Calbindin D28K expression in relation to granule cell dispersion, mossy fibre sprouting and memory impairment in hippocampal sclerosis: A surgical and post mortem series
Article Outline
Summary
Reorganisation of the dentate gyrus, including granule cell dispersion (GCD) and mossy fibre sprouting, typically accompany hippocampal sclerosis (HS) in temporal lobe epilepsy. Calbindin (CB) expression in granule cells increases during infancy, influences granule cell excitability, vulnerability to excitotoxicity in addition to important physiological functions in memory. Our aim was to study CB patterns in relation to dentate gyrus re-organisation, epilepsy history and memory function.
Forty-five surgical cases and 11 post mortems were examined from patients with drug-resistant epilepsy in addition to three post mortem controls. In the surgical cases, CB expression, and the degree of GCD and mossy fibre sprouting were measured. In post mortem cases, CB expression was assessed in relation to the pattern of HS along the rostral–caudal axis of the hippocampus, and compared to PM controls.
Three patterns were identified. In Group 1 (40%), the most dispersed granule cells were CB-positive and basal cells negative imitating developmental patterns. In Group 2 (47%), normal CB expression was retained and in Group 3 (13%), CB was predominantly lost in granule cells. These patterns correlated with GCD, the presence of mossy fibre sprouting and the pattern of HS. Group 1 was associated with early onset of seizures but not independently predictive of outcome. In post mortem cases, altered CB expression lateralised to the side of HS and persisted despite seizure remission in some patients. No significant correlation between CB expression and memory function was identified.
CB expression patterns in HS may indicate developmental dysmaturation and are associated with the extent of GCD and mossy fibre sprouting in HS. The functional significance of CB loss, in terms of epileptogenesis and effects on memory, remain uncertain.
Keywords: Calbindin, Granule cell dispersion, Memory, Hippocampal sclerosis
Introduction
Hippocampal sclerosis (HS) is a common pathology identified in drug-resistant temporal lobe epilepsy in large surgical series (Blumcke, 2009). Neuronal loss is typically accompanied by changes in the dentate gyrus including granule cell dispersion (GCD) (Houser, 1990), mossy fibre sprouting and synaptic reorganisation (Sutula et al., 1989) leading to disruption of hippocampal circuitry (Blumcke, 2009). These cellular changes, as well as influencing hippocampal excitability, may have additional effects on learning and memory (Helmstaedter and Elger, 2009, Pauli et al., 2006).
In contrast to other subfields, the granule cells of the dentate gyrus have remarkable capacity to survive which may be dependent on several factors including calbindin D28K (CB) content. This calcium binding protein has diverse physiological roles in various mammalian species and human tissues acting as a calcium buffer and censor as well as a neuromodulator essential to memory formation (Blaustein, 1988, Molinari et al., 1996). During post-natal development progressive expression of CB in granule cells appears in an outside-in pattern in the granule cell layer reaching full maturation throughout the mossy fibre pathway by 11 years (Abraham et al., 2009).
Previous studies in human and experimental epilepsy have shown loss of CB expression in granule cells as a relatively consistent finding in HS (Abraham et al., 2011, Arellano et al., 2004, Magloczky et al., 1997). How loss of CB expression relates to patterns of GCD, mossy fibre sprouting as well as clinical measures including pre-operative memory dysfunction, effects on surgical outcome and long term seizure control have not been explored. Through the study of a surgical and post mortem epilepsy series we aim to further evaluate patterns and extent of CB loss in relation to these clinico-pathological measures to investigate its importance in HS.
Materials and methods
Cases were selected from the neuropathology archives at the Institute of Neurology and the National Hospital for Neurology and Neurosurgery, London (UK). The study was approved by the Joint Research Ethics Committee of the Institute of Neurology and the National Hospital for Neurology and Neurosurgery. In all surgical cases, patients underwent therapeutic temporal lobe resection for refractory temporal lobe epilepsy and the pathological tissue was surplus to diagnostic requirement. Clinical and pathology data are presented in Table 1. The post mortem (PM) cases were from patients with long histories of epilepsy and era-appropriate consent for retention of tissue was granted by the next of kin. Clinical and pathology data is presented in Table 2.
Table 1. Clinical and pathology data on the 45 surgical cases.
| Calbindin pattern in granule cell layer | Group 1 (partial loss CB) | Group 2 (normal CB) | Group 3 (loss CB) | Significance | |
|---|---|---|---|---|---|
| Number of cases | 18 | 21 | 6 | ||
| Hippocampal sclerosis patterns | CHS | 17 | 8 | 5 | p <0.001 |
| EFS | 1 | 2 | 0 | ||
| CA1p | 0 | 6 | 0 | ||
| No HS | 0 | 5 | 1 | ||
| Granule cell dispersion pattern (predominant) | bGCD | 7 | 1 | 1 | p<0.01 |
| GCD | 11 | 7 | 4 | ||
| No GCD | 0 | 13 | 1 | ||
| Mossy fibre sprouting (dynorphin) | Grade 0 | 1 | 9 | 0 | p<0.001 |
| Grade 1 | 1 | 8 | 1 | ||
| Grade 2 | 16 | 4 | 5 | ||
| Granule cell dispersion (μm) (SD) | 244 (64.4) | 162 (88.6) | 156 (31.1) | p<0.01 | |
| Age onset (SD) (years) | 7.4 (5.7) | 12.4 (8.1) | 11.6 (5.4) | p<0.05 | |
| Age at surgery (years) | 34.7 (9.5) | 32.5 (9.9) | 37.5 (8.01) | p=0.5 | |
| Duration of epilepsy (prior to surgery) (SD) (years) | 25.4 (8.4) | 22.2 (10.5) | 27 (7.2) | p=0.181 | |
| Pre-operative memory scores Verbal (SD):Visual (SD) | 41 (11):33 (9.6) | 42 (13.3):34 (9.6) | 46 (8.1):26 (13.8) | p=0.8:p=0.3 | |
| % ILAE grade 1 at 2 years post-operative | 83.3% | 42.9% | 50% | p<0.05 | |
Table 2. Clinical and pathology details of 11 epilepsy and 3 control cases examined at post mortem.
 |
 |
Forty-five surgical temporal lobectomy specimens were selected from the archives and included cases with a spectrum of cell dispersion patterns in the granule cell layer (Blumcke et al., 2009). One hippocampal block from the mid body of each case was included which showed the representative pathology and all hippocampal subfields and dentate gyrus. In four patients a second pathology was present in the temporal neocortex (1 cavernoma, 1 old scar and 2 old cortical infarcts). Eleven PM cases were also selected, 3 with bilateral HS, 3 with bilateral asymmetrical HS, 2 with unilateral HS and 3 with No HS. In the PM cases, paired blocks of right and left hippocampus were sampled from the archival formalin-fixed tissue at five different levels along the longitudinal hippocampal axis: subthalamic nucleus/red nucleus, lateral geniculate nucleus, pulvinar (rostral), pulvinar (caudal) and subsplenial (hippocampal tail). In addition, 3 PM control cases with no history of epilepsy and no neuropathology were similarly sampled. Tissue was not available for both hemispheres at all levels, with a mean number of 7 hippocampal samples per post mortem case from a mean of 4 different levels obtained (Table 2).
Immunohistochemistry
7μm-thick paraffin wax-embedded sections were dewaxed and rehydrated through graded alcohols and taken to water. Endogenous peroxidase was blocked with 3% hydrogen peroxide in deionized water for 15min. Sections were microwaved in Vector Antigen Retrieval Solution (Vector: Burlington, CA, USA) then incubated with polyclonal markers against calcium binding protein: calbindin D28K (1:10,000, Swant, Switzerland), mossy fibre sprouting: dynorphin (1:100 AbD Serotec MorphoSys UK Ltd., Endeavour House, Kidlington, Oxford, UK) and monoclonal neuronal marker: NeuN (1:1500, Chemicon International; Temecula, CA USA). Labelling was detected with Dako Envision horse radish peroxidase (DAKO: Glostrup, Denmark). Staining was visualized with Dako DAB+ Chromogen. Negative controls were treated identically except that the primary antibody was omitted. Between all steps sections were washed with PBS and 0.05% Tween 20. Antibodies were diluted in Dako ChemMate Diluent. Dynorphin and NeuN studies were not carried out on the PM series, as the long fixation times precluded adequate staining.
Confocal immunofluorescence
7μm sections from selected cases were dewaxed and rehydrated and washed in water, endogenous peroxidase was quenched with 3% hydrogen peroxide and sections were microwaved in antigen retrieval buffer (Vector). Protein blocking was done with normal horse serum (Vector) followed by incubation of primary antibodies overnight at 4°C. The combinations were polyclonal dynorphin (1:100 AbD Serotec MorphoSys, Park Langford Lane, Kidlington, Oxford, UK) with monoclonal calbindin (CB) (1:30 Sigma–Aldrich, St. Louis, MO, USA). Sections were washed and incubated with secondary anti rabbit ImmPRESS (Vector Laboratories) for polyclonal dynorphin followed by fluorescein tyramide signal amplification (TSA) (PerkinElmer Life and Analytical Sciences, Boston, MA). Sections were then washed and quenched in hydrogen peroxide again in order to prevent any deposited tyramide combining with the second tyramide signal that followed. The sections were incubated with anti mouse ImmPRESS (Vector) for monoclonal CB followed by Cy3 TSA. All sections were protected from light throughout and all sections were washed with PBS in between each step. Sections were mounted on Vectashield with DAPI (Vector Laboratories) and visualized with Zeiss LSM 510 Meta confocal laser microscope.
Qualitative analysis
In both surgical and PM cases the pattern of hippocampal sclerosis was classified as classical HS, atypical HS (end-folium sclerosis or CA1-predominant HS) or No HS (no significant neuronal loss in any subfield as previously described) (Thom et al., 2010). The pathology of the dentate gyrus granule cell layer was qualitatively classified, based on a recent proposal (Blumcke et al., 2009), and according to the predominant pattern as (i) granule cell dispersion with cells evenly spread into the molecular layer (GCD), (ii) GCD with a bilaminar pattern or cell-free gap between granule cells (bGCD) or (iii) normal with minimal dispersion (No GCD) (Fig. 1d–f). The grading of the mossy fibre sprouting (MFS) was carried out on dynorphin stained sections as previously described: Grade 0 (normal pattern), Grade 1 (focal mossy fibre sprouting in inner molecular layer) and Grade 2 (wide band of sprouting throughout molecular layer) (Fig. 1g–i).
Figure 1.
Surgical cases labelled with calbindin (CB) (top row), NeuN (middle row) and dynorphin (bottom row). The three patterns of calbindin expression (which formed the three groups) are shown. (a) Pattern 1: The basal granule cells are predominantly CB negative and the outermost granule cells, dispersed in the molecular layer are CB positive (inset shows very occasional CB positive cell with basal axon: arrowhead). (b) Pattern 2: The majority of granule cells are positive with strong labelling of the apical dendrites in the molecular layer and in CA4 mossy fibres (normal pattern). (c) Pattern 3: A virtual absence of CB expression in the granule cells. (d) NeuN showing a bilaminar pattern of granule cell dispersion with a cell free zone between the basal and most migrated granule cells (bGCD), (e) granule cell dispersion (GCD) and (f) no granule cell dispersion with a compact cell layer. (g) Dynorphin with Grade 2 mossy fibre sprouting with a broad band of axonal labelling in the molecular layer, (h) Grade 1 mossy fibre sprouting with more patchy labelling in the inner molecular layer and (i) Grade 0 no mossy fibre sprouting with positivity for dynorphin only in CA4. Bar
=35μm.
Quantitative methods
The width of the granule cell layer and extent of GCD was measured in each surgical case on NeuN stained sections with Image Pro Plus software (Media Cybernetics, UK) as previously described (Thom et al., 2010). In brief, images were captured using a 20× objective lens; the 10 most distal granule cells in the outer molecular layer were tagged and best fit straight line between these points was drawn. A similar process was carried out with innermost cells at the hilar border of the GCL. The average distance between the two drawn lines was measured. This measurement was repeated in eight regions along the GCL, to include both the inner and outer blades of the cell layer, and the mean distance per case calculated. In 20 cases quantitative analysis of the position of calbindin-positive granule cells was carried out. Between 3 and 6 images were captured per case at 20× magnification and the position of all calbindin-positive and calbindin-negative granule cells in relation to the inner border of the GCL was recorded and the mean distances calculated per case. The position of 1708 CB-positive and 1998 CB-negative cells were measured in total.
Clinical data
Data were retrieved from the clinical record for surgical patients regarding pre-operative epilepsy history as well as seizure outcome following surgery. Preoperatively all patients underwent an IQ assessment and completed the List Learning and Design Learning tasks from the Adult Memory and Information Processing Battery (AMIPB). These tests have been shown to be sensitive to temporal lobe pathology and have been described previously (Baxendale et al., 2006, Baxendale et al., 2008). The List Learning Test is similar to the California Verbal Learning Test but it has been standardised on a British population. In the AMIPB the patient is read a list of 15 common words and asked to recall as many as possible. The total number of words recalled over five trials is recorded. In the Design Learning Test the patient is shown a design comprised of nine connected lines on a 4×4 dot matrix. The design is presented for 10s and the patient is required to reproduce it on a blank grid after each presentation. The total number of correctly drawn lines over five trials is recorded. For the post mortem patients detail of epilepsy history and any cognitive or memory impairments were retrieved from clinical records.
Statistical analysis between groups was carried out with SPSS version 16 using Kruskal Wallis test for non parametric data and linear regression analysis.
Results
Surgical cases
Three different patterns of expression were noted for CB in the granule cells in the 45 surgical cases. In 18 cases CB expression was evident in the most dispersed or outer granule cells with the majority of basal cells CB-negative (Group 1, partial CB expression; Fig. 1a). CB highlighted branching apical dendrites of granule cells, with occasional basal dendrites noted (Fig. 1a). Residual CB positivity was present in the mossy fibre axons in the CA4/3 region in this group (Fig. 3a). In 21 cases, near normal CB expression patterns were retained (Sloviter et al., 1991), with positive labelling of the majority of granule cells, a dense band of staining in the molecular layer corresponding to apical dendrites and in CA4/3 corresponding to the mossy fibre pathway (Group 2, normal CB; Fig. 1b). In 6 cases, the majority of granule cells were CB-negative with residual CB-positive cells often in a more dispersed location rather than basal; immunolabelling of hippocampal and cortical interneurons in this group remained intense (Group 3, CB loss; Fig. 1c). Any loss of CB expression was present in both blades of the dentate gyrus. The qualitative impressions of the distribution of CB-positive cells were confirmed through quantitative analysis. The mean distances of CB-positive cells from the basal layer was 84.91μm (SD 49.5) compared to 45.7μm (SD 16.92) for CB-negative cells overall cases, which was significantly different (p<0.01). When similar data was compared between the three groups, these differences only remained significant for Group 1 (p<0.001) (Fig. 2). The surgical cases were grouped according to these three CB staining patterns for further analysis (Table 1).
Figure 2.
Bar chart showing the mean distances of CB-positive and CB-negative granule cells in relation to the basal cell layer. Over all cases the distances were significantly greater for the positive cells (p
<0.01) but between the groups this finding remained significant for Group 1 only (p<0.001). (Numbers in brackets indicate the number of cases in each group in which measurements were made; significant measurement indicated with *.)
Group 2 (normal CB expression) were more often associated with No HS or atypical HS (62% of cases compared to 5% in Group 1 and 16% in Group 3) (p<0.001). The four patients in our surgical series with additional extrahippocampal neocortical pathologies were all in Group 2. Qualitative assessments also showed differences in the patterns of GCD between the groups with Group 1 showing GCD or bGCD in all cases compared to only 8 of 21 cases in Group 2 (Fig. 1). This was supported by significant difference in the mean measurements of GCD between the three groups with higher values in Group 1 (Table 1; p<0.01). Dynorphin staining for mossy fibre sprouting was also significantly different between groups with 89% and 83% of groups 1 and 3 respectively showing grade 2 sprouting compared to 19% in Group 2 (Table 1; p<0.001). Confocal studies confirmed no co-localisation between CB (mainly in the apical dendrites and cell bodies) and dynorphin (mainly in sprouted mossy fibres) in the molecular layer (Fig. 3e) although co-localisation in mossy fibre axons in the hilus was noted (Fig. 3a, c and f). Furthermore, there were few dynorphin-positive sprouted fibres beyond the outermost dispersed calbindin-positive granule cells in the molecular layer (Fig. 3f). The observations all support the impression that the recurrent collaterals or ‘sprouted’ mossy fibres in the ML arise primarily from more basal and CB-negative granule cells.
Figure 3.
A surgical case with hippocampal sclerosis from Group 1. (a) Calbindin and dynorphin (c) at low magnification show similar distribution of labelling in the molecular layer and the CA4/3 region (arrows). Pattern 1 of calbindin expression (b) and mossy fibre sprouting (d) is confirmed at high magnification. (e) Confocal microscopy confirms a lack of co-localisation in the molecular layer with calbindin expressed in cell bodies and apical dendrites of granule cells and dynorphin in the sprouted mossy fibres. In addition there are less dynorphin positive sprouted fibres beyond the outermost calbindin positive cells (arrow head) (f) but there is co-localisation between dynorphin and calbindin in residual CA4 mossy fibres (arrow). Bar in a and c
=10μm, in b, d and f=35μm and in e=50μm.
A younger age of seizure onset was noted for Group 1 with mean age of 7.4 years compared to 12.4 and 11.6 years for Groups 2 and 3 (p<0.05). There were no significant differences between ages at surgery or duration of habitual seizures between the groups. In addition, a significantly better post-operative outcome was noted for Group 1 with 83% seizure-free compared to 42% and 50% in Groups 2 and 3 respectively (p<0.05). A multivariate statistical analysis however confirmed that CB was not an independent predictor of outcome when pattern of HS was considered (p=0.52). Group 3 with CB loss showed lowest pre-operative visual memory scores, although differences between groups were not significant (Table 1). A significant positive correlation between visual memory score and the width of GCD was observed (p<0.05).
Post mortem cases
In control cases, CB positivity was seen in granule cells, their apical dendrites in the molecular layer and in processes in CA3 and CA4 corresponding to the mossy fibre pathway, throughout the serial blocks taken along the axis of the hippocampal body and tail (Sloviter et al., 1991) (Fig. 4a and b and Table 2). In epilepsy cases, CB loss lateralised to the side of greater HS in asymmetrical cases and was observed in both hippocampi in bilateral HS cases. In one unilateral HS case CB loss was only present on the side of sclerosis (Fig. 4c and d) and in epilepsy patients without HS no CB loss was seen (Table 2). The CB loss associated with HS was noted to extend throughout the hippocampal body and extended into the hippocampal tail in six cases (Table 2 and Fig. 4c and d (insets)). CB loss was present in patients with long seizure histories of 70 years as well as in a patient whose seizures had been in remission for up to 30 years before death. Where clinical data were available, 5 patients with age of onset of epilepsy at three years or less showed pattern 1 of CB expression (basal cells negative, outer dispersed cells positive) (Fig. 4e) but this pattern was not seen in any of 4 patients with later onset of epilepsy (3–17 years). There was a history of cognitive decline in 5 patients; two showed normal CB and three showed loss of CB in the granule cells.
Figure 4.
Post mortem cases. (a) Left and (b) right hippocampus from control cases with normal calbindin expression. Unilateral hippocampal sclerosis (case EP200 in Table 2) showing asymmetry of calbindin expression with a normal pattern observed on the side without sclerosis (c) and absence of staining in the contralteral sclerosed hippocampus (d). This pattern of calbindin was retained in all blocks into the hippocampal tail (insets in c and d). (e) In post mortem cases relatively restricted calbindin expression, present mainly in the outermost granule cells, was also observed. Bars in a–d
=10μm and e=25μm.
Discussion
Alterations to the organisation of the dentate gyrus are a common observation in patients with hippocampal sclerosis and temporal lobe epilepsy. We have demonstrated distinct patterns of CB expression in our series including a group with partial loss of expression in granule cells associated with cell dispersion. The most dispersed neurons more often retained CB expression and we confirmed this through quantitative analysis. Loss of CB in granule cells has been previously reported in a study of 10 surgical patients with TLE and HS (Magloczky et al., 1997); CB loss was noted to be complete in 4 and partial in 6 cases. It was also noted in this study that greater CB loss was often seen in cases with a compact, narrow granule cell layer without dispersion. In a further study of 14 cases with reorganisation in the epileptic sclerotic hippocampus, CB loss was reported in five and to preferentially affect the granule cells in the hilar aspect (Arellano et al., 2004). A more recent study has also showed greater proportion of granule cells to be CB positive in the outer part of the molecular layer in HS (Abraham et al., 2011).
These patterns of CB expression in HS draw comparison to studies of ontogeny of the granule cells which shows an outside-in pattern of maturation. CB expression is first observed at 22–23 weeks gestation in the most superficial granule cells in the dorsal blade and by birth is present through granule cells of both blades bar the deepest cells at the hilar border. At 2–3 years CB is present in all neurons through the layer (Abraham et al., 2009) and throughout the mossy fibre axons by 8–11 years. As such, neurochemical maturation of the dentate gyrus is ongoing in the first years of life. Although our study group is small we noted a significantly earlier age of onset of epilepsy of 7.4 years in the surgical group with HS and expression of CB restricted to the most dispersed granule cells. This was also the finding in a recent study (Abraham et al., 2011). Additionally in the PM cases where clinical data were available, 5 patients with age of onset of epilepsy at three years or less showed similar partial CB expression whereas this was not observed in 4 patients with later onset of epilepsy. These findings support the notion of an age-dependent effect on CB expression in granule cells in epilepsy. Possible explanations for the lack of CB in more basal cells in temporal lobe epilepsy include an arrested developmental maturation of basal cells, their greater susceptibility to reduced CB expression or the CB-negative basal cells represent an excess of regenerated neurons.
The cause of GCD in HS has been long debated in humans. In experimental models, ongoing neurogenesis in the dentate gyrus, stimulated by seizures, with aberrant migration of newly generated cells has been demonstrated (Parent et al., 1997, Scharfman et al., 2007). We, and others, have previously noted increased numbers of cells with proliferative potential in surgical patients with GCD in support of enhanced neurogenesis (Crespel et al., 2005, Thom et al., 2005). However, current theories favour that GCD is a result of migration of the most mature granule cells along a radial glial scaffold influenced by local reelin deficiency (Fahrner et al., 2007, Haas and Frotscher, 2010) which is supported by experimental data (Duveau et al., 2010). If CB expression is interpreted as a marker of granule cell maturity, our findings in the present study would seem to support this concept with positive cells having migrated further into the molecular layer, particularly in Group 1. It could also suggest that the propensity of a subset of granule cells to migration is influenced by their intrinsic neurochemical properties.
Our study also highlights a relationship between loss of CB expression in granule cells and the presence and extent of mossy fibre sprouting. During development, CB expression in the mossy fibre pathway only reaches the mature adult pattern at 8–11 years, where strong immunoreactivity is seen in both the CA3 as well as the hilus of the hippocampus in mossy fibre terminals (Abraham et al., 2009). Loss of CB expression in granule cells in our study was coupled by progressive loss of labelling in axons in CA4 and CA3. Axonal sprouting from both excitatory granule cells as well as interneurons in temporal lobe epilepsy is an early and well recognised phenomena, likely to be as functionally significant as hippocampal neuronal loss (Magloczky, 2010). Sprouted mossy fibre axons in the molecular layer in our series were only visualized with dynorphin and not with calbindin immunohistochemistry, in keeping with previous observations (Magloczky et al., 1997, Magloczky et al., 2000) and we confirmed no overlap between the expression of these proteins using confocal microscopy in this region. Although a previous study using EM demonstrated both CB-positive and negative mossy fibre terminals in the molecular layer (Magloczky et al., 1997) our study supports the notion that mossy fibre sprouting arises primarily from CB-negative granule cells. Furthermore the most dispersed CB-positive granule cells appeared beyond the main zone of dynorphin positivity in the molecular layer suggesting a lesser contribution of these mature neurons to the process of axonal sprouting.
It has been proposed that loss of CB could cause changes in the intrinsic excitability of granule cells mediated through loss of calcium binding or buffering capacity. This in turn could modulate any overall hyperexcitability of the dentate gyrus (Bouilleret et al., 2000). A study of the firing pattern of granule cells in relation to CB content in human epilepsy tissues, however, did not find any clear relationship (Selke et al., 2006). An investigation of functional membrane properties of granule cells has demonstrated that more dispersed neurons in the dentate gyrus are less excitable with reduced input resistance and enhanced leak conductance which was interpreted as a neuroprotective effect (Stegen et al., 2009). Although they did not correlate findings with CB cell expression, through comparison with our current data this could suggest the more dispersed, CB-positive cells contribute less and are more resistant to seizures which was also the interpretation in a previous histological study (Abraham et al., 2011). There is also experimental data showing enhanced vulnerability of CB negative, and immature granule cells to apoptosis following induced seizures (Lopez-Meraz et al., 2010). In our series, CB patterns as a potential measure of hippocampal excitability was not independently predictive of outcome following surgery, as also shown for GCD in previous studies (Blumcke et al., 2009, Thom et al., 2010). However differences in outcome between Groups 1 and 3, which were similar in aspects of the patterns of HS and MFS but differing in the CB expression, remain at present unexplained and require further investigation.
In addition, as well as influencing excitability and vulnerability to excitotoxicity, CB levels can affect normal hippocampal function. Experimental CB over-expression in the dentate gyrus has been shown to reduce long-term potentiation (LTP) in the hippocampus and impair spatial memory (Dumas et al., 2004). Similarly, CB deficient mice do not maintain LTP and show memory deficits (Jouvenceau et al., 1999, Jouvenceau et al., 2002, Molinari et al., 1996). Reduction of CB in granule cells has been reported in severely demented patients with Alzheimer's disease (Palop et al., 2003) and a functional role of CB in cognitive deficit in models of Alzheimer's disease proposed (Odero et al., 2010) as well as age-related hippocampal decline (Moreno et al., 2011). Cognitive impairment, particularly memory disruption, is a major complicating feature in temporal lobe epilepsy (Bell et al., 2011, Helmstaedter and Elger, 2009). Granule cell loss in the internal limb of the dentate gyrus has been previously correlated with poor memory performance (Pauli et al., 2006) and more recently loss of hippocampal regenerative capacity in the dentate gyrus has been correlated with memory dysfunction (Coras et al., 2010). In the current series, although patients with CB loss showed lower pre-operative memory scores, overall there was no significant correlation between CB pattern in the surgically resected specimens and memory deficit. This was also the finding in the PM cases although acknowledging the study numbers are small.
Several physiological factors regulate CB expression in granule cells. Loss of CB has been shown in experimental animals with increasing age (Choi et al., 2009, Molinari et al., 1996, Palop et al., 2003) as well as stress (Nowak et al., 2010). Our post mortem human studies confirm that normal CB expression can be maintained into the 9th decade, including in the mossy fibre pathway of the most caudal samples from hippocampal tail, in contradiction to previous reports (Lim et al., 1997). Drugs may modulate CB expression in granule cells with reduction shown with serotonergic antidepressant drugs (Kobayashi et al., 2010). The loss of CB associated with experimental seizures has been shown to be a permanent alteration for up to 2 years (Carter et al., 2008) with decreased gene expression noted. CB restoration has been demonstrated following cell grafting into CA3 and restitution of the disrupted intrinsic circuitry (Shetty and Hattiangady, 2007) suggesting potential reversibility of the process. Through our post mortem studies we were able to confirm that CB loss mainly lateralised to the side of HS and therefore unlikely to be a generalized effect of seizures or anti-epileptic drugs. CB loss was also present in patients with long seizure histories as well as seizure remission periods of up to 30 years prior to death. In contrast to a previous paper reporting that CB loss correlated with duration of epilepsy (Abraham et al., 2011) we did not demonstrate this in our series. Our data supports that CB alteration may be long-lasting but, in parallel to our recent observations of mossy fibre sprouting in post mortem cases (Thom et al., 2009), is not always coupled with the continuation of clinical seizures.
In summary we have identified specific patterns of CB expression in the dentate gyrus which conform to specific clinico-pathological types of HS with granule cell dispersion, may indicate developmental dysmaturation but its functional significance is less clear. Routine CB immunohistochemistry may aid in the evaluation of HS, particularly in limited surgical samples.
Acknowledgements
We are grateful to Jane de Tissi for her help. This work was supported by a research grant from the Medical Research Council (Grant 0600934). This work was undertaken at UCLH/UCL who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme.
References
- . Degree and pattern of calbindin immunoreactivity in granule cells of the dentate gyrus differ in mesial temporal sclerosis, cortical malformation- and tumor-related epilepsies. Brain Res. 2011;
- . Ontogeny of calbindin immunoreactivity in the human hippocampal formation with a special emphasis on granule cells of the dentate gyrus. Int. J. Dev. Neurosci. 2009;27(2):115–127
- . Histopathology and reorganization of chandelier cells in the human epileptic sclerotic hippocampus. Brain. 2004;127(Pt 1):45–64
- . Predicting memory decline following epilepsy surgery: a multivariate approach. Epilepsia. 2006;47(11):1887–1894
- . Improvements in memory function following anterior temporal lobe resection for epilepsy. Neurology. 2008;71(17):1319–1325
- . The neurobiology of cognitive disorders in temporal lobe epilepsy. Nat. Rev. Neurol. 2011;7(3):154–164
- . Calcium transport and buffering in neurons. Trends Neurosci. 1988;11(10):438–443
- . Neuropathology of focal epilepsies: a critical review. Epilepsy Behav. 2009;15(1):34–39
- Towards a clinico-pathological classification of granule cell dispersion in human mesial temporal lobe epilepsies. Acta Neuropathol. 2009;117(5):535–544
- . Neurodegenerative and morphogenic changes in a mouse model of temporal lobe epilepsy do not depend on the expression of the calcium-binding proteins parvalbumin, calbindin, or calretinin. Neuroscience. 2000;97(1):47–58
- . Long-term decrease in calbindin-D28K expression in the hippocampus of epileptic rats following pilocarpine-induced status epilepticus. Epilepsy Res. 2008;79(2–3):213–223
- . Reduction of calbindin D-28k-immunoreactive neurons in the dog dentate gyrus. J. Vet. Med. Sci. (The Japanese Society of Veterinary Science). 2009;71(8):1125–1128
- Low proliferation and differentiation capacities of adult hippocampal stem cells correlate with memory dysfunction in humans. Brain: A Journal of Neurology. 2010;133(11):3359–3372
- . Increased number of neural progenitors in human temporal lobe epilepsy. Neurobiol. Dis. 2005;19(3):436–450
- . Overexpression of calbindin D(28k) in dentate gyrus granule cells alters mossy fiber presynaptic function and impairs hippocampal-dependent memory. Hippocampus. 2004;14(6):701–709
- . Impaired reelin processing and secretion by Cajal-Retzius cells contributes to granule cell dispersion in a mouse model of temporal lobe epilepsy. Hippocampus. 2010;
- . Granule cell dispersion is not accompanied by enhanced neurogenesis in temporal lobe epilepsy patients. Exp. Neurol. 2007;203(2):320–332
- . Reelin deficiency causes granule cell dispersion in epilepsy. Exp. Brain Res. (Experimentelle Hirnforschung. Experimentation Cerebrale), 2010;200(2):141–149
- . Chronic temporal lobe epilepsy: a neurodevelopmental or progressively dementing disease?. Brain: A Journal of Neurology. 2009;132(Pt 10):2822–2830
- . Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy. Brain Res. 1990;535(2):195–204
- . Glutamatergic synaptic responses and long-term potentiation are impaired in the CA1 hippocampal area of calbindin D(28k)-deficient mice. Synapse. 1999;33(3):172–180
- . Decrease in calbindin content significantly alters LTP but not NMDA receptor and calcium channel properties. Neuropharmacology. 2002;42(4):444–458
- . Reversal of hippocampal neuronal maturation by serotonergic antidepressants. Proc. Natl. Acad. Sci. U.S.A. 2010;107(18):8434–8439
- . Connections of the hippocampal formation in humans: I. The mossy fiber pathway. J. Comp. Neurol. 1997;385(3):325–351
- . Vulnerability of postnatal hippocampal neurons to seizures varies regionally with their maturational stage. Neurobiol. Dis. 2010;37(2):394–402
- . Sprouting in human temporal lobe epilepsy: excitatory pathways and axons of interneurons. Epilepsy Res. 2010;89(1):52–59
- . Loss of calbindin-D28K immunoreactivity from dentate granule cells in human temporal lobe epilepsy. Neuroscience. 1997;76(2):377–385
- . Changes in the distribution and connectivity of interneurons in the epileptic human dentate gyrus. Neuroscience. 2000;96(1):7–25
- Deficits in memory and hippocampal long-term potentiation in mice with reduced calbindin D28K expression. Proc. Natl. Acad. Sci. U.S.A. 1996;93(15):8028–8033
- . The absence of the calcium-buffering protein calbindin is associated with faster age-related decline in hippocampal metabolism. Hippocampus. 2011;
- . Alterations in hippocampal calcium-binding neurons induced by stress models of depression: a preliminary assessment. Pharmacol. Rep. 2010;62(6):1204–1210
- Evidence for the involvement of calbindin D28K in the presenilin 1 model of Alzheimer's disease. Neuroscience. 2010;169(1):532–543
- . Neuronal depletion of calcium-dependent proteins in the dentate gyrus is tightly linked to Alzheimer's disease-related cognitive deficits. Proc. Natl. Acad. Sci. U.S.A. 2003;100(16):9572–9577
- . Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. J. Neurosci. (The Official Journal of the Society for Neuroscience). 1997;17(10):3727–3738
- . Deficient memory acquisition in temporal lobe epilepsy is predicted by hippocampal granule cell loss. Neurology. 2006;67(8):1383–1389
- . Ectopic granule cells of the rat dentate gyrus. Dev. Neurosci. 2007;29(1–2):14–27
- . Firing pattern and calbindin-D28K content of human epileptic granule cells. Brain Res. 2006;1120(1):191–201
- . Restoration of calbindin after fetal hippocampal CA3 cell grafting into the injured hippocampus in a rat model of temporal lobe epilepsy. Hippocampus. 2007;17(10):943–956
- . Calcium-binding protein (calbindin-D28K) and parvalbumin immunocytochemistry in the normal and epileptic human hippocampus. J. Comp. Neurol. 1991;308(3):381–396
- . Increased leak conductance in dentate gyrus granule cells of temporal lobe epilepsy patients with Ammon's horn sclerosis. Epilepsia. 2009;50(4):646–653
- . Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol. 1989;26(3):321–330
- . Reliability of patterns of hippocampal sclerosis as predictors of postsurgical outcome. Epilepsia. 2010;51(9):1801–1808
- . Bilateral reorganization of the dentate gyrus in hippocampal sclerosis: a postmortem study. Neurology. 2009;73(13):1033–1040
- . Cell proliferation and granule cell dispersion in human hippocampal sclerosis. J. Neuropathol. Exp. Neurol. 2005;64(3):194–201
PII: S0920-1211(11)00235-X
doi:10.1016/j.eplepsyres.2011.08.011
© 2011 Elsevier B.V. All rights reserved.
