- Top of page
Summary: Purpose: Heat shock protein-27 (HSP-27) belongs to the group of small heat shock proteins that become induced in response to various pathologic conditions. HSP-27 has been shown to protect cells and subcellular structures, particularly mitochondria, and serves as a carrier for estradiol. It is a reliable marker for tissues affected by oxidative stress. Oxidative stress and related cellular defence mechanisms are currently thought to play a major role during experimentally induced epileptic neuropathology. We addressed the question whether HSP-27 becomes induced in the neocortex resected from patients with pharmacoresistant epilepsy.
Methods: Human epileptic temporal neocortex was obtained during neurosurgery, and control tissue was obtained at autopsy from subjects without known neurologic diseases. The tissues were either frozen for Western blot analysis or fixed in Zamboni's fixative for the topographic detection of HSP-27 at the cellular level by means of immunohistochemistry.
Results: HSP-27 was highly expressed in all epilepsy specimens and in the cortex of a patient who died in the final stage of multiple sclerosis (positive control), whereas only low amounts of HSP-27 were detectable in control brains. In epilepsy patients, HSP-27 was present in astrocytes and in the walls of blood vessels. The intracortical distribution patterns varied strongly among the epilepsy specimens.
Conclusions: These results demonstrate that HSP-27 becomes induced in response to epileptic pathology. Although the functional aspects of HSP-27 induction during human epilepsy have yet to be elucidated, it can be concluded that HSP-27 is a marker for cortical regions in which a stress response has been caused by seizures.
Heat shock proteins (HSPs) belong to a highly conserved family of proteins, some of which are inducible by various noxious stimuli and are part of the cellular defense system (1,2). HSPs act as chaperones transporting proteins to cell organelles or participate in protein-folding or both (3–5). In addition, HSP-70 and HSP-27 serve as binding proteins for vitamin D and estradiol, respectively (6–8). Both HSP-70 and HSP-27 are induced in the CNS during heat shock, ischemia, hypoxia, and seizures, where they contribute to neuroprotection and the phenomenon of preconditioning (2,5). For HSP-27 and the mouse homolog HSP-25, it is known that they inhibit apoptotic neuronal death (9,10) by acting at the permeability transition pore of mitochondria (11). In a more general view, HSPs also are considered to represent a reliable marker for tissues affected by oxidative stress (5).
HSP-70, which becomes rapidly induced by epileptic activity in the kainate model, as well as in human patients (12–14), is, however, not expressed after application of the γ-aminobutyric acid (GABA)A-receptor antagonist pentylenetetrazol (PTZ), N-methyl-d-aspartate (NMDA), or lindane, despite convulsive activity (12). In contrast, the PTZ model revealed a strong induction of HSP-72 (15,16). Physiologic cerebral HSP-27 expression in humans occurs during gestational week 17 (17). HSP-27 occurrence during human neuropathologic conditions has been studied only in resected glioblastomas (18–20) and in multiple sclerosis (21). In addition, Erdamar et al. (22) reported a nonselective presence of HSP-27 in patients with Ammon's horn sclerosis.
Recently, free radicals and oxidative stress have been implicated to play a major role in epilepsy, where they are involved mainly in cell loss (23–30). HSP-27 induction has been proposed as marker for cells affected by oxidative stress in vitro as well as in animal models. To elucidate whether an oxidative stress response takes place in temporal cortices of patients with pharmacoresistant epilepsy, we studied the distribution of HSP-27 in resected tissue by means of immunohistochemistry and compared these observations with those found in temporal lobes of human control specimens obtained at various stages of postmortem delay as well as with postmortem tissue obtained from a patient with multiple sclerosis.
- Top of page
Parts of human temporal neocortex were resected during neurosurgery at the Epilepsy Center Bethel, Bielfeld, Germany (Table 1). The patients had pharmacoresistent epilepsy. For details, refer to Speckmann et al. (31). The resected tissue was immediately placed in ice-cold artificial cerebrospinal fluid (aCSF; 32,33). After documentation of the topography of the resected tissue block, tissue designated for in vitro functional analysis by using either conventional neurophysiologic techniques or fast optical recording system after application of voltage-sensitive dyes was cut in slices (500 μm thick) and transferred into a transportation chamber (34). Approximately 70% of the slices showed one or more spontaneously active epileptic foci independent of each other (Table 1). Tissue designated for immunohistochemistry was either directly frozen on dry-ice for Western blot analysis or fixed in Zamboni fixative containing 0.05% glutaraldehyde for 5 to 7 h. Tissue was then transferred to freshly prepared Zamboni, fixed overnight at 4°C, cryoprotected in 25% sucrose in phosphate-buffered saline (PBS, pH 7.4) and frozen at –40°C in isopentane. Frozen sections (50 μm) were cut with a microtome Frigomobil (Leica, Bensheim, Germany). Sections were collected in PBS and either used immediately or stored at 4°C in PBS containing 0.001% sodium azide until further processing. Control tissue from human temporal cortices was obtained at autopsy by taking a 4- to 5-mm-thick tissue slice from the temporal lobe of freshly dissected brains of subjects with no known neurologic disorders (Table 2). This tissue slice was fixed, cryoprotected, frozen, and sectioned as described earlier. In addition, seven frozen specimens of human temporal cortex were taken from brains that had been collected before the onset of this current study and that had been stored frozen (–70°C) at the C. & O. Vogt Institute of Brain Research (Table 2). These tissues had postmortem delays ranging from 4 to 18.5 h and also were used for Western blot analysis. The brain (control case 9, Table 2) from the patient with multiple sclerosis was obtained at autopsy, and parts of the temporal lobes were cut into 4-mm-thick slices, fixed, and prepared for immunohistochemistry as described earlier. Additionally, some alternate tissue slices were frozen for analysis by Western blot (see later). It is generally agreed that studies focusing on living human brain tissue cannot guarantee a perfect age matching between the different groups of investigations. The ages of our patients are listed in Tables 1 and 2.
Table 1. Data of patients concerning seizure history, antiepileptic drugs, histopathologic findings and sharp waves occurring spontaneously under control conditions in corresponding neocortical slices
| Case|| Gender|| Age (yr)||Seizure frequency (per mo)||Seizures for n yr and last before surgery in (days)|| AED|| Pathology||Tissue type||Slice activity|
|1||m||21||10||10; (>3d)||1||ac spg||TL H l||nsp|
|2||f||22||40||17; (>3d)||1, 6, 10, 12, 17||spg||TL H r||nsp|
|3||f||20||16||7; (1d)a||1, 3, 6, 17||gc spg||TL H l||sp|
|4||f||54|| 3||32; (>3d)||1, 3, 5, 6,13, 17||g a,||TL r||sp|
|5||m||36||20||31; (1d)a ||1, 3, 5, 6||ac focal dysplasia||TL r||nsp|
|6||m||36|| 2||9; (1d)a||14, 6, 8, 14, 17,||g ac gc||TL H r||sp|
|7||m||42||14||18; (1d)a ||1, 4, 5, 6, 8, 9, 10, 11, 14, 15, 17||gg a||TL H r||sp|
|8||m||41|| 2||8||1,5, 6, 17||ac||TL r||sp|
|9||f||41|| 5||34; (>4d)||1, 3, 6, 11|| ||TL H l||sp|
|10||m||35||20||34; (>5d)||1, 3, 5, 17||g (severe)||TL r|| |
|11||f||20|| ||16; (>3d)||1, 6, 7, 10, 15, 17||g ld||TL H l||nsp|
|12||m|| 0|| ||0; (3d)b||3, 6, 8, 9, 12, 15,16, 17||ld||TL H l|| |
|13||m||42|| 4||12; (>3d)||1, 3||ac||TL H l||sp|
|14||f||18|| ||6; (3d)c||8||g gg a||TL r||sp|
|15||m||24|| 1||22||1, 9||Focal dysplasia||TL H l||sp|
|16||m||43|| 1||16||1, 3, 6, 8, 14, 15, 17||g ac||TL r||sp|
|17||f||34|| 3||33; (>6d)||1, 17||Focal dysplasia gg a||TL r||nsp|
|18||m|| 4||60||4; (1d)a||8, 9, 17||Focal dysplasia gg a||TL H l||sp|
|19||f||48|| 3||25; (>3d)||1, 3, 4, 5, 6, 8, 9,14, 17||gg a||TL r||sp|
|20||m||28|| 2|| 7; (>4d)||3||g (severe)||TL l||sp|
|21||m||20|| 7||18; (>4d)||1, 9, 11, 14,15, 17||Focal dysplasia||TL r||sp|
|22||m||52||10||52; (1d)a ||1, 3, 6, 17||g (severe)||TL l||nsp|
|23||f||66||70||2; (1d)a|| ||gg a||TL r||nsp|
|24||m||16|| 2|| 7; (>1d)||1, 2, 6, 8, 17||Focal dysplasia||TL H r||sp|
|25||f||20|| ||19; (>3d)||1, 3, 6, 10,15, 17|| ||TL H r|| |
|26||f||39|| 4||30; (1d)a ||1, 6, 9, 17||ac||TL H l||sp|
|27||m||38|| 4||32; (3d)b ||1, 8,11, 15,17||g (fiber)||TL H l||sp|
|28||m||47|| 4|| ||1, 6, 8, 14,17|| ||TL r||sp|
Table 2. Control tissue
|Case (cc)||Gender||Age (yr)||Topography of studied tissue||Diagnosed disease||Postmortem delay (h)|
|1||m||75||Left occiput. temp. lobe gts, gtm||Respiratory insufficiency, pulmonary carcinoma||16.5 |
|2||f||80||Right frontal temp. lobe gtm, gti||Cardiac arrest, diabetes||4.0|
|3||f||62||Right frontal temp. lobe gtm, gti||Cardiac myopathy||5.0|
|4||m||56||Right occiput. temp. lobe gtm, gti||Cardiac arrest||15.0 |
|5||m||79||Left rostral. temp. lobe gtm, gti||Pancreas carcinoma||9.0|
|6||m||77||Left rostral temp. lobe gts, gtm||Arteriosclerosis, hypertension||10.0 |
|7||m||62||Left rostral temp. lobe gts, gtm||Rectal carcinoma||18.5 |
|8||m||72||Left frontal, medial temp. lobe gtm, gti||Cardiac arrest||8.0|
|9||f||53||Temp. lobe gts-gti||Multiple sclerosis||4.0|
|10||f||74||Right left temp. lobe gts-gti, hippoc||Respiratory insufficiency||9.5|
|11||m||72||Right left frontal, medial temp. lobe, gts-gti, hippoc||Hepatic failure, rectal carcinoma||7.5|
All subjects had given written consent or had been included in the body-donor program of the C. & O. Vogt Institute of Brain Research or both. In addition, extra pieces of tissue from a few epilepsy patients from whom enough resected tissue was obtained were placed in aCSF at 32°C for ∼3 h, followed by 3 to 9 h at room temperature, and were then frozen for Western blot analysis to check for epitope stability.
All procedures concerning epilepsy patients have been approved by the ethical commission of the Ärtzekammer Westfalen-Lippe and the Medical Faculty of the University of Münster, Germany. For additional information, refer to Speckmann et al. (31).
Western blot analysis
Tissue samples were lysed at 4°C with 10 mM Tris-HCl buffer (pH 7.4) containing 1% Triton X-100, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM EGTA, 20 mM NaF, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 0.5% Nonidet P-40 (NP-40). After homogenization, the lysates were centrifuged at 20.000 g at 4°C. For sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis, the supernatant was added to an identical volume of ×2 gel loading buffer containing 200 mM dithiothreitol (DTT) (pH 6.8) and heated to 95°C for 5 min. The probes were subject to gel electrophoresis (10% gels) using 30 μg protein per lane. Thereafter, gels were equilibrated with transfer buffer (39 mM glycine, 48 mM Tris-HCl, 0.03% SDS, and 20% methanol. Proteins were transferred to nitrocellulose membranes using a semidry transfer apparatus (Pharmacia, Freiburg, Germany). Blots were blocked overnight in 5% bovine serum albumin (BSA) solubilized in 20 mM Tris-HCl, pH 7.5, containing 150 mM NaCl and 0.1% Tween 20, and incubated for 2 h with antiserum against monoclonal antiHSP-27 (SPA-800; StressGen) diluted 1:5,000 at 4°C. After several rinses and incubation with horseradish peroxidase–coupled antimouse immunoglobulin G (IgG; Sigma, Deisenhofen) diluted 1:10,000 at 4°C for 2 h, blots were washed and developed by using enhanced chemiluminescent (ECL) detection (Amersham, Braunschweig, Germany).
In addition, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was monitored for control of protein loading of the Western blot. Blots were probed with an antiserum against GAPDH (Biosource, Cologne, Germany) diluted 1:5,000 for 2 h at 4°C. After washing and incubation with horseradish peroxidase–coupled anti-mouse IgG antibody diluted 1:10,000 at 4°C for 2 h, the blots were washed and developed as described earlier.
Sections were stained free-floating under constant gentle shaking starting with three rinses in PBS (pH 7.4) followed by inhibition of endogenous peroxidase with 3% H2O2 in PBS for 20 min, followed by four rinses in PBS. Sections were then incubated for 1 h in PBS containing 0.1% Triton X-100 and 10% normal serum from the species in which the second antibody was raised [normal goat serum, (NGS) or normal horse serum (NHS), Dianova, Hamburg, Germany]. Afterward, sections were incubated in antiserum against polyclonal HSP-25 (SPA-801, Stressgen Canada) at a final dilution of 1:500, and additional sections were incubated with monoclonal antibody to HSP-27 (SPA-800, Stressgen) at a dilution of 1:200 at 4°C for 48 h. After four rinses of 15 min each in PBS, sections were incubated with biotinylated second antibody goat antirabbit or goat antimouse (Dianova) at a final dilution of 1:200 for 24 h at 4°C. After additional four rinses, sections were incubated at room temperature with AB-complex (Vectastain) for 90 min, and antibody binding was visualized with 3,3'-diaminobenzidine tetrahydrochloride according to standard protocols (35). Control sections were stained either without primary antibody or with another monoclonal antibody such as SMI-311 (Sternberger, Monoclonals) with the same protocol.
For colocalization studies with double immunofluorescence, we used the following combinations of primary antibodies: monoclonal HSP-27 (SPA-800, 1:100) with polyclonal glial fibrillary acid protein (GFAP; Sigma, 1:400) and polyclonal MAP-2 (gift from Dr. V. Jancsik, Budapest) or with the endothelial cell marker, von Willebrand factor (Chemicon, AB7356; 1:600) with monoclonal HSP-27. After incubation in blocking serum, sections were incubated for 48 h at 4°C under gentle shaking in the primary antibodies diluted in PBS containing 0.1% Triton X-100. After four rinses in PBS, they were incubated with the secondary antibodies Cy-3 coupled goat anti-rabbit (1:150) and fluorescein isothiocyanate (FITC)-coupled goat antimouse (1:150, Dianova) for 24 h at 4°C. Sections were rinsed four times and mounted with Fluoromount G (Southern Biotechnology Assoc., Burmingham, AL, U.S.A.). Fluorescence-labeled sections were evaluated by using a laser-scanning microscope (Leica, Bensheim, Germany).
- Top of page
In all cases, in the living tissue adjacent to that processed for histochemical analyses, functional in vitro investigations have been carried out with conventional electrophysiologic techniques or voltage-sensitive dyes together with a fast optical recording system. In nearly 70% of the slices, one or more spontaneously epileptic foci, characterized preferentially by sharp waves, were present (Table 1; 36,37). This demonstrates that the tissue under investigation can be classified as being primary epileptic. Such epileptic activity failed to occur in living control tissue obtained from tumor surgery when patients did not show signs of epileptogenicity.
In human temporal cortex of epilepsy patients, immunoreactivity for HSP-27 was consistently present in all cases and was confined mainly to the walls of blood vessels and to glial cells (Figs. 1, 2, and 5). Colocalization studies using HSP-27 in combination with the endothelial cell marker von Willebrand factor revealed a much stronger HSP-27 immunostaining in the astrocytic processes that contacted the walls of the blood vessels than in the endothelial cells (Fig. 5C and D). HSP-27 was still clearly detectable within endothelial cells (Fig. 5D and D” insets). In sections from control specimens obtained at autopsy with short postmortem delay, almost no glial immunoreactivity for HSP-27 was present, whereas SMI-311–positive neurofilaments were clearly labeled in adjacent sections (Fig. 1A and B).
Figure 1. Sections through the temporal neocortex (medial temporal gyrus) of subjects without neuropathologic alterations (control case, cc-11; A-A″, B), with multiple sclerosis (cc-9; C-C′) and from epilepsy patients (case 12, 14, 25, 26, 28; D–H). In the temporal neocortex of control brains, immunoreactivity for heat-shock protein (HSP)-27 is absent (A). At a higher resolution, only background-like staining is present, and some dark spots are confined to clotted blood vessels in cortex (A″) and white matter (A″). Layer-specific distribution of nonphosphorylated neurofilaments (SMI-311), however, is detectable in these tissue samples (B). Within and around a multiple sclerosis lesion, several cells express HSP-27 (C–C′). The whole network of blood vessels is HSP-27 immunopositive in some epilepsy specimens (D). Focal clusters of HSP-27–immunoreactive glial cells are associated with individual blood vessels (F). In some regions of c-26, only small foci of HSP-27–positive glial cells are present (E–E′). HSP-27–immunopositive glial cells are visible in cortical layer I and in the white matter (c-25), whereas the walls of blood vessels are strongly labeled (cortex G, white matter G′) in deeper cortical layers. In some cases (e.g., c-28), a patchy distribution pattern of HSP-27–immunoreactive cells is present (H, overview H′). For further information about the examined brains, see Tables 1 and 2. Cortical layers, I, II/III, IV, V, VI; pial surface, P. Bars: 1 mm (A and B); 100 μm, A′–A″, C–H′).
Download figure to PowerPoint
Figure 2. Distribution of heat-shock protein (HSP)-27–immunoreactive glial cells and blood vessels of the vertex (A–A′) and fundus (B–B′) of the medial temporal gyrus (c-17; Table 1). Note the inhomogeneous and in part focus-like distribution of HSP-27–positive glial cells in grey and white matter (B-B′ enlargement). A′: Enlargement of the transition zone between HSP-27–positive and normal cerebral tissue, containing some hypertrophic astrocytes (arrow). Blood vessels, bv cortical layers, LI–VI, white matter, WM. Bars: 1 mm.
Download figure to PowerPoint
Figure 5. Expression of heat-shock protein (HSP)-27 and of the neuronal marker MAP-2 in gray matter (A) and white matter (B) and HSP-27 in combination with the endothelial cell marker von Willebrand factor (C–D). HSP-27 is present in glial cells and blood vessels (A′, B′), whereas MAP-2–positive neurons and their processes (A″, B″) do not colocalize with HSP-27. HSP-27 is present in the blood vessel–associated astrocytic processes (C–C′) that are in tight contact with the endothelial lining, which stains positive for von Willebrand factor (C″). D: An enlarged wall of a blood vessel. Arrowheads, A part of the endothelial lining where some astrocytic processes are in very close contact to endothelial cells (D–D″). The von Willebrand factor–positive endothelial cells (D″) show much weaker immunostaining for HSP-27 when compared with the contacting astrocytic processes (D′). The almost immunonegative narrow space between the contacting astrocytic endfeet and the endothelial lining arrow is clearly depicted in the insets (D–D″). Bars: 40 μm (A–C); 10 μm (D).
Download figure to PowerPoint
For Western blot analysis, unfixed tissue from the epilepsy specimen and additional small frozen tissue pieces from the temporal neocortex of human control brains were used. The postmortem delay of the control samples varied between 4 and 18.5 h (Table 2). Similar to the results obtained with immunohistochemistry, all samples from epilepsy patients contained HSP-27 immunoreactivity, whereas no or a very faintly stained band was present in the control tissue (Fig. 3).
Figure 3. Western blot analysis showing the presence of the heat-shock protein (HSP)-27–immunoreactive band in protein extracts of the epilepsy patients and control patients listed in Tables 1 and 2.
Download figure to PowerPoint
Because it is not possible for ethical reasons to obtain nonpathologic human control tissue that has been fixed or frozen immediately after death (short postmortem delay) or resection, we performed a series of experiments to exclude that the lack or minimal presence of HSP-27 in controls was simply a result of degradation during postmortem delay:
Cortical tissue resected from epilepsy patients was stored for ∼3 h in aCSF at 32°C without oxygen supply. Tissue probes were subsequently stored at room temperature for an additional 3, 5, and 9 h. As shown by Western blot analysis, HSP-27 immunoreactivity was detectable after a total postmortem delay of 12 h (Fig. 4
Cortical tissue from a patient who had died in the final stage of multiple sclerosis was obtained at autopsy after a postmortem delay of 4 h (Table 2
). In this tissue, strong HSP-27 immunoreactivity was found histologically (Fig. 1C and C′
), and Western blot analysis revealed a clear immunoreactive band (Fig. 4
Figure 4. Western blot analysis of protein extracts, obtained from epilepsy patients, which were stored under anoxic conditions to mimic conditions during postmortem delay, and extracts from the brain affected by multiple sclerosis. Heat-shock protein (HSP)-27 is still detectable after 12 h of artificial anoxic conditions as well as after 4 h of postmortem delay in the brain affected by multiple sclerosis.
Download figure to PowerPoint
In the resected temporal lobe cortices from epilepsy patients, the HSP-27–labeled cells were not homogeneously distributed throughout the cortex. Instead, a wide variety of staining patterns for HSP-27 immunoreactive cells was found, some of which are depicted in Figs. 1 and 2. In many cases, astrocytes of the white matter and those in cortical layer I were strongly labeled, as were the walls of many blood vessels (Fig. 1D and G-G′). In other cases, only the walls of blood vessels were immunopositive in the cortical gray matter, whereas glial cells were rarely seen, except in white matter. Another case contained immunoreactive blood vessels with attached foci of immunoreactive glial cells (Fig. 1F). In several cases, large and very small foci of immunoreactive glial cells were found surrounded by completely normal-appearing, immunonegative tissue (Figs. 1E-E′; 2A and B). Two cases showed a net-like pattern of HSP-27–positive glial cells around immunonegative spots in the whole cortex (Fig. 1H-H′). As shown by double immunofluorescence, the majority of the parenchymal HSP-27–positive cells were astrocytes (Fig. 5). In a few cases of epilepsy from which hippocampal tissue was studied, a similarly increased expression of HSP-27 was observed (not shown).
- Top of page
To date, most reports of HSP-27 induction in human brain tissue refer to multiple sclerosis (21), glioblastomas (19, 38), and Alexander disease (39). Assimakopoulou and Varakis (20) described a selective increase of HSP-27 in astrocytomas that coexpress AP-1.
The present study demonstrated a considerably increased expression of HSP-27 in human epilepsy temporal cortex, supporting previous work by Erdamar et al. (22), who described the nonselective, homogeneous presence of HSP-27 outside of corpora amylacea, including the hippocampal formation as well as the mesial and lateral temporal cortex. In our samples of epilepsy tissue specimen from the lateral temporal cortex, HSP-27 was present, but its distribution was rather heterogeneous, indicating that the induction of HSP-27 expression may be caused by region-specific stimuli. Because HSP-27 is cytoprotective (10), its induction occurs in viable cells, which become affected by pathology-related mechanisms strong enough for the induction of a so-called stress response (7). Therefore our findings indicate that HSP-27 could be used as a marker for regions that become specifically affected during the course of the disease. In resected specimens, the borders of such regions may be of considerable interest, because they could offer the opportunity to study the beginning of ongoing pathology-related alterations in the expression of neurotransmitter receptors, multidrug transporters, or ion channels.
Studies carried out in rats after kainic acid–induced seizures (40,41) reported the existence of focal HSP-27 distribution patterns in the grey matter, in part similar to those described here for human epileptic tissue. The focal induction of HSP-27 in the rat (40,41) and its focal presence in our resected epilepsy human cortices demonstrated that HSP-27 expression was not caused by the neurosurgical procedure including the cauterization of epicortical blood vessels some minutes (5–15 min) before the final resection of the tissue. This also is in agreement with results obtained in animals in which HSP-27 induction is first found 3 to 5 h after a 60- or 90-min hyperthermic stimulus (42,43). In addition, HSP-27 expression was not degraded during the first 12 h of experimentally mimicked postmortem delay in our experiments. HSP-27 was clearly detectable by Western blotting and in immunohistochemistry after usual postmortem conditions in a case of multiple sclerosis. Such a pathological case can be used as a positive control, according to Aquino et al. (21).
From other experimentally induced neuropathologic conditions, it is known that HSP-27 induction is transient but lasts for several days and up to several weeks (7,41,44,45). This long-lasting upregulation could be the reason for the consistent presence of increased HSP-27 immunoreactivity in all our patients, independent of seizure frequency.
Our results are partly in agreement with previously reported data obtained in rats treated with kainate. Kainate injection induced either no HSP-27 upregulation (13) or only weak HSP-27 immunoreactivity during the first 24 h (40,41), whereas messenger RNA (mRNA) was upregulated much earlier and in a more widespread manner than that known from immunohistochemical studies (46). In contrast to these findings in experimental seizures, no HSP-27 immunoreactivity was found in human neurons, which is in agreement with previously reported data (22). This result is comparable to the findings reported for Alexander disease (39). Whether this reflects a species-specific difference or depends on the strength, the type, or the duration of seizures remains unclear. However, neuronal HSP-27 has been clearly shown in animal models of hyperthermia (42), traumatic cerebral injury (7,45), and in rats treated with kainic acid (41). Because kainic acid induces not only seizures, but also a longer-lasting mild hyperthermia (47), both seizure activity and hyperthermia may contribute to HSP-27 induction in this animal model.
For the kainate model, HSP-27 induction seems to be related to the appearance of apoptotic cell death (10). In addition, it is known that α,β-crystallin, another small heat shock protein that becomes co-induced in the kainate model (41), is induced by elevated extracellular potassium concentrations, depending on the activity of heat shock factor 2 (48,49). This mechanism of induction also may contribute to HSP-27 induction during seizure activity in human epilepsy cortex, because K+ concentration in the extracellular space was found to be increased with elevated activity in humans, as currently revealed during an interruption of the epilepsy surgical intervention (50).
Furthermore, the described glial responses to neuronal activity could be due to a seizure-related enhanced glial glutamate uptake and enhanced glutamate-receptor expression (51,52), as well as glutamate-induced cytosolic calcium oscillations (53,54) and disturbances in astrocyte-specific glutamate metabolism via glutamine synthetase, which is in turn affected by epileptic activity (55,56). As a whole, it is generally agreed that neuronal seizure activity represents a remarkable stress for the function of glial cells and of the blood–brain barrier (BBB). In addition, the astrocytic environment is an essential modulator of the glutamate-mediated intersynaptic cross-talk (57). Furthermore, neuron-to-astrocyte signaling is important for the regulation of cerebral microcirculation and BBB functioning (58). This may explain why a pathology-associated HSP-27 expression affects astrocytes and endothelial cells, as found in our epileptic human specimens. Similar findings are known from rats after PTZ-induced tonic–clonic seizure activity (24,59), thus indicating that HSP-27 in endothelial and mainly blood vessel–associated astrocytes may become initially induced in regions where the BBB is disturbed in response to seizure activity (60–62). This is further supported by the findings that HSP-27 contributes to a functional stabilization of the endothelial cytoskeleton and protection against stress oxidative (63–65) indicating that our observed endothelial HSP-27 induction in epileptic temporal cortex serves similar functions. Interestingly, Goldman et al. (61) reported that a subconvulsive dose of PTZ affects the BBB more severely than a higher dose. In the temporal cortex of epilepsy patients and in the PTZ model, blood vessels and associated glial cells were HSP-27 positive. This suggests that epileptic activity affects blood supply and vessel integrity, which may include hypoxia (66,67).
Hypoxia is known to induce seizures (67,68). Such hypoxic episodes may cause the formation of reactive oxygen species (ROS), potentiating oxidative stress and inhibiting mitochondrial metabolism. The formation of hydroxyl radicals and increasing NOx levels during epileptic activity have been reported (30,69–71). Kato et al. (41) described the loss of the antioxidant glutathione after kainic acid–induced seizures. Furthermore, a recent study showed that kindling with PTZ downregulates the endogenous defense mechanisms against oxidative stress, an effect that could be prevented by the application of taurine (72). It is well known that HSP-27 protects mitochondria (11) during cellular stress and also blocks fatty acid synthetase/apolipoprotein (FAS/APO)-1 and tumor necrosis factor (TNF)-α–induced cell death by increasing glutathione levels (73,74). Therefore, HSP-27 expression can be understood as a cellular defense mechanism by cells affected by seizure induced stressors, but which are still not damaged up to the point of no return.
In summary, a potent protective role of HSP-27 has been demonstrated in general (75) and by means of kainate-induced seizures in HSP-27–overexpressing animals (10). Therefore HSP-27 induction highlights cerebral regions in which seizures induce the initiation of a cellular stress–response program involving HSP-27, warranting further experimental studies to explore the therapeutic potential of HSP-27 modulation.