Differential vulnerability of interneurons in the epileptic hippocampus

Abstract

The loss of hippocampal interneurons has been considered as one reason for the onset of temporal lobe epilepsy (TLE) by shifting the excitation-inhibition balance. Yet, there are many different interneuron types which show differential vulnerability in the context of an epileptogenic insult. We used the intrahippocampal kainate (KA) mouse model for TLE in which a focal, unilateral KA injection induces status epilepticus (SE) followed by development of granule cell dispersion (GCD) and hippocampal sclerosis surrounding the injection site but not in the intermediate and temporal hippocampus. In this study, we characterized the loss of interneurons with respect to septotemporal position and to differential vulnerability of interneuron populations. To this end, we performed intrahippocampal recordings of the initial SE, in situ hybridization for glutamic acid decarboxylase 67 (GAD67) mRNA and immunohistochemistry for parvalbumin (PV) and neuropeptide Y (NPY) in the early phase of epileptogenesis at 2 days and at 21 days after KA injection, when recurrent epileptic activity and GCD have fully developed. We show that SE extended along the entire septotemporal axis of both hippocampi, but was stronger at distant sites than at the injection site. There was an almost complete loss of interneurons surrounding the injection site and expanding to the intermediate hippocampus already at 2 days but increasing until 21 days after KA. Furthermore, we observed differential vulnerability of PV- and NPY-expressing cells: while the latter were lost at the injection site but preserved at intermediate sites, PV-expressing cells were gone even at sites more temporal than GCD. In addition, we found upregulation of GAD67 mRNA expression in dispersed granule cells and of NPY staining in ipsilateral granule cells and ipsi- and contralateral mossy fibers. Our data thus indicate differential survival capacity of interneurons in the epileptic hippocampus and compensatory plasticity mechanisms depending on the hippocampal position.

Keywords: glutamic acid decarboxylase; kainate injection; neuropeptide Y; parvalbumin; septotemporal axis; temporal lobe epilepsy.

FIGURE 1

Representative LFP recording during status epilepticus (SE). (A, B) LFPs were recorded at 4 positions (1: most septal, 4: most temporal) in the ipsilateral and at one position in the contralateral (c) hippocampus (for illustration of electrode positions see schematic brain drawing) starting from ~2 h after awakening from surgery. SE was characterized by a repetitive pattern of spike-and-wave discharges with short intermittent depression periods, accompanied by behavioral signs such as convulsive movements, rotation or immobility. (B) Enlargement of (A). SE activity was strongest in the intermediate and contralateral hippocampus at this time point.

FIGURE 2

Status epilepticus is followed by prominent neuronal death in the septal ipsilateral hippocampus. (A–E) Representative Fluoro-Jade B-stained sections of a KA-injected mouse at one day after KA injection to monitor neuronal degeneration. (A, C) The contralateral hippocampus was devoid of any cell death in the septal (A) and intermediate and temporal hippocampus (C). (B) In the ipsilateral hippocampus prominent Fluoro-Jade B staining occurred in the pyramidal cell layer of CA3 and CA1 and in the hilus, but single Fluoro-Jade B-positive cells were also located in strata oriens and radiatum of CA3 and CA1 and in the inner and outer portion of the granule cell layer. An enlargement of the hilus and granule cell layer, displaying the shape of Fluoro-Jade B-positive cells is shown in (E). (D) Fluoro-Jade B-positive cells were also visible in the intermediate hippocampus but the temporal hippocampus was devoid of dying cells. The arrow marks the most temporally located Fluoro-Jade B-positive cell group. (F) Densitometric analysis of Fluoro-Jade B-positive areas relative to the area of the whole hippocampus in the same section at four positions along the septotemporal axis (see scheme in Figure 1) in the ipsilateral and contralateral hippocampus. Values are displayed as mean ± SEM. A significant difference was observed at position 2 (n = 5, p < 0.001, ANOVA, ***p < 0.001 Tukey’s post-test). Scale bars: A–D, 100 μm; E, 50 μm. GCL, granule cell layer; H, hilus; CA1, CA2, CA3, cornu ammonis; I1, ipsilateral position 1; C1, contralateral position 1.

FIGURE 3

Septotemporal distribution and quantification of GAD67 mRNA expressing neurons at 2 and 21 days after KA injection. In situ hybridization for GAD67 mRNA was performed as described in Section “Materials and Methods.’ Representative coronal sections of four septotemporal levels are shown. (A–D) Control mouse, septal (A, B), intermediate (C) and temporal (D) hippocampus. GAD67 mRNA-expressing cells were distributed throughout the hippocampus but accumulated close to the granule cell layer and pyramidal cell layer. This pattern was comparable at all septotemporal levels. Granule cells in the dentate gyrus show weak GAD67 mRNA expression. (E–H) Contralateral hippocampus, 2 days after KA injection, septal (E, F), intermediate (G) and temporal level (H). The distribution and density of GAD67 mRNA-positive cells was comparable to controls at all septotemporal levels. In some mice GAD67 mRNA-positive cells were slightly reduced in distal CA1 of the septal hippocampus (E). GAD67 mRNA expression in granule cells was comparable to controls or slightly upregulated in the septal hippocampus only. (I–L) Ipsilateral hippocampus, 2 days after KA injection. (I, J) In the septal hippocampus most GAD67 mRNA-expressing cells were lost except for a group of cells in CA3/CA2 (arrow). GAD67 mRNA expression in granule cells was upregulated. (K) In the more dorsal part of the intermediate hippocampus GAD67 mRNA-expressing interneurons were lost, whereas in the more ventral parts they were preserved. GAD67 mRNA expression in granule cells was upregulated. (L) In the temporal hippocampus GAD67 mRNA-expressing interneurons were only lost in the very dorsal parts of the section but preserved elsewhere, GAD67 mRNA expression in granule cells was comparable to controls. (M–P) Ipsilateral hippocampus, 21 days after KA injection. (M, N) In the septal hippocampus only a few GAD67 mRNA-expressing cells were preserved. Note that cells in CA3, which were present earlier, were lost at the late time point indicating a progressive interneuron loss. The apparent reduction of GAD67 mRNA upregulation in granule cells is most likely due to their reduced density owing to the prominent dispersion of the granule cell layer. (O) In the intermediate hippocampus GAD67 mRNA-expressing cells were lost to a larger extent than at 2 days after KA and GAD67 mRNA expression in the granule cells was still strongly increased. (P) In the temporal hippocampus, GAD67 mRNA-expressing interneurons were lost in the dorsal part of the section indicating the progressive loss of interneurons. GAD67 mRNA was upregulated only in dispersed granule cells (arrow marks the transition zone). In the ventral parts of the section GAD67 mRNA expression was comparable to controls. (Q) Quantification of GAD67 mRNA-expressing cells at four positions along the septotemporal axis in controls and at 2 and 21 days after KA. Cells were counted in the whole hippocampus and cell density is given as GAD67-positive cells/mm²(n = 5 each group). Values are displayed as mean ± SEM, statistical comparison was made with an ANOVA and Tukey’s multiple comparison test, significance values were set as follows: *p < 0.05, **p < 0.01, ***p < 0.001. (R) As in (Q), but only the hilus was marked as region of interest. Note that differences at position three are only significant when only the area of the hilus is considered. (S) Enlargement of the transition zone, see box marked in (P). A few GAD67 mRNA-positive cells appeared dorsal to the transition zone (arrows) but the full density of GAD67 mRNA-positive cells was only restored ventral to this area. Note that GAD67 mRNA upregulation affected only dispersed granule cells. (T) Quantification of GAD67 mRNA-expressing cells only in the areas of the hippocampus where GCD was present at 21 days (in contrast to (Q), where the whole hippocampus was considered). The loss was only significant at 21 days, indicating the progressive loss of interneurons in the temporal hippocampus (p = 0.015, ANOVA, p < 0.05, Tukey’s post-test, n = 5 each group). (U) Same as in T but only for the area of the hilus (p = 0.008, ANOVA, p < 0.01, Tukey’s post-test, n = 5 each group). Scale bars: 200 μm, S: 50 μm. GCL, granule cell layer; H, hilus; CA1, CA2, CA3, cornu ammonis.

FIGURE 4

Spatial distribution of GAD67 mRNA-positive interneurons. (A–L) Density of GAD67 mRNA-expressing interneurons in representative hippocampal sections along the septotemporal axis displayed as heatmaps in false-color, relative to the maximal density within each position. The outer border of the hippocampus and the hilus are marked with white traces. Brighter color indicates higher, darker color lower relative cell density. (A–D) Control mouse, septal (A, B), intermediate (C) and temporal hippocampus (D). (E–H) Ipsilateral hippocampus, 2 days after KA injection, septal (E, F), intermediate (G) and temporal level (H). A gradient of cell loss along the septotemporal axis is visible. (I–L) Ipsilateral hippocampus, 21 days after KA injection, septal (I, J), intermediate (K) and temporal level (L). Note that the loss of GAD67 mRNA-expressing cells extends to further temporal areas than at 2 days after KA and the reduction of interneuron density can be observed beyond the transition zone from GCD to normal granule cell layer width (L, arrows). Scale bars: 200 μm.

FIGURE 5

Parvalbumin-positive cells are lost in the septal and intermediate hippocampus at 2 days after KA injection. (A–N) Representative sections of an immunocytochemical staining for PV. (A–C) Control mouse, septal (A), intermediate (B), and temporal hippocampus (C). PV-positive cells were mainly located in or close to the granule cell layer or the pyramidal cell layer and in stratum oriens and were equally distributed along the septotemporal axis. The PV-positive axon plexus surrounding the principal cell layers is visible. (D–N) Ipsilateral hippocampus, 2 days after KA. (D–F) In the septal hippocampus most PV-positive cells were lost, except for a few cells in CA3 (white box, enlarged in E, arrow marks PV-positive cell), and the PV-positive axon plexus was reduced particularly in the dentate gyrus and disorganized in CA1 (yellow box, enlarged in F). (G) The more dorsal areas of the intermediate hippocampus showed a cell loss pattern in CA1 comparable to the septal hippocampus (white arrow) whereas in the more ventral areas PV-positive cells were preserved (yellow arrow). In the dorsal areas of the dentate gyrus residual PV expression was observed close to the granule cell layer (white box, enlarged in H and I; with DAPI to highlight the granule cell layer). In the ventral areas of the intermediate dentate gyrus and in CA3 PV-expressing cells were preserved (yellow box, enlarged in J and K with DAPI). (L) In the temporal hippocampus PV-expressing cells were only reduced in the very dorsal parts of the section (arrow) but preserved elsewhere (white box, enlarged in M and N with DAPI). (O) Quantification of PV-positive interneurons at four positions along the septotemporal axis in controls and at 2 and 21 days after KA. Cells were counted in the whole hippocampus and cell density is given as PV-positive cells/mm². Values are displayed as mean ± SEM (n = 6 controls, n = 4 KA 2 days, n = 5 KA 21 days), statistical comparison was made with an ANOVA and Tukey’s multiple comparison test, significance values were set as follows: *p < 0.05, **p < 0.01, ***p < 0.001. (P) As in (O), but only the hilus was marked as region of interest. Note that at position two PV-positive cells in the hilus were nearly completely lost after KA injection. Scale bars: A–D, G, L, 200 μm; E, F, H–K, M, N, 100 μm.

FIGURE 6

Progressive loss of parvalbumin-positive cells at 21 days after KA injection. (A–D) Ipsilateral hippocampus, 21 days after KA. (A) In the septal hippocampus not only the PV-positive cells but also the PV-positive axon plexus almost completely vanished, indicating a progressive loss. A few dysmorphic cells remained in CA3 in some slices (white box, enlarged in B). Note the prominent GCD. (C) In the intermediate hippocampus PV-positive cells and the PV-positive axon plexus were lost in the dorsal part (white box, enlarged in inset with white frame) but preserved at more ventral sites (yellow box, enlarged in inset with yellow frame). The loss extended to more ventral areas than at 2 days after KA. The transition zone from GCD to normal granule cell layer width is marked (arrows). Note that the loss of PV-positive cells extends slightly beyond this zone. (D) In the temporal hippocampus PV-expressing cells and the PV-positive plexus were gone in the dorsal parts but preserved elsewhere. The transition zone is marked with arrows. Scale bars: A, C, D, 200 µm; B, insets in C, 50 µm.

FIGURE 7

Ectopic pattern of NPY staining at 2 days after KA injection. (A–L) Representative sections of immunocytochemical staining for NPY (red) with DAPI counterstaining (blue) at 2 days after KA injection. (A–D) Control mouse, septal (A, B), intermediate (C) and temporal hippocampus (D). NPY-expressing cells were located mainly in the hilus and in stratum oriens and their connections were visible throughout the hippocampus. (B) Cutout of (A) with hilus and granule cell layer enlarged. NPY-positive cells were mostly located in the hilus and their axons were visible throughout the dentate gyrus but granule cells and mossy fibers were not stained. (E–H) Contralateral hippocampus. (E) In the septal hippocampus an upregulation of NPY expression in interneurons was visible in the CA region in strata oriens, pyramidale and radiatum. In addition, NPY was strongly upregulated in the mossy fibers. NPY-positive interneurons in the hilus were preserved but mostly outshined by the strong mossy fiber staining in the whole hilus (see enlarged in F). (G, H) In the intermediate and temporal hippocampus staining of NPY-positive interneurons was slightly upregulated in the dorsal part of the section but comparable to controls in the ventral part. NPY expression in the mossy fibers was strongly increased in the hilus, CA3 and CA2 at all septotemporal levels. (I–L) Ipsilateral hippocampus. (I) In the septal hippocampus NPY-expressing interneurons were lost except for a few remaining, strongly stained cells in CA3 stratum oriens. NPY expression was upregulated in the mossy fibers. In contrast to the contralateral side, NPY imunostaining was strongly enhanced in a large portion of granule cell bodies and dendrites (see enlarged in J, arrow). (K) In the intermediate hippocampus NPY-expressing cells were preserved in the hilus and expression in the CA region was comparable to controls. Upregulation of NPY in the mossy fibers was only visible in the very dorsal parts of the section. (L) In the temporal hippocampus NPY expression was comparable to controls. Scale bars: 200 μm; B, F, J, insets 100 μm.

FIGURE 8

NPY staining pattern changes between 2 and 21 days after KA injection. (A–J) Representative sections of an immunocytochemical staining for NPY (red) with DAPI counterstaining (blue) at 21 days after KA injection. (A–E) Contralateral hippocampus, septal (A), intermediate (B) and temporal hippocampus (C–E). Upregulation of NPY in the CA region was no longer visible but NPY-expressing interneurons were distributed comparable to controls. NPY was upregulated in the mossy fibers at all septotemporal levels (white and yellow box in C, enlarged in D and E, respectively). Some terminals of NPY-positive fibers were seen in the inner molecular layer of the septal hippocampus (arrow in A). (F–J) Ipsilateral hippocampus. (F) In the septal hippocampus most NPY-positive interneurons were lost.