Parvalbumin interneurons and calretinin fibers arising from the thalamic nucleus reuniens degenerate in the subiculum after kainic acid-induced seizures

Neuroscience 189 (2011) 316 –329

PARVALBUMIN INTERNEURONS AND CALRETININ FIBERS ARISING FROM THE THALAMIC NUCLEUS REUNIENS DEGENERATE IN THE SUBICULUM AFTER KAINIC ACID-INDUCED SEIZURES M. DREXEL,* A. P. PREIDT, E. KIRCHMAIR AND G. SPERK*

ing from the nucleus reuniens thalami significantly contributes to the rearrangement of neuronal circuitries in the subiculum and entorhinal cortex during epileptogenesis. © 2011 IBRO. Published by Elsevier Ltd. All rights reserved.

Department of Pharmacology, Innsbruck Medical University, PeterMayr-Str. 1a, 6020 Innsbruck, Austria

Key words: status epilepticus, temporal lobe epilepsy, epileptogenesis, entorhinal cortex, epilepsy models

Abstract—The subiculum is the major output area of the hippocampus. It is closely interconnected with the entorhinal cortex and other parahippocampal areas. In animal models of temporal lobe epilepsy (TLE) and in TLE patients it exerts increased network excitability and may crucially contribute to the propagation of limbic seizures. Using immunohistochemistry and in situ-hybridization we now investigated neuropathological changes affecting parvalbumin and calretinin containing neurons in the subiculum and other parahippocampal areas after kainic acid-induced status epilepticus. We observed prominent losses in parvalbumin containing interneurons in the subiculum and entorhinal cortex, and in the principal cell layers of the pre- and parasubiculum. Degeneration of parvalbumin-positive neurons was associated with significant precipitation of parvalbumin-immunoreactive debris 24 h after kainic acid injection. In the subiculum the superficial portion of the pyramidal cell layer was more severely affected than its deep part. In the entorhinal cortex, the deep layers were more severely affected than the superficial ones. The decrease in number of parvalbumin-positive neurons in the subiculum and entorhinal cortex correlated with the number of spontaneous seizures subsequently experienced by the rats. The loss of parvalbumin neurons thus may contribute to the development of spontaneous seizures. On the other hand, surviving parvalbumin neurons revealed markedly increased expression of parvalbumin mRNA notably in the pyramidal cell layer of the subiculum and in all layers of the entorhinal cortex. This indicates increased activity of these neurons aiming to compensate for the partial loss of this functionally important neuron population. Furthermore, calretinin-positive fibers terminating in the molecular layer of the subiculum, in sector CA1 of the hippocampus proper and in the entorhinal cortex degenerated together with their presumed perikarya in the thalamic nucleus reuniens. In addition, a significant loss of calretinin containing interneurons was observed in the subiculum. Notably, the loss in parvalbumin positive neurons in the subiculum equaled that in human TLE. It may result in marked impairment of feed-forward inhibition of the temporo-ammonic pathway and may significantly contribute to epileptogenesis. Similarly, the loss of calretinin-positive fiber tracts originat-

Mesial temporal lobe epilepsy (TLE) is the most frequent form of epilepsies in adults (Engel, 2001). It often develops years after an initial insult like febrile seizures, head trauma, cerebral infections, or a status epilepticus (SE) (Baulac et al., 2004). TLE is often refractory to drug therapy and associated with severe neurodegeneration in the entorhinal cortex (EC), the dentate gyrus and most prominently in sectors CA1 and CA3 of the hippocampus, termed hippocampal or Ammon’s horn sclerosis (Sommer, 1880; Babb et al., 1984). On the other hand, the subiculum remains largely preserved (Bratz, 1899; Furtinger et al., 2001). Accumulating experimental and clinical evidence suggests a crucial role of the subiculum and other parahippocampal regions in the pathophysiology of TLE. Thus, in animal models of TLE, network hyperexcitability was demonstrated in the subiculum (Knopp et al., 2005; de Guzman et al., 2006) and in deep and superficial layers of the EC (Kobayashi et al., 2003; Wozny et al., 2005a; de Guzman et al., 2008; Bragin et al., 2009). Injection of the sea weed toxin kainic acid (KA) in rats induces a SE with neurodegeneration reflecting those observed in human TLE (Ben-Ari et al., 1980; Sperk et al., 1983; Sperk, 1994). The initial SE is followed by a seizurefree period of several days or weeks. After this “silent period” the rats develop spontaneously recurrent limbic seizures (Cavalheiro et al., 1982). KA injections in rats therefore have been not only considered as model for studying the pathophysiology of SE but also as an animal model of TLE (Leite et al., 2002). During the last decades, the neurochemical and neuropathological changes related to KA-induced SE have been extensively investigated. However, most studies focused on the dentate gyrus and the hippocampus proper but did not include parahippocampal areas. Using immunohistochemistry and in situ hybridization for the calcium binding proteins parvalbumin (PV) and calretinin (CR), we now investigated the epilepsyassociated changes in parahippocampal regions (subiculum, presubiculum, parasubiculum, and EC) after KA-induced SE and related these changes to the frequency of spontaneous seizures.

*Corresponding author. Tel: ⫹43-(0)-512-9003-71210; fax: ⫹43-(0)512-9003-73200. E-mail address: [email protected] (G. Sperk) or meinrad. [email protected] (M. Drexel). Abbreviations: CR, calretinin; EC, entorhinal cortex; -ir, immunoreactive; KA, kainic acid; NeuN, neuron specific nuclear protein; O-LM, oriens-lacunosum moleculare; PV, parvalbumin; ROD, relative optical densities; SE, status epilepticus; TBS, tris-buffered saline; TLE, temporal lobe epilepsy.

0306-4522/11 $ - see front matter © 2011 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2011.05.021

316

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

EXPERIMENTAL PROCEDURES Animals Male Sprague–Dawley rats (210 –260 g; Forschungsinstitut für Versuchstierzucht, Himberg, Austria) were used. They had free access to food and water and were housed in individually ventilated cages at a temperature of 22–23 °C, a relative humidity of 50 – 60%, and a 12 h light/dark cycle. All animal experiments were conducted according to national guidelines and European Community laws. They were approved by the Committee for Animal Protection of the Austrian Ministry of Science.

Kainic acid injection and rating of status epilepticus Thirty nine rats were injected with 10 mg/kg KA (5 mg/ml in saline, pH 7.0, i.p.) and 14 with saline. To reduce mortality and severity of the neuropathological outcome, rats were treated with diazepam (10 mg/kg i.p., Gewacalm, Nycomed Austria GmbH, Linz, Austria) 2 h after the first generalized seizure. The dose of KA, in combination with anticonvulsant treatment was based on our previous studies resulting in strong SE in the majority of rats with limited tissue necrosis (Sperk et al., 1983; Sperk, 1994; Tsunashima et al., 1997). Seizure behavior of all rats was observed for at least 3 h and rated using a 5-stage rating scale described previously (Sperk et al., 1983). Rats without any obvious behavioral changes were rated as stage 0, rats exposing wet dog shakes only as stage 1, rats with chewing, head bobbing and forelimb cloni as stage 2, rats with generalized seizures and rearing as stage 3, rats with generalized seizures, rearing and falling over (loss of postural tone) as stage 4, and rats that died during SE were rated as stage 5. Among the 39 rats subjected to KA, 1 rat developed stage 2 seizures, 26 rats stage 3 and 7 rats stage 4 seizures. Two rats did not respond and 3 rats died. Rats exhibiting rating 3 or 4 were used in the study.

Video-assisted telemetric EEG-monitoring In a separate experiment video-assisted EEG recordings were performed in 51 rats for up to 3 months. Two stainless steel screws (M1*2, Hummer und Rie␤ GmbH, Nürnberg, Germany) were set in an epidural position (4.0 mm posterior and 3.0 mm left and right to the bregma) and connected with a biopotential transmitter (TA10EA-F20, Data Sciences International, Arden Hills, USA) placed in a s.c. pocket at the back of the rats as described in detail elsewhere (submitted for publication). Thirty-seven of the 51 rats were treated with KA as described above and 14 rats received saline. EEG activity was recorded using a telemetry system (Dataquest A.R.T. Data Acquisition 4.0 for telemetry systems, Data Sciences International, Arden Hills, USA) and behavior was monitored using Axis 221 network infrared sensitive video cameras (Axis communications AB, Lund, Sweden) with infrared illumination during the dark phase.

Preparation of tissue and of tissue sections For immunohistochemistry, KA-injected rats were killed either 24 h (n⫽4), 8 days (n⫽3), 1 month (n⫽7), or 3 months (n⫽15) after KA-induced SE. Controls were killed 8 days (n⫽7) or 3 months (n⫽8) after saline injection. These intervals were chosen for assessing changes related to the direct consequences of the SE (24 h), to changes in the presumed silent phase (8 days), and changes due to the chronic epilepsy syndrome (1 and 3 months). For immunohistochemistry, the rats were perfused transcardially with 40 ml phosphate buffered saline pH 7.4 (PBS) followed by 100 ml of ice-cold 4% paraformaldehyde (Merck, Darmstadt, Germany) in 50 mM phosphate buffer, pH 7.4. Their brains were removed from the sculls and post-fixed in ice-cold paraformaldehyde for 90 min. They were then immersed in ice-cold 20% sucrose in 50 mM phosphate buffer for 24 h and then snap-frozen

317

in ⫺70 °C isopentane (Merck, Darmstadt, Germany, 3 min) and stored at ⫺70 °C. Horizontal 30 ␮m sections were cut with a cryostat-microtome (Carl Zeiss AG, Vienna, Austria), collected in 50 ␮M Tris-buffered saline (TBS, pH 7.2) containing 0.1% NaN3 (Merck, Darmstadt, Germany) and stored at 4 °C. Every 11th section was stained with Cresyl Violet, dehydrated, cleared in butyl acetate, and coverslipped with Eukitt mounting medium (O. Kindler GmbH, Freiburg, Germany). These sections were used for collecting sections from the different animals at the same anatomical level for using them in concomitant histochemistry. For in situ hybridization, KA-treated rats were killed by exposure to CO2 gas after 24 h (n⫽5), 8 days (n⫽6), 1 month (n⫽5) or 3 months (n⫽3). Controls were killed 24 h (n⫽3), 1 month (n⫽2) or 3 months (n⫽3) after saline injection, respectively. Brains were rapidly removed and snap-frozen in ⫺70 °C isopentane. Horizontal 20 ␮m sections were cut using a cryostat-microtome, thawmounted on silane-coated slides and stored at ⫺70 °C.

Indirect immunohistochemistry Immunohistochemistry for Neuron specific nuclear protein (NeuN), PV, and CR was performed on free-floating sections. The sections were incubated in 50 mM Tris buffered saline with 0.4% Triton X-100 (TBS-Triton, Sigma, St. Louis, USA) for 30 min and for 90 min in blocking serum (10% normal goat or rabbit serum in TBS-Triton). They were then incubated over night at room temperature with monoclonal mouse antibodies for NeuN (1:10,000; MAB 377, Chemicon, USA) or PV (1:2,000; P3088, Sigma-Aldrich Handels GmbH, Vienna, Austria), or a CR antiserum raised in rabbits (1:5,000; 7698, SWANT, Bellinzona, Switzerland) in 10% blocking serum containing 0.1% NaN3. After rinsing in TBS-Triton three times for 5 min, the sections were incubated with horse radish peroxidase coupled secondary antibodies (P0260 rabbit anti-mouse 1:200 for NeuN and PV, P0448 goat anti-rabbit 1:250 for CR; both DAKO, Glostrup, Denmark) in 10% blocking serum/ TBS-Triton (150 min, room temperature). The sections were then washed three times (5 min) in TBS buffer, and reacted with 0.001% of diaminobenzidine tetrahydrochloride dihydrate (DAB, Fluka, Sigma-Aldrich Handels GmbH, Vienna, Austria) in TBSbuffer containing 0.005% H2O2 (30%, Merck, Darmstadt, Germany). Sections were mounted on gelatin-coated glass slides in 55% ethanol and allowed to dry overnight. After dehydration in ethanol and clearing in butyl acetate, they were coverslipped using Eukitt mounting medium. Omission of the primary antisera eliminated all specific immunoreactivity.

In situ hybridization The following custom-synthesized oligonucleotides (Microsynth, Balgach, Switzerland) complementary to the respective mRNAs were used as probes: CR, bases 391– 435: 5=-GAT GTA GCC ACT TCT GTC TGT GTC ATA CTT CCG CCA AGC CTC CAT-3= (GenBank accession no. NM_053988.1, Gabrielides et al., 1991), PV, bases 248 –291: 5=-GTC CTT GTC TCC AGC AGC CAT CAG CGT CTT TGT TTC CTT AGC AG-3= (NM_022499, Berchtold et al., 1982), and GAD65, bases 348 –384: 5=-CTC CTT CAC AGG CTG GCA GCA GGT CTG TTG CGT GGA G-3= (NM_012563, Chang and Gottlieb, 1988). The oligonucleotides (2.5 pmol) were labeled at the 3=-end with [35S] ␣-thio-dATP (1,300 Ci/mmol; New England Nuclear, Boston, USA) by reaction with terminal deoxynucleotidyltransferase (Roche Austria GmbH, Vienna, Austria) and in situ hybridization was performed as described previously (Furtinger et al., 2001).

Evaluation of neuronal loss Cell counts of parvalbumin- and GAD65-containing neurons. For each rat, two 30 ␮m thick horizontal sections (immunolabeled for PV) located about 6.8 mm and 5.1 mm ventral to bregma,

318

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

respectively (Paxinos and Watson, 1998), were used for counts of PV-ir neurons and area measurements. The area of the parahippocampal region was digitalized at 200-fold magnification using a Zeiss AxioCam (Carl Zeiss AG, Vienna, Austria) digital camera. The images were imported into the image analysis software ImageJ and the following parahippocampal areas were manually outlined for area measurements: they included the pyramidal cell layer of the subiculum, the principal cell layers of the presubiculum (layers II and III) and parasubiculum, as well as layers II, III and V–VI of the medial and lateral EC. The area of the pyramidal cell layer of the subiculum was divided into two equal parts defined as deep (adjacent to the alveus) and superficial (adjacent to the molecular layer of the subiculum) parts. Moreover, the pyramidal cell layer of the subiculum was divided into a proximal (adjacent to sector CA1) and a distal part (adjacent to the presubiculum). The areas of the respective regions were measured and PV-ir perikarya were counted by two blinded observers using a manual counting device. The number of PV-labeled neurons per mm2 was calculated for each subregion and values obtained from the left and right hemispheres of two sections (total of four values) were averaged. Sections dipped with photoemulsion after in situ hybridization were used for counting by the same procedure PV and GAD65 mRNA containing interneurons in the subiculum at 400fold magnification. Cell counts of calretinin- and NeuN-positive neurons. CR-ir neurons were counted in the pyramidal cell and molecular layers of the subiculum and in the nucleus reuniens of the thalamus in 30 ␮m horizontal sections at levels of 5.3 mm to 5.6 mm (subiculum) and 7.1 mm to 7.6 mm ventral to bregma (thalamic nucleus reuniens), respectively. An ocular counting grid (500⫻500 ␮m) was used at 200-fold magnification. CR-ir cells and NeuN-ir cells were counted under the area of one grid in adjacent brain sections and the numbers of CR-ir and NeuN-ir cells per mm2 were assessed. Values obtained from the left and right hemispheres were averaged and mean values were calculated for each group of rats. It has to be noted, however, that we used non-stereological methods for quantification of NeuN-, PV- and CR-positive neurons. The main bias resulting from this approach may be that it does not account for changes in volume due to tissue shrinkage and therefore may somewhat underestimate (not overestimate) losses of the respective cells. Another bias may result from the increased expression of PV and CR at the late intervals after KA as shown by increase expression of the mRNAs. Thus at these intervals neurons immunoreactive for the respective calcium binding proteins may be detected more clearly and may eventually lead to a partial overestimation of cell counts.

Semi-quantitative densitometric analysis of calretinin-ir fibers Semi-quantitative densitometric analysis of CR-ir fibers and terminals in the subiculum was performed in 8 bit images (100-fold magnification) of sections concomitantly processed for CR immunohistochemistry. Images were digitized using the OpenLab 4 program (PerkinElmer, USA) and were then imported into the image analysis software ImageJ. A rectangle area (length: 300 ␮m, width: 150 ␮m) was placed over the layer of interest and the mean gray value of all pixels within this area was determined. Nonspecific background labeling was determined in the corpus callosum. Relative optical densities (ROD) were calculated from the mean gray values of the areas using the following formula: ROD⫽log[256/(255⫺gray value)]. RODs of the left and right hemispheres were averaged and the ROD of the nonspecifically labeled background was subtracted.

Expression of parvalbumin mRNA in interneurons Expression of PV mRNA in the subiculum and EC was quantified at 400-fold magnification in darkfield images obtained from photoemulsion dipped sections after in situ hybridization and after counterstaining with Cresyl Violet. The area covered by PV mRNA expressing neurons including the close vicinity were outlined by a 30 ␮m diameter circle using the ImageJ program. The density of silver grains within individual circles was measured. As threshold, the density of silver grains in the black background was used. In total, 1467 neurons in the EC and 336 neurons in the subiculum were examined (Figs. 2G–I and 3D–F).

Statistical analyses Statistical analyses were carried out using GraphPad Prism 5.0a for Macintosh (GraphPad Software, San Diego, CA, USA). ANOVA with Dunnett’s multiple comparison as post hoc test were used for determining between-group differences among multiple data sets. For determining statistical differences between mean values of two groups, the Student t-test was used. Correlations were determined by using the Pearson correlation coefficient. Statistical significance was defined as P⬍0.05. All data are presented as mean⫾SEM.

RESULTS Area specific degeneration of parvalbumin neurons and increased expression of parvalbumin mRNA in surviving neurons In controls, PV-positive interneurons were present throughout the principal cell layer of the subiculum (Fig. 1A), in layers II, III and V/VI of the EC (Fig. 3A) and in the pre- and parasubiculum (Fig. 4A). In the subiculum, the number of PV-ir neurons was decreased at all time intervals after KA injection (Fig. 1E, F, H, I, Table 1). The superficial part of the pyramidal cell layer was most severely affected (Fig. 1E, F; compare also the proximal vs. distal part in Table 1). There, cell losses were evident already 24 h after KA-injection (loss of 64.6% of neurons). At this interval PV-ir debris originating from degenerating PV neurons was visible (Fig. 1E, H). Furthermore, marked de-arrangement and loss of the respective PV-positive dendrites was observed in the molecular layer at the same interval (Fig. 1E). At the later time intervals cell losses of 55 to 75% were observed in the superficial part of the pyramidal cell layer (Table 1). Degeneration of PV-ir neurons was by far less severe in the deep part of the pyramidal cell layer at all investigated intervals (losses between 8% and 34%, Table 1). The extent of neurodegeneration also differed along the proximo-distal axis of the subiculum and was more severe in the proximal part of the pyramidal cell layer (Table 1). Accordingly, decreased numbers of PV mRNA-expressing neurons were detected by in situ hybridization in the subiculum at all intervals (Fig. 2C, E). The reduction of PV mRNA containing neurons in the superficial pyramidal cell layer was paralleled by a concomitant reduction of GAD65 mRNA containing neurons (Fig. 2F; see also Knopp et al., 2008). This decrease in numbers of both (overlapping) neuron populations was statistically significant already after 24 h. It was more pronounced for PV mRNA than for GAD65 mRNA contain-

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

319

Fig. 1. Loss of parvalbumin (PV)-ir interneurons in the subiculum after KA-induced seizures. In (A–C) PV-ir is shown at low magnification in the subiculum of a control rat (A), 24 h (B) and 3 mon (C) after KA-induced SE. In controls, PV-ir interneurons were distributed in the principal cell layer of the subiculum (A). After KA-induced SE, PV-ir interneurons in the subiculum were mainly lost in the superficial and to a lesser extent in the deep part of the pyramidal cell layer of the subiculum (B, C). Labeling of the surviving cells, however, appears to be increased (C). Dashed lines indicate the borders of the subiculum to sector CA1, the dentate gyrus (DG) and the presubiculum (PrS). The rectangle in (A) denotes the area shown in (D–F), depicting at a higher magnification PV-ir in the principal cell layer of the subiculum and the field of apical dendrites in the molecular layer of the subiculum. Note the loss of PV neurons after 24 h (E) and 3 mon (F) especially in the superficial part of the pyramidal cell layer, as well as the striking de-arrangement and extensive loss of the PV-ir dendrites (arrows in E and F). In (G–I) the superficial pyramidal cell layer of the subiculum is shown at larger magnification. Note the loss of PV-ir neurons 24 h (H) and 3 mon (I) after KA injection. The PV-ir debris indicates progressing neurodegeneration at the 24 h interval (arrows in H). Abbreviations: Co, control; DG, dentate gyrus; PrS, presubiculum; Sub, subiculum; sup, superficial part of the principal cell layer of the subiculum.

ing neurons indicating that GAD65 mRNA containing neurons may comprise also other less sensitive populations of GABA interneurons. At late the intervals (1 and 3 months after KA-induced SE), surviving PV neurons expressed increased concentrations of PV mRNA in the subiculum (Fig. 2G–I). In the EC, the number of PV-expressing interneurons was significantly reduced in the deep layers (layers V and VI); no significant reduction was observed in the superficial layers (layers II and III; Fig. 3B, C, Table 1). The degree of cell losses was equal in the medial and

lateral parts of the EC (Table 1). Already 24 h after KA-induced seizures, losses of 53 and 52% in PV-ir neurons were observed in the deep layers of the medial and lateral EC, respectively. These decreases persisted also at the later intervals. PV-ir debris were also observed in layer V of the EC 24 h after KA (not shown). In surviving interneurons of the superficial layers of the EC, expression of PV mRNA was increased 8 days and 1 and 3 months after KA, indicating constant activation of these neurons (Fig. 3E, F). Similarly, surviving PV neurons in the deep layers expressed significantly in-

320

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

Table 1. Loss of PV-ir neurons in parahippocampal regions at different time intervals after KA-induced status epilepticus. Values represent numbers of PV-immunolabeled neurons per mm2 in the principal cell layers of parahippocampal subregions (mean⫾SEM). Percent values of controls are shown in brackets Region

Controls (n⫽8)

KA 24 h (n⫽4)

KA 8 d (n⫽3)

KA 1 mon (n⫽4)

KA 3 mon (n⫽4)

Proximal pcl subiculum Distal pcl subiculum Superficial pcl subiculum Deep pcl subiculum Presubiculum Parasubiculum Medial EC Layer II Layer III Layers V–VI Lateral EC Layer II Layer III Layers V–VI

188⫾8.9 229⫾12.0 145⫾7.9 279⫾9.8 345⫾12.2 252⫾8.7

111⫾11.8*** (59.1%) 186⫾15.9 (81.2%) 51⫾6.1*** (35.4%) 221⫾10.8* (79.1%) 252⫾25.2* (73.0%) 181⫾23.1 (71.8%)

101⫾6.9*** (53.6%) 153⫾9.2** (66.8%) 58⫾7.3*** (40.1%) 246⫾11.8 (88.0%) 238⫾28.3** (68.9%) 262⫾21.4 (104.1%)

132⫾15.0** (70.4%) 175⫾8.8* (76.4%) 65⫾7.3*** (45.0%) 257⫾21.5 (92.0%) 290⫾40.5 (84.3%) 185⫾51.1 (73.5%)

101⫾3.4*** (53.4%) 138⫾8.8*** (60.3%) 36⫾3.3*** (24.9%) 184⫾6.3*** (65.9%) 262⫾16.5* (75.9%) 195⫾16.0 (77.6%)

199⫾8.1 223⫾6.6 160⫾9.6

199⫾14.7 (100.1%) 205⫾7.3 (91.8%) 75⫾7.0*** (47.0%)

179⫾12.3 (89.9%) 233⫾22.1 (104.1%) 72⫾10.1*** (44.8%)

169⫾4.2 (85.1%) 220⫾15.9 (98.3%) 92⫾12.8*** (57.2%)

166⫾7.3 (83.3%) 221⫾7.2 (98.9%) 75⫾3.7*** (46.7%)

166⫾11.9 248⫾11.5 167⫾7.9

185⫾16.1 (111.4%) 248⫾17.7 (99.7%) 81⫾6.8*** (48.5%)

162⫾9.3 (97.5%) 258⫾1.6 (103.7%) 91⫾8.5*** (54.9%)

146⫾9.8 (88.0%) 230⫾20.5 (92.6%) 76⫾8.6*** (45.6%)

139⫾7.4 (83.6%) 206⫾11.4 (82.8) 70⫾6.1*** (42.2%)

Abbreviation: pcl, pyramidal cell layer. Statistical analysis was done by ANOVA with Dunnett’s multiple comparison post hoc test; * P⬍0.05; ** P⬍0.01; *** P⬍0.001.

creased levels of PV mRNA 1 and 3 months after KAinjection (Fig. 3D–F). Less severe losses in PV-ir interneurons (up to 31% of controls) were observed in the principal layers of the pre- and parasubiculum (Fig. 4, Table 1). Correlation of loss of parvalbumin-containing neurons with the number of spontaneous seizures As described in detail elsewhere (submitted for publication), among the rats subjected to telemetric EEG and video monitoring, all 27 animals with initial rating 2 to 4 developed spontaneous seizures after 3 to 36 days (mean: 15.3⫾1.54 days). Then rats in average showed 7.6⫾3.21 spontaneous seizures per week. As shown in Fig. 5, cumulative numbers of spontaneous recurrent seizures experienced by the rats over 11 weeks were negatively correlated with the numbers of PV-ir interneurons in the superficial (r2⫽0.4107, P⫽0.0247) and deep principal cell layer of the subiculum (r2⫽0.4679, P⫽0.0142) and in layer III of the medial EC (r2⫽0.3566, P⫽0.0404). No correlation was found in other parahippocampal regions (legend to Fig. 5). Loss of calretinin-positive local neurons in the subiculum In the rat hippocampus, CR-ir neurons comprise a subset of non-pyramidal interneurons (Gulyas et al., 1992). As shown in Fig. 6A, B, they were present at low density in the subiculum of controls (57.2⫾2.42 neurons/mm2). The numbers of CR-labeled neurons were significantly reduced in the subiculum 24 h (by 23.5⫾11.48%; P⬍0.05; Fig. 6D, E, J) and 3 months after KA injection (by 21.1⫾3.67%; P⬍0.01; Fig. 6G, H, J). The reduction in CR-ir cells did not reach statistical significance at the 8 days (by 15.0⫾5.52%) and 1 month (by 1.4⫾3.55%) intervals (Fig. 6J).

Loss of afferent calretinin-ir fibers in the subiculum and entorhinal cortex In controls, a dense plexus of CR-labeled fibers was present in the outer molecular layer of the (proximal) subiculum (see Fig. 6A, B), in stratum lacunosum-moleculare of sector CA1 (Fig. 6A), and in layer I of the EC (Fig. 6C; see also: Gulyas et al., 1992; Miettinen et al., 1997). As shown in Fig. 6D, E, immunolabeling was slightly increased in these layers 24 h after SE (arrow in Fig. 6D), but was decreased at the later intervals (arrow in Fig. 6G, H). Similarly in layer I of the EC, CR fiber labeling (Fig. 6C, F, I) was decreased at later time intervals after KA injection (8 days, 1 and 3 months; Fig. 6I). We substantiated these findings by densitometric measurements and found significantly decreased ROD values in the outer molecular layer of the subiculum 8 days (by 41.1⫾3.2%; P⬍0.05), 1 month (by 39.6⫾8.26%; P⬍0.01) and 3 months (by 47.0⫾6.65%; P⬍0.001) after KA injection (Fig. 6K). Degeneration of calretinin-expressing neurons in the thalamic nucleus reuniens Well-established glutamatergic projections from the nucleus reuniens of the thalamus innervate the outer molecular layer of the subiculum, the stratum lacunosummoleculare of sector CA1, and layer I of the EC (Wouterlood et al., 1990; Dolleman-Van der Weel and Witter, 2000). These projections contain also CR (Bokor et al., 2002; Wouterlood et al., 2008a). We therefore hypothesized that concomitantly with the loss of CR-ir nerve terminals in the subiculum, CA1 and the EC CR-containing neurons in the thalamic nucleus reuniens may also degenerate. Indeed, the numbers of both NeuN-ir (Fig. 7A, B) and of CR-ir neurons (Fig. 7C, D) significantly decreased in the nucleus reuniens 3 months after KAinjection by 30.9⫾4.16% (P⬍0.0001) and 62.4⫾5.27% (P⬍0.0001), respectively. In situ hybridization for CR mRNA revealed a similarly reduced number of CR-ex-

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

321

Fig. 2. Loss of PV mRNA containing neurons and increased PV mRNA expression in surviving neurons. In (A) and (B) film autoradiographs of hippocampal sections after in situ hybridization with a 35S-labeled probe for PV mRNA are shown for a control (A) and 3 mon (B) after KA injection, respectively. Although the loss in PV mRNA containing neurons cannot be envisaged at this magnification, the markedly increased expression of PV mRNA in surviving neurons becomes obvious 3 mon after KA (B). (D) and (E) show dark field images of the subiculum in photoemulsion dipped sections of a control rat (D) and 3 mon after KA (E). Note the preferential loss of PV mRNA containing neurons in the superficial pyramidal cell layer (pcl) after KA. In (G) and (H), PV mRNA containing neurons of the pyramidal cell layer are depicted at a higher magnification of a control rat (arrows in G) and 3 mon after KA (H). Note the increased density of silver grains in surviving neurons reflecting increased PV mRNA expression (arrows in H). Panels (C) and (F) show cell counts of PV and GAD65 mRNA containing neurons, respectively, in the deep and superficial pyramidal cell layers. Note the significant reduction of interneurons in the superficial part of the pyramidal layer already 24 h after KA injection. The portion of PV mRNA containing neurons degenerating (C) exceeded that of GAD65 positive neurons (F). In panel (I) the densities of silver grains (representing PV transcript) over surviving PV mRNA positive neurons is shown. Significantly increased PV mRNA expression was observed in surviving neurons 8 d up to 3 mon after KA injection (I). Data in (C) and (F) represent mean values⫾SEM from 9 controls, and 3 to 6 for KA treated rats. Data in panel (I) represent the area over the PV mRNA positive neurons covered by silver grains and are presented as mean values ⫾ SEM from 117 neurons of controls and 36 to 88 neurons for KA treated rats. Statistical analysis was done by ANOVA with Dunnett’s multiple comparison as post hoc test. * P⬍0.05; ** P⬍0.01; *** P⬍0.001. Abbreviations: alv, alveus; Co, control; DG, dentate gyrus; ml, molecular layer; PaS, parasubiculum; pcl, pyramidal cell layer; PrS, presubiculum; Sub, subiculum.

pressing neurons in the nucleus reuniens of the epileptic rats (Fig. 7E, F).

DISCUSSION Loss of parvalbumin-ir interneurons in parahippocampal regions In all parahippocampal regions, the Ca2⫹-binding protein PV is expressed in subgroups of GABAergic interneurons. Among these, PV-expressing basket- and axo-axonic cells are considered as the most powerful inhibitory interneu-

rons (Katsumaru et al., 1988; Sloviter, 1989; Somogyi and Klausberger, 2005). They innervate the perisomatic area and axon initial segments of principal neurons and thus control their output (Fig. 8). In the hippocampus proper, they extend their dendrites into the stratum lacunosummoleculare where they receive glutamatergic input originating in layer III of the EC (temporo-ammonic path; Kiss et al., 1996). Hence, electrical stimulation of the temporoammonic pathway in control rats evokes prominent inhibition in rat CA1 pyramidal cells (Colbert and Levy, 1992; Buzsaki et al., 1995). Also in the subiculum, dendrites of

322

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

Fig. 3. Parvalbumin-ir and parvalbumin mRNA containing neurons in the medial entorhinal cortex after KA-induced seizures. PV-ir neurons were distributed throughout the superficial (layers II and III) and the deep EC (layers IV to VI) of controls (A). Notably in the deep layers their number was reduced, both 24 h (B) and 3 mon (C) after KA-induced SE (refer to cell counts in Table 1). In (D) and (E), dark field images of photoemulsion dipped section after in situ hybridization of PV mRNA is depicted. Note the increased density of silver grains (reflecting PV transcript) 3 mon after KA (arrows in E) compared with the control (arrows in D). In panel (F) the density of silver grains (representing PV transcript) is shown. Using ImageJ, it was measured as the area over the PV mRNA positive neurons covered by silver grains and is presented as mean values⫾SEM from 475 neurons of controls and 139 to 432 neurons for KA treated rats. The increase in PV mRNA concentrations in surviving PV neurons became significant 8 d after KA in the superficial EC and 1 and 3 months after KA in all layers of the EC (F). Data are shown as mean⫾SEM. Statistical analysis was done by ANOVA with Dunnett’s multiple comparison as post hoc test. ** P⬍0.01; *** P⬍0.001. Abbreviations: a, alveus; Co, control; I–VI, layers I–VI of the EC.

GABAergic interneurons represent a considerable fraction (about 18% according to Baks-Te Bulte et al., 2005) of postsynaptic targets of the temporo-ammonic path, presumably mediating powerful feed-forward inhibition (Baks-Te Bulte et al., 2005). Wouterlood et al. (2008b) reported that 16% of boutons of temporo-ammonic axons made contact with PV-ir dendrites in the subiculum, indicating that among the GABAergic interneurons of the subiculum, the population of PV-ir interneurons most likely represents the main recipient of entorhinal input and mediates powerful feed-forward inhibition. Subiculum. The number of PV-ir interneurons was primarily reduced in the superficial portion of the pyramidal cell layer of the subiculum. Accordingly, we observed pronounced PV-ir debris and numerous degenerating PV-ir

dendrites in the pyramidal cell layer and in the molecular layer of the subiculum already 24 h after KA-injection. Taken together with the time course of PV-ir cell losses this indicates early degeneration of PV neurons related to the initial SE and is consistent with the high number of FluoroJade-C labeled neurons and microglia activation observed in the same animals especially at the early intervals (submitted for publication). A consequence of the loss of PV cells was the dramatic reduction of PV-ir dendrites in the molecular layer of the subiculum at the later intervals after KA. These observations are in line with studies in different animal models and in human TLE. Decreased numbers of PV-ir interneurons were reported in the subiculum (by 40 to 72%) after pilocarpine-induced seizures (de Guzman et al.,

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

323

Fig. 4. Parvalbumin (PV)-ir and PV containing neurons in the pre- and parasubiculum after KA-induced seizures. Only minor losses of PV-ir neurons were observed in the pre- and parasubiculum of rats 1 mon after KA-injection (B) compared to controls (A). Refer to Table 1 for cell counts. Abbreviations: Co, control; PaS, parasubiculum; PrS, Presubiculum.

2006; Knopp et al., 2008) and after electrically induced SE (van Vliet et al., 2004). These former studies, however, did not discriminate between the deep and superficial pyramidal cell layer of the subiculum that are differently affected after KA-induced seizures. This is important since, similar as in our study on KA-treated rats, also in specimens from TLE patients the loss of PV-ir interneurons was particularly prominent (by 55%) in the superficial aspect of the subicular pyramidal cell layer (Andrioli et al., 2007). The reason may be that in the deep pyramidal cell layer PV is also expressed in somatostatin containing “stratum oriens-lacunosum moleculare” cells (O-LM cells) that are less vulnerable since they are not innervated by the temporo-ammonic pathway (Fig. 8) (Klausberger et al., 2003; Ferraguti et al., 2004). Entorhinal cortex. In the EC the number of PV-ir interneurons was primarily reduced in the deep layers and to a lesser extent in its superficial layers. A similar reduction in PV-ir neurons also was observed in the deep layers of the medial EC after electrically induced SE (van Vliet et al., 2004) and in the pilocarpine model (⫺40% according to de Guzman et al., 2008). We now report abundant PV-ir debris in the deep entorhinal layers already 24 h after KA injection, indicating actual neurodegeneration of PV-ir neurons. Vulnerability of hippocampal PV-ir interneurons in epilepsy. A general vulnerability of PV-ir interneurons in different areas of the epileptic hippocampus, however, is sometimes debated. Alternatively it has been suggested that suppressed PV synthesis may be the cause for the apparent seizure-induced reduction of PV-positive cells (Sloviter et al., 1991). Thus Scotti et al. (1997) found a reduction of PV-ir interneurons in sector CA1 of epileptic gerbils, although the number of GABAergic neurons was unchanged. In contrast, our experiments revealed a concomitant reduction of GAD65- and PV-ir neurons in the subiculum of KA treated rats. The portion of PV-ir lost, however, exceeded the number of degenerating GAD65 neurons, presumably due to the existence of less vulner-

able GAD65 positive neurons. Also in the dentate gyrus and hippocampus of TLE patients apparently decreased numbers of PV-ir interneurons were observed (Wittner et al., 2001, 2005). The symmetrical synaptic coverage of perikarya of principal neurons (mainly comprising PV terminals), however, remained unchanged, indicating that the respective neurons stayed intact in spite of a loss in PVlabeling (Wittner et al., 2001, 2005, 2009). Degeneration of PV-ir interneurons, notably in the hippocampal sector CA1, has been reported in different animal models (Best et al., 1993; Andre et al., 2001; Dinocourt et al., 2003; Kobayashi et al., 2003) and in tissue of TLE patients (Zhu et al., 1997; Andrioli et al., 2007). On the ultrastructural level, Best et al. (1994) demonstrated many degenerating PV-ir interneurons in sector CA1 and an almost entire loss of PV-ir terminals around the soma of CA1 pyramidal neurons after unilateral i.c.v. KA-injection. In their study, PV-ir terminals around the axon initial segments of CA1 pyramidal neurons, however, were resistant to KA-induced neurodegeneration, indicating a preferred initial loss of basket cells and preservation of axo-axonic cells. In contrast, degeneration of PV-ir interneurons and a concomitant loss of axo-axonic coverage were reported in sector CA1 in the pilocarpine model (Dinocourt et al., 2003). Possible contribution of the loss in PV-ir neurons in the subiculum and entorhinal cortex to the development of spontaneous seizures There was a high correlation of losses in PV-ir interneurons in the superficial and deep principal cell layer of the subiculum and in layer III of the medial EC and the number of spontaneous seizures. The loss in PV-ir (after 24 h) preceded the average onset of spontaneous seizures (mean 15 days, range 3–36 days after the initial SE (submitted for publication)). This indicates that the loss in PV-ir interneurons in the subiculum and EC may be causatively related to the generation of spontaneous seizure activity.

324

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

may therefore contribute to the enhanced network excitability of the subiculum reported in the pilocarpine model (Knopp et al., 2005; de Guzman et al., 2006) and in human tissue (Cohen et al., 2002; Wozny et al., 2005b; see Fig. 8). Also the loss of PV-ir in the deep EC may crucially alter its function and may contribute to epileptogenesis. The deep layers of the EC are among the main targets of projections arising from pyramidal neurons of the subiculum. Loss of these PV-positive, GABAergic interneurons may contribute to the reported hyperexcitability and to the reduced inhibition demonstrated in the deep layers of the EC recently (de Guzman et al., 2008). It is, however, interesting to note that surviving PV neurons showed markedly increased PV mRNA expression. Since this increase in mRNA concentration occurs especially at the late intervals after KA seizures, it may be triggered by increasing numbers of spontaneous seizures and may reflect general activation of residual PV neurons. This mechanism could aim to compensate for the loss of other PV neurons. Increased expression of PV, however may also protect these neurons from further cell death (Leranth and Ribak, 1991). Loss of calretinin-containing interneurons in the subiculum

Fig. 5. Correlation analysis of loss of parvalbumin-ir neurons and number of spontaneous seizures. The number of PV-ir neurons in the superficial (A) and deep (B) pyramidal cell layer (pcl) of the subiculum and in layer III of the medial EC (C) determined 3 mon after KA injection negatively correlated with the number of spontaneous seizures experienced. Pearson correlation (r2) and P-values were determined using the GraphPad Prism statistic program. No correlation was observed between the number of spontaneous seizures and loss of PV-ir neurons in other parahippocampal subregions (presubiculum, r2⫽0.1879, P⫽0.1592; parasubiculum, r2⫽0.0556, P⫽0.4606; medial EC: layer II, r2⫽0.2785, P⫽0.0778; layers V/VI, r2⫽0.0193, P⫽0.6666; lateral EC: layer II, r2⫽0.3197, P⫽0.0554; layer III, r2⫽0.3106, P⫽0.0598; layers V/VI, r2⫽0.1106, P⫽0.2908).

The high vulnerability of PV-ir interneurons may certainly be related to the presumably high number of excitatory synapses (⬎15,000 excitatory synapses per neuron, e.g. in sector CA1 of the hippocampus) compared with a low number of inhibitory synapses (⬃ 1050) converging on these interneurons, and also compared with the considerably lower number of excitatory synapses on other subclasses of interneurons (containing CR, calbindin, or cholecystokinin; Gulyas et al., 1999; Matyas et al., 2004). The loss of PV-ir GABAergic interneurons in the subiculum can considerably alter the function of the temporo-ammonic path. Degeneration of PV-ir interneurons in the subiculum presumably leads to a loss of feed-forward inhibition and

Losses in CR containing neurons were observed in sectors CA1 and CA3 in animal models and in human TLE (Magloczky and Freund, 1993; Suckling et al., 2000; Andre et al., 2001; Toth et al., 2010). Recently Knopp et al. (2008) also reported a significant decrease of CRpositive neurons notably in the molecular layer of the subiculum of pilocarpine treated rats. In the CA1 region of the rat, CR-containing cells selectively terminate on interneurons mostly also containing calbindin (Gulyas et al., 1996). The CR-containing inhibitory cells are probably responsible for the synchronization of dendritic inhibitory cells and thus for the inhibitory control of excitatory synaptic input of principal cell dendrites (Gulyas et al., 1996; Miles et al., 1996). The loss of these mechanisms may lead to the formation of principal cell assemblies connected by abnormally potentiated synapses, which may be involved in epileptogenesis and seizure generation (Toth et al., 2010). Loss of calretinin-containing afferent projections originating in the thalamus CR-containing glutamatergic projections from the thalamic nucleus reuniens to the EC (Wouterlood et al., 2008a) and to stratum lacunosum-moleculare of sector CA1 of the hippocampus have been demonstrated recently (Bokor et al., 2002). Also the molecular layer of the subiculum receives glutamatergic projections from the thalamic nucleus reuniens (Wouterlood et al., 1990; see Fig. 8). We now observed reduced density of CR-ir fibers in the subiculum, in sector CA1 and in layer I of the EC that coincides with the degeneration of CR- and NeuN-ir neurons in the nucleus reuniens. Taken together these findings indicate that

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

325

Fig. 6. Loss of CR-ir interneurons and decreased fiber density in the subiculum and EC of KA-injected rats. In controls, CR-ir interneurons were located in all layers of the subiculum (A). A dense CR-ir fiber plexus was present in the outer molecular layer of the subiculum (arrow in (A, B) and in layer I of the EC (C). CR-labeling in the outer molecular layer of the subiculum was slightly increased 24 h after KA-induced SE (arrow in D), presumably reflecting activation of CR synthesis prior to neurodegeneration. CR-ir labeling in the outer molecular layer, however, significantly decreased after 3 mon (arrow in G). High magnification microphotographs show the transiently increased CR-labeling in the outer molecular layer of the subiculum (E) and in layer I of the medial EC (F) 24 h after KA-induced SE. Note the significant decrease of CR-labeling intensity in these areas 3 mon after SE (H and I). Panels (J) and (K): Bar graphs comparing the mean density of CR-ir interneurons in the subiculum (J) and the intensity of fiber labeling in the outer molecular layer of the subiculum (K) in controls and at different time intervals after KA-induced SE (mean⫾SEM, ANOVA with Dunnett’s post hoc test, * P⬍0.05; ** P⬍0.01; *** P⬍0.001; numbers of animals are given in columns). Abbreviations: Co, control; DG, dentate gyrus; PrS, presubiculum; Sub, subiculum.

also the projections from the nucleus reuniens to the subiculum, like those to sector CA1 and to the EC, contain CR and degenerate after KA-induced SE (Fig. 8). The excitatory innervation of hippocampal and parahippocampal areas arising from the nucleus reuniens thalami may have a crucial role in the epileptogenesis after KA-induced seizures. Bertram et al. (2001) found a 40% loss of Nissl stained neurons in the nucleus reuniens of chronic epileptic rats in a model of electrically induced SE

and demonstrated that the midline thalamus was consistently involved in limbic seizure activity. They suggested that the seizure onset may be simultaneous in the hippocampus and the midline thalamus. The excitatory drive of the nucleus reuniens projections to the stratum lacunosum-moleculare of sector CA1 is much stronger on dendrites of GABAergic neurons than on those of pyramidal cells (Dolleman-Van der Weel et al., 1997; see, however, Bertram and Zhang, 1999). Assuming similar functional

326

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

Fig. 7. Loss of NeuN- and CR-ir neurons in the nucleus reuniens of the thalamus. Panels (A) and (C) illustrate the density of NeuN- and CR-ir neurons respectively. In the nucleus reuniens of controls. Note the decreased density of NeuN- (arrow in B) and CR-labeled neurons (arrows in D) in the nucleus reuniens 3 mon after KA-induced SE. The number of neurons expressing CR mRNA was decreased in the nucleus reuniens 3 mon after SE (arrows in F) compared to controls (E). Panel (G) illustrates the borders of the nucleus reuniens on a Nissl-stained horizontal section of a control rat. Abbreviations: PVA, paraventricular thalamic nucleus (anterior); RE, nucleus reuniens. Bar graphs in (H) compare mean densities of NeuN- and CR-labeled neurons in the nucleus reuniens of controls (white bars) and of epileptic rats 3 mon after SE (black bars). Note the significantly reduced neuron numbers in epileptic rats (mean⫾SEM, Student’s t-test, *** P⬍0.001, n⫽5 for controls, n⫽15 for KA treated rats).

differentiation of terminals of the nucleus reuniens thalami in the outer molecular layer of the subiculum, the loss of inhibition in the subiculum may also prevail that of excitation after epilepsy-induced neurodegeneration of this pathway. Conclusions Losses in PV containing neurons affect the entire subiculum and even more the deep layers of the EC. Interestingly, although the general neurodegeneration is by far

less severe in the subiculum of TLE patients than in the rat model, losses of PV-positive neurons are equally extensive. They result in an impairment of feed forward inhibition and very likely contribute to increased excitability of parahippocampal regions and by this to epileptogenesis. This is also indicated by the correlation between losses of PV neurons and subsequent spontaneous seizures. Furthermore, also the prominent loss of excitatory CR-positive neurons projecting from the thalamic nucleus reuniens to the subiculum, to sector CA1 and to the EC may funda-

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

Fig. 8. Proposed organization of basket/axo-axonic cells and orienslacunosum-moleculare (O-LM) cells in the subiculum. The subiculum receives a main excitatory input from layer III of the medial EC in the outer molecular layer (oml). These glutamatergic terminals impinge on dendrites of pyramidal cells (PC) and on dendrites of PV-positive GABAergic basket/axo-axonic cells (BC) (Wouterlood et al., 2008b). They are located in the deep and superficial portions of the pyramidal cell layers (pcl) and control the output of the PC. Basket/axo-axonic cells receive a strong excitatory innervation by the temporo-ammonic pathway arising from layer III of the EC that presumably mediates seizure-induced neurodegeneration in the respective target neurons (see Fig. 1, Table 1). Notably the surviving neurons express more PV mRNA indicating that they become constantly activated. On the other hand, PV (and somatostatin) -positive GABAergic interneurons, corresponding to O-LM-cells of hippocampal sector CA1, are located in the deep pcl of the subiculum and innervate the distal dendrites of pyramidal neurons in the outer molecular layer and mediate feed-back inhibition. Since their dendrites are not innervated by the temporoammonic pathway, they are resistant to seizure-induced damage (see Fig. 1E, F, Table 1). CR-positive fibers originating from the nucleus reuniens thalami build up a strong excitatory input on dendrites of GABAergic (solid line) and a weaker input on glutamatergic neurons (dotted line). Together with their perikarya in the nucleus reuniens they degenerate after KA-induced SE (see Fig. 6G, H), presumably resulting in increased excitation. Neurons and projections lost in epilepsy are indicated by a cross. For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.

mentally contribute to the altered neuronal network in the hippocampus and EC. Acknowledgments—The authors thank Anna Wieselthaler-Hölzl and Elisabeth Gasser for excellent technical assistance. This research was supported by the Austrian Research Funds (grant number P19464) and by the European Union Grant FP6 EPICURE (LSH-CT-2006-037315). The funding sources had no involvement in study design, in the collection, analysis and interpretation of data, in the writing of the report, or in the decision to submit the paper for publication.

REFERENCES Andre V, Marescaux C, Nehlig A, Fritschy JM (2001) Alterations of hippocampal GABAergic system contribute to development of

327

spontaneous recurrent seizures in the rat lithium-pilocarpine model of temporal lobe epilepsy. Hippocampus 11:452– 468. Andrioli A, Alonso-Nanclares L, Arellano JI, Defelipe J (2007) Quantitative analysis of parvalbumin-immunoreactive cells in the human epileptic hippocampus. Neuroscience 149:131–143. Babb TL, Brown WJ, Pretorius J, Davenport C, Lieb JP, Crandall PH (1984) Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia 25:729 –740. Baks-Te Bulte L, Wouterlood FG, Vinkenoog M, Witter MP (2005) Entorhinal projections terminate onto principal neurons and interneurons in the subiculum: a quantitative electron microscopical analysis in the rat. Neuroscience 136:729 –739. Baulac S, Gourfinkel-An I, Nabbout R, Huberfeld G, Serratosa J, Leguern E, Baulac M (2004) Fever, genes, and epilepsy. Lancet Neurol 3:421– 430. Ben-Ari Y, Tremblay E, Ottersen OP, Meldrum BS (1980) The role of epileptic activity in hippocampal and “remote” cerebral lesions induced by kainic acid. Brain Res 191:79 –97. Berchtold MW, Heizmann CW, Wilson KJ (1982) Primary structure of parvalbumin from rat skeletal muscle. Eur J Biochem 127:381–389. Bertram EH, Mangan PS, Zhang D, Scott CA, Williamson JM (2001) The midline thalamus: alterations and a potential role in limbic epilepsy. Epilepsia 42:967–978. Bertram EH, Zhang DX (1999) Thalamic excitation of hippocampal CA1 neurons: a comparison with the effects of CA3 stimulation. Neuroscience 92:15–26. Best N, Mitchell J, Baimbridge KG, Wheal HV (1993) Changes in parvalbumin-immunoreactive neurons in the rat hippocampus following a kainic acid lesion. Neurosci Lett 155:1– 6. Best N, Mitchell J, Wheal HV (1994) Ultrastructure of parvalbuminimmunoreactive neurons in the CA1 area of the rat hippocampus following a kainic acid injection. Acta Neuropathol 87:187–195. Bokor H, Csaki A, Kocsis K, Kiss J (2002) Cellular architecture of the nucleus reuniens thalami and its putative aspartatergic/glutamatergic projection to the hippocampus and medial septum in the rat. Eur J Neurosci 16:1227–1239. Bragin DE, Sanderson JL, Peterson S, Connor JA, Muller WS (2009) Development of epileptiform excitability in the deep entorhinal cortex after status epilepticus. Eur J Neurosci 30:611– 624. Bratz E (1899) Ammonshornbefunde bei Epileptikern. Arch Psychiatr Nervenkrankh 31:820 – 835. Buzsaki G, Penttonen M, Bragin A, Nadasdy Z, Chrobak JJ (1995) Possible physiological role of the perforant path-CA1 projection. Hippocampus 5:141–146. Cavalheiro EA, Riche DA, Le Gal La Salle G (1982) Long-term effects of intrahippocampal kainic acid injection in rats: a method for inducing spontaneous recurrent seizures. Electroencephalogr Clin Neurophysiol 53:581–589. Chang YC, Gottlieb DI (1988) Characterization of the proteins purified with monoclonal antibodies to glutamic acid decarboxylase. J Neurosci 8:2123–2130. Cohen I, Navarro V, Clemenceau S, Baulac M, Miles R (2002) On the origin of interictal activity in human temporal lobe epilepsy in vitro. Science 298:1418 –1421. Colbert CM, Levy WB (1992) Electrophysiological and pharmacological characterization of perforant path synapses in CA1: mediation by glutamate receptors. J Neurophysiol 68:1– 8. de Guzman P, Inaba Y, Baldelli E, de Curtis M, Biagini G, Avoli M (2008) Network hyperexcitability within the deep layers of the pilocarpine-treated rat entorhinal cortex. J Physiol 586:1867–1883. de Guzman P, Inaba Y, Biagini G, Baldelli E, Mollinari C, Merlo D, Avoli M (2006) Subiculum network excitability is increased in a rodent model of temporal lobe epilepsy. Hippocampus 16:843– 860. Dinocourt C, Petanjek Z, Freund TF, Ben-Ari Y, Esclapez M (2003) Loss of interneurons innervating pyramidal cell dendrites and axon initial segments in the CA1 region of the hippocampus following pilocarpine-induced seizures. J Comp Neurol 459:407– 425.

328

M. Drexel et al. / Neuroscience 189 (2011) 316 –329

Dolleman-Van der Weel MJ, Lopes da Silva FH, Witter MP (1997) Nucleus reuniens thalami modulates activity in hippocampal field CA1 through excitatory and inhibitory mechanisms. J Neurosci 17:5640 –5650. Dolleman-Van der Weel MJ, Witter MP (2000) Nucleus reuniens thalami innervates gamma aminobutyric acid positive cells in hippocampal field CA1 of the rat. Neurosci Lett 278:145–148. Engel J Jr. (2001) Mesial temporal lobe epilepsy: what have we learned? Neuroscientist 7:340 –352. Ferraguti F, Cobden P, Pollard M, Cope D, Shigemoto R, Watanabe M, Somogyi P (2004) Immunolocalization of metabotropic glutamate receptor 1alpha (mGluR1alpha) in distinct classes of interneuron in the CA1 region of the rat hippocampus. Hippocampus 14:193–215. Furtinger S, Pirker S, Czech T, Baumgartner C, Ransmayr G, Sperk G (2001) Plasticity of Y1 and Y2 receptors and neuropeptide Y fibers in patients with temporal lobe epilepsy. J Neurosci 21:5804 –5812. Gabrielides C, McCormack AL, Hunt DF, Christakos S (1991) Brain calbindin-D28k and an Mr 29,000 calcium binding protein in cerebellum are different but related proteins: evidence obtained from sequence analysis by tandem mass spectrometry. Biochemistry 30:656 – 662. Gulyas AI, Hajos N, Freund TF (1996) Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J Neurosci 16:3397–3411. Gulyas AI, Megias M, Emri Z, Freund TF (1999) Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19:10082–10097. Gulyas AI, Miettinen R, Jacobowitz DM, Freund TF (1992) Calretinin is present in non-pyramidal cells of the rat hippocampus—I. A new type of neuron specifically associated with the mossy fibre system. Neuroscience 48:1–27. Katsumaru H, Kosaka T, Heizmann CW, Hama K (1988) Immunocytochemical study of GABAergic neurons containing the calciumbinding protein parvalbumin in the rat hippocampus. Exp Brain Res 72:347–362. Kiss J, Buzsaki G, Morrow JS, Glantz SB, Leranth C (1996) Entorhinal cortical innervation of parvalbumin-containing neurons (Basket and Chandelier cells) in the rat Ammon’s horn. Hippocampus 6:239 –246. Klausberger T, Magill PJ, Marton LF, Roberts JD, Cobden PM, Buzsaki G, Somogyi P (2003) Brain-state- and cell-type-specific firing of hippocampal interneurons in vivo. Nature 421:844 – 848. Knopp A, Frahm C, Fidzinski P, Witte OW, Behr J (2008) Loss of GABAergic neurons in the subiculum and its functional implications in temporal lobe epilepsy. Brain 131:1516 –1527. Knopp A, Kivi A, Wozny C, Heinemann U, Behr J (2005) Cellular and network properties of the subiculum in the pilocarpine model of temporal lobe epilepsy. J Comp Neurol 483:476 – 488. Kobayashi M, Wen X, Buckmaster PS (2003) Reduced inhibition and increased output of layer II neurons in the medial entorhinal cortex in a model of temporal lobe epilepsy. J Neurosci 23:8471– 8479. Leite JP, Garcia-Cairasco N, Cavalheiro EA (2002) New insights from the use of pilocarpine and kainate models. Epilepsy Res 50:93–103. Leranth C, Ribak CE (1991) Calcium-binding proteins are concentrated in the CA2 field of the monkey hippocampus: a possible key to this region’s resistance to epileptic damage. Exp Brain Res 85:129 –136. Magloczky Z, Freund TF (1993) Selective neuronal death in the contralateral hippocampus following unilateral kainate injections into the CA3 subfield. Neuroscience 56:317–335. Matyas F, Freund TF, Gulyas AI (2004) Convergence of excitatory and inhibitory inputs onto CCK-containing basket cells in the CA1 area of the rat hippocampus. Eur J Neurosci 19:1243–1256.

Miettinen M, Pitkanen A, Miettinen R (1997) Distribution of calretininimmunoreactivity in the rat entorhinal cortex: coexistence with GABA. J Comp Neurol 378:363–378. Miles R, Toth K, Gulyas AI, Hajos N, Freund TF (1996) Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16:815– 823. Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates, 4th edition. New York: Academic Press. Scotti AL, Kalt G, Bollag O, Nitsch C (1997) Parvalbumin disappears from GABAergic CA1 neurons of the gerbil hippocampus with seizure onset while its presence persists in the perforant path. Brain Res 760:109 –117. Sloviter RS (1989) Calcium-binding protein (calbindin-D28k) and parvalbumin immunocytochemistry: localization in the rat hippocampus with specific reference to the selective vulnerability of hippocampal neurons to seizure activity. J Comp Neurol 280:183–196. Sloviter RS, Sollas AL, Barbaro NM, Laxer KD (1991) Calcium-binding protein (calbindin-D28K) and parvalbumin immunocytochemistry in the normal and epileptic human hippocampus. J Comp Neurol 308:381–396. Sommer W (1880) Erkrankung des Ammonshornes als ätiologisches Moment der Epilepsie. Arch Psychiatr Nervenkrankh 10:631– 675. Somogyi P, Klausberger T (2005) Defined types of cortical interneurone structure space and spike timing in the hippocampus. J Physiol 562:9 –26. Sperk G (1994) Kainic acid seizures in the rat. Prog Neurobiol 42:1–32. Sperk G, Lassmann H, Baran H, Kish SJ, Seitelberger F, Hornykiewicz O (1983) Kainic acid induced seizures: neurochemical and histopathological changes. Neuroscience 10:1301–1315. Suckling J, Roberts H, Walker M, Highley JR, Fenwick P, Oxbury J, Esiri MM (2000) Temporal lobe epilepsy with and without psychosis: exploration of hippocampal pathology including that in subpopulations of neurons defined by their content of immunoreactive calcium-binding proteins. Acta Neuropathol 99:547–554. Toth K, Eross L, Vajda J, Halasz P, Freund TF, Magloczky Z (2010) Loss and reorganization of calretinin-containing interneurons in the epileptic human hippocampus. Brain 133:2763–2777. Tsunashima K, Schwarzer C, Kirchmair E, Sieghart W, Sperk G (1997) GABA(A) receptor subunits in the rat hippocampus III: altered messenger RNA expression in kainic acid-induced epilepsy. Neuroscience 80:1019 –1032. van Vliet EA, Aronica E, Tolner EA, Lopes da Silva FH, Gorter JA (2004) Progression of temporal lobe epilepsy in the rat is associated with immunocytochemical changes in inhibitory interneurons in specific regions of the hippocampal formation. Exp Neurol 187:367–379. Wittner L, Eross L, Czirjak S, Halasz P, Freund TF, Magloczky Z (2005) Surviving CA1 pyramidal cells receive intact perisomatic inhibitory input in the human epileptic hippocampus. Brain 128:138 –152. Wittner L, Huberfeld G, Clemenceau S, Eross L, Dezamis E, Entz L, Ulbert I, Baulac M, Freund TF, Magloczky Z, Miles R (2009) The epileptic human hippocampal cornu ammonis 2 region generates spontaneous interictal-like activity in vitro. Brain 132:3032–3046. Wittner L, Magloczky Z, Borhegyi Z, Halasz P, Toth S, Eross L, Szabo Z, Freund TF (2001) Preservation of perisomatic inhibitory input of granule cells in the epileptic human dentate gyrus. Neuroscience 108:587– 600. Wouterlood FG, Aliane V, Boekel AJ, Hur EE, Zaborszky L, BarrosoChinea P, Hartig W, Lanciego JL, Witter MP (2008a) Origin of calretinin-containing, vesicular glutamate transporter 2-coexpressing fiber terminals in the entorhinal cortex of the rat. J Comp Neurol 506:359 –370. Wouterlood FG, Boekel AJ, Aliane V, Belien JA, Uylings HB, Witter MP (2008b) Contacts between medial and lateral perforant pathway fibers and parvalbumin expressing neurons in the subiculum of the rat. Neuroscience 156:653– 661.

M. Drexel et al. / Neuroscience 189 (2011) 316 –329 Wouterlood FG, Saldana E, Witter MP (1990) Projection from the nucleus reuniens thalami to the hippocampal region: light and electron microscopic tracing study in the rat with the anterograde tracer Phaseolus vulgaris-leucoagglutinin. J Comp Neurol 296:179 –203. Wozny C, Gabriel S, Jandova K, Schulze K, Heinemann U, Behr J (2005a) Entorhinal cortex entrains epileptiform activity in CA1 in pilocarpine-treated rats. Neurobiol Dis 19:451– 460.

329

Wozny C, Knopp A, Lehmann TN, Heinemann U, Behr J (2005b) The subiculum: a potential site of ictogenesis in human temporal lobe epilepsy. Epilepsia 46 (Suppl 5):17–21. Zhu ZQ, Armstrong DL, Hamilton WJ, Grossman RG (1997) Disproportionate loss of CA4 parvalbumin-immunoreactive interneurons in patients with Ammon’s horn sclerosis. J Neuropathol Exp Neurol 56:988 –998.

(Accepted 11 May 2011) (Available online 18 May 2011)