Abstract
Temporal lobe epilepsy is prevalent and can be difficult to treat effectively. Granule cell axon (mossy fiber) sprouting is a common neuropathological finding in patients with mesial temporal lobe epilepsy, but its role in epileptogenesis is unclear and controversial. Focally infused or systemic rapamycin inhibits the mammalian target of rapamycin (mTOR) signaling pathway and suppresses mossy fiber sprouting in rats. We tested whether long-term systemic treatment with rapamycin, beginning 1 d after pilocarpine-induced status epilepticus in mice, would suppress mossy fiber sprouting and affect the development of spontaneous seizures. Mice that had experienced status epilepticus and were treated for 2 months with rapamycin displayed significantly less mossy fiber sprouting (42% of vehicle-treated animals), and the effect was dose dependent. However, behavioral and video/EEG monitoring revealed that rapamycin- and vehicle-treated mice displayed spontaneous seizures at similar frequencies. These findings suggest mossy fiber sprouting is neither pro- nor anti-convulsant; however, there are caveats. Rapamycin treatment also reduced epilepsy-related hypertrophy of the dentate gyrus but did not significantly affect granule cell proliferation, hilar neuron loss, or generation of ectopic granule cells. These findings are consistent with the hypotheses that hilar neuron loss and ectopic granule cells might contribute to temporal lobe epileptogenesis.
Introduction
Granule cell axon (mossy fiber) sprouting is a common neuropathological finding in patients with mesial temporal lobe epilepsy (de Lanerolle et al., 1989; Sutula et al., 1989; Houser et al., 1990) and laboratory animal models (Nadler et al., 1980; Lemos and Cavalheiro, 1995). Normally, granule cells in the dentate gyrus form very few synapses with one another, but sprouted mossy fibers form recurrent, excitatory synapses (Babb et al., 1991; Represa et al., 1993; Franck et al., 1995; Okazaki et al., 1995; Zhang and Houser, 1999; Wenzel et al., 2000b; Buckmaster et al., 2002; Cavazos et al., 2003). Hippocampal slice experiments have shown that after mossy fiber sprouting, granule cells monosynaptically excite one another (Wuarin and Dudek, 1996, 2001; Molnár and Nadler, 1999; Lynch and Sutula, 2000; Scharfman et al., 2003), and seizure-like events are more easily evoked (Patrylo and Dudek, 1998; Gabriel et al., 2004; Winokur et al., 2004). Computer simulations suggest that even modest levels of mossy fiber sprouting substantially promote spread of seizure-like activity (Santhakumar et al., 2005; Dyhrfjeld-Johnsen et al., 2007). Together, these findings support the recurrent excitation hypothesis, which contends that mossy fiber sprouting is epileptogenic (Tauck and Nadler, 1985). In contrast, some have argued that mossy fiber sprouting suppresses seizure activity by activating inhibitory interneurons (Sloviter, 1992; Sloviter et al., 2006).
However, although some in vivo studies found correlations between the extent of mossy fiber sprouting and seizure frequency (Mathern et al., 1993, 1997; Lemos and Cavalheiro, 1995; Wenzel et al., 2000b; Pitkänen et al., 2005; Kharatishvili et al., 2006), many have not (Cronin and Dudek, 1988; Masukawa et al., 1992; Mello et al., 1993; Buckmaster and Dudek, 1997; Spencer et al., 1999; Timofeeva and Peterson, 1999; Pitkänen et al., 2000; Wenzel et al., 2000a; Gorter et al., 2001; Nissinen et al., 2001; Zhang et al., 2002; Raol et al., 2003; Jung et al., 2004; Lynd-Balta et al., 2004; Williams et al., 2004; Harvey and Sloviter, 2005; Rao et al., 2006; Kadam and Dudek, 2007). The lack of consistently replicable, statistically significant correlations with in vivo seizure activity suggests mossy fiber sprouting might be an epiphenomenon without major pro- or anti-convulsant effects (Gloor, 1997).
Rapamycin might be useful to test the role of mossy fiber sprouting in temporal lobe epileptogenesis. Prolonged, focal infusion of rapamycin into the dentate gyrus inhibits the mammalian target of rapamycin (mTOR) signaling pathway and suppresses mossy fiber sprouting (Buckmaster et al., 2009). Systemic treatment with rapamycin also suppresses mossy fiber sprouting and was reported to reduce seizure frequency in kainate-treated rats (Zeng et al., 2009), which supports the recurrent excitation hypothesis. We sought to further test the recurrent excitation hypothesis and asked the following questions. Does systemic rapamycin suppress mossy fiber sprouting in a mouse model of temporal lobe epilepsy? If so, does it affect the frequency or severity of spontaneous seizures? Finally, does rapamycin affect other epilepsy-related abnormalities of dentate gyrus circuitry, including granule cell proliferation, hilar neuron loss, and generation of ectopic granule cells?
Materials and Methods
Animals.
All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Stanford University Institutional Animal Care and Use Committee. When they were 48 ± 1 d old, male and female mice of the FVB strain (The Jackson Laboratory) were treated with pilocarpine (300 mg/kg, i.p.) 51 ± 6 min (mean ± SEM) after atropine methylbromide (5 mg/kg, i.p.). Diazepam (10 mg/kg, i.p.) was administered 2 h after the onset of stage 3 or greater seizures (Racine, 1972) and repeated as needed to suppress convulsions. During recovery, mice were kept warm and received lactated Ringer's solution with dextrose. Control mice included animals that were treated identically but did not develop status epilepticus, as well as naive mice.
Rapamycin treatment.
Beginning 24 h after pilocarpine treatment, rapamycin was administered systemically using a previously reported dosage regimen that inhibits mTOR activity in the hippocampus (Zeng et al., 2008), except in the present study, treatment was daily instead of 5 d per week. Rapamycin (LC Laboratories) was dissolved initially in 100% ethanol to 20 mg/ml stock solution, which was stored at −20°C. Immediately before injection, stock solution was diluted in 5% Tween 80 and 5% polyethyleneglycol 400 to final concentrations of 4% ethanol and 0.5 or 1 mg/ml rapamycin. Mice were treated daily (intraperitoneally) with 1.5 mg/kg rapamycin, 3 mg/kg rapamycin, or vehicle alone.
Anatomical processing.
Mice were killed by urethane overdose (2 g/kg, i.p.) and perfused through the ascending aorta at 15 ml/min for 2 min with 0.9% sodium chloride, 5 min with 0.37% sodium sulfide, 1 min with 0.9% sodium chloride, and 30 min with 4% paraformaldehyde in 0.1 m phosphate buffer (PB; pH 7.4). Brains were postfixed overnight at 4°C. Then, one hippocampus was isolated, cryoprotected in 30% sucrose in PB, gently straightened, frozen, and sectioned transversely with a microtome set at 40 μm. In the present study, the microtome stage was set to advance automatically, which produced sections with an actual thickness very close to 40 μm. In a previous study, the microtome stage was advanced manually, which inadvertently generated thicker sections (Zhang et al., 2009). Subsequently, volume estimates of the present study are more accurate and larger, but relative differences between control and epileptic groups are similar. Volumes of the granule cell layer plus molecular layer of epileptic mice are ∼160% of control values in both studies (see below). Sections were collected in 30% ethylene glycol and 25% glycerol in 50 mm PB and stored at −20°C until they were processed together after rinsing in PB.
Starting at a random point near the septal pole, a 1-in-12 series of sections from each hippocampus was mounted, dried, and developed for 45 min in 120 ml of 50% gum arabic, 20 ml of 2 m citrate buffer, 60 ml of 0.5 m hydroquinone, and 1 ml of 19% silver nitrate. After rinsing, sections were exposed to 5% sodium thiosulfate for 4 min before dehydration and coverslipping with DPX. An adjacent 1-in-12 series of sections was processed for Nissl staining with 0.25% thionin. In some cases, a third 1-in-12 series was processed for prox-1 immunoreactivity using a protocol from McCloskey et al. (2006) with slight modifications. Sections were rinsed in PB and treated with 1% H2O2 for 2 h. After rinses in PB and 0.1 m Tris-buffered saline (TBS; pH 7.4), sections were treated with blocking solution consisting of 3% goat serum, 2% bovine serum albumin (BSA), and 0.3% Triton X-100 in 0.05 m TBS for 2 h. Sections were rinsed in TBS and incubated for 7 d at 4°C in rabbit anti-prox1 serum (1:40,000, cat. no. PRB-238C, Covance) diluted in 1% goat serum, 0.2% BSA, and 0.3% Triton X-100 in 0.05 M TBS. After rinses in TBS, sections incubated for 2 h in biotinylated goat anti-rabbit serum (1:500; Vector Laboratories) in secondary diluent consisting of 2% BSA and 0.3% Triton X-100 in 0.05 m TBS. After rinses in TBS, sections incubated for 2 h in avidin–biotin–horseradish peroxidase complex (1:500; Vector Laboratories) in secondary diluent. After rinses in TBS and 0.1 m Tris buffer (TB; pH 7.6), sections were placed for 5 min in chromogen solution consisting of 0.02% diaminobenzidine, 0.04% NH4Cl, and 0.015% glucose oxidase in TB and transferred to fresh chromogen solution with 0.1% β-d-glucose for 13 min. The reaction was stopped in rinses of TB. Sections were mounted and dried on gelatin-coated slides, dehydrated, cleared, and coverslipped with DPX. A similar protocol was used to process tissue from a subset of mice for zinc transporter-3 (ZnT-3) immunoreactivity, except sections incubated for 66 h at 4°C in rabbit anti-ZnT-3 serum (1:5000, catalog # 197002; Synaptic Systems).
Anatomical analyses.
Mossy fiber sprouting was measured by estimating the fraction of the entire volume of the granule cell layer plus molecular layer that was black after Timm staining. A 1-in-12 series of sections sampled from the entire septotemporal length of the hippocampus was evaluated (average, 14 sections per mouse). Timm-staining analysis and all other analyses of this study were done by an investigator who was blind as to whether mice had been treated with vehicle versus rapamycin. From each section, an image of the dentate gyrus was obtained with a 10× objective using identical microscope (Axiophot; Zeiss) and camera (Axiocam; Zeiss) settings. NIH ImageJ (1.42q) was used to measure the area of a contour drawn around the granule cell layer plus molecular layer (Fig. 1D2). The Timm-positive area within the contour was determined by converting color images to black and white and adjusting a darkness threshold setting until the selected area appeared to match the black Timm-positive area of the same section (Fig. 1D1,D2). Volumes were calculated by summing areas and multiplying by 12 (section sampling) and 40 μm (section thickness). The percentage of the granule cell layer plus molecular layer that was Timm positive for the entire hippocampus was calculated for each mouse.
Numbers of Nissl-stained granule cells, Nissl-stained hilar neurons, and prox1-immunoreactive hilar neurons per dentate gyrus were estimated using the optical fractionator method and Stereo Investigator (MBF Bioscience). In each case, a 1-in-12 series of sections, randomly and systematically selected from the entire septotemporal length of the hippocampus, was evaluated. A 10× objective was used to draw contours around regions of interest, and a 100× objective was used to count cells. Parameters of stereological analyses are reported in Table 1. For Nissl-stained granule cells, the granule cell layer was analyzed, and granule cell nuclei were counted if they were not cut at the superficial surface of the section. For Nissl-stained hilar neurons, the hilus was outlined along its border with the granule cell layer and straight lines drawn from the ends of the granule cell layer to the proximal end of the CA3 pyramidal cell layer. Nuclei were counted if the maximum diameter of their soma was ≥10 μm and they were not cut at the superficial surface of the section. For prox1-immunopositive hilar neurons, the entire hilar area was sampled, except the region within 25 μm of the granule cell layer, which was excluded. Immunopositive somata were counted if they were not cut at the superficial surface of the section. For each cell type, mean coefficients of error were substantially less than coefficients of variation (Table 1), suggesting sufficient within-animal sampling.
Parameters of stereological analyses
Seizure monitoring.
Mice were evaluated for spontaneous seizures by behavioral analysis alone and by video/EEG recording. For behavioral analysis, mice were videotaped 9 h every day. Daily monitoring may be important because, as we observed (data not shown) and others have reported (Goffin et al., 2007; Williams et al., 2009), seizures in individual rodents tend to occur in clusters followed by longer seizure-free periods. Therefore, without daily monitoring for sufficient periods, seizure frequency might be overestimated or underestimated depending on whether monitoring periods coincide with clusters or not. Videotapes were reviewed manually, and motor seizures of stage 3 or greater (Racine, 1972) were counted.
For video/EEG analysis, field potential recordings were obtained from the right, dorsal hippocampus. To implant electrodes, mice were sedated with diazepam (10 mg/kg, i.p.), anesthetized with isoflurane (1–3%), placed in a stereotaxic frame, maintained on a heating pad, and prepared for aseptic surgery. A hole was drilled 1.8 mm posterior and 1.5 mm right of bregma. An insulated, 50-μm-diameter stainless steel wire (California Fine Wire) was implanted 1.7 mm below the surface of the brain. The reference electrode was placed in the cerebellum. The ground was a screw placed rostrally in the skull. Electrode leads were connected to pins that were inserted into a strip connector, which was attached to the skull with cranioplastic cement. After surgery, mice were kept warm and received lactated Ringer's solution with dextrose, antibiotic (enrofloxacin, 10 mg/kg, s.c.), and analgesic (buprenorphine, 0.05 mg/kg, s.c.). After recovering for 7 d, mice were monitored 9 h every day by video/EEG recording. Recordings were obtained with AI402 probes and a Cyberamp 380 (Molecular Devices). Signals were amplified 500×, filtered (1–100 Hz), and sampled at 200 Hz (PClamp; Molecular Devices). Electrographic seizures were manually identified as bursts of high-frequency, high-amplitude rhythmic activity that lasted at least 15 s, evolved over time, and exhibited an abrupt onset and offset. All electrographic seizures were evaluated for behavioral correlates, and false-positives attributable to grooming or other artifacts were excluded. Behavioral correlates of electrographically identified seizures were scored according to Racine (1972). Recording electrode locations were verified in Nissl-stained coronal sections.
Unless specified otherwise, chemicals were obtained from Sigma-Aldrich. Statistical analyses were performed using SigmaStat (Systat Software) with p < 0.05 considered significant. Values are expressed as mean ± SEM.
Results
Rapamycin suppresses mossy fiber sprouting
In rats, rapamycin suppresses mossy fiber sprouting, even when initiated after an epileptogenic treatment (Buckmaster et al., 2009; Zeng et al., 2009). In control mice, daily 1 mg/kg rapamycin (intraperitoneally) significantly reduces mTOR activity in the hippocampus (Zeng et al., 2008). We tested whether daily 1.5 or 3 mg/kg rapamycin (intraperitoneally) would suppress mossy fiber sprouting in mice that experienced pilocarpine-induced status epilepticus. Rapamycin treatment began 24 h after status epilepticus. Mossy fiber sprouting develops gradually over a period of weeks to months after pilocarpine-induced status epilepticus (Mello et al., 1993). Therefore, daily rapamycin treatment continued for 2 months, at which time mice were perfused for Timm staining. Mossy fiber sprouting was measured as the fraction of the granule cell layer plus molecular layer that exhibited black Timm staining (Fig. 1D). In individual untreated control mice that had not experienced status epilepticus, only 2–4% of the entire granule cell layer plus molecular layer was Timm positive (Fig. 1A). The mild level of black Timm staining in control animals is primarily attributable to a projection of mossy fibers into the inner molecular layer at the temporal pole of the hippocampus (Haug, 1974) (Fig. 1E). Mice that had experienced status epilepticus and had been treated with vehicle for 2 months displayed a prominent band of black Timm staining in the inner molecular of the dentate gyrus (Fig. 1B). Aberrant mossy fiber sprouting in vehicle-treated mice occurred all along the septotemporal axis of the hippocampus (Fig. 1E). In individual vehicle-treated mice, 10–34% of the entire granule cell layer plus molecular layer was Timm positive, and the group average (23 ± 2%) was over 10 times greater than that of controls (Fig. 1F). Mossy fiber sprouting in rapamycin-treated mice was less than in vehicle-treated animals but more than in controls (Fig. 1C,E,F). In individual mice treated with 1.5 mg/kg rapamycin, 9–19% of the entire granule cell layer plus molecular layer was Timm positive (average, 13 ± 1%). In individual mice treated with 3 mg/kg rapamycin, 7–13% of the granule cell layer plus molecular layer was Timm positive (average, 10 ± 1%), which was less than half (42%) of the amount of mossy fiber sprouting in vehicle-treated mice. All groups were significantly different from each other (p < 0.05, ANOVA with Student–Newman–Keuls method) (Fig. 1F).
Rapamycin suppresses mossy fiber sprouting in a mouse model of temporal lobe epilepsy. A–C, Timm-stained sections of the dentate gyrus from a control (A), a mouse after status epilepticus and treatment for 2 months with vehicle (B), and a mouse after status epilepticus and treatment for 2 months with daily, systemic 3 mg/kg rapamycin (C). Images on the right (A2, B2, C2) are high-magnification views of areas indicated by arrows in the left images. The inner molecular layer is devoid of black Timm staining in the control mouse and displays a prominent band in the vehicle-treated mouse, which is reduced in the rapamycin-treated mouse. m, Molecular layer; g, granule cell layer; h, hilus; CA3, CA3 pyramidal cell layer. D, The method used to quantify mossy fiber sprouting involved making a contour (D2, black line) around the granule cell layer plus molecular layer and adjusting a darkness threshold setting until the selected area (D2, red) matched the black Timm-positive area of the same section (D1). E, Mossy fiber sprouting was suppressed all along the septotemporal axis of the hippocampus by rapamycin treatment. Values represent mean ± SEM. F, Group averages for mossy fiber sprouting in the entire hippocampus. All groups were significantly different from all other groups (p < 0.05, ANOVA with Student–Newman–Keuls method). gcl, Granule cell layer; ml, molecular layer.
To test the validity of using Timm staining to measure mossy fiber sprouting, adjacent sections from a subset of mice were processed for another marker of mossy fibers. ZnT-3 is expressed most abundantly on synaptic vesicles of granule cell mossy fibers (Wenzel et al., 1997). ZnT-3 immunoreactivity revealed similar patterns of staining as Timm-positive regions of adjacent sections in control mice (n = 8) and mice that experienced status epilepticus and were then treated with vehicle (n = 7) or 3 mg/kg rapamycin (n = 7) (Fig. 2). These findings suggest Timm staining was a reliable label for quantifying extent of mossy fiber sprouting. Together, these findings confirm that even when treatment initiates after an epileptogenic injury, long-term systemic rapamycin suppresses mossy fiber sprouting in rodent models of temporal lobe epilepsy.
ZnT-3 immunocytochemistry and Timm staining reveal similar patterns of mossy fiber labeling. ZnT-3 immunocytochemistry in the dentate gyrus at low (left) and high (middle) magnification is shown. Arrows in A1, B1, and C1 indicate regions illustrated in A2, B2, and C2. A3, B3, and C3 show a similar region of the dentate gyrus from an adjacent or nearby section in the same hippocampus that was processed for Timm staining. ZnT-3 immunocytochemistry and Timm staining both reveal minimal labeling in the inner molecular layer of the control mouse (A), a prominent band of labeling in the mouse that experienced status epilepticus and was treated for 2 months with vehicle (B), and reduced labeling in the mouse that experienced status epilepticus and was treated systemically with daily 3 mg/kg rapamycin for 2 months (C). m, Molecular layer; g, granule cell layer; h, hilus; CA3, CA3 pyramidal cell layer.
Rapamycin does not affect seizure frequency
Systemic rapamycin treatment in a rat model of temporal lobe epilepsy was reported to suppress seizure frequency, even when initiated after status epilepticus (Zeng et al., 2009; Huang et al., 2010). We tested whether daily 3 mg/kg rapamycin (intraperitoneally) would suppress the frequency or severity of spontaneous seizures in mice that experienced pilocarpine-induced status epilepticus. Rapamycin treatment began 24 h after status epilepticus and continued for 2 months. During the second month of treatment, mice were video monitored 9 h every day for spontaneous motor seizures of grade 3 or greater (Racine, 1972). Over 12,600 mouse-hours of data were analyzed, and 2080 motor seizures were identified. Vehicle-treated mice that had experienced status epilepticus (n = 29) had 0.02–0.56 seizures per hour. Rapamycin-treated mice that had experienced status epilepticus (n = 26) had 0.02–0.32 seizures per hour. Group averages were similar in vehicle- and rapamycin-treated mice (0.168 ± 0.023 and 0.166 ± 0.018 seizures per hour, respectively; p = 0.7, t test). Seizure severity ranged from forelimb clonus (stage 3), rearing (stage 4), and loss of postural control (stage 5) to wild running (stage 6). Group averages of seizure severity were similar in vehicle- and rapamycin-treated mice (4.64 ± 0.01 and 4.71 ± 0.04, respectively; p = 0.3, t test). These findings suggest rapamycin does not affect seizure frequency or severity in pilocarpine-treated mice.
To test rapamycin's effect on epileptogenesis more rigorously, we used video/EEG monitoring, evaluated two doses, delayed onset of monitoring for 2 months to allow more time for mossy fiber sprouting to develop, and perfused mice for Timm staining after recording was complete (Fig. 3A). Rapamycin treatment began 24 h after status epilepticus, and after almost 2 months, mice were implanted for recording field potentials from the hippocampus (n = 18 and 21 for 1.5 and 3 mg/kg rapamycin, respectively). Rapamycin treatment continued while seizure data were collected, because previous findings suggest continued treatment is necessary to keep sprouting suppressed (Buckmaster et al., 2009). When video/EEG recording started during the third month, mice that had been treated with vehicle (n = 15) began treatment with 3 mg/kg rapamycin to control for any acute anticonvulsant or pro-convulsant effect that rapamycin might have. Previous findings suggest treating the vehicle group with rapamycin beginning 2 months after status epilepticus is unlikely to reverse already established mossy fiber sprouting (Buckmaster et al., 2009; Zhou et al., 2009). In addition, a group of control mice (n = 10) was treated with 3 mg/kg rapamycin and video/EEG monitored to test whether rapamycin was pro-convulsant.
Rapamycin treatment suppresses mossy fiber sprouting but does not affect the frequency of spontaneous seizures in a mouse model of temporal lobe epilepsy. A, Experimental design. B, Spontaneous electrographic seizures recorded in the hippocampus of an epileptic vehicle- and 3 mg/kg rapamycin-treated mouse. C–E, Timm-stained sections of the dentate gyrus from video/EEG-monitored mice: a control (C), a mouse after status epilepticus and treatment for 2 months with vehicle (D), and a mouse after status epilepticus and treatment for 2 months with daily 3 mg/kg rapamycin (E). The inner molecular layer is devoid of black Timm staining in the control mouse and displays a prominent band in the vehicle-treated mouse, which is reduced in the rapamycin-treated mouse. m, Molecular layer; g, granule cell layer; h, hilus; CA3, CA3 pyramidal cell layer. F, Seizure frequency was similar in vehicle- and rapamycin-treated mice that experienced status epilepticus (p = 0.8, ANOVA). Rapamycin-treated control mice were never observed to have seizures. Values represent mean ± SEM. G, Mossy fiber sprouting was suppressed by rapamycin. All groups are significantly different from each other (p < 0.05, ANOVA with Student–Newman–Keuls method), except there is no significant difference between 1.5 and 3 mg/kg rapamycin-treated mice.
Video/EEG recording lasted up to 1 month, but electrode implants detached early in some cases, and the average duration of recording was 18 d per mouse. Over 10,600 mouse-hours of data were analyzed, and 1480 electrographic seizures were identified. Seizures appeared similar in vehicle- and rapamycin-treated mice and consisted of bursts of high-frequency, high-amplitude rhythmic activity with abrupt onsets and offsets, lasting at least 15 s (Fig. 3B). None of the rapamycin-treated control mice that had not experienced status epilepticus displayed seizures. Seizure frequency was similar in vehicle- and rapamycin-treated mice that had experienced status epilepticus (p > 0.8, ANOVA) (Fig. 3F). Motor seizures of stage 3 or greater (Racine, 1972) accounted for 93% of all electrographic seizures identified. Behavioral seizure severity was similar in vehicle- and 3 mg/kg rapamycin-treated mice (4.60 ± 0.16 and 4.58 ± 0.07, respectively) and slightly more severe in 1.5 mg/kg rapamycin-treated mice (4.93 ± 0.04; p < 0.05, ANOVA with Student–Newman–Keuls method).
Timm staining of the dentate gyrus in video/EEG-monitored mice (Fig. 3C–E) revealed qualitatively similar results to that of mice perfused 2 months after status epilepticus (Fig. 1A–C). The average percentage of the entire granule cell layer plus molecular layer that was Timm positive was minimal in the control group, largest in the epileptic vehicle-treated mice, and intermediate in the epileptic rapamycin-treated mice. However, differences between vehicle- and rapamycin-treated groups were smaller after video/EEG recording and 3 months of rapamycin treatment (Fig. 3G) than in mice perfused at the 2 month time point (Fig. 1F). In video/EEG-monitored 3 mg/kg rapamycin-treated mice, mossy fiber sprouting was reduced to 75% of the level in video/EEG-monitored vehicle-treated animals, whereas, in mice perfused after 2 month treatment with rapamycin, mossy fiber sprouting was reduced to 42% of the level of vehicle-treated animals. To test whether reduced suppression of mossy fiber sprouting might be attributable to surgery for EEG recording versus rapamycin treatment duration or time since status epilepticus, another set of mice (n = 9) was treated with daily 3 mg/kg rapamycin for 3 months beginning 1 d after pilocarpine-induced status epilepticus, but they were not implanted for recording field potentials. They displayed a similar level of mossy fiber sprouting (14.9 ± 1.6% of the granule cell layer plus molecular layer) as mice that were treated with 3 mg/kg rapamycin for 3 months and video/EEG recorded (13.7 ± 0.9%; p = 0.48, t test). These findings suggest rapamycin's suppressive effect on mossy fiber sprouting decreases during the third month of treatment. During video/EEG monitoring, therefore, the average level of mossy fiber sprouting in mice treated with 3 mg/kg rapamycin probably was reduced to between 42% (at 2 months after status epilepticus, when video/EEG monitoring began) to 75% of that in vehicle-treated epileptic mice (at 3 months after status epilepticus when video/EEG monitoring ceased). In summary, both behavioral and video/EEG seizure monitoring data suggest rapamycin does not affect seizure frequency in pilocarpine-treated mice.
Rapamycin suppresses dentate gyrus hypertrophy
Mouse models of temporal lobe epilepsy display hypertrophy of the dentate gyrus (Bouilleret et al., 1999; Zhang et al., 2009), which might be attributable, at least in part, to mTOR-mediated enlargement of granule cells (Backman et al., 2001; Kwon et al., 2001, 2003, 2006). We evaluated sections from mice described above that had been treated with vehicle or rapamycin for 2 months beginning 24 h after pilocarpine-induced status epilepticus to test whether rapamycin suppresses dentate gyrus hypertrophy. In vehicle-treated mice that had experienced status epilepticus, the average volume of the granule cell layer plus molecular layer in Timm-stained sections was 2.73 ± 0.19 mm3, which was 158% of controls (1.72 ± 0.07 mm3; p < 0.001, ANOVA with Student–Newman–Keuls method) (Fig. 4D). In mice treated with 1.5 mg/kg rapamycin, the average volume of the granule cell layer plus molecular layer was reduced to 2.41 ± 0.06 mm3 (1.5 mg/kg rapamycin vs control, p < 0.001; 1.5 mg/kg rapamycin vs vehicle, p < 0.03). In mice treated with 3 mg/kg rapamycin, the average volume of the granule cell layer plus molecular layer was reduced to 2.01 ± 0.07 mm3 (3 mg/kg rapamycin vs control, p = 0.12; 3 mg/kg rapamycin vs vehicle, p < 0.001; 3 mg/kg rapamycin vs 1.5 mg/kg rapamycin, p = 0.007). These findings suggest rapamycin suppresses dentate gyrus hypertrophy in epileptic pilocarpine-treated mice.
Dentate gyrus volume, granule cell numbers, and hilar neuron numbers in control and epileptic vehicle- and rapamycin-treated mice. A–C, Nissl-stained sections of the dentate gyrus from a control (A), a mouse after status epilepticus and treatment for 2 months with vehicle (B), and a mouse after status epilepticus and treatment for 2 months with daily 3 mg/kg rapamycin (C). m, Molecular layer; g, granule cell layer; h, hilus; CA3, CA3 pyramidal cell layer. D, Epileptic mice display hypertrophy of the granule cell layer (gcl) plus molecular layer (ml), which is suppressed by rapamycin. All groups are significantly different from each other (p < 0.03, ANOVA with Student–Newman–Keuls method), except there is no significant difference between control and 3 mg/kg rapamycin-treated groups (p = 0.12). Values represent mean ± SEM. E, Vehicle- and rapamycin-treated mice that experience status epilepticus have more granule cells per dentate gyrus than controls (p < 0.001, ANOVA with Student–Newman–Keuls method). There is no significant difference between vehicle- and rapamycin-treated groups (p = 0.15). F, Numbers of Nissl-stained hilar neurons (with somata >10 μm maximum diameter) per dentate gyrus. Hilar neuron loss is similar in epileptic vehicle- and rapamycin-treated mice (p > 0.8, ANOVA with Student–Newman–Keuls method). All groups had significantly fewer hilar neurons than controls (p < 0.001).
Granule cell somata and dendrites are the major components of the granule cell layer plus molecular layer. Hypertrophy of the region could be attributable to larger granule cells. Another possibility is increased numbers of granule cells, which occurs in epileptic pilocarpine-treated mice (Zhang et al., 2009). We used stereological methods to estimate numbers of granule cells. Control mice had 416,000 ± 13,000 granule cells per hippocampus (Fig. 4E). Both vehicle- and 3 mg/kg rapamycin-treated mice that had experienced status epilepticus had more granule cells than controls (141% and 132% of control levels, respectively; p < 0.001, ANOVA with Student–Newman–Keuls method). The difference between vehicle- and rapamycin-treated mice was not statistically significant (p = 0.15). These findings are consistent with rapamycin's lack of effect on neurogenesis in kainate-treated rats (Zeng et al., 2009). In addition, these findings suggest rapamycin suppresses epilepsy-related dentate gyrus hypertrophy by reducing granule cell enlargement, not by inhibiting granule cell proliferation.
Rapamycin does not affect hilar neuron loss
Hilar neuron loss is a common neuropathological feature in patients with temporal lobe epilepsy (Margerison and Corsellis, 1966) and in rodent models (Nadler et al., 1980), including those treated with pilocarpine (Lemos and Cavalheiro, 1995). We used stereological methods to estimate numbers of Nissl-stained hilar neurons with a maximum soma diameter > 10 μm. Control mice (n = 6) had 12,500 ± 400 hilar neurons per hippocampus (Fig. 4F). Vehicle-treated (n = 13), 1.5 mg/kg rapamycin-treated (n = 14), and 3 mg/kg rapamycin-treated (n = 15) mice that had experienced status epilepticus had fewer hilar neurons (53, 53, and 54% of controls, respectively; p < 0.001, ANOVA with Student–Newman–Keuls method). The differences between vehicle- and rapamycin-treated mice were not statistically significant (p > 0.8). These findings suggest hilar neuron loss is not affected by rapamycin treatment that begins 24 h after status epilepticus.
Rapamycin does not affect generation of ectopic granule cells
Epileptogenic injuries generate granule cells that become ectopically located in the hilus of rodent models (Houser, 1990; Parent et al., 1997; Jiao and Nadler, 2007) and in patients with temporal lobe epilepsy (Parent et al., 2006). Prox1 is expressed by granule cells (Pleasure et al., 2000; Elliott et al., 2001) including ectopic granule cells but not other hilar neurons (McCloskey et al., 2006). We used stereological methods to estimate numbers of prox1-immunoreactive hilar neurons that were at least 25 μm away from the granule cell layer. Hilar prox1-immunoreactive cells were rare in controls but more common in epileptic vehicle- and rapamycin-treated mice (Fig. 5A–C). Control mice (n = 9) had 977 ± 71 prox1-positive hilar neurons per hippocampus (Fig. 5D). Both vehicle-treated (n = 9) and 3 mg/kg rapamycin-treated (n = 10) mice that had experienced status epilepticus had over 2.5 times more prox1-positive hilar neurons than controls (p < 0.001, ANOVA with Student–Newman–Keuls method). The difference between vehicle- and rapamycin-treated mice was not statistically significant (p = 0.9). These findings suggest epilepsy-related generation of ectopic granule cells is not affected by rapamycin treatment that begins 1 d after status epilepticus.
Ectopic granule cells in control and epileptic vehicle- and rapamycin-treated mice. A–C, Prox1-immunostained sections of the dentate gyrus from a control (A), a mouse after status epilepticus and treatment for 2 months with vehicle (B), and a mouse after status epilepticus and treatment for 2 months with daily 3 mg/kg rapamycin (C). m, Molecular layer; g, granule cell layer; h, hilus; CA3, CA3 pyramidal cell layer. D, Numbers of prox1-immunoreactive hilar neurons (≥25 μm away from the granule cell layer) per dentate gyrus. Epileptic vehicle- and rapamycin-treated mice have more prox1-immunopositive hilar neurons than controls (p < 0.001, ANOVA with Student–Newman–Keuls method). There is no significant difference between the vehicle- and rapamycin-treated groups (p = 0.9). Values represent mean ± SEM.
Discussion
The principal findings of this study are the following. Systemic treatment with rapamycin beginning 24 h after status epilepticus suppressed mossy fiber sprouting in pilocarpine-treated mice. However, the frequency of spontaneous seizures in rapamycin-treated mice was similar to that of vehicle-treated mice. Treatment with rapamycin diminished epilepsy-related hypertrophy of the dentate gyrus but not granule cell proliferation, hilar neuron loss, or generation of ectopic granule cells. These findings are consistent with hilar neuron loss and ectopic granule cells as epileptogenic mechanisms but raise doubts about the role of mossy fiber sprouting.
Rapamycin suppresses mossy fiber sprouting
Accumulating evidence suggests an important role for the mTOR signaling pathway in mossy fiber sprouting. Treatments that trigger mossy fiber sprouting, including traumatic brain injury (Chen et al., 2007) and status epilepticus (Shacka et al., 2007; Buckmaster et al., 2009; Zeng et al., 2009), activate the mTOR signaling pathway. Activation of the mTOR signaling pathway stimulates axon growth in cultured sensory (Grider et al., 2009) and hippocampal (Choi et al., 2009) neurons. Even without a preceding epileptogenic injury, mossy fiber sprouting develops in transgenic animals with excessive mTOR activity (Kwon et al., 2006), and rapamycin, an inhibitor of mTOR, suppresses axon regeneration (Verma et al., 2005; Park et al., 2008) and guidance of growth cones (Campbell and Holt, 2001).
In the present study, systemic treatment with rapamycin in mice beginning 1 d after pilocarpine-induced status epilepticus significantly reduced mossy fiber sprouting measured 2–3 months later, which is consistent with previous rat studies (Buckmaster et al., 2009; Zeng et al., 2009). Rapamycin appears to be one of the first successful treatments to reduce mossy fiber sprouting when applied after a precipitating injury. Previously, treatment with the protein inhibitor cycloheximide just before an epileptogenic injury was reported to suppress mossy fiber sprouting (Longo and Mello, 1997, 1998; Silva and Mello, 2000). However, cycloheximide protects hilar neurons from excitotoxicity during status epilepticus (Schreiber et al., 1993; Bengzon et al., 1997; Covolan et al., 2000), and extent of mossy fiber sprouting is correlated with the degree of hilar neuron loss (Cavazos and Sutula, 1990; Houser et al., 1990; Babb et al., 1991; Masukawa et al., 1992; Buckmaster and Dudek, 1997; Gorter et al., 2001; Nissinen et al., 2001; Jiao and Nadler, 2007). Therefore, cycloheximide probably reduces mossy fiber indirectly by protecting hilar neurons from excitotoxic damage (Longo et al., 2003). Although rapamycin can protect hilar neurons from excitotoxicity (Zeng et al., 2009), that did not occur in the present study probably because rapamycin treatment was delayed until 1 d after status epilepticus.
Rapamycin is neither anticonvulsant nor antiepileptogenic in the mouse pilocarpine model of temporal lobe epilepsy
Seizure frequency was similar in vehicle- and rapamycin-treated mice, suggesting rapamycin is neither pro-convulsant nor anticonvulsant. Previous reports on rapamycin's effects on seizure activity are mixed. Rapamycin does not alter evoked field potentials in CA1 of hippocampal slices from rats (Daoud et al., 2007) and has either no effect or inhibits spontaneous action potential firing in hippocampal neuron cultures from rats (Rüegg et al., 2007). Rapamycin was reported to be anticonvulsant when administered to pilocarpine-treated rats after they had developed chronic, spontaneous, recurrent seizures (Huang et al., 2010). However, severity of chemically induced status epilepticus is similar in rats with and without prior rapamycin treatment (Zeng et al., 2009; Huang et al., 2010).
In rodent models of conditions such as tuberous sclerosis that involve mutations causing excessive activity of the mTOR signaling pathway, seizures are common (Backman et al., 2001; Kwon et al., 2001, 2006; Uhlmann et al., 2002; Meikle et al., 2007), and rapamycin (Zeng et al., 2008; Ljungberg et al., 2009; Zhou et al., 2009) and other inhibitors of mTOR (Kwon et al., 2003) reduce seizure frequency. In rodent models of temporal lobe epilepsy, rapamycin's effects are less clear. In contrast to pilocarpine-treated mice in the present study, rapamycin reduced seizure frequency in kainate-treated rats (Zeng et al., 2009). Possible explanations for the apparently contradictory results include differences in the timing and methods of seizure monitoring or a confounding anticonvulsant effect of rapamycin specifically in rats (Huang et al., 2010).
Is mossy fiber sprouting epileptogenic?
Rapamycin significantly reduced mossy fiber sprouting without affecting the frequency of spontaneous seizures in pilocarpine-treated mice. This finding suggests that mossy fiber sprouting might be an epiphenomenon without major antiepileptogenic or pro-epileptogenic effects; however, there are caveats. Ideally, to evaluate its role in epileptogenesis, mossy fiber sprouting would be blocked completely. In the present study, mossy fiber sprouting was reduced to 42–75% of levels found in vehicle-treated epileptic mice. There may be a threshold effect such that beyond a critical level, mossy fiber sprouting dramatically increases the probability of seizures. Computer simulations suggest relatively small amounts of mossy fiber sprouting might be sufficient to increase seizure susceptibility (Santhakumar et al., 2005; Dyhrfjeld-Johnsen et al., 2007). However, it is unlikely that higher doses of rapamycin would completely block mossy fiber sprouting. Focal infusion of up to 10 mm rapamycin directly into the dentate gyrus did not entirely block mossy fiber sprouting (Buckmaster et al., 2009). The 3 mg/kg dose used in the present study virtually eliminated hippocampal mTOR activity in control mice of a previous study (Zeng et al., 2008) and reduced dentate gyrus hypertrophy to levels not significantly different from nonepileptic controls. Signaling pathways that control processes such as axon growth, pathfinding, and synaptogenesis are likely to be redundant and resistant to complete suppression by inhibiting a single important node (Bromberg et al., 2008).
Another caveat is that epileptogenesis might have been maintained by rapamycin-related side effects or other “compensatory” changes. For example, GABAergic interneuron axons in the dentate gyrus sprout and form more synapses with granule cells after epileptogenic injuries (Babb et al., 1989; Zhang et al., 2009; Thind et al., 2010). If rapamycin also suppresses GABAergic axon sprouting, it might counteract the effect of reduced mossy fiber sprouting. In addition, other recurrent excitatory circuits with granule cells might compete with sprouted mossy fibers for synaptic targets on granule cell dendrites. If competing axons were resistant to rapamycin, they might develop excessively when mossy fiber sprouting is reduced. Possibilities include axon sprouting by surviving mossy cells (Namgung et al., 1997; Del Turco et al., 2003; Prang et al., 2003) and CA3 pyramidal cells (Bausch and McNamara, 2004; Siddiqui and Joseph, 2005; Rao et al., 2006).
Finally, the lack of effect of reduced mossy fiber sprouting on epileptogenesis might be attributable to limitations of the model. Status epilepticus-based rodent models replicate many neuropathological features found in patients with temporal lobe epilepsy, including mossy fiber sprouting. However, there are potentially important differences. For example, excitotoxic damage appears to involve extrahippocampal structures to a greater degree in rodent models compared with patients (Chen and Buckmaster, 2005). Some have argued that chemoconvulsant rodent models are invalid for this experimental question, because their seizures might initiate by different mechanisms and in different brain regions than in patients with temporal lobe epilepsy (Mello et al., 1996; Harvey and Sloviter, 2005; Sloviter et al., 2006). However, fos labeling reveals early activation of granule cells during spontaneous seizures in epileptic pilocarpine-treated mice (Peng and Houser, 2005), suggesting they are a reasonable model in which to test the role of mossy fiber sprouting.
Other potentially epileptogenic changes
Some epilepsy-related abnormalities observed in the present study are unlikely to contribute to temporal lobe epileptogenesis (e.g., granule cell hypertrophy). Transgenic mice with excessive activation of the mTOR signaling pathway display enlarged granule cell somata (Backman et al., 2001; Kwon et al., 2001, 2003; Way et al., 2009) and dendrites (Kwon et al., 2006), which can be reversed by rapamycin (Zhou et al., 2009). In the present study, rapamycin reduced hypertrophy nearly to control levels without affecting seizure frequency. An epilepsy-related net increase in granule cell numbers seems unlikely to be epileptogenic, because it does not occur in epileptic pilocarpine-treated rats (Thind et al., 2010) or patients with temporal lobe epilepsy (Mouritzen Dam, 1980; Mathern et al., 1995).
On the other hand, results of the present study are consistent with other potential epileptogenic mechanisms. Almost half the hilar neuron population was killed presumably by status epilepticus in vehicle- and rapamycin-treated mice. In control animals, ∼60% of hilar neurons are excitatory mossy cells, and ∼40% are inhibitory GABAergic interneurons (Buckmaster and Jongen-Rêlo, 1999; Jiao and Nadler, 2007). Loss of mossy cells (Sloviter et al., 2003) and interneurons (de Lanerolle et al., 1989) has been proposed as epileptogenic mechanisms. Ectopic granule cells receive aberrant excitatory synaptic input from other granule cells (Dashtipour et al., 2001; Pierce et al., 2005). They display abnormal circuitry (Scharfman et al., 2000) and abnormal intrinsic electrophysiological properties (Zhan and Nadler, 2009). Their numbers correlate with seizure frequency in epileptic pilocarpine-treated rats (McCloskey et al., 2006), and reducing the generation of ectopic granule cells was reported to suppress epileptogenesis (Jung et al., 2004). Results of the present study are consistent with an epileptogenic role for ectopic granule cells. In summary, rapamycin's effects are complicated but helpful for prioritizing epilepsy-related changes that might affect seizure generation.
Footnotes
This work was supported by the National Institute of Neurological Disorders and Stroke and National Center for Research Resources. We thank Dr. Xiling Wen for assistance with anatomy.
- Correspondence should be addressed to Paul S. Buckmaster, Department of Comparative Medicine, R321 Edwards Building, 300 Pasteur Drive, Stanford University, Stanford, CA 94305 psb{at}stanford.edu