SEARCH

SEARCH BY CITATION

Keywords:

  • mTOR;
  • Synaptogenesis;
  • GABA;
  • Dentate gyrus;
  • Hippocampus

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References

Purpose: In temporal lobe epilepsy many somatostatin interneurons in the dentate gyrus die. However, some survive and sprout axon collaterals that form new synapses with granule cells. The functional consequences of γ-aminobutyric acid (GABA)ergic synaptic reorganization are unclear. Development of new methods to suppress epilepsy-related interneuron axon sprouting might be useful experimentally.

Methods: Status epilepticus was induced by systemic pilocarpine treatment in green fluorescent protein (GFP)-expressing inhibitory nerurons (GIN) mice in which a subset of somatostatin interneurons expresses GFP. Beginning 24 h later, mice were treated with vehicle or rapamycin (3 mg/kg intraperitoneally) every day for 2 months. Stereologic methods were then used to estimate numbers of GFP-positive hilar neurons per dentate gyrus and total length of GFP-positive axon in the molecular layer plus granule cell layer. GFP-positive axon density was calculated. The number of GFP-positive axon crossings of the granule cell layer was measured. Regression analyses were performed to test for correlations between GFP-positive axon length versus number of granule cells and dentate gyrus volume.

Key Findings: After pilocarpine-induced status epilepticus, rapamycin- and vehicle-treated mice had approximately half as many GFP-positive hilar neurons as did control animals. Despite neuron loss, vehicle-treated mice had over twice the GFP-positive axon length per dentate gyrus as controls, consistent with GABAergic axon sprouting. In contrast, total GFP-positive axon length was similar in rapamycin-treated mice and controls. GFP-positive axon length correlated most closely with dentate gyrus volume.

Significance: These findings suggest that rapamycin suppressed axon sprouting by surviving somatostatin/GFP-positive interneurons after pilocarpine-induced status epilepticus in GIN mice. It is unclear whether the effect of rapamycin on axon length was on interneurons directly or secondary, for example, by suppressing growth of granule cell dendrites, which are synaptic targets of interneuron axons. The mammalian target of rapamycin (mTOR) signaling pathway might be a useful drug target for influencing GABAergic synaptic reorganization after epileptogenic treatments, but additional side effects of rapamycin treatment must be considered carefully.

Loss of somatostatin interneurons in the dentate gyrus is common in patients with temporal lobe epilepsy (de Lanerolle et al., 1989; Sundstrom et al., 2001). Paradoxically, in the same region, somatostatin-immunoreactive axons persist and appear exuberant, which suggests that surviving interneurons sprout axons and form new synapses with granule cells (Mathern et al., 1995). Support for this view recently came from a mouse model of temporal lobe epilepsy. In the molecular layer of the dentate gyrus, somatostatin-immunoreactivity of axons is weak in mice compared to that of other species (Buckmaster et al., 1994). However, in GIN mice a subset of somatostatin interneurons express enhanced green fluorescent protein (GFP), which is a superior marker for axons (Oliva et al., 2000). Epileptic pilocarpine-treated GIN mice display fewer hilar GFP-positive neurons, consistent with loss of somatostatin interneurons, but greater GFP-immunoreactive axon length and more connectivity to granule cells, suggesting axon sprouting and synaptogenesis (Zhang et al., 2009).

The functional consequences of epilepsy-related somatostatin axon sprouting are unclear. It might compensate for the loss of some interneurons by restoring and strengthening feedback inhibition of granule cells. But there also are proepileptic alternatives. Possibilities include hypersynchronization of excitatory hippocampal neurons (Babb et al., 1989), generation of depolarizing γ-aminobutyric acid (GABA) responses in principal cells (Staley et al., 1995; Fujiwara-Tsukamoto et al., 2010), and development of excessively connected surviving somatostatin interneurons that might become network hubs and drive seizure activity in the dentate gyrus. A potential role for aberrant hub neurons in epilepsy has been suggested by computer simulations (Morgan & Soltesz, 2008) and recordings from immature hippocampi (Bonifazi et al., 2009).

To further evaluate the consequences of somatostatin axon reorganization it would be helpful to experimentally manipulate its development after epileptogenic treatments. Recently, rapamycin was discovered to inhibit granule cell axon (mossy fiber) sprouting after chemoconvulsant-induced status epilepticus in rats (Buckmaster et al., 2009; Zeng et al., 2009) and mice (Buckmaster & Lew, 2011). Rapamycin inhibits the mTOR (mammalian target of rapamycin) signaling pathway (Sabatini et al., 1994), which is activated after epileptogenic treatments (Chen et al., 2007; Shacka et al., 2007; Huang et al., 2010). The present study processed tissue from mice in a previous study (Buckmaster & Lew, 2011) to test whether treating GIN mice with rapamycin for 2 months after pilocarpine-induced status epilepticus affects total length of GFP-positive axon per dentate gyrus. If so, it would suggest that rapamycin might be a useful tool for evaluating functional roles of GABAergic axon sprouting in future experiments. If not, it would suggest that the effects of rapamycin on mossy fiber sprouting might be selective.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References

Animals

The present study utilized extra tissue sections collected from a subset of mice in a previous study (Buckmaster & Lew, 2011). 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 45 ± 2 days old, male and female GIN mice (FVB-Tg[GadGFP]4570Swn/J; The Jackson Laboratory, Bar Harbor, ME, U.S.A.) (Oliva et al., 2000) were treated with pilocarpine (300 mg/kg, i.p.) 73 ± 1 min [mean ± standard error of the 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 ringers with dextrose. Controls were age- and gender-matched naive mice (n = 6).

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 days/week. Rapamycin (LC Laboratories, Woburn, MA, U.S.A.) 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% polyethylene glycol 400 to a final concentration of 4% ethanol and 1 mg/ml rapamycin. Mice were treated daily (i.p.) with 3 mg/kg rapamycin (n = 6) or vehicle alone (n = 6).

Green fluorescent protein (GFP)-immunocytochemistry

After 2 months of treatment with rapamycin or vehicle, mice were killed by urethane overdose (2 g/kg, i.p.) perfused through the ascending aorta at 15 ml/min for 2 min with 0.9% sodium chloride, 5 min with 0.37% sodium sulfide (for the Timm stain that was part of another study), 1 min with 0.9% sodium chloride, and 30 min with 4% formaldehyde 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 for which the actual thickness was very close to 40 μm. In a previous study, the microtome stage was advanced manually, which inadvertently generated thicker sections (Zhang et al., 2009). Consequently, volume estimates of the present study are more accurate and larger, but relative differences between experimental groups are similar (see Results). Sections were collected in 30% ethylene glycol and 25% glycerol in 50 mm PB and stored at −20°C.

Sections from all experimental groups were processed together. Beginning at random points near the septal pole of the hippocampus and extending through the entire septotemporal length, 1-in-6 series of 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 (Vector Laboratories, Burlingame, CA, U.S.A.), 2% bovine serum albumin (BSA), and 0.3% Triton X-100 in 0.05 m TBS for 1 h. Sections were rinsed in TBS and incubated for 40 h at 4°C in rabbit anti-GFP serum (1:2,000; Invitrogen, Carlsbad, CA, U.S.A.) diluted in 1% goat serum, 0.2% BSA, and 0.3% Triton X-100 in 0.05 m TBS. After rinses in TBS, sections were 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 were 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 then transferred to fresh chromogen solution with 0.1%β-d-glucose for 13 min. The reaction was stopped in rinses of TB, and sections were mounted and dried on gelatin-coated slides. Sections were dehydrated and cleared in a series of ethanols and xylenes and were coverslipped with DPX.

Data analysis

The total number of GFP-positive hilar neurons per dentate gyrus was estimated using the optical fractionator method (West et al., 1991). Investigators were blinded to experimental groups during analysis. Starting at random points near the septal pole of the hippocampus, 1-in-6 series of GFP-immunostained sections were sampled. Using Neurolucida (MBF Bioscience, Williston, VT, U.S.A.) and a microscope equipped with a 10× objective, contours were drawn around the hilus, which was defined by its border with granule cell layer and straight lines drawn from the ends of granule cell layer to the proximal end of the CA3 pyramidal cell layer. The entire hilar area was sampled. Total section thickness was used for dissector height, and using a 100× objective only labeled somata not cut at upper surfaces of sections were counted. An average of 127 GFP-positive somata were counted in 25 sections per mouse. The total number of GFP-positive hilar neurons per dentate gyrus was estimated by multiplying counted neurons by a factor for section sampling (6). For GFP-positive hilar neuron counts of all mice, the coefficient of variation (0.40) was much larger than the mean coefficient of error (0.07), suggesting sufficient sampling within subjects (West et al., 1991).

The total length of GFP-positive axon in the granule cell layer plus molecular layer per dentate gyrus was estimated using a stereologic approach. Starting at random points near the septal pole of the hippocampus, 1-in-12 series of GFP-immunostained sections were sampled. Contours were drawn around the granule cell layer plus molecular layer in each section, and areas were recorded. Using Stereo Investigator (MBF Bioscience), sample points were determined randomly and systematically. Counting grids were 350 × 350 μm, and counting frames were 25 × 25 μm. All GFP-positive axons within counting frames and throughout the entire depth of sections were reconstructed using a 100× objective, and cumulative length at each sample site was recorded. An average of 19,074 μm of axon at 38 sample sites was measured in 13 sections per mouse. Total axon length per dentate gyrus (granule cell layer plus molecular layer) was estimated by multiplying measured axon length by factors for section sampling (12) and area sampling, which was the total area of the granule cell layer plus molecular layer divided by the analyzed area (number of counting frames per section times counting frame area). For GFP-positive axon lengths of all mice, the coefficient of variation (0.38) was much larger than the mean coefficient of error (0.07), suggesting sufficient sampling within subjects (West et al., 1991).

Numbers of GFP-positive axon crossings of the granule cell layer were measured in a section from the mid-septotemporal level of the hippocampus from each mouse. Using Neurolucida and a microscope equipped with a 100× objective, a line was drawn along the middle of the granule cell layer starting at the tip of the superior blade and extending to the tip of the inferior blade. All GFP-positive axon crossings were counted. Number of crossings per length of the granule cell layer was calculated. A total of 3,056 crossings were counted in 18 sections from 18 mice.

Chemicals were from Sigma-Aldrich (St. Louis, MO, U.S.A.) unless specified otherwise. SigmaStat (Systat Software, Chicago, IL, U.S.A.) was used for statistical analyses.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References

The pattern of GFP-immunoreactivity in the dentate gyrus of control GIN mice was similar to that reported previously (Oliva et al., 2000; Zhang et al., 2009). GFP-positive somata were located almost exclusively in the hilus (Fig. 1A1), which also contained axons and dendrites covered with long, simple spines. Axon collaterals concentrated most in the outer molecular layer (Fig. 1A2). Within the molecular layer, GFP-positive dendrites were occasionally observed but were distinguished from axons by their larger diameter and spines (Fig. 2). The general pattern of GFP-immunoreactivity was similar in mice that had experienced status epilepticus, but there were differences compared to controls. In both vehicle- and rapamycin-treated mice there appeared to be fewer GFP-positive hilar neurons than in control mice (Fig. 1B1,C1). And in vehicle-treated mice the dentate gyrus appeared larger and GFP-positive axons distributed more evenly across the inner and outer molecular layer (Fig. 1B) than in rapamycin-treated and control mice.

image

Figure 1.   GFP-immunostaining of the dentate gyrus in a control mouse (A) and in mice that had experienced pilocarpine-induced status epilepticus and then were treated daily for 2 months with vehicle (B) or 3 mg/kg rapamycin (C). Asterisks in left panels indicate areas shown at higher magnification in right panels. m = molecular layer, g = granule cell layer, h = hilus, CA3 = CA3 pyramidal cell layer.

Download figure to PowerPoint

image

Figure 2.   GFP-positive axons could be distinguished from a dendrite (arrows) in the dentate gyrus molecular layer in a vehicle-treated mouse that experienced status epilepticus.

Download figure to PowerPoint

Numbers of GFP-positive hilar neurons per dentate gyrus were estimated using stereologic methods (Fig. 3A). Control mice had 1,130 ± 69 (mean ± SEM). In mice that had experienced status epilepticus and were treated with vehicle or rapamycin, the average number of GFP-positive hilar neurons was reduced to only 48% and 55% of controls, respectively [p < 0.001, analysis of variance (ANOVA) with Student-Newman-Keuls method]. There was no significant difference between vehicle- and rapamycin-treated mice (p = 0.39). These findings suggest that in both vehicle- and rapamycin-treated mice that experienced pilocarpine-induced status epilepticus, many hilar somatostatin-immunoreactive interneurons were killed.

image

Figure 3.   Number of GFP-positive hilar neurons, length of GFP-positive axon in the dentate gyrus, and related parameters of control mice and mice that experienced pilocarpine-induced status epilepticus and then were treated daily for 2 months with vehicle or 3 mg/kg rapamycin. (A) Vehicle- and rapamycin-treated mice had fewer GFP-positive hilar neurons compared to controls (*p < 0.001, ANOVA with Student-Newman-Keuls method). There was no significant difference between vehicle- and rapamycin-treated mice. Error bars indicate SEM. Each group consisted of six mice. (B) Vehicle-treated mice had more GFP-positive axon per dentate gyrus compared to control and rapamycin-treated mice (*p < 0.001). There was no significant difference between control and rapamycin-treated mice. (C) Average GFP-positive axon density was slightly elevated in vehicle-treated mice but not significantly different among experimental groups (p = 0.068). (D) Average number of GFP-positive axon crossings of the granule cell layer was similar among experimental groups. (E) GFP-positive axon length correlated with number of granule cells per dentate gyrus (p = 0.006, R2 = 0.385). (F) GFP-positive axon length correlated with volume of the dentate gyrus (p < 0.001, R2 = 0.871).

Download figure to PowerPoint

Total length of GFP-positive axons in the molecular layer plus granule cell layer was estimated using stereological methods (Fig. 3B). Control mice had 33.9 ± 2.9 m (range = 23.5–42.3 m). Mice that experienced status epilepticus and were treated for 2 months with vehicle had over twice that of controls (range = 61.9–87.7 m) (p < 0.001, ANOVA with Student-Newman-Keuls method). These findings are consistent with a previous report that surviving somatostatin/GFP-positive hilar interneurons sprout axons after pilocarpine-induced status epilepticus (Zhang et al., 2009). Absolute values of GFP-positive axon length are larger in the present study, probably because tissue was sectioned more thinly (see Methods), which facilitated immunostaining and axon reconstruction for measurement. However, relative increases of pilocarpine-treated versus control mice were similar in both studies (220% of controls in the present analysis; 192% of controls in the previous study). Mice that experienced status epilepticus and were treated with rapamycin for 2 months had 37.2 ± 2.2 m (range 30.6–43.1 m) of GFP-positive axon per molecular layer plus granule cell layer, which was not significantly different from controls (p = 0.48) but was less than that of vehicle-treated mice (p < 0.001). These findings suggest rapamycin suppressed axon sprouting by GFP-positive interneurons after pilocarpine-induced status epilepticus.

Changes in total axon length per dentate gyrus could be attributable to changes in axon density, changes in total volume of the dentate gyrus, or both. To measure axon density, the average length of GFP-positive axon in a volume of tissue 1 μm2 and the thickness of a section (40 μm) was calculated. Average axon density was 0.76 ± 0.04 in control mice, 0.91 ± 0.07 in mice that experienced status epilepticus and were then treated with vehicle, and 0.74 ± 04 in mice that experienced status epilepticus and were then treated with rapamycin (Fig. 3C). The average in vehicle-treated mice was approximately 1.2 times that of the other groups, but the difference was not significant (p = 0.068, ANOVA). Average numbers of GFP-positive axon crossings of the granule cell layer in sections from the mid-septotemporal level were similar in control (0.11 ± 0.01 crossings per 1-μm-length of granule cell layer), vehicle-treated (0.11 ± 0.01 crossings), and rapamycin-treated mice (0.10 ± 0.01 crossings) (Fig. 3D), despite fewer interneurons in mice that had experienced status epilepticus.

The dentate gyrus enlarges after pilocarpine-induced status epilepticus in mice, and this effect is suppressed by rapamycin (Zhang et al., 2009). Tissue utilized in the present study was collected as part of a previous study that used stereologic methods to measure dentate gyrus volume (Buckmaster & Lew, 2011). For mice in the present study, the volume of the molecular layer plus granule cell layer was 1.81 ± 0.07 mm3 in controls, 3.16 ± 0.15 mm3 in mice that experienced status epilepticus and were then treated with vehicle, and 2.00 ± 0.09 mm3 in mice that experienced status epilepticus and then were treated with rapamycin. The value for the vehicle-treated group was larger than that of rapamycin-treated and control mice (p < 0.001, ANOVA with Student-Newman-Keuls method). There was no significant difference between control and rapamycin-treated groups (p = 0.22). Increased dentate gyrus volume could be attributable to granule cell proliferation and/or enlargement. Use of previous data (Buckmaster & Lew, 2011) for the animals in the present study revealed that control mice had 416,000 ± 13,000 granule cells per dentate gyrus. Vehicle- and rapamycin-treated mice that had experienced pilocarpine-induced status epilepticus had 586,000 ± 22,000 and 565,000 ± 26,000 granule cells, respectively, which was >135% of controls (p < 0.001, ANOVA with Student-Newman-Keuls method). There was no significant difference between vehicle- and rapamycin-treated groups (p = 0.51). Regression analyses revealed that GFP-positive axon length correlated with granule cell number (p = 0.006, ANOVA) (Fig. 3E) and dentate gyrus volume (p < 0.001) (Fig. 3F). The correlation was stronger with dentate gyrus volume (R2 = 0.871) than with granule cell number (R2 = 0.385).

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References

The principal finding of the present study is that in GIN mice, rapamycin suppressed the increase in GFP-immunoreactive axon length that normally occurs after pilocarpine-induced status epilepticus. This result suggests epilepsy-related axon sprouting by surviving somatostatin-positive interneurons depends in part on activation of the mTOR signaling pathway.

In GIN mice, 16% of somatostatin-positive neurons in the dentate hilus express GFP, and virtually all GFP-immunoreactive neurons express somatostatin (Oliva et al., 2000). Somatostatin-positive hilar neurons normally form many inhibitory synapses with granule cells. They account for more than half of the GABAergic neurons in the hilus (Houser & Esclapez, 1996; Buckmaster & Jongen-Rêlo, 1999; Austin & Buckmaster, 2004). Their axons concentrate in the outer two thirds of the molecular layer (Bakst et al., 1986) and synapse with granule cell dendrites (Leranth et al., 1990; Katona et al., 1999). Individual somatostatin hilar neurons have axon projections that cover long septotemporal spans (Boyett & Buckmaster, 2001), are dense, and form many synapses (Buckmaster et al., 2002). Therefore, loss of somatostatin-positive hilar neurons in temporal lobe epilepsy is likely to substantially reduce inhibitory synaptic input to granule cells. However, surviving interneurons sprout axons (Mathern et al., 1995; Zhang et al., 2009), resulting in even more extensive interneuron-to-granule cell connectivity than under control conditions (Babb et al., 1989; Thind et al., 2010). Rapamycin’s suppressive effect on GFP-positive axon length (but not granule cell proliferation) probably reduced the number of GABAergic synapses per granule cell compared to vehicle-treated mice.

GFP-positive axon length per dentate gyrus more than doubled after pilocarpine-induced status epilepticus in vehicle-treated mice, despite loss of approximately half of the hilar GFP-positive neurons. In contrast, GFP-positive axon length in rapamycin-treated mice was similar to that of controls. These findings suggest that the suppressive effect of rapamycin on GFP-positive axon extension after pilocarpine-induced status epilepticus is substantial. However, as with mossy fiber sprouting (Buckmaster & Lew, 2011), rapamycin treatment did not appear to completely block interneuron axon sprouting. If it had, one would expect less not equal amounts of axon as in controls. In addition, numbers of GFP-positive axon crossings per length of granule cell layer were not reduced in rapamycin-treated mice compared to controls as one would expect after interneuron loss, suggesting that some axon sprouting persisted.

It is unclear how rapamycin inhibited reactive axon sprouting. One possibility is a direct effect on somatostatin interneurons. Another possibility is an indirect mechanism, for example through the target of interneuron axons. Rapamycin did not reduce granule cell proliferation but inhibited dentate gyrus enlargement. Therefore, granule cell dendritic area probably was reduced in rapamycin- versus vehicle-treated mice that had experienced status epilepticus. Average axon density was slightly but not significantly increased in vehicle-treated mice that had experienced status epilepticus compared to rapamycin-treated and control animals. In addition, GFP-positive axon length correlated closely with dentate gyrus volume. Together, these findings are consistent with a homeostatic mechanism whereby surviving somatostatin interneurons adjust their level of axon sprouting according to their target area (i.e., granule cell dendrites).

Clearly, the mTOR signaling pathway regulates multiple intracellular processes (Swiech et al., 2008), and rapamycin’s suppressive effects on epilepsy-related inhibitory axon sprouting are not selective. In other neurons, rapamycin inhibits axon regeneration (Verma et al., 2005; Park et al., 2008), and conversely, activation of the mTOR signaling pathway stimulates axon growth (Choi et al., 2009; Grider et al., 2009). Together with results of the current study and previous work on mossy fiber sprouting (Buckmaster et al., 2009; Zeng et al., 2009; Buckmaster & Lew, 2011), these findings suggest that activation of the mTOR signaling pathway might contribute to axon sprouting by many different types of neurons. “Side effects” of rapamycin will require careful interpretation in future experiments. Nevertheless, results of the present study suggest that the mTOR signaling pathway might be a useful drug target for manipulating and learning more about consequences of epilepsy-related changes in GABAergic circuits.

Disclosure

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Disclosure
  8. References