Seizures, cell death, and mossy fiber sprouting in kainic acid-treated organotypic hippocampal cultures*

Mossy fiber sprouting in vitro

Pergamon PII: S0306-4522(99)00358-9

Neuroscience Vol. 94, No. 3, pp. 755–765, 1999 755 Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00

SEIZURES, CELL DEATH, AND MOSSY FIBER SPROUTING IN KAINIC ACID-TREATED ORGANOTYPIC HIPPOCAMPAL CULTURES* M. J. ROUTBORT,*† S. B. BAUSCH*†‡ and J. O. McNAMARA†‡§ k ¶ Departments of †Neurobiology, ‡Medicine (Neurology), §Pharmacology and kMolecular Cancer Biology, Duke University Medical Center, Durham, NC 27710, U.S.A.

Abstract—Sprouting of the mossy fiber axons of the dentate granule cells is a structural neuronal plasticity found in the mature brain of epileptic humans and experimental animals. Mossy fiber sprouting typically arises in experimental animals after repeated seizures and may contribute to the hyperexcitability of the epileptic brain. Investigation of the molecular triggers and spatial cues involved in mossy fiber sprouting has been hampered by the lack of an optimal in vitro model for studying this rearrangement. For an in vitro model to be feasible, the circuitry and receptors involved in convulsant-induced mossy fiber sprouting would have to be localized near the granule cells, rather than being dependent on long-range brain interconnections. However, it is not known whether this is the case. We report here that that application of the convulsant, kainic acid, to organotypic hippocampal explant cultures induces seizures, neuronal cell death, and subsequent dramatic mossy fiber sprouting with a similar laminar preference and time-course to that seen in intact animals. Prolonged (48 h) but not transient (4 h) kainic acid treatment caused regionally selective neuronal cell death. Cultures treated with kainic acid for a prolonged period displayed a time- and dose-dependent increase in supragranular Timm staining reflective of increased mossy fiber innervation to this area. Direct visualization of mossy fiber axons with neurobiotinlabeling revealed that mossy fibers in kainic acid-treated cultures exhibited a dramatic increase in supragranular axonal branch points and synaptic boutons. The cellular and molecular determinants required for kainic acid-induced cell death and subsequent mossy fiber reorganization thus appear to be intrinsic to the hippocampal slice preparation, and are preserved in culture. Given the ease with which functional inhibitors or pharmacological agents may be utilized in this system, slice cultures may provide a powerful model in which to study the molecular components involved in triggering mossy fiber outgrowth and underlying its laminar specificity. Elucidation of these molecular pathways will likely have both specific utility in clarifying the functional consequences of mossy fiber sprouting, as well as general utility in understanding of synaptic reorganization in the mature central nervous system. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: mossy fibers, sprouting, seizures, kainic acid, slice cultures, organotypic.

Synaptic reorganization following brain injury is likely to contribute to the expression of hyperexcitability and epilepsy. A robust and well-characterized structural plasticity—mossy fiber sprouting—has been identified in the hippocampal formation of humans with temporal lobe epilepsy as well as in many animal seizure models. 12,16,23 Mossy fiber synaptic terminals contain large quantities of zinc and can be readily identified at the light microscopic level by a histochemical method, Timm staining. A striking increase of Timm-stained terminals of the mossy fiber axons of the granule cells is evident in the molecular layer of the dentate gyrus in epileptic but not in normal brains; 16,20,23 the laminar specificity of this projection is exquisite in that it is mainly confined to the inner portion of the molecular layer. This represents an enormous expansion of a normally very small projection. 9,17 When examined with methods that directly reveal axonal morphology, recurrent mossy fiber collaterals are frequently found in the dentate inner molecular layer in epileptic brains. 8,17 The functional consequences of mossy fiber sprouting are incompletely understood, but some evidence indicates that the abnormal circuits created by mossy fiber sprouting may

contribute to the hyperexcitability found in epilepsy and animal seizure models. 6,25 The reorganization evident in mossy fiber sprouting is one of the most striking examples of structural plasticity seen in the mature brain. Unfortunately, while mossy fiber sprouting has been wellcharacterized anatomically, very little is known about the molecular determinants involved in triggering the sprouting or establishing the marked laminar specificity of the reorganized axons. In part, this has been due to the difficulty of inhibiting the function of specific molecular pathways in the intact brain short of generating knock-out animals. A relatively simple and pharmacologically accessible model to examine mossy fiber sprouting would facilitate the identification and testing of such candidate molecules. In particular, a tissue culture model with limited diffusion constraints would facilitate testing of candidate molecules through the use of inhibitory reagents such as blocking antibodies, peptides, or antisense oligonucleotides. Organotypic hippocampal slice cultures can display an impressive degree of morphological and synaptic preservation, in a system that allows easy pharmacological manipulation. 7 However, it has not been established whether seizure-associated mossy fiber sprouting occurs in hippocampal slice cultures. A critical issue is whether the spatial and cellular mechanisms involved in coordinated convulsant-induced mossy fiber rearrangement are local to the hippocampal formation or require longer-range neuronal interactions. In intact animals, administration of kainic acid (KA) is

*M. J. Routbort and S. B. Bausch contributed equally to the work described in this paper. ¶To whom correspondence should be addressed. Abbreviations: BSA, bovine serum albumin; DAB, 3,3 0 -diaminobenzidine; DIV, days in vitro; EGTA, ethyleneglycolbis(aminoethylether)tetraacetate; HEPES, N-2-hydroxyethylpiperazine-N 0 -2-ethanesulfonic acid; KA, kainic acid; PB, phosphate buffer; PBS, phosphate-buffered saline. 755

756

M. J. Routbort et al.

under each insert in a six-well plate. Medium composition was 50% minimum essential medium, 25% Hanks’ balanced salt solution, 25% horse serum, 25 mM HEPES, pH 7.2, supplemented with glutamine [0.125 ml GlutaMaxII (Gibco) per 100 ml] and 6.5 mg/ml glucose. Medium was exchanged every three days. For KA treatment, different concentrations of KA (Sigma) diluted in medium were applied to cultures at 10 days in vitro (DIV). Unless noted in the text, a concentration of 6–7 mM KA was used. After treatment, inserts were transferred to fresh medium and subsequently transferred to fresh medium at 1 h and 24 h intervals thereafter to ensure wash-out of residual KA. Subsequently, both vehicle and KA-treated cultures were returned to the normal three-day feeding schedule. Toluidine Blue and Timm staining

Fig. 1. Acute application of KA to hippocampal slice cultures causes seizures in the granule cell layer. Extracellular field potentials were recorded in the dentate granule cell layer of hippocampal slice cultures (10 DIV) as described in the Experimental Procedures. (A, top) Traces from a representative recording show the transient increase in spike frequency that was observed in all cultures (n ˆ 5) 4.8 ^ 0.5 to 10.9 ^ 1.2 min following the beginning of acute KA (6 mM) application. (B) The same trace with an expanded time-scale reveals that this spiking pattern was ictal-like in frequency and progression. (A, bottom) Spiking resumed following a “quiet” period of 17 ^ 4.8 min and persisted for the remainder of the recording. (C) The same trace with an expanded time-scale shows a repeating pattern of epileptiform activity. Symbols indicate a break in the trace: *11 min; 150 s; #100 s. Scale bars in C apply to B, C.

associated with seizures, prominent neuronal cell death, and subsequent mossy fiber sprouting. 1,14,16 Using interface-style cultures, which generally exhibit excellent preservation of hippocampal anatomy, we confirm and extend previous observations that KA causes seizures and neurotoxicity in slice cultures. 2,21 We also demonstrate that KA treatment triggers robust mossy fiber sprouting in these cultures with laminar specificity similar to that evident in vivo. Thus, the phenomenon of seizure-associated mossy fiber sprouting can be induced and potentially manipulated in a relatively simple in vitro culture preparation. EXPERIMENTAL PROCEDURES

For Toluidine Blue staining, cultures on membrane inserts were fixed for 15 min in 3% glutaraldehyde in 100 mM phosphate buffer, pH 7.4 (PB). After rinsing with tap water, cultures on inserts were permeabilized for 5 min with 0.5% Triton X-100, rinsed again with tap water, and stained for 5 min in 0.5% Toluidine Blue. After graded alcohol washes, inserts were briefly dipped in 95% ethanol containing 1% glacial acetic acid. Cultures on inserts were then rehydrated and stored in 10% glycerol until mounting. For Timm staining, cultures on membrane inserts were initially treated with 1 ml 1% sodium sulfide in PB for 10 min. Following gentle addition of 2 ml of 0.3% glutaraldehyde in PB to the sodium sulfide solution, cultures on inserts were fixed for 5 min. Inserts were then washed with PB and postfixed overnight at 48C in 70% ethanol. Cultures on inserts were washed with PB and incubated in Timm development solution for 30–40 min in the dark. Timm development solution contained, per 50 ml: 30 ml of a 50% gum arabic solution, 15 ml of a 0.5 M hydroquinone solution, 5 ml citrate buffer solution (1.2 M citric acid, 0.8 M trisodium citrate), and 250 ml of a 1 M silver nitrate solution. After washing in tap water, cultures were stored in 10% glycerol until mounting. After either Toluidine Blue or Timm staining, the membrane was removed from the tissue culture inserts using a scalpel, and the membrane was applied culture-side down onto gelatin-subbed slides and these assemblies were then allowed to completely air dry. Once dry, slides were placed into 95% ethanol and membranes were pulled off, leaving the cultures attached to the glass slides. Cultures on slides were then dehydrated, cleared, and coverslipped. Timm scoring Both dentate gyrus blades were scored independently by a blinded observer using a Nikon SMZ-1 stereomicroscope and these two values were then averaged to yield a single score for each culture. Individual cultures were excluded from analysis if there was a complete or nearly total loss of the general mossy fiber pathway staining pattern or if there was obviously aberrant cell layer loss (other than loss of CA3 in kainate-treated cultures). Because granule cell toxicity and overall faintness of the Timm staining pattern were often observed in KA-treated cultures, the density of hilar staining was used as a reference in scoring supragranular mossy fiber staining according to the following scheme: 0—no granules in supragranular region; 1—faint hint of granules in supragranular region (subconfluent to barely confluent), much lighter than hilus; 2—distinct band of granules, light in comparison to hilus; 3—obvious band of confluent granules, beginning to approach the density of the hilus; 4—dense band of confluent granules, nearly as dark as the hilus; 5—dense, thick band of confluent granules, as dark as the hilus. In related experiments that were quantitated by two independent observers to assess the reproducibility of this scoring protocol (M. R. and J. M.), correlation in Timm scores was quite high, with over 98% of Timm scores made by independent observers within one point of each other (256/261 culture scores, Spearman non-parametric rank order correlation r ˆ 0.89, P , 0.0001).

Organotypic hippocampal cultures and kainic acid treatment Slice cultures were prepared and maintained using the interface method. 22 Hippocampi from postnatal day 10–12 rat pups (P10– P12) were dissected and cut into 400-mm slices using a McIlwain tissue chopper. Only slices from the middle third of the hippocampus were used for culturing. These slices were placed on top of semipermeable membrane inserts (Millipore) and 1.2 ml of medium was placed

Cell loss scoring Dentate granule cell and pyramidal cell layers were scored by a blinded observer using an Axiovert 135 microscope. Cell layer density was scored using the following scheme: 0—no stained neurons in cell layer; 1—very sparse number of scattered stained neurons; 2—sparse number of stained neurons; 3—many prominently stained neurons.

Mossy fiber sprouting in vitro

757

or “spike” which immediately followed the stimulus artifact with a response threshold # 100 mA. The basis for concluding that this waveform was a field action potential was that the waveform could be abolished with tetrodotoxin (1 mM; Calbiochem; data not shown). Given the very short latency and lack of an underlying field excitatory postsynaptic potential, this field action potential was most likely due to the antidromic stimulation and the subsequent synchronous firing of a population of dentate granule cells. Data were collected using an Axopatch 1D amplifier (2 kHz analog filter) and Axotape (extracellular field recordings, 3.33–10 kHz acquisition rate) or pClamp software. For analyses of electrophysiological recordings, the following definitions were used: (i) interictal spikes—single spikes or spikes superimposed on a positive field potential shift, which were # 80 ms in duration (see Fig. 2A); (ii) electrographic seizures—the abrupt onset of a burst of rhythmic activity ( $ 3 s in duration) whose waveform changed over time and terminated abruptly; and (iii) epileptiform activity—bursts of rhythmic spikes or spikes superimposed on positive field potential shifts which were $ 80 ms in duration, but that did not fit the criteria for seizures (see Fig. 1C).

Dentate granule cell morphology

Fig. 2. Twenty-four to 48 h application of KA to hippocampal slice cultures causes interictal spikes and epileptiform activity in the granule cell layer. Extracellular field potentials were recorded in the dentate granule cell layer of hippocampal slice cultures (A, 11 DIV or B, 12 DIV) treated with 6 mM KA for 24 h (A) or 48 h (B) as described in the Experimental Procedures. Cultures were taken from culture media containing 6 mM KA and placed directly into recording buffer containing 6 mM KA. (A1) A trace from a representative recording shows the spiking pattern that was observed in all cultures (n ˆ 6) approximately 24 h after the beginning of KA application. (A2) The same trace with an expanded time-scale reveals an interictal spiking pattern. (B1) A trace from a representative recording shows the extended periods of high-frequency spiking followed by extended periods of low-frequency spiking that were observed in 33% of cultures (n ˆ 6) approximately 48 h after the beginning of KA application. (B2) The same trace with an expanded time-scale reveals epileptiform activity. *Indicates a 30 min break in the trace. Vertical scale bar in A1 applies to all traces. Horizontal scale bar in A1 applies to A1, B1; in A2 applies to A2, B2.

Neurons were differentiated from glia based upon their larger nuclear size. The CA3c region was defined as the CA3 pyramidal cell layer located between the blades of the dentate granule cell layer; the CA3a/b region was defined as the CA3 pyramidal cell layer excluding the CA3c region. Electrophysiological recording Recordings were performed immediately prior to or during treatment with KA (at 10–12 DIV). A portion of the tissue culture insert membrane containing a single slice culture was cut and the membrane and slice culture were placed into a recording chamber mounted to a Zeiss Axioskop microscope. The membrane was held down with a platinum wire and slice cultures were superfused at room temperature with a recording buffer composed of (in mM): NaCl 120, KCl 3.5, MgSO4 1.3, CaCl2 2.5, NaH2PO4 1.24, NaHCO3 25.6, glucose 10 equilibrated with 95% O2/5% CO2. KA (6 mM) was diluted immediately before use in recording buffer and applied by bath superfusion. Recording pipettes were pulled on a Flaming-Brown puller and filled with 3 M NaCl for extracellular field potential recordings. Dentate granule cell layer field potential recordings were deemed acceptable if hilar stimulation (0.3 ms square pulse, 0.03 Hz, 20– 700 mA) using a concentric bipolar electrode (MCE-100, Rhodes Medical Supply) and a Grass stimulator yielded an action potential

Individual granule cells from the middle portion of the suprapyramidal blade of the dentate gyrus were filled with neurobiotin (Vector) during whole-cell current-clamp using the “blind” technique. 3 Methods were similar to those described above for electrophysiological recordings, except that (i) electrodes (8–12 MV) were filled with (in mM): K-gluconate 100, HEPES 40, EGTA 0.5, MgCl2 5, Na2ATP 2, Na2GTP 0.3 and 0.5% neurobiotin (pH 7.2 with KOH, 260 mOsm before addition of neurobiotin), and (ii) recording buffer was composed of (in mM) NaCl 119, KCl 2.5, MgCl2 1.3, CaCl2 2.5, NaH2CO3 26.2, glucose 11 equilibrated with 95% O2/5% CO2. Granule cells were filled for 5 min after establishing whole-cell configuration. The pipette was then rapidly withdrawn, and cultures were placed back onto a tissue culture insert with growth medium, and put into a 378C, 5% CO2 incubator for 25 min to allow the neurobiotin to be transported axonally. Cultures were then fixed for 30 min in 4% paraformaldehyde in phosphate buffered-saline (PBS; PB containing 0.9% NaCl). After washing 30 min in PBS and subsequent cryoprotection with 30% sucrose in PBS, cultures were frozen and stored at 2708C until development. To visualize neurobiotin, cultures were thawed, washed in PBS, and endogenous peroxidase was quenched for 10 min in 10% methanol containing 0.2% H2O2. Cultures were then permeabilized in 2% bovine serum albumin (BSA) and 0.75% Triton X-100 in PBS for 1 h, rinsed in 2% BSA in PBS, and incubated with ABC reagent (Vectastain elite [Vector] prepared according to kit instructions in 2% BSA in PBS) overnight at 48C. Cultures were washed, pre-incubated for 10 min in a 3,3 0 -diaminobenzidine solution (DAB solution; 10 mg/ml 3,3 0 -DAB [Sigma], 0.03% cobalt chloride, 0.02% nickel ammonium sulfate in PBS), and developed in DAB solution containing 0.0008% H2O2 until labeling was evident. Granule cells exhibiting complete dendritic and axonal labeling without obvious filling defects were included in a blinded data analysis. As a quantitative index of mossy fiber sprouting, axonal branch points and varicosities (defined as areas of swelling with a diameter greater than twofold larger than the axonal diameter) were counted with the aid of a camera lucida attached to a Zeiss Axioskop microscope. The hilar/granule cell layer boundary was generally very well defined and visible even in cultures without counterstaining, whereas the granule cell layer/molecular layer boundary was usually much less distinct. Branch points and varicosities within the granule cell layer and molecular layer were therefore counted together.

Statistical analysis Numbers represent the mean or the median and error bars represent S.E.M. in the stated number of slice cultures. All statistical analyses were performed with Sigma Stat or InStat software. Data fitting a nonparametric distribution were tested for significance using the Kruskal– Wallis ANOVA by ranks test with Dunn’s post hoc comparison when comparing multiple groups or Mann–Whitney Rank Sums test when comparing two experimental groups. Data fitting a normal parametric distribution were tested for significance using a t-test.

758

M. J. Routbort et al.

KA for 48 h showed more variable results: 50% of granule cell layer field recordings (n ˆ 6) showed interictal spiking similar to that seen at 24 hours. Thirty-three percent of cultures showed extended periods of frequent epileptiform activity followed by extended periods of low frequency spiking (Fig. 2B); granule cell layer recordings from the remaining 17% of cultures displayed epileptiform activity and multiple electrographic seizures (Fig. 3). No epileptiform activity or seizures were observed in granule cell field layer recordings from vehicle-treated cultures in normal recording buffer without KA present (n ˆ 8). In summary, application of KA to hippocampal slice cultures at 10–12 DIV caused both seizures and other forms of epileptiform activity immediately after treatment and caused diverse forms of epileptiform activity when studied after 24 and 48 h of exposure. Long-term kainic acid treatment caused neuronal death

Fig. 3. Forty-eight hour application of KA to hippocampal slice cultures causes seizures in the granule cell layer. Extracellular field potentials were recorded in the dentate granule cell layer of hippocampal slice cultures (12 DIV) treated with 6 mM KA for 48 h as described in the Experimental Procedures. Cultures were taken from culture media containing 6 mM KA and placed directly into recording buffer containing 6 mM KA. (A) A trace from a representative recording shows the brief periods of highfrequency spiking followed by extended periods of very low-frequency spiking that were observed in 17% of cultures (n ˆ 6) approximately 48 h after the beginning of KA application. (B) One period of high frequency spiking from the same trace shown on an expanded time-scale. (C, D) An expanded time-scale of the high-frequency spiking shown in B reveals two ictal-like events representing the tonic part of the seizure. (E) An expanded time-scale of the trace B (taken 30 s following the trace in D) reveals epileptiform activity representing the clonic part of the seizure. (F) An expanded time-scale of what appeared to be an interictal spike in trace A reveals epileptiform activity. Scale bars in C apply to C–F. RESULTS

Acute application of kainic acid to hippocampal slice cultures caused seizures Administration of KA in vivo causes intense limbic and tonic-clonic motor seizures with characteristic electroencephalographic patterns of seizure activity. To examine the electrophysiological effects of acutely applied KA on slice cultures, extracellular field recordings were performed at three points after KA application: immediately, or after 24 or 48 h of exposure. Electrographic seizure activity was observed in the granule cell layer of all cultures (n ˆ 5) within 5 min of acute KA application (6 mM) and persisted for approximately 5 min (Fig. 1A top, 1B). Following a 15– 20 min “quiet” period with little spike activity, epileptiform activity appeared and generally persisted to the end of the recording (Fig. 1A bottom, 1C). After 24 h of treatment, a relatively frequent occurrence of interictal spikes or epileptiform activity (per definition in Experimental Procedures) was observed in granule cell layer field recordings from all slice cultures (n ˆ 6) incubated in KA and placed directly into recording buffer containing KA (Fig. 2A). Recordings from slice cultures treated with

In intact animals, KA treatment causes profound neuronal cell death. 18 CA3 pyramidal cells are particularly sensitive to KA-induced neurotoxicity. 14 Toluidine Blue staining was used to assess KA-induced cell death in slice cultures. Cultures treated with vehicle for 48 h at 10 DIV and then examined at 17–60 DIV generally (96%; n ˆ 102 slice cultures) exhibited a normal distribution of hippocampal pyramidal cell layers (Fig. 4A, Table 1). The remaining vehicle-treated cultures displayed a sparser population of either CA1 (2% of cultures) or CA3a/b (2% of cultures) pyramidal cells. Vehicle-treated cultures also exhibited many prominently stained neurons in the dentate granule cell layer (Fig. 4A). However, the infrapyramidal granule cell layer and granule cell layer at the apex were generally sparser than the suprapyramidal layer (Table 1). A 4 h treatment with KA at 10 DIV caused no obvious neuronal cell death (Fig. 4B, cell loss scores not significantly different from vehicle-treated cultures; data not shown). In contrast, a 48 h treatment with KA caused profound neuronal cell death (Fig. 4C). The CA3a/b pyramidal cell layer was completely destroyed in 92% of cultures treated with KA for 48 h (n ˆ 102, Table 1). Cell loss in the CA3c region was much more variable, and was completely destroyed in only 48% of cultures treated with KA for 48 h (n ˆ 105). A modest loss was observed in 21% of cultures and no apparent loss of CA3c pyramidal cells (as determined by the presence of many prominently stained neurons) was seen in 33% of cultures treated with KA for 48 h. Long-term KA treatment also decreased granule cell density when compared to vehicletreated cultures (Table 1). However, similarly to vehicletreated cultures, cultures treated with KA for 48 h displayed a region-specific pattern of dentate granule cell density; the infrapyramidal granule cell layer and granule cell layer at the apex were sparser than the suprapyramidal layer (Table 1). No significant cell loss was observed in the CA1 pyramidal cell layer of cultures treated with KA for 48 h. In summary, long-term (48 h) but not short-term (4 h) treatment of slice cultures with KA caused regionally selective neuronal cell death that was most prominent in CA3a/b. Long-term kainic acid treatment caused progressive mossy fiber sprouting Animals that have experienced KA-induced status epilepticus

Mossy fiber sprouting in vitro

759

Fig. 4. Forty-eight hour but not 4 h application of KA to hippocampal slice cultures causes neuronal death and mossy fiber sprouting. Vehicle-treated cultures (A, D) and cultures treated with KA for 4 h (B, E) or 48 h (C, F) were stained with Toluidine Blue (A–C) or the Timm stain (D–F) as described in Experimental Procedures. Representative vehicle (A) and 4 h KA-treated (B) cultures showing normal slice culture neuronal laminar organization. (C) Representative 48 h KA-treated culture showing nearly complete destruction of CA3. Representative vehicle (D) and 4 h KA-treated (E) cultures showing normal slice culture Timm staining pattern involving dark mossy fiber staining in the hilus and along stratum lucidum, with a lighter amount of supragranular staining in the inner portion of the molecular layer (arrowhead). (F) Representative 48 h KA-treated culture showing strongly increased Timm staining in the inner molecular layer. Scale bar ˆ 100 mm. GCL, dentate granule cell layer; H, hilus; HF, hippocampal fissure; IML, inner molecular layer; OML, outer molecular layer; SL, stratum lucidum.

subsequently develop profound mossy fiber sprouting. 1,5,16,24 To determine whether this form of morphological plasticity is also seen in slice cultures, vehicle-treated cultures and cultures treated with KA for 4 or 48 h were processed with the Timm stain to visualize mossy fiber boutons two to three weeks after treatment (at 25–30 DIV). Timm staining of vehicle-treated cultures revealed the majority of staining to be confined to the expected distribution of the mossy fiber pathway, with dense large granular staining along stratum lucidum of CA3, and dense smaller granular staining in the dentate hilus (Fig. 4D, granular nature of staining pattern not visible at this magnification). In addition to this intense staining, a smaller amount of mossy fiber-like staining was evident in the inner molecular layer. This appearance is consistent with the presence of supragranular mossy fibers that have Table 1. Long-term kainic acid treatment in vitro caused cell loss Subfield Vehicle CA1 CA3

a/b c

GCL supra apex infra

3 3 3 3 3 2† 2†

Median Score KA (48 h) (102) (102) (102) (109)

3 0** 2** 2* 3 2† 2†

(105) (102) (105) (105)

Slice cultures treated with vehicle or 6 mM KA for 48 h at 10 DIV were stained at 17–60 DIV with Toluidine Blue as described in the Experimental Procedures. Cell layers were scored blindly as: 3, many prominently stained neurons; 2, sparser number of stained neurons; 1, very sparse number of scattered stained neurons; 0, no stained neurons in cell layer. Statistics: Scores represent the median from the number of slice cultures indicated in parentheses. *Significant difference from vehicle-treated cultures, P , 0.01; **P , 0.001 (Mann–Whitney Rank Sums test). †Significant difference from suprapyramidal granule cell layer, P , 0.05 (Kruskal–Wallis ANOVA by ranks with Dunn’s post hoc comparison). GCL, dentate granule cell layer; infra, infrapyramidal dentate granule cell layer; supra, suprapyramidal dentate granule cell layer.

been found in slice cultures, and to a lesser extent, in normal animals. 17,28 Lighter and more diffuse non-granular Timm staining was also evident throughout other, non-mossy fiber-associated areas of vehicle-treated cultures (e.g., CA1 stratum oriens and radiatum, dentate outer molecular layer; Fig. 4D, E). Treatment of slice cultures with KA for 48 h produced robust mossy fiber sprouting several weeks later as gauged by Timm staining (Fig. 4F). The overall original pattern of mossy fiber staining in the hilus and stratum lucidum persisted in cultures treated for 48 h with KA. However, the intensity of Timm staining in this areas was often attenuated, and the staining in CA3 stratum lucidum consisted of much smaller, irregular granules than were present in vehicletreated cultures. Furthermore, the non-mossy fiber associated Timm staining was often attenuated or absent. Despite decreased Timm staining in other areas of the cultures, the inner molecular layer of most cultures exhibited a clear increase in Timm staining, implying increased mossy fiber sprouting to this area. A semiquantitative scoring system was used to gauge the intensity of this inner molecular layer staining based on a single “Timm score” for each culture. Cultures treated for 48 h with KA had a median Timm score of 3.5, compared to a median score of 1 for vehicletreated cultures (P , 0.001, Kruskal–Wallis ANOVA by ranks test with Dunn’s post hoc comparison, n ˆ 49 and 35 for vehicle and 48 h KA groups, respectively). Although some heterogeneity was noted in the magnitude of the Timm scores in both the vehicle and KA-treated cultures, this semiquantitative scoring based upon a blinded analysis disclosed clear evidence of mossy fiber sprouting. The relative contribution of cell death and abnormal neuronal activity to mossy fiber sprouting is uncertain. To begin to address this issue, the effects of a briefer application of KA on mossy fiber sprouting were examined. KA uniformly induced the onset of electrographic seizures and intense epileptiform activity beginning almost immediately after application; to our surprise, KA treatment for as long as 4 h produced no overt neuronal death as detected by

760

M. J. Routbort et al.

Fig. 5. Dose-dependence and time-course of KA-induced mossy fiber sprouting. Slice cultures were treated with varying concentrations of KA at 10 DIV and Timm stained two to three weeks after treatment (A), or treated with a fixed concentration of KA and allowed to recover for varying periods of time before Timm staining (B). Timm scores were calculated as described in the Experimental Procedures. Statistics: Median Timm scores for each group are shown. *Significant difference between vehicle- (0 mM KA) and KA-treated cultures, P , 0.001 (Kruskal–Wallis ANOVA by ranks with Dunn’s post hoc comparison). n ˆ 8–49 for each data point in A and B, except for KA-treated cultures one day after treatment, where n was only 3 for reasons noted in the text.

inspection of Toluidine Blue-stained explant cultures (Fig. 4B). The dissociation afforded by a 4 h (instead of 48 h) application of KA provided an opportunity to examine the relative contributions of cell death and intense epileptiform activity to mossy fiber sprouting. Treatment with KA for 4 h produced no measurable increase of mossy fiber sprouting compared to vehicle-treated cultures as gauged by Timm scores (Fig. 4E, D). Both vehicle-treated cultures and cultures treated for 4 h with KA had a median Timm score of 1 (not significant, Kruskal–Wallis ANOVA by ranks test with Dunn’s post hoc comparison, n ˆ 49 and 11 for vehicle and 4 h KA groups, respectively). Because 4 h of KA treatment was not associated with cell death or increased mossy fiber sprouting, a 48 h treatment period was used for the remainder of experiments examining the effects of KA treatment on slice cultures. Unless otherwise specified, the term “KA treatment” will be used to refer to this longer treatment period. It is important to note that there was clear heterogeneity in the Timm scores of KA-treated cultures, and that the Timm scores of vehicle- and KA-treated cultures overlapped. For instance, the range of Timm scores in this sample of KA-treated cultures was from 1.5 to 5, while the range of Timm scores in vehicle-treated cultures was from 0 to 3. However, no vehicle-treated cultures ever had a Timm

score higher than 3, while 61% (22/36) of KA-treated cultures had Timm scores higher than 3 (data not shown). Thus, not all cultures treated with KA demonstrated intense inner molecular layer Timm staining. However, vehicle-treated cultures never showed an amount of staining comparable to the subset of KA-treated cultures that did have high Timm scores. In order to examine the concentration-dependence of KAinduced Timm score changes, cultures were treated with different concentrations of KA ranging from 0 (vehicle) to 9 mM and Timm stained two to three weeks after KA treatment. KA concentrations of 4 mM and above were associated with significantly elevated Timm scores compared to vehicletreated cultures (Fig. 5A). There was no statistically significant trend of increasing Timm scores at KA concentrations higher than 4 mM. At the highest concentration tested, 9 mM, there was a substantial increase in the number of cultures excluded from scoring due to a near total loss of Timm staining (data not shown). In Toluidine Blue-stained cultures at this higher dose of KA, the granule cell layer was occasionally dramatically damaged or nearly completely destroyed, which likely accounts for the near total loss of Timm staining in a subset of cultures treated with 9 mM KA (data not shown). Six to 7 mm KA was used for the remainder of these experiments, because it produced reliable mossy fiber sprouting and more consistent cell death than lower KA concentrations (data not shown). Mossy fiber sprouting after KA status epilepticus in intact animals progresses over time and takes weeks to months to reach its maximal levels. 16 To examine the time-course of mossy fiber sprouting after KA-treatment in slice cultures, vehicle- and KA-treated cultures were Timm stained at various time-points. The median Timm scores of vehicletreated slice cultures at different ages in vitro were consistently around 1, and did not change with time. In contrast, the median Timm scores of cultures treated with KA increased with time after treatment (Fig. 5B). Immediately after treatment with KA, only a small subset of Timm-stained cultures could be scored due to dramatic lightening of the Timm stain (data not shown), presumably due to depletion of zinc released from active synaptic terminals. The cultures that could be analysed at this time-point did not have elevated Timm scores (Fig. 5B). One week after treatment, the median Timm score of KA-treated cultures was slightly but not significantly increased from 1 to 2 (Fig. 5B). At two to three weeks and beyond KA-treatment, there was a significant increase in the Timm score of KA-treated cultures compared to time-matched controls. The highest median Timm score in these time-course experiments occurred at five weeks after treatment. In summary, long-term (48 h) but not short-term (4 h) KA treatment caused increased supragranular Timm staining in a concentration- and time-dependent manner that was consistent with the induction of profound and progressive mossy fiber sprouting. Individual granule cell morphology is markedly altered in kainic acid-treated cultures Because Timm staining is a histochemical procedure that labels a number of heavy metals, changes in synaptic zinc localization or homeostasis could cause changes in the pattern of Timm staining without reflecting an underlying axonal rearrangement such as mossy fiber sprouting. For instance,

Mossy fiber sprouting in vitro

761

Fig. 6. Representative granule cell morphology in vehicle- and KA-treated cultures. Individual granule cells were filled with neurobiotin as described in the Experimental Procedures. Representative granule cells from vehicle-treated cultures (A, B) reconstructed with computerized camera lucida system. Dendrites are shown in hatched pattern, with axons as fine black lines. Regional landmark labels are superimposed upon the reconstruction of the neuron. Synaptic varicosities are not shown. Representative granule cells from KA-treated culture that display some (C) or a very high level (D) of molecular layer sprouting. The profound degree of sprouting displayed by the granule cell in D was seen in about 1/3 of granule cells in KA-treated cultures. Scale bar ˆ 100 mm. GCL, dentate granule cell layer; HF, hippocampal fissure; ML, molecular layer.

as noted previously, Timm staining was dramatically attenuated immediately following KA treatment. It was therefore important to confirm the impression of mossy fiber sprouting revealed by Timm staining in KA-treated cultures with an independent method that would actually visualize mossy fibers. To this end, individual granule cells in slice cultures were filled with neurobiotin to visualize mossy fiber morphology directly. Neurons from the middle portion of the suprapyramidal blade of the dentate gyrus of vehicle- and KAtreated slice cultures were filled at 21–50 DIV (11–40 days after treatment). In both vehicle- and KA-treated cultures, a single main mossy fiber axon generally emerged from the basal pole of the neuron (or, rarely, from a proximal dendritic shaft) and gave off numerous branches in the hilus (Fig. 6). In vehicletreated cultures, the main mossy fiber axon usually coursed through the hilus and extended to stratum lucidum of CA3 (in 88% of neurons, n ˆ 16), where giant varicosities were present in 88% of neurons (varicosities are not shown in Fig. 6). Occasional axon collaterals in vehicle-treated cultures entered into the granule cell layer (Fig. 6A) or extended through the granule cell layer and into the molecular layer (Fig. 6B). The mean number of axon collaterals in vehicletreated cultures that penetrated into the granule cell or molecular layer was 1.1 ^ 0.25. These recurrent processes usually were simple and unbranched, with relatively few varicosities (Fig. 7A). Mossy fiber axons in KA-treated cultures had a dramatically different and more heterogeneous distribution. Mossy

fiber giant varicosities were generally absent (present in only 17% of neurons, n ˆ 18), and the main mossy fiber axon was truncated before reaching CA3a/b in 39% of KAtreated neurons (Fig. 6C). Granule cells in KA-treated cultures frequently had several different mossy fiber branches passing through the entire granule cell layer and terminating in the inner portion of the molecular layer (Fig. 6D). The mean number of axon collaterals that penetrated into the granule cell or molecular layer in KA-treated cultures was 3.7 ^ 0.35, which was significantly more than found in vehicle-treated cultures (P , 0.0002, Mann–Whitney Rank Sums Test). In some cases, a mossy fiber axon originating from the suprapyamidal blade of the dentate gyrus traversed the entire width of the hilus to penetrate through the infrapyramidal granule cell layer and molecular layer (28% of neurons, see Fig. 6C for example). This was never observed in granule cells from vehicle-treated cultures at these timepoints in culture. Recurrent mossy fibers from KA-treated cultures in the granule and molecular layers often branched prolifically and displayed a large number of varicosities (Fig. 7B). Quantitation of axonal branch points and varicosities within the granule cell and molecular layer confirmed the subjective impression of dramatic morphological changes in granule cells from KA-treated cultures. The mean number of mossy fiber axonal branch points in the granule cell/molecular layer was increased about 11-fold (from 0.94 to 10.2) in KA-treated cultures compared to vehicle-treated cultures (P , 0.0001). The mean number of mossy fiber axonal

762

M. J. Routbort et al.

Fig. 7. Individual granule cells in KA-treated cultures display marked supragranular sprouting. Individual granule cells were filled with neurobiotin and axonal branch points and varicosities in the granule cell/molecular layer were quantitated as described in the Experimental Procedures. High-power camera lucida drawing of individual mossy fibers sprouted to the granule cell/molecular layer in a vehicle-treated culture (A) and a KA-treated culture (B). The mean number of mossy fiber axonal branch points in the granule cell/molecular layer increased from 0.94 to 10.2 in KA-treated cultures compared to vehicle-treated cultures*. The mean number of mossy fiber axonal varicosities in the granule cell/molecular layer increased from 20 to 134 in KA-treated cultures compared to vehicle-treated cultures*. *Significant difference between vehicle- and KA-treated cultures, P , 0.0001 (Mann–Whitney Rank Sums test). n ˆ 16 cells from vehicle-treated cultures, n ˆ 18 cells from KA-treated cultures. Average culturing time for vehicle-treated cultures was 29 ^ 6 DIV, average age for KA-treated cultures was 27 ^ 6 DIV (not significantly different).

varicosities in the granule cell/molecular layer was increased about sevenfold (from 20 to 134) in KA-treated cultures compared to vehicle-treated cultures (P , 0.0001, Mann– Whitney Rank Sums test). Recurrent mossy fibers in either vehicle- or KA-treated cultures usually terminated within the inner half of the molecular layer (Fig. 6), although occasionally collaterals extending into the outer half of the molecular layer were observed (not shown). Specifically, 88% of axon collaterals entering the granule cell/molecular layer in vehicle-treated cultures terminated in the granule cell layer or the inner half of the molecular layer without penetrating to the outer half of the molecular layer (n ˆ 17 collaterals). In KA-treated cultures, 85% of these axon collaterals did not penetrate past the inner half of the molecular layer (n ˆ 68 collaterals). In one KAtreated neuron (1/18), a mossy fiber axon traversed the entire molecular layer, passed the hippocampal fissure, and continued on until terminating in CA1 stratum radiatum (not shown). However, in general, the mossy fiber sprouting was strikingly laminar in that it was generally confined to the inner half of the dentate molecular layer (Fig. 6). In both vehicle- and KA-treated cultures, granule cells had a well-developed conical arborization of apical dendrites (Fig. 6). Dendrites were found in the molecular layer and in many cases extended past the hippocampal fissure into stratum lacunosum-moleculare of CA1. Small basal dendrites were found in a subset of cells (31% of vehicle-treated neurons, 17% of KA-treated neurons). Although dendritic morphological parameters were not quantitated in this study, there were no obvious differences in the dendritic morphology of vehicle- and KA-treated neurons. In summary, individual granule cells in KA-treated cultures exhibited axonal rearrangements that included a dramatic increase in axonal branches and varicosities in the granule cell and molecular layers. DISCUSSION

The goal of these studies was to develop an in vitro model

of a structural plasticity evident in animal models and human epilepsy, namely sprouting of the mossy fiber axons of dentate granule cells, and to test whether the determinants required for convulsant-induced mossy fiber reorganization would be present in the hippocampal slice preparation. Four principal findings emerged. (1) Application of the limbic convulsant, KA, to cultured hippocampal explants produced intense electrographic seizures. (2) Prolonged (48 h) application of KA produced regionally selective cell death, with the CA3a/b pyramidal neurons being most vulnerable. (3) Prolonged (48 h) application of KA produced robust sprouting of mossy fiber axons in a laminar-specific manner into the inner molecular layer of the dentate gyrus. (4) Electrographic seizure activity produced by briefer (4 h) applications of KA was not sufficient to induce either cell death or measurable mossy fiber sprouting. Thus, the molecular and synaptic circuitry required for KA toxicity and subsequent regionally specific mossy fiber reorganization are present even in the relatively simple, isolated, and largely two-dimensional lamellar organotypic hippocampal culture. Kainic acid-induced seizure activity and cell death in hippocampal slice cultures The present study confirms and extends previous reports of KA-induced cell death and epileptiform activity in hippocampal slice culture. 2,21 Application of KA to slice cultures initiated the immediate appearance of electrographic seizure activity; epileptiform activity was evident at additional times during the 48 h treatment. The present study also characterized the time dependence of KA-induced neurotoxicity. A 48 h application of KA exerted profound neurotoxic effects; surprisingly, a briefer yet quite prolonged (4 h) application of KA failed to induce robust neurotoxic effects. The relatively indolent pace of the neurotoxic effects is in accord with studies of neurons in primary culture which describe “slowness” of KA toxicity. 27 Overall, the neurotoxic effects of KA were primarily

Mossy fiber sprouting in vitro

evident in the CA3 field of slice cultures. CA3c was relatively less sensitive to KA neurotoxicity than CA3a/b, and some degree of granule cell toxicity was also apparent. These findings closely agree with the hierarchy of vulnerability of the hippocampal cell types to KA in vivo. 14,15 Although the precise explanation for the selective vulnerability of the CA3a/b pyramidal cells to KA is unclear, the presence of the GluR6 subunit of KA receptors may be contributory. It has been recently reported that glutamate receptor (GluR)6deficient mice exhibit a reduction of [ 3H]KA binding in stratum lucidum as well as reduced sensitivity to KA-induced seizures and astrogliosis. 13 The present results demonstrate that actions of KA intrinsic to the hippocampus are sufficient to account for the selective vulnerability of CA3 pyramidal neurons to the neurotoxic actions following systemic administration of KA in vivo. Spontaneous and kainic acid-induced mossy fiber sprouting in hippocampal slice cultures While the slice culturing process itself is associated with some mossy fiber sprouting as evident in vehicle-treated cultures, treatment of slice cultures with KA induces a dramatic reorganization of mossy fibers that is similar to the sprouting identified in animal models and human temporal lobe epilepsy. Timm staining in roller-tube hippocampal slice cultures by prior investigators has suggested the presence of a significant amount of “spontaneous” supragranular mossy fiber sprouting. 4,28 This spontaneous reorganization may arise as a consequence of the massive denervation granule cells experience when the hippocampus is chopped into 400-mm-thick slices. Some increased supragranular Timm staining consistent with spontaneous mossy fiber sprouting was also observed in the vehicle-treated interfacetype slice cultures used in this study, although the degree of spontaneous sprouting seen in the interface cultures used in this study appears to be much less than what has been shown in roller-tube cultures. 4,28 Importantly, KA treatment resulted in the development over several weeks of very intense supragranular Timm staining that was markedly more prominent than the amount seen spontaneously in vehicle-treated cultures. Spontaneous mossy fiber sprouting in slice cultures might be used as a model system in its own right to examine the functional role of molecules in the sprouting process. However, the potential mechanistic differences between spontaneous deafferentation-induced sprouting and seizure-induced sprouting, as well as the much more dramatic sprouting seen in KA-treated cultures, argue instead for the utility of examining KA-induced mossy fiber sprouting in slice cultures. Correspondence between Timm staining and individual granule cell morphology The Timm stain-based impression of dramatic mossy fiber reorganization in KA-treated cultures as well as some spontaneous mossy fiber sprouting in vehicle-treated cultures corresponded very well to alterations of mossy fiber morphology in neurobiotin-labeled individual granule cells. In agreement with prior results, 17,24 this finding provides support for the idea that Timm staining is an accurate histochemical marker of mossy fiber morphological changes. One simple morphological endpoint reflective of mossy fiber sprouting is the percentage of granule cells that have

763

axons that penetrate into the supragranular molecular layer. In the current study, the intragranular and supragranular layers were not distinguished for morphological analyses, due to the fact that the granule cell layer/molecular layer boundary was not always well defined. In the comparison between vehicle- and KA-treated cultures in the current study, intra/supragranular mossy fiber axons were found in 63% (10/16) of vehicle-treated cells, but 96% (24/25) of KAtreated cells. This finding of an increasing percentage of KAtreated granule cells having intra/supragranular mossy fibers is consistent with a recent report by Sutula et al. examining individual granule cell morphology in acute slices from control rats and rats that had previously experience KA-induced status epilepticus. 24 Although methodological differences in the analysis preclude a direct quantitative comparison between the two studies, granule cells from control rats in the study by Sutula et al. did not show any evidence of supragranular sprouting. Together, these data support the idea that granule cells in vehicle-treated slice cultures display spontaneous mossy fiber sprouting compared to granule cells in the intact hippocampal formation, but that KA treatment both in vivo and in vitro causes a dramatic mossy fiber reorganization compared to the control conditions in either system. Mossy fiber sprouting retains a laminar preference in vitro One of the remarkable findings about both the spontaneous and the KA-induced mossy fiber sprouting observed in hippocampal slice cultures was its relative laminar specificity. Sprouted mossy fibers appeared to display a strong preference for the inner portion of the molecular layer as opposed to the outer portion. This is largely consistent with what has been reported from in situ studies of mossy fiber reorganization. In intact animals, sprouting mossy fibers that penetrate the granule cell layer display a nearly absolute preference for the inner portion of the molecular layer, although rare mossy fiber collaterals are observed to penetrate throughout the entire molecular layer and extend to the hippocampal fissure in pilocarpine-treated rats that display a large degree of mossy fiber sprouting. 17 The degree of preference of mossy fibers for the inner molecular layer in slice cultures, while probably not as strong as that seen in the intact brain, remains striking. What is particularly striking about this preference in slice cultures is that the outer molecular layer remains relatively unstained for mossy fibers, even though it is almost completely denervated during the preparation of slice cultures. The outer molecular layer receives the vast majority of its innervation from the entorhinal cortex in situ, and this innervation was completely removed in these slice cultures. Thus, sprouting mossy fibers in vitro appear to retain a selective preference of the inner molecular layer despite the plentiful nearby potential targets in the outer molecular layer. Unless these potential targets are either lost early on in culture or quickly occupied by other in-growing neurons, it seems likely that such selectivity is mediated at a molecular level, either by attractive cues (recruiting or enabling ingrowth into the inner molecular layer) and/or by repulsive cues (preventing ingrowth into the outer molecular layer). Insight into conditions required for induction of mossy fiber sprouting The relative contribution of denervation and pathologic

764

M. J. Routbort et al.

activity to the sprouting of mossy fiber axons remains controversial. Denervation, presumably secondary to cell death, has been advanced as the initial trigger for mossy fiber reorganization. 9 However, mossy fiber reorganization has also been observed in the presence of recurrent seizures without detectable cell death. 24,26 In the present study, treatment with KA for 4 h clearly evoked pathologic activity in the form of intense seizure activity yet failed to induce either overt cell death or measurable mossy fiber sprouting; moreover this treatment in slice cultures is sufficient to trigger cascades of gene expression such as increased expression of neurotrophin gene expression. 10 Instead, mossy fiber sprouting was observed here only in association with overt cell death evoked by prolonged KA treatment. Thus, in this model, transient seizure activity is not sufficient to induce mossy fiber sprouting, suggesting that prolonged seizures and neuronal death may be critical factors in producing robust mossy fiber reorganization. While it seems plausible that cell death-induced denervation is the likely trigger for the sprouting, denervation or “synaptic space” alone is not sufficient to induce sprouting of mossy fibers. That is, preparation of the explants removes the entorhinal cortex which provides 86% of the synapses in the outer two thirds of the molecular layer, yet sprouting of mossy fibers into the outer molecular layer was minimal for both vehicle and KA-treated cultures. Laminar cues may be important in this minimal sprouting to the outer molecular layer. We favor an explanation in which denervation of the inner molecular layer, probably by death of mossy cells or other afferents of the inner molecular layer, triggers sprouting of mossy fibers into the inner molecular layer. The model presented here should permit control of variables required to further test the relative contribution of the denervation and pathologic activity in mossy fiber sprouting. Usefulness of this in vitro model of mossy fiber sprouting in hippocampal slice cultures The present results establish the feasibility of inducing robust sprouting of the mossy fiber axons of the dentate granule cells in a chronic in vitro model. The sprouting of mossy fibers is arguably the most thoroughly studied axonal reorganization of the mature mammalian nervous system. The ease of detecting sprouted mossy fibers with a histochemical stain

together with their striking laminar specificity have led to their extensive investigation. The potential contribution of the synapses formed by sprouted mossy fibers to the hyperexcitability of the epileptic brain provides additional rationale for their study. Analysis of the underlying molecular mechanisms in chronic experimental animals in vivo is not only time consuming but is also limited by difficulties controlling access and concentrations of molecules such as antibodies, antisense oligonucleotides, etc. The rationale outlined in the preceding paragraph has led previous investigators to develop various models of granule cell axon growth in vitro. However, prior models using disassociated granule cells in primary culture or neurite outgrowth from dentate explants into a collagen matrix do not preserve the complex, potentially laminar cues provided to axonal growth cones by cell bound and soluble factors in the extracellular mileu. 11,19 The present preparation provides a simple preparation in which the striking preference for sprouted mossy fibers for the inner portion of the dentate molecular layer is retained. The laminar specificity of the sprouted axons is evident in the Timm stain and in the morphology of neurobiotin-labeled individual granule cells. The changes in the mossy fiber morphology in KA-treated slice cultures reported here is strikingly similar to that described by Sutula et al. in rats that had experienced status epilepticus in vivo. 24 Although the morphological end-points and boundaries used in these studies differed slightly, KA treatment resulted in a striking increase in the granule cell axons in the molecular layer of the dentate gyrus in each study, thereby validating the in vitro model. The ease of preparation together with ease of control of access, timing, and concentrations of “interventional molecules” (e.g., functional antibodies, antisense nucleotides, etc.) will permit functional assessment of the rapidly expanding numbers of recognized molecules underlying axonal outgrowth, guidance, and synapse formation following injury of the mature nervous system. Such studies should shed light on the molecular mechanisms underlying mossy fiber sprouting.

Acknowledgements—This work was supported by NIH grant NS32334 (J. O. M.), NS-07370 (S.B.B.), and NS-10387 (S.B.B.)

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Ben-Ari Y. (1985) Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience 14, 375–403. Benedikz E., Casaccia-Bonnefil P., Stelzer A. and Bergold P. J. (1993) Hyperexcitability and cell loss in kainate-treated hippocampal slice cultures. NeuroReport 5, 90–92. Blanton M. G., Lo Turco J. J. and Kriegstein A. R. (1989) Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex. J. Neurosci. Meth. 30, 203–210. Coltman B. W., Earley E. M., Shahar A., Dudek F. E. and Ide C. F. (1995) Factors influencing mossy fiber collateral sprouting in organotypic slice cultures of neonatal mouse hippocampus. J. comp. Neurol. 362, 209–222. Cronin J. and Dudek F. E. (1988) Chronic seizures and collateral sprouting of dentate mossy fibers after kainic acid treatment in rats. Brain Res. 474, 181–184. Dudek F. E. and Spitz M. (1997) Hypothetical mechanisms for the cellular and neurophysiologic basis of secondary epileptogenesis: proposed role of synaptic reorganization. [published erratum appears in J. clin. Neurophysiol. 1997 Sep; 14(5):459]. J. clin. Neurophysiol. 14, 90–101. Gahwiler B. H., Capogna M., Debanne D., McKinney R. A. and Thompson S. M. (1997) Organotypic slice cultures: a technique has come of age. Trends Neurosci. 20, 471–477. Isokawa M., Levesque M. F., Babb T. L. and Engel J. Jr. (1993) Single mossy fiber axonal systems of human dentate granule cells studied in hippocampal slices from patients with temporal lobe epilepsy. [published erratum appears in J. Neurosci. 1993 Jun; 13(6): following table of contents]. J. Neurosci. 13, 1511–1522. Laurberg S. and Zimmer J. (1981) Lesion-induced sprouting of hippocampal mossy fiber collaterals to the fascia dentata in developing and adult rats. J. comp. Neurol. 200, 433–459. Lauterborn J. C., Rivera S., Stinis C. T., Hayes V. Y., Isackson P. J. and Gall C. M. (1996) Differential effects of protein synthesis inhibition on the activity-dependent expression of BDNF transcripts: evidence for immediate-early gene responses from specific promoters. J. Neurosci. 16, 7428–7436.

Mossy fiber sprouting in vitro

765

11. Lowenstein D. H. and Arsenault L. (1996) Dentate granule cell layer collagen explant cultures: spontaneous axonal growth and induction by brainderived neurotrophic factor or basic fibroblast growth factor. Neuroscience 74, 1197–1208. 12. Mello L. E., Cavalheiro E. A., Tan A. M., Kupfer W. R., Pretorius J. K., Babb T. L. and Finch D. M. (1993) Circuit mechanisms of seizures in the pilocarpine model of chronic epilepsy: cell loss and mossy fiber sprouting. Epilepsia 34, 985–995. 13. Mulle C., Sailer A., Perez-Otano I., Dickinson-Anson H., Castillo P. E., Bureau I., Maron C., Gage F. H., Mann J. R., Bettler B. and Heinemann S. F. (1998) Altered synaptic physiology and reduced susceptibility to kainate-induced seizures in GluR6-deficient mice. Nature 392, 601–605. 14. Nadler J. V., Perry B. W. and Cotman C. W. (1978) Intraventricular kainic acid preferentially destroys hippocampal pyramidal cells. Nature 271, 676–677. 15. Nadler J. V., Perry B. W. and Cotman C. W. (1978) Preferential vulnerability of hippocampus to intraventricular kainic acid. In Kainic Acid as a Tool in Neurobiology (eds McGeer E., Olney J. and McGeer P.), pp. 219–238. Raven, New York. 16. Nadler J. V., Perry B. W. and Cotman C. W. (1980) Selective reinnervation of hippocampal area CA1 and the fascia dentata after destruction of CA3–CA4 afferents with kainic acid. Brain Res. 182, 1–9. 17. Okazaki M. M., Evenson D. A. and Nadler J. V. (1995) Hippocampal mossy fiber sprouting and synapse formation after status epilepticus in rats: visualization after retrograde transport of biocytin. J. comp. Neurol. 352, 515–534. 18. Olney J. W., Rhee V. and Ho O. L. (1974) Kainic acid: a powerful neurotoxic analogue of glutamate. Brain Res. 77, 507–512. 19. Patel M. N. and McNamara J. O. (1995) Selective enhancement of axonal branching of cultured dentate gyrus neurons by neurotrophic factors. Neuroscience 69, 763–770. 20. Represa A., Le Gall La Salle G. and Ben-Ari Y. (1989) Hippocampal plasticity in the kindling model of epilepsy in rats. Neurosci. Lett. 99, 345–350. 21. Rimvall K., Keller F. and Waser P. G. (1987) Selective kainic acid lesions in cultured explants of rat hippocampus. Acta neuropath. 74, 183–190. 22. Stoppini L., Buchs P. A. and Muller D. (1991) A simple method for organotypic cultures of nervous tissue. J. Neurosci. Meth. 37, 173–182. 23. Sutula T., Cascino G., Cavazos J., Parada I. and Ramirez L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol. 26, 321–330. 24. Sutula T., Zhang P., Lynch M., Sayin U., Golarai G. and Rod R. (1998) Synaptic and axonal remodeling of mossy fibers in the hilus and supragranular region of the dentate gyrus in kainate-treated rats. J. comp. Neurol. 390, 578–594. 25. Tauck D. L. and Nadler J. V. (1985) Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. J. Neurosci. 5, 1016–1022. 26. Watanabe Y., Johnson R. S., Butler L. S., Binder D. K., Spiegelman B. M., Papaioannou V. E. and McNamara J. O. (1996) Null mutation of c-fos impairs structural and functional plasticities in the kindling model of epilepsy. J. Neurosci. 16, 3827–3836. 27. Weiss J. H., Hartley D. M., Koh J. and Choi D. W. (1990) The calcium channel blocker nifedipine attenuates slow excitatory amino acid neurotoxicity. Science 247, 1474–1477. 28. Zimmer J. and Gahwiler B. H. (1984) Cellular and connective organization of slice cultures of the rat hippocampus and fascia dentata. J. comp. Neurol. 228, 432–446. (Accepted 29 June 1999)