcholine receptor-mediated activation of extracellular signal-regulated kinase 1/2 in pilocarpine-induced seizures


Address correspondence and reprint requests to Allan I. Levey, Emory Center for Neurodegenerative Disease, Emory University, Whitehead Biomedical Research Building, 615 Michael Street, 5th Floor, Atlanta, Georgia 30322, USA. E-mail:


Pilocarpine-induced seizures are mediated by the M1 subtype of muscarinic acetylcholine receptor (mAChR), but little is known about the signaling mechanisms linking the receptor to seizures. The extracellular signal-regulated kinase (ERK) signaling cascade is activated by M1 mAChR and is elevated during status epilepticus. Yet, the role of ERK activation prior to seizure has not been evaluated. Here, we examine the role of pilocarpine-induced ERK activation in the induction of seizures in mice by pharmacological and behavioral approaches. We show that pilocarpine induces ERK activation prior to the induction of seizures by both western blot and immunocytochemistry with an antibody to phosphorylated ERK. In addition, we show that the ERK pathway inhibitor SL327 effectively blocks the pilocarpine-induced ERK activation. However, SL327 pretreatment has no effect on the initiation of seizures. In fact, animals treated with SL327 had higher seizure-related mortality than vehicle-treated animals, suggesting activated ERK may serve a protective role during seizures. In addition, ERK inhibition had no effect on the development of the long-term sequelae of status epilepticus (SE), including mossy fiber sprouting, neuronal death and spontaneous recurrent seizures.

Abbreviations used



extracellular signal-regulated kinase


muscarinic acetylcholine receptor


mitogen activated protein kinase


normal goat serum


phosphate buffer


phenylmethylsulfonyl fluoride




status epilepticus


spontaneous recurrent seizures.

Generalized limbic seizures and subsequent status epilepticus (SE) can be induced by the muscarinic acetylcholine receptor (mAChR) agonist pilocarpine (Turski et al. 1984). The pilocarpine seizure model is frequently used because it mimics several features of human epilepsy. Following pilocarpine-induced seizures, animals experience an approximately two week latency period followed by spontaneous recurrent seizures (SRS) (Cavalheiro et al. 1996; Arida et al. 1999). In the hippocampi of both pilocarpine-treated rodents and human epileptics, there is cell death in the dentate hilus and pyramidal neurons (Liu et al. 1994; Fountain 2000). There is also abnormal mossy fiber sprouting into the molecular layer of the dentate gyrus (Babb et al. 1991; Mello et al. 1993).

Though the pilocarpine model is used to evaluate anti-epileptics, the mechanisms of seizure induction are unclear. A recent study demonstrated that M1 is the mAChR subtype responsible for pilocarpine-induced seizures (Hamilton et al. 1997). Interestingly, M1 is also the mAChR subtype responsible for extracellular signal-regulated kinase 1/2 (ERK) activation in the hippocampus (Berkeley et al. 2001). ERK is a member of the mitogen-activated protein kinase (MAPK) family of kinases, which are activated by many cell-surface receptors. ERK is involved in many neuronal processes, including LTP formation (English and Sweatt 1997), neurite outgrowth (Lloyd and Wooten 1992), and immediate early gene expression (Vanhoutte et al. 1999). Given that ERK can modulate synaptic plasticity and neuronal excitability, it may also mediate seizure induction though this possibility has not been explored. Alternatively, M1 could initiate seizures via an ERK-independent pathway, and simultaneous ERK activation could play either a protective role or no role during seizure. In addition, the long-term sequelae of SE might also be modulated by ERK. As ERK can regulate anti-apoptotic signaling and neurite outgrowth, its activation may affect either the cell death or mossy fiber sprouting that follow SE.

Here, we test the hypothesis that mAChR activation of ERK mediates pilocarpine-induced seizures in mice. Consistent with this hypothesis, we predict that pilocarpine will cause ERK activation prior to, and independent of, seizure induction. Previous studies examined ERK activation during established SE, but little is known about ERK activation prior to seizure (Kim et al. 1994; Garrido et al. 1998). To evaluate the role of ERK in pilocarpine-induced seizures, we use SL327, an inhibitor of MEK, the kinase directly upstream of ERK. SL327 has been shown to decrease basal levels of ERK phosphorylation and several forms of hippocampus-dependent learning, but its ability to attenuate stimulated ERK activation in brain has not been evaluated (Atkins et al. 1998; Selcher et al. 1999). We show that pilocarpine causes ERK activation prior to seizure induction, and SL327 blocks this activation. Surprisingly, we show that ERK inhibition does not prevent pilocarpine-induced seizures, suggesting that other mAChR-mediated signaling pathways are responsible. In fact, seizures in SL327-treated animals appear more severe and result in higher mortality than in controls. Moreover, we demonstrate that ERK inhibition during initial seizures has no effect on long-term changes, including cell death, mossy fiber sprouting, and frequency of SRS.

Materials and methods

Drug treatments

All animal treatment and care followed the National Institutes of Health guidelines and was approved by the Animal Care and Use Committee of Emory University. Five to six week old C57Bl6 mice (Charles River Laboratories, Wilmington, MA) were injected i.p. with 325 mg/kg pilocarpine (Sigma, St Louis, MO) dissolved in 0.9% NaCl. Thirty min prior to pilocarpine, animals were injected i.p. with 1 mg/kg scopolamine methylnitrate (Sigma) to reduce peripheral effects of mAChR agonism. Animals were also injected i.p. with vehicle, 50 mg/kg SL327 (Dupont, Wilmington, DE) dissolved in DMSO or 10 mg/kg diazepam (Sigma) dissolved in saline with tween-80 30 min before pilocarpine. Animals were observed and seizures were noted. Three behavioral endpoints were recorded: at least one stage 5 seizure (indicated by bilateral clonus and loss of posture), status epilepticus (SE), and death. Comparisons between conditions were made by Chi-square analysis using SPSS software (SPSS, Inc., Chicago, IL, USA).


Fifteen minutes before time of perfusion, animals were injected i.p. with 100 i.u. heparin sulfate (Sigma) followed 15 min later by 200 mg/kg sodium pentobarbital. Animals were perfused transcardially by gravity with 0.005% sodium nitroprusside (30 s) followed by 4% paraformaldehyde in 0.1 m phosphate buffer (PB, 10 min). Brains were removed and sectioned coronally using a vibratome. Fifty µm sections were rinsed in PB, followed by several rinses in TBS. Sections were immersed in 0.3% H2O2 for 10 min at RT, rinsed then blocked in TBS containing 4% normal goat serum (NGS), 0.1% triton-X and 10 µg/mL avidin for 1 h at 4°C. Sections were rinsed twice and placed in TBS containing 2% NGS and the primary antibody (phospho-p42/44 MAPK, 1 : 1000, Cell Signaling Technology, Beverly MA; NeuN, 1 : 1000, Chemicon, Temecula, CA, USA) overnight at 4°C. The next day, slices were rinsed and placed in TBS containing 2% NGS and biotinylated secondary antibody (goat anti-rabbit or anti-mouse 1 : 200; Vector Laboratories, Inc., Burlingame, CA) for 1 h at 4°C. Slices were rinsed and incubated in horseradish peroxidase-conjugated avidin-biotin complex (ABC, Vector Laboratories, Inc.) for 1 h at 4°C. Slices were rinsed, developed using diaminobenzidine (Sigma) and mounted onto glass slides. For comparisons, all images were captured with a Hamamatsu Orca 100 digital camera and MetaVue software (Universal Imaging Corp., Downingtown, PA, USA) with identical settings and were processed identically in Adobe Pmotoshop. Cell death was determined by counting the number of NeuN positive neurons in an equally sized region placed in the middle of the hilus of the dentate gyrus in each image. 2–3 sections were counted per animal for 3 animals per condition. motoshops were performed with LSD post hoc test.


Cellular morphology was evaluated in animals perfused three days after seizure induction by counting the number of shrunken acidophilic nuclei per high-power field (HPF, 40×) in hematoxylin and eosin (H&E) stained sections. In CA1, 2 HPFs were counted per hippocampus and in the dentate hilus one HPF per hippocampus was counted. 3 sections per animal (6 hippocampi) were counted for 3 animals per condition. All counts were made by an examiner blinded to the treatment conditions. motoshops were performed and conditions were compared with LSD post hoc test.

Western blots

Following cervical dislocation, brains were removed and placed in ice-cold saline for 3–5 min. Hippocampi were dissected and placed in labeled tubes on dry ice then stored at −80°C until further use. Samples were homogenized by sonication in 150 µL homogenization buffer (in mM: 50 Tris-HCl, pH 7.5, 50 NaCl, 10 EGTA, 5 EDTA, 2 sodium pyrophosphate, 4 para-nitrophenylphosphate (pNPP), 1 sodium orthovanadate, 1 phenylmethylsulfonyl fluoride (PMSF), 20 µg/mL leupeptin, and 4 µg/mL aprotinin). Protein concentration was determined using BCA protein assay (Pierce, Rockford, IL, USA) Samples were prepared by addition of 2× sample buffer followed by heating to 95°C for 5 min. Twenty µg were loaded onto 12% polyacrylamide gels. Proteins were transferred electrophoretically to imobilon-p PVDF membranes (Millipore Co., Bedford, MA, USA). Blots were blocked 20 min in TBS + Tween 20 (TBST) with 5% dry milk. Milk was omitted for phospho-MAPK blots. Blots were then incubated overnight at 4°C in TBST ± milk and primary anti-body (phospho-p42/44 MAPK, 1 : 1000; total p42/44 MAPK, 1 : 500; Cell Signaling Technology). Blots were rinsed 3 times and incubated in TBST with milk and horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (1 : 10 000; Biorad, Hercules, CA). Blots were developed using Renaissance® Western Blot Chemiluminescence Reagent (PerkinElmer Life Sciences, Boston, MA, USA) and visualized using a Kodak Image Station 440CF. Blots were quantified using Kodak 1D software using net intensities for each band. For each experiment, parallel blots were probed for phospho-ERK and total ERK. Values for phospho-ERK were normalized for loading by dividing by those obtained for total ERK. Values obtained for the left and right hippocampus were averaged to obtain a single value for each animal. All experiments were repeated at least 3 times. motoshops were performed with LSD post hoc test.

Long-term behavioral observations

Seizures were induced as described above. Animals were allowed to recover from SE, which usually lasted 4–6 h. Animals were monitored closely for 72 h following SE induction for resumption of grooming behavior and food intake. All animals were given oral Gatorade and moistened food pellets on the cage floor. Dehydrated animals were given 1.5 mL i.p. injections of Ringer's Lactate as needed. Control animals for all long-term studies consisted of animals given the same combination of drugs (e.g. SL327 + pilocarpine), but which did not reach SE. Animals were grouped 3 per cage according to drug treatment and whether SE was induced. Beginning 16 days after initial seizure induction, animals were videotaped 6 h per day between 9 a.m. and 5 p.m. for 10 days. The total number of seizures per cage was recorded for each day. Conditions were compared using t-test for independent samples.

Neo-Timm's stain

On the 27th day after initial seizure induction, animals were perfused for Neo-Timm's staining. Animals were heparinized and anesthetized as described above. Animals were perfused transcardially with 0.1% sodium sulfide prepared in modified Millonig's buffer (0.1 motoshop NaH2PO4H2O, 0.5% sucrose, 0.002% CaCl2, pH 7.3 with 10 N NaOH) for 10 min, followed by 3% glutaraldehyde in 0.1 motoshop PB for 4 min. Brains were removed and postfixed and cryoprotected in 2% glutaraldehyde in 30% sucrose in 0.1 motoshop PB overnight. The next day, 40 µm sections were cut on a freezing microtome. Sections were mounted on glass slides and allowed to dry. Slides were developed for 1 h in the dark in Timm's stain: 80 mL 50% gum arabic, 10 mL citrate buffer (3.42 g citric acid and 3.15 g sodium citrate in H2O), 20 mL hydroquinone (1.24 g in H2O) and 0.666 mL silver nitrate (0.17 g in 1 mL H2O; added in dark). Slides were then rinsed in running tap water in dark for 5 min. Slides were dehydrated and coverslipped. 3 sections per animal (6 hippocampi) spanning the rostral to caudal axis were captured and quantified using MetaVue software. For each section, staining density was quantified by selecting 4 equally sized regions along the granule cell layer/molecular layer border: 2 each from the upper and lower leaflets of the dentate gyrus. An equally sized region from the outer portion of the molecular layer was used to calculate background. For each region, average pixel intensity was determined, and regions were averaged to obtain a value for each hippocampus. 3–6 animals were counted for each condition. motoshops were performed using LSD post hoc test.


Pilocarpine induces ERK activation independent of seizure induction

The hypothesis that an ERK-dependent signaling pathway mediates pilocarpine-induced seizures is predicated on the assumption that pilocarpine injections increase levels of activated ERK prior to and independent of seizure induction. We tested this assumption in several ways. First, we injected animals with 300 mg/kg pilocarpine, a subthreshold dose for seizure induction and examined the brains for activated ERK. Sixty min after pilocarpine injections, we observed increased neuronal phospho-ERK immunoreactivity in several limbic brain regions, including cingulate cortex, amygdala, piriform cortex and hippocampus (Fig. 1a). While this dose did not cause seizures, these brain regions are associated with seizure induction (Barone et al. 1993). Immunoreactivity was most prominent in the hippocampus, particularly in the granule cells of the dentate gyrus, and slightly less in the pyramidal neurons of CA1. There was relative sparing of pilocarpine-induced phospho-ERK immunoreactivity in CA3 pyramidal neurons. However, the surrounding neuropil regions corresponding to the mossy fiber terminals demonstrated substantial immunoreactivity without stimulation that increased with pilocarpine treatment. As ERK is activated by phosphorylation, an increase in immunoreactivity with a phospho-specific ERK antibody indicates an increase in activated ERK (Heasley et al. 1994; Schramm and Limbird 1999).

Figure 1.

Pilocarpine-induced ERK activation. (a) Mice were injected with 300 mg/kg pilocarpine or saline 1 h before perfusion. There is an increase in immunoreactivity with a phospho-ERK specific antibody in the cingulate cortex, the piriform cortex, the endopiriform nucleus, the amygdala, and the hippocampus. In the hippocampus, the activation is most prominent in the dentate gyrus and CA1. Cg, cingulum; End, Endopiriform nucleus; Pir, piriform cortex; CA, central nucleus of the amygdala; BLA, basolateral amygdala; LA, lateral amygdala; DG, dentate gyrus. Scale bar is 500 µm. (b) Western blot analysis shows a robust increase in phospho-ERK immunoreactivity in hippocampal homogenates (top blot) 15 min and 1 h after injection with 325 mg/kg pilocarpine. The bottom blot is a parallel blot that shows that there is no change in total ERK expression and that the lanes are loaded equally. (c) Western blot of hippocampal homogenates from mice pretreated with diazepam (10 mg/kg) 30 min before pilocarpine (325 mg/kg) or saline. 15 min and 1 h after pilocarpine injection, there is an increase in phospho-ERK immunoreactivity (top blot). Bottom blot shows total ERK.

Secondly, we examined whether pilocarpine-induced ERK activation preceded seizures in animals injected with 325 mg/kg pilocarpine, the dose used to induce seizures and SE. In animals killed 15 min after pilocarpine injections, a time point generally prior to seizure initiation, we observed a robust increase in ERK activation by Western blot analysis of the hippocampus (Figs 1b and 2c). In initial experiments, we examined several brain regions including hippocampus, frontal cortex and inferior temporal lobe containing the amygdala (not shown). We observed no differences in the patterns of ERK activation among the various regions examined, and so for this, and all future quantitative experiments by western blot, we examined only the hippocampus because of its central importance in seizure induction. While we observed no behavioral indications of seizure prior to being killed at 15 min, it is possible that seizures had begun prior to motor manifestations. To assess this possibility, we pretreated the animals with the anticonvulsant diazepam (10 mg/kg), a dose shown to prevent seizures in the pilocarpine model (Turski et al. 1984). In animals pretreated with diazepam, we still observed a robust increase in pilocarpine-induced ERK activation (Fig. 1c). However, basal levels of ERK phophorlyation varied greatly in diazepam-pretreated animals resulting in a non-statistically significant increase in phospho-ERK immunoreactivity with pilocarpine treatment (quantitation not shown). Combined, the data from these three sets of experiments indicate that pilocarpine activates ERK in brain independent of its ability to induce seizures. These results are consistent with those of previous studies demonstrating mAChR-induced ERK activation in PC12 cells (Berkeley and Levey 2000), hippocampal slices (Berkeley et al. 2001) and brain (Rosenblum et al. 2000).

Figure 2.

SL327 inhibits pilocarpine-induced ERK activation. (a) Western blot of hippocampal homogenates of mice treated pretreated with vehicle or SL327 (50 mg/kg) 30 min before pilocarpine (325 mg/kg) or saline injections. SL327 pretreatment inhibits the increase in phospho-ERK immunoreactivity induced by pilocarpine alone. (b) Immunocytochemistry of mice perfused 30 min after pilocarpine (325 mg/kg) or saline administration. Mice pretreated with SL327 (50 mg/kg) show markedly reduced pilocarpine-induced phospho-ERK immunoreactivity in the hippocampus. A few granule cells of the dentate gyrus remain immunoreactive. Immunoreactivity in the mossy fiber terminal region is reduced below basal levels in the SL327-pretreated animal. DG, dentate gyrus; Hi, dentate hilus; GC, granule cell layer; Mo, molecular layer. Scale bars are 500 µm for the left column and 100 µm for the right column. (c) Quantitation of western blots of hippocampal homogenates from animals killed 15 min and 2 h after pilocarpine (325 mg/kg) administration. SL327 (50 mg/kg) pretreatment blocks ERK activation at both 15 min and 2 h. By 2 h, phospho-ERK levels in animals injected with pilocarpine, but who did not go into SE, returned to basal levels. All animals received pilocarpine. Values are fold increase over phospho-ERK immunoreactivity in control (no pilocarpine) animals. 15 min, vehicle: 1.76 + 0.34 (mean + SEM., n = 4). 15 min, SL327: 0.96 + 0.13 (n = 3). 2 h, vehicle: 1.42 + 0.11 (n = 5). 2 h, SL327: 0.94 + 0.18 (n = 4). 2 h, vehicle no SE 0.69 + 0.09 (n = 2) *p < 0.05 compared animals treated with pilocarpine (and vehicle).

SL327 prevents pilocarpine-induced ERK activation

SL327 is an inhibitor of MEK, the kinase directly upstream of ERK and responsible for its activation, that penetrates the blood–brain barrier. Previous studies have demonstrated that SL327 is specific; it has an IC50 for MEK inhibition of 0.27 µmotoshop, and has no effect on PKC, αCaMKII or PKA at concentrations up to 10 µmotoshopin vitro (Atkins et al. 1998; Selcher et al. 1999). Pretreatment of animals with 50 mg/kg SL327 prevents pilocarpine-induced ERK activation in the hippocampus 15 min after pilocarpine injection (325 mg/kg), as determined by Western blot (Fig. 2a), and 30 min after injection as determined by immunohistochemistry (Fig. 2b). By immunocytochemistry, SL327′s effect on ERK phosphorylation appears more dramatic than by western blot. This is likely due to the fact that processing tissue for immunoblotting increases basal/background phospho-ERK immunoreactivity making the changes appear less significant. As with the lower dose of pilocarpine, ERK activation in the hippocampus is most prominent in the dentate gyrus and CA1. Activation in these regions occurs in both cell bodies (including nuclei) and dendrites. SL327 pretreatment nearly abolishes ERK activation throughout the hippocampus, except for a few granule cells in the dentate gyrus and some faint staining of hilar interneurons. Interestingly, SL327 reduced phospho-ERK immunoreactivity below basal levels in the mossy fiber terminal region near CA3. Previous studies have shown that ERK activation remains elevated for the duration that an animal is in SE (Kim et al. 1994; Garrido et al. 1998). However, this sustained elevation is most likely due to secondary release of other neurotransmitters, such as glutamate. Two hours after pilocarpine injection, ERK activation returns to basal in those animals whose seizures do not progress to SE, though it remains elevated in those who are in SE (Fig. 2c). In addition, at 2 h after injection, SL327 continues to suppress ERK activation. Combined, these data indicate that SL327 blocks the entire phase of pilocarpine-mediated increase of ERK activation and it continues to inhibit SE-induced ERK activation as well.

ERK inhibition increases seizure severity and seizure-associated mortality

In animals injected with SL327 30 min prior to pilocarpine administration, we observed an overall increase in seizure severity. We analyzed seizure severity using three distinct end points (Fig. 3). First, the percentage of animals that had at least one stage 5 motor limbic seizure, defined as bilateral clonus with a loss of postural stability, was roughly the same in SL327- and vehicle-treated animals. The percentage of SL327- and vehicle-treated animals whose seizures progressed to SE was also statistically equivalent, though there was a trend towards more SE in SL327-treated animals. Finally, greater than 6 times the number of SL327-treated animals died as a result of their seizures than vehicle treated animals. SL327 alone was not fatal, as no animals treated with SL327 followed by saline injection died. In addition, no animals given both SL327 and pilocarpine died other than during a seizure. Often, death in the SL327-treated animals was immediately preceded by dramatic seizures that sometimes included running and hopping. These violent seizures in the SL327-treated animals were often the first seizure, suggesting that they bypassed the earlier less severe stages in the progression towards SE. Vehicle treated animals usually exhibited less severe seizures (e.g. head bobbing, single limb clonus) prior to their first stage 5 seizure. The running and hopping behaviors were relatively rare in vehicle-treated animals and when they did occur, they were less frequently fatal. Interestingly, preliminary experiments examining EEG recordings from SL327- and vehicle-treated animals did not reveal any obvious differences between the conditions (data not shown). Overall, these data demonstrate that SL327 pre-treatment results in a more severe seizure phenotype with higher mortality, suggesting that ERK might play a protective role during seizures.

Figure 3.

SL327 pretreatment increases seizure-associated mortality. Mice were pretreated with vehicle or SL327 (50 mg/kg) 30 min before pilocarpine treatment (325 mg/kg). Seizure was defined as at least one stage 5 seizure. Animals were observed for 2 h after pilocarpine injection. For each end-point, seizure behavior was more severe in SL327 treated mice, but only for death was the difference statistically significant. ≥ 1 seizure: vehicle = 91%, SL327 = 96%. χ2 (1) = 0.46. SE vehicle = 73%, SL327 = 87.5%. χ2 (1) = 1.59. Death: vehicle = 9%, SL327 = 58%. χ2 (1) = 12.27. ***p < 0.001.

ERK inhibition during initial seizures does not alter long-term sequelae of SE

In animals that survive SE, there are several well-described long-term pathologic sequelae. These include the death of hilar interneurons followed by the pyramidal neurons of CA3 and CA1, mossy fiber sprouting and spontaneous recurrent seizures. We tested whether inhibiting ERK during the initial induction of seizures and SE would alter these downstream effects. To examine cell death, we perfused animals 3 days after SE and examined immunoreactivity with the neuronal nuclear marker NeuN. In all animals in which SE was induced, there was a dramatic decrease in the number of hilar interneurons (Figs 4a and b). There were also decreases in immunoreactivity in the CA3 pyramidal layer, and less dramatically in CA1 pyramidal layer. However, there was no difference in the loss of hilar interneurons between animals that were pretreated with vehicle or SL327.

Figure 4.

SL327 does not alter cell death following SE. (a) Animals were perfused 3 days after injection with pilocarpine (325 mg/kg) preceded by vehicle or SL327 (50 mg/kg) administration. Immunocytochemistry with the neuronal nuclear marker NeuN shows a reduction in hilar interneurons in animals who exhibited SE, though there is no difference between those that received vehicle or SL327. Scale bar is 100 µm. (b) Quantitation of hilar interneurons in SL327- and vehicle-treated animals. NeuN immunoreactive hilar interneurons were counted in an equally sized region in 2–3 sections (4–6 hippocampal sections) each for 3 animals per condition. Vehicle control: 11.31 ± 0.61 (mean ± SEM., n = 3 mice). SL327 control: 12.13 ± 1.12 (n = 3 mice). Vehicle SE 2.72 ± 1.15 (n = 3 mice). SL327 SE 3.11 ± 0.45 (n = 3 mice). ***p < 0.001 compared to either non-SE control. (c) Hematoxylin and eosin staining of sections from the same animals as used in (a) shows a loss of normal cellular morphology and an increase in shrunken acidophilic nuclei in CA1 of animals who exibited SE. There is no apparent difference between vehicle- and SL327-treated animals. Scale bar is 50 µm. D. Quantitation of shrunken acidophilic nuclei in CA1. The number of acidophilic nuclei was counted in 2 high power fields (HPF, 40x) containing CA1 per hippocampus in 3 sections (6 hippocampi) per animal. Three animals were examined for each condition. Vehicle control: 0.80 ± 0.17 (mean ± SEM., n = 3 mice). SL327 control: 0.77 ± 0.15 (n = 3 mice). Vehicle SE 72.90 ± 7.43 (n = 3 mice). SL327 SE 75.83 ± 15.46 (n = 3 mice). ***p < 0.001 compared to either non-SE control.

NeuN staining reveals only remaining neurons at the time point examined (three days). To obtain an impression of ongoing cell death 3 days after seizure induction, we examined H&E stained sections and counted shrunken acidophilic nuclei representing dead or dying cells. In the dentate hilus, there are significantly more acidophilic neurons in the SE animals compared to the controls, though pretreatment with vehicle or SL327 has no effect in the number of dead or dying cells. [Vehicle control: 0.18 ± 0.10 (mean ± SEM., n = 3 mice). SL327 control: 0.11 ± 0.11 (n = 3 mice). Vehicle SE 7.70 ± 0.76 (n = 3 mice). SL327 SE 7.50 ± 0.17 (n = 3 mice). p < 0.001 for both SE conditions compared to either control.] Generally, there are relatively few acidophilic cells in the hilus in any condition. This is likely because the cell death following SE occurs earliest in the hilus and by 3 days, most of the dead cells have been cleared. In CA1, NeuN staining shows little change in animals in which SE was induced (not shown). However, CA1 shows the most dramatic increase in number of shrunken acidophilic neurons, but again SL327-pretreatment has no effect (Figs 4c and d). The large number of acidophilic cells in CA1 indicates that cell death is actively occurring in this region three days following SE induction. As a whole, these data indicate that there is no difference between vehicle and SL327-treated animals in either ongoing cell death or the number of remaining neurons three days following seizure induction.

To examine mossy fiber sprouting, 27 days after initial injections we perfused animals that survived the initial induction of SE ± SL327 pretreatment and performed neo-Timm's staining to detect the mossy fibers. Interestingly, there was a slight, but statistically insignificant, increase in mossy fiber density in SL327 treated animals that did not go into SE. We observed a significant increase in mossy fibers in the granule cell and inner molecular layers in all SE animals, but no difference in sprouting between the SL327- and vehicle-treated animals (Fig. 5).

Figure 5.

SL327 does not alter mossy fiber sprouting following SE. (a) Neo-Timm's staining of dentate gyrus of animals perfused 27 days after pilocarpine (325 mg/kg) preceded by vehicle or SL327 (50 mg/kg) administration. All animals who exhibited SE show mossy fiber sprouting compared to controls who received the same injections but did not show SE. Scale bar is 100 µm. (b) Quantitation of Timm's staining at the granule cell/molecular layer border (see methods). 3 sections (6 hippocampal sections) each for 3–6 animals were quantified. Values are arbitrary units of pixel intensity. Control, vehicle: 74.16 ± 24.97 (mean ± SEM, n = 3 mice). Control, SL327: 123.65 ± 11.34 (n = 6 mice). SE, vehicle: 253.92 ± 36.46 (n = 6 mice). SE, SL327: 251.37 ± 29.46 (n = 6 mice). **p < 0.01 compared to either non-SE control.

Similarly, we observed no difference in the frequency of spontaneous recurrent seizures between SE animals that received pretreatment with SL327 or vehicle. The average number of spontaneous recurrent seizures per 6 h observation period in animals that received either vehicle or SL327 (50 mg/kg) prior to induction of SE with pilocarpine (325 mg/kg) was determined by observing the animals from day 16 to day 26 after induction of SE (Vehicle: 0.28 ± 0.07 [mean ± SEM, n = 6 mice]; SL327: 0.27 ± 0.06 [n = 6 mice]).


The pilocarpine model of epilepsy is frequently used to study seizures and the efficacy of anti-epileptic drugs, but relatively little is known about the mechanisms by which mAChR agonism causes seizures. Here, we examined the role of ERK activation by pilocarpine in seizure induction and in the development of other pathological features of the pilocarpine model. We hypothesized that ERK may play a role in the signaling pathway between mAChR and seizure because it is activated by the M1 mAChR subtype (Berkeley et al. 2001), the same subtype involved in initiating pilocarpine seizures (Hamilton et al. 1997). In addition, ERK has been shown to phosphorylate Kv4.2 (Adams et al. 2000), an A-type potassium channel that plays a role in regulating neuronal excitability and whose expression in hippocampus is altered by seizures (Francis et al. 1997). Our data, however, do not support a role for ERK activation in initiation of pilocarpine-induced seizures. On the contrary, our results suggest that ERK may play a protective role during seizures.

Here, we show that pilocarpine injections increase ERK activation prior to the initiation of seizures, allowing for the possibility that it plays a role in the signal transduction cascade between M1 mAChR activation and seizures. Consistent with this possibility, we observe the earliest and most dramatic increase in ERK activation in the hippocampus, specifically the dentate gyrus, a region critical for the initiation of pilocarpine-induced seizures (Turski et al. 1989). However, we also show that SL327 effectively blocks pilocarpine-induced ERK activation, yet does not prevent seizures. These data indicate that ERK activation is not necessary for M1 mAChR to initiate seizures. By what other mechanisms might M1 initiate seizures? In sympathetic ganglia (Hamilton et al. 1997) but not in hippocampal pyramidal neurons (Rouse et al. 2000), M1 has been shown to suppress the M-current, a tonically active voltage-dependent potassium current important in regulating neuronal excitability. The role of M1 in regulating the M-current in dentate granule cells remains unclear, but M-current suppression could participate in increasing the firing rate of these cells to initiate seizures. Alternatively, M1 mAChR have been shown to potentiate NMDA currents through a PKC-dependent mechanism (Calabresi et al. 1998; Marino et al. 1998). While the signaling cascades between M1 and regulation of either the M-current or NMDA channels are not fully defined and could involve ERK, they may also be ERK-independent.

While ERK activation may not play a role in seizure induction, ERK is clearly activated by both pilocarpine and by seizure activity itself. Our data suggest that this ERK activation may play a protective role during seizures given that when ERK activation is inhibited by SL327 pretreatment, mortality during seizures increases approximately 6 fold. However, we cannot exclude the possibility that the increased mortality with SL327 pretreatment is not due exclusively to the inhibitory effects of SL327 on ERK activation. The selectivity of SL327 for MEK has been demonstrated in vitro (Selcher et al. 1999), yet little is known about its selectivity in vivo. However, SL327 is a structural analog of the highly specific MEK inhibitor U0126, suggesting that it shares a similar activity and specificity profile (Atkins et al. 1998; Favata et al. 1998).

How might ERK be protective during seizures? One possibility is through phosphorylation of Kv4.2. As a fast-inactivating A-type potassium channel localized postsynaptically in dentate granule cell and pyramidal neuron cell bodies and dendrites, Kv4.2 channels are in a prime location to regulate action potential firing frequency, spike repolarization and integration of signals received on dendrites (Hoffman et al. 1997; Martina et al. 1998; Serodio and Rudy 1998). In demonstrating that Kv4.2 is differentially localized in the hippocampus depending on its phosphorylation state, Varga et al. suggested that ERK phosphorylation of Kv4.2 may target it to subcellular regions where it is needed (Varga et al. 2000). Hence, when ERK activation is blocked, Kv4.2 channels are unable to localize to areas of hyperactivity to perform their normal functions of spike repolarization and limiting action potential firing, resulting in uncontrolled hyperexcitability and cell firing, eventually leading to lethal seizures. Supporting this role for Kv4.2 in seizures is the fact that heterotopic cells that lack functional Kv4.2 channels display hyperexcitable firing, and animals with these cells have a markedly decreased seizure threshold (Castro et al. 2001). In addition, 24 h after SE Kv4.2 RNA levels are elevated suggesting a protective response to seizure (Francis et al. 1997). Thus, ERK phosphorylation of Kv4.2 may represent a protective mechanism by which hyperexcitability in hippocampal neurons is limited. Figure 6 is a proposed model for the role of ERK in pilocarpine-induced seizures, and summarizes other potential mechanisms by which mAChR activation leads to seizure initiation.

Figure 6.

Model for the role of ERK activation in pilocarpine-induced seizures. (a) Inhibition of ERK activation with the MEK inhibitor SL327 does not prevent mAChR-induced seizures, which are mediated through alternative signaling pathways. (b) Activated ERK may reduce the severity of seizures by phosphorylating Kv4.2 channels and thereby reducing neuronal excitability. Mechanisms by which mAChR cause seizures may include potentiation of NMDA currents and suppression of the M-current, both of which result in increased neuronal excitability.

The brains of animals that survive SE display two major pathological changes: cell death and mossy fiber sprouting. Because of ERK's integral role in similar functions, we tested whether ERK inhibition during seizure-induction would alter these sequelae of SE. We first examined cell death, as ERK plays known roles in anti-apoptotic signaling pathways (Xia et al. 1995; Anderson and Tolkovsky 1999; Hetman et al. 1999). On the other hand, ERK inhibition protects against apoptotic cell death in a culture-based seizure model (Murray et al. 1998), demonstrating a pro-apoptotic effect of ERK activation. In our study, however, we saw no change in the amount of cell death in vehicle- or SL327-treated animals, indicating that blocking ERK activation during seizure induction neither abrogates nor increases cell death and brain damage following SE. There are several explanations for the difference in results between the in vitro and in vivo seizure models. First, seizures in brain may initiate other signals that compensate for the absence of activated ERK and lead to neuronal death. The cell culture model may not fully mimic seizures in terms of the pathways activated. Some of these other pathways may include other MAPKs, such as p38 and JNK, both of which have been implicated in apoptotic signaling. It would be interesting to examine whether inhibitors of these kinases affect cell death in seizure models, as both pathways are activated during seizures, though more transiently than ERK (Mielke et al. 1999; Jeon et al. 2000). Secondly, we have demonstrated that SL327 inhibits pilocarpine-induced activation of ERK, but have not examined ERK activation during the entire course of SE. It is possible that ERK activation would need to be inhibited for a much longer period to alter the cell death that occurs maximally 8–24 h after SE (Covolan and Mello 2000).

We observed a similar lack of effect of SL327 on mossy fiber sprouting. One of the earliest defined functions for ERK in neuronal systems was that it promoted differentiation, visible as neurite or process extension (Lloyd and Wooten 1992; Qui and Green 1992). More recently, as ERK's role in synaptic plasticity has been elucidated, its role in synapse rearrangement, including new spine formation and stabilization, has been demonstrated (Wu et al. 2001). Mossy fiber sprouting encompasses both of these phenomena: outgrowth of new processes and the creation of new synapses (Tauck and Nadler 1985; Wuarin and Dudek 1996; Ribak et al. 2000). Regardless, inhibition of ERK does not appear to affect mossy fiber sprouting following SE Again, this is probably due to the fact that mossy fiber outgrowth occurs for weeks after SE (Mello et al. 1993) and we inhibited ERK activation only during the initial insult. Given that we saw no effect of SL327 on either cell death or mossy fiber outgrowth, it is not surprising that we also saw no change in the frequency of spontaneous recurrent seizures (SRS). SRS are thought to occur as a result of cell loss and, more controversially, by the creation of recurrent excitatory circuits by mossy fiber sprouting (Tauck and Nadler 1985; Wuarin and Dudek 1996). See also (Longo and Mello 1997, 1998).

In conclusion, our data show a protective effect of ERK during pilocarpine-induced seizures but no long-term effect of ERK inhibition on animals that survive SE. These results indicate that M1 mAChR activation initiates seizures through a non-ERK-dependent pathway, but that the simultaneous activation of ERK is beneficial. Long-term inhibition of ERK may provide some benefit following seizure in terms of preventing cell death and mossy fiber sprouting. However, as this would also block synaptic plasticity, learning and memory, it would not prove an effective prevention for SRS. Once systemic inhibitors of the p38 MAPK and JNK pathways become available, it will be interesting to determine if these MAPKs have more dramatic effects on the downstream effects of SE, cell death in particular.


We thank James M. Trzaskos, Janice L. Hytrek, A. Christine Tabaka, James S. Piecara and Christopher A. Teleha of DuPont Research Pharmaceuticals for the kind gift of SL327. This work was supported by a PhRMA Foundation Advanced Predoctoral Fellowship (JLB), NS 30454 (AIL), NS 40221 (MJD) and the Alzheimer's Association.