The calpain inhibitor MDL-28170 and the AMPA/KA receptor antagonist CNQX inhibit neurofilament degradation and enhance neuronal survival in kainic acid-treated hippocampal slice cultures

Organotypical slice cultures

Hippocampal slice cultures were prepared from 6-7-day-old [postnatal day (P)6-7] Sprague-Dawley rats using the modified method of Stoppini et al. (1991), and as described recently in detail by Holopainen et al. (2001). After decapitating the rat, the brain was placed immediately in cold Gey's balanced salt solution (Gibco, Invitrogen, Paisley, UK) supplemented with glucose (6.5 mg/mL). Hippocampal slices (400 µm) were cut perpendicular to the septotemporal axis using a McIlwain tissue chopper, and placed on top of semipermeable membrane inserts (Millipore Corporation, Bedford, MA, USA) in a six-well plate containing culture medium (50% minimum essential medium, 25% Hanks's balanced salt solution, 25% heat-inactivated horse serum, 25 mm HEPES, supplemented with GlutaMaxII and 6.5 mg/mL glucose; pH adjusted to 7.2]. Slices were cultured in an incubator (37 °C, 5% CO₂) for 7 days in vitro (DIV) with a change of medium carried out twice a week. No antibiotics were used. All treatments of animals were carried out according to the European Community Council directives 86/609/EEC and had the approval of the Animal Use and Care Committee of the University of Turku.

Pharmacological treatments

To determine the time course of the KA-induced degradation of NF proteins, hippocampal slices (7 DIV) were incubated with KA (5 µm; OPIKA-1, Ocean Produce International, Shelburne, NS, Canada) for 3, 6, 12, 24 and 48 h, and thereafter prepared for Western blotting. In addition, the 7-DIV cultures were treated with the following compounds for 24 h: KA (5 µm); the AMPA/KA selective antagonist CNQX (10 µm; Tocris Cookson Ltd, Avonmouth, UK); the l-type Ca²⁺-channel blocker nifedipine (10 µm; Tocris), the calpain protease inhibitor MDL-28170 (0.5, 5, 25 and 50 µm; Biomol, Plymouth Meeting, PA, USA); and the NMDA-receptor antagonist MK-801 (0.5 µm; Tocris). Treated and control slices were then used for Western blotting, immunocytochemistry and neuronal death studies. In all experiments, 7-DIV cultures were used as KA treatment at this in vitro age induces selective CA3 nerve cell damage while sparing the other main hippocampal nerve cell types (Holopainen et al., 2004).

Antibodies

Three different monoclonal antibodies were used for the Western blotting and immunocytochemical studies. Clones N52, NN18, and NR4 (all from Sigma, St. Louis, MO, USA) were used to detect phosphorylation-independent epitopes of NF-H, NF-M, and NF-L, respectively. All three antibodies were mouse IgG isotypes. As secondary antibodies, an HRP-conjugated goat anti-mouse IgG antibody (Sigma) was used in Western blots, and a biotin SP-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) was used in the immunocytochemical studies.

Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) immunoblotting

Western blot analysis was performed to determine the changes in the NF protein levels in response to different treatments as indicated above. Immunoblotting was carried out as described in detail for NF proteins in OHCs (Holopainen et al., 2001) with some modifications. Briefly, control and treated groups of slices from two different culture batches (24 slices for each time point and treatment group in both batches) were collected in ice-cold homogenization buffer containing 50 mm TRIS-HCl (pH 7.4), 1% SDS, 2 mm EDTA, 1 mm phenyl methyl sulphonyl chloride and 0.7 mm dithiothreitol, homogenized using an Ultra-Turrax T25 homogenizer (Janke and Kunkel, Staufen, Germany), homogenates immediately boiled, and then centrifuged at 10000 g for 30 min at 4 °C. Supernatants were collected, frozen and stored at −80 °C until required for use. The protein concentration of the samples was measured using the Lowry-based Biorad DC Protein assay (Biorad, Hercules, CA, USA). Equal amounts of protein were applied to each lane for SDS-PAGE, separated by electrophoresis with a 7.5% acrylamide minigel using Mini-protean III (Biorad), and then transferred to a polyvinylidene fluoride Immobilon-P (Millipore) membrane using the Mini Trans-blot Cell system (Transblot SD; Biorad). Membranes were incubated at 4 °C overnight with the monoclonal primary antibodies for NF-L (1 : 3500), NF-M (1 : 4000), and NF-H (1 : 3000). Thereafter, samples were incubated at room temperature (RT) with the horseradish peroxidase-conjugated secondary antibody for 1 h. The signal was obtained using the chemiluminiscence ECL system (Amersham, Bucks, UK) and Hyperfilm ECL (Amersham). The optical signals were quantified with Image J 1.33 (NIH, USA). Immunoblotting studies were repeated 4–6 times for each batch of slices and experimental conditions. Because of the rapid developmental changes in the expression of the NF protein (Lopez-Picon et al., 2004), a separate control group was used for the short (3, 6 and 24 h) and for the long (48 h) treatment groups.

Immunocytochemistry

Immunocytochemistry was used to detect possible regional changes in the expression and localization of NF proteins after the treatments. After 7 DIV, slices were transferred to fresh culture medium containing KA alone (5 µm), or combined with CNQX (10 µm), MDL-28170 (0.5 µm), nifedipine (10 µm) and MK-801 (0.5 µm), and cultured for a further 24 h. The immunocytochemistry was carried out as described recently in detail for NF proteins in OHCs (Holopainen et al., 2001), with some modifications. Briefly, slices were first washed with 0.1 m phosphate-buffered saline (PBS; pH 7.4), fixed with 4% paraformaldehyde (PFA) for 1 h at RT, and then processed with an antigen retrieval to enhance epitope detection and to reduce the background, as recently published by Fritschy et al. (1998). Following this, slices were detached mechanically from the insert, and the further staining steps were carried out with free-floating slices. Slices were first incubated in blocking solution (BS) containing 2% bovine serum albumin, 2% goat serum and 0.1% Triton X-100 in Tris-saline (TS-Triton) buffer (pH 7.4) for 1 h at RT, and thereafter with the primary antibodies for 24 h at 4 °C in BS at the following dilutions: NF-L (1 : 1000); NF-M (1 : 1500); and NF-H (1 : 1500). After washing, the slices were incubated with the biotin-conjugated secondary antibody (1 : 4000) in BS, and finally with the avidin-peroxidase conjugate (Vectastain ABC Kit, Vector Laboratories, Burlingame, CA, USA). The staining was detected with 3,3′-diaminobenzidinetetrahydrochloride (DAB; Sigma) as a chromogen, and processed further as described previously (Lopez-Picon et al., 2003). Slices, in which the primary antibody was omitted but that were otherwise treated as above, served as the negative controls for the immunostaining. Slices (n = 12–15) from at least three different culture batches were used in all experimental groups, and for each antibody. The immunoreactivity was scored visually by using the following scheme: -, negative; -/+, weakly positive; +, positive; + +, moderately positive; + + +, strongly positive.

Thionin stainings

Thionin staining was used to determine whether KA treatment (24 h) alone or combined with CNQX (10 µm) and MDL-28170 (0.5, 5, 25 and 50 µm) resulted in neuronal loss in slices. After treatment, the slices were cultured for an additional 48 h in normal medium (recovery phase), as our preliminary time course experiments showed that 24 h in normal medium after the treatment was not long enough to verify the long-term consequence of the treatments. For the staining, hippocampal slices were removed from semipermeable membranes to gelatin-coated glass slides and briefly stained in 0.1% thionin, dehydrated in an alcohol series, cleared in xylene and coverslipped. Slices from at least three different culture batches (n = 7–12 in each treatment) and stained in different experiments were used for each treatment. Neurons in the CA3 regions were scored using the following scheme: 0, no stained neurons in the cell layer (all neurons dead in terms of loss of Nissl staining); 1, some stained neurons; 2, a sparse number of stained neurons; 3, many stained neurons, but with slightly disturbed cell layer integrity; 4, numerous stained neurons with good cell layer integrity (regarded as normal).

Light microscopy

After immunocytochemical and thionin stainings, slices were examined with a Leica DM R microscope (Heerbrugg, Switzerland) under bright field optics. A digital camera (Olympus U-TV1 X; Olympus Optical Co., Ltd, Tokyo, Japan) was used to capture pictures using an Olympus BX60 microscope (Olympus), and images were processed further using Adobe Photoshop (version 6.0; Adobe Systems Inc. San Jose, CA, USA) and Corel Draw (version 11.0).

Fluoro-Jade B

Fluoro-Jade B (FJB) was used to study the KA-induced neuronal degeneration, as described in detail by Holopainen et al. (2004). This dye is an anionic tribasic fluoresce derivative with excitation peaks of 362 nm and 390 nm and an emission peak of 550 nm, and it stains degenerating neurons (Schmued & Hopkins, 2000). The FJB staining was studied in 7-DIV slices treated for 24 h with KA (5 µm) alone, or combined with CNQX (10 µm) and MDL-28170 (0.5, 5, 25 and 50 µm). Slices, attached to the membrane, were first rinsed with cold PBS, fixed with 4% PFA for 1 h at RT, washed twice with PBS and once with water, transferred to 0.06% potassium permanganate (Sigma) for 2 min, washed with water, and then transferred to 0.001% FJB solution for 30 min. Finally, the slices were detached from membranes, transferred to gelatin-coated glass slides, dried, immersed in xylene and coverslipped.

Confocal microscopy and verification of FJB staining

All specimens stained with FJB were examined with a Leica TCS SP confocal microscopy system (Leica, Heidelberg, Germany) equipped with an argon-krypton laser (Omnichrome, Melles Griot, Carlsbad, CA, USA). The laser wavelength used for excitation of FJB was 488 nm, and the emission detection window was 500–600 nm. Occasionally, the emission detection settings were changed slightly for optimal performance. Confocal image stacks were acquired at 2-µm steps by using 10 × objective (HC PL APO 10 ×/0.40), and processed with Leica TCS NT/SP Scanware (version 1.6.587) software. The algorithm used was 2-D maximum intensity projection, which determines the maximum of all intensity values in a stack of sections, and displays them in one single image. Reconstruction of the images of the whole hippocampal slice was accomplished afterwards from 1-mm² maximum projections using Adobe Photoshop. As our recent study indicated that KA treatment mainly damaged the pyramidal CA3a/b region (Holopainen et al., 2004), we focused our studies on this specific region. For our scoring analyses, the area of stained neurons (i.e. degenerating neurons) was measured blinded (T.-K. and K.-L.) from maximum projections of FJB-stained hippocampi (n = 10–15 in each experiment group) using ImageJ software (NIH, USA), as described recently in detail (Kukko-Lukjanov et al., 2006). The following scoring system, in which the scoring numbers indicate the area (in mm²) of FJB-stained CA3a/b neurons, was used to evaluate the extent of damage: 0, no FJB-stained neurons (regarded as normal); 1, up to 0.100 mm²; 2, 0.101–0.200 mm²; 3, 0.201–0.300 mm²; and 4, > 0.300 mm².

Statistical analysis

The overall differences in the signal intensity in Western blots of NF-L, NF-M, NF-H after KA treatment of different durations and treatments, and the overall group differences in the score numbers after the FJB and thionin stainings were assessed with one-way analysis of variance (anova) with the Tukey-Kramer multiple comparison test as a post hoc test. The difference between two experimental groups was assessed with Student's two-tailed t-test. The level of significance was set at P < 0.05. The statistical analyses were performed using GraphPad Prism software (version 4.0; GraphPad Program, San Diego, CA, USA).

Time course of NF degradation

Western blotting was used to assess the time course of KA-induced NF protein degradation in hippocampal slices at 7 DIV. Figure 1A shows representative Western blots, and Fig. 1B shows the semiquantitative analysis of the NF-L, NF-M and NF-H proteins in control cultures and in cultures after KA treatment for 3, 6, 12, 24 and 48 h. As early as 3 h after treatment, the signal for all the NF proteins started to decrease, and this decline was significant (P < 0.01) after 6 h. The levels of all the NF proteins were then ∼ 30% lower in KA-treated slices than in control slices. The signals continued to decrease significantly (P < 0.001) with the treatment time, and after the 24-h KA treatment the values were ∼ 40% of those detected in control cultures. By 48 h small, but not further significant, changes compared with the 24 h treatment were detected.

Effect of pharmacological treatments on NF stability

To elucidate the contribution of different glutamate receptor subtypes, the significance of l-type Ca²⁺-channels and calpains in the regulation of KA-induced NF degradation, additional Western blots were carried out in hippocampal slices (7 DIV) treated for 24 h with 5 µm KA together with the AMPA/KA selective antagonist CNQX (10 µm), the calpain inhibitor MDL-28170 (0.5 µm), the Ca²⁺-channel blocker nifedipine (10 µm) and the NMDA receptor antagonist MK-801 (0.5 µm). Figure 2A shows representative Western blots of NF-L, NF-M and NF-H in control and treated cultures, and Fig. 2B shows the semiquantitative analysis of the levels of the NF protein after the treatments. In KA-treated cultures, the signal for all the NF proteins decreased by about 60%, whereas CNQX and the calpain inhibitor MDL-28170 effectively blocked the KA-induced degradation, resulting in the NF levels similar to those of control slices (Fig. 2B). Nifedipine and MK-801 did not significantly block the KA-induced NF protein degradation.

Immunocytochemistry

Immunocytochemistry was carried out to detect any changes in the expression of NF-L, NF-M and NF-H at the cellular level in 7-DIV slices treated for 24 h either with KA (5 µm) alone or combined with CNQX (10 µm), MDL-28170 (0.5 µm), nifedipine (10 µm) and MK-801 (0.5 µm). Table 1 shows the visual scoring of NF-L, NF-M and NF-H immunoreactivity in control and treated slices, and representative images of NF-L and NF-H staining are shown in 3, 4, respectively.

In control cultures, the NF-L staining intensity varied region-specifically (Table 1). The CA3 pyramidal cell bodies were moderately NF-L immunoreactive (Fig. 3A and B), whereas mainly neuronal processes were stained in the CA1 region (Fig. 3C). In KA-treated slices, the NF-L immunoreactivity decreased in all hippocampal subregions, but more markedly in the CA3 pyramidal cells (Fig. 3D and E), whereas heavily immunopositive neurons were found in the CA1; on the basis of their morphology, these cells were regarded as interneurons (Fig. 3F). In the presence of KA + MDL-28170 (0.5 µm; Fig. 3G–I), the intensity of staining within the hippocampal subregions either remained at the control level, or even increased (Table 1), specifically in the CA3 pyramidal neurons and in their proximal processes (Fig. 3H). Moreover, the demarcation between immunopositive CA3 neurons and immunonegative/lightly stained CA1 neurons was surprisingly abrupt (Fig. 3G), and numerous heavily stained interneurons appeared in the CA1 region (Fig. 3I). In KA + CNQX-treated slices (Fig. 3J and K), the staining pattern was remarkably similar to that of KA + MDL-28170, being enhanced in the CA3 pyramidal neurons, the DG hilar region and also in the molecular layers. Moreover, heavily stained interneurons appeared in the CA1 region (data not shown) whereas DG granule cell bodies remained lightly stained (Fig. 3J).

In general, the NF-H staining resembled that of NF-L (Table 1). Representative images show that the moderate NF-H immunostaining in control slices (Fig. 4A–C) decreased markedly in KA-treated slices in many hippocampal subregions (Fig. 4D–F), particularly in the CA3 pyramidal cell layer (Fig. 4D and E). In the presence of KA + MDL-28170 (0.5 µm), staining was enhanced in the CA3 cell bodies (Fig. 4G and H) and in their proximal processes. Moreover, the inner molecular layer staining was more pronounced in KA- and KA + MDL-28170-treated slices (Fig. 4G and I) than in control slices (Fig. 4C), whereas granule cell bodies remained lightly stained.

The NF-M immunoreactivity resembled that of NF-L and NF-H in control and KA-treated slices, but in the presence of KA + CNQX and KA + MDL-28170, the staining intensity of the CA3 cell bodies and interneurons remained at the control level (Table 1). After KA + nifedipine and KA + MK-801 treatments, the immunoreactivity of all the NF proteins remained unaltered or even slightly decreased (Table 1).

Neuronal death

Neuronal death was studied by using FJB staining (Fig. 5A–D) in 7-DIV cultures treated for 24 h with KA (5 µm) alone, or combined with CNQX (10 µm), and MDL-28170 (0.5, 5, 25 and 50 µm). In control cultures, no FJB-stained neurons were detected in any of the hippocampal subregions (Fig. 5A), indicating a good survival of neurons. In contrast, the CA3a/b region was stained massively with FJB in KA-treated slices, indicating pronounced neuronal damage (Fig. 5B). The area of FJB-stained neurons was decreased significantly in the presence of KA + CNQX (Fig. 5C and I), and decreased concentration-dependently in slices treated with KA + MDL-28170 (50 µm) (Fig. 5D and I), whereas nifedipine and MK-801 (data not shown) had no effect. As FJB-staining indicated neuronal damage after the 24-h treatment with KA, we were interested to know whether or not stained neurons were permanently damaged and would later disappear, and the conventional thionin staining was carried out after the 48-h culture period in normal medium. The staining confirmed that the main hippocampal cell layers were well-preserved in control cultures (Fig. E), whereas KA treatment resulted in a massive loss of CA3 neurons (Fig. 5F and J), which was alleviated significantly by CNQX (Fig. 5G and J), and dose-dependently by MDL-28170 (Fig. 5H and J). It is noteworthy that CA1 pyramidal neurons and DG granule cells were not affected by KA treatment.

Calpain- and KA-induced neurofilament degradation

Our present results showed degradation of all three NF proteins in response to KA treatment in OHCs, a finding in accordance with recent studies in various models of brain insults. For example, NF-M and NF-H degradation has been detected within 1 h of hypoxic insult in the optic nerve (Stys & Jiang, 2002), and shortly after oxygen and glucose deprivation in acute brain slices prepared from adult rats (Tekkok et al., 2005). Moreover, NF-H degradation occurs within 1–2 h of traumatic spinal cord injury (Schumacher et al., 1999), and within hours of a glutamate-induced insult in cultured cortical neurons (Chung et al., 2005). Also, our earlier study indicated that KA-treatment (5 µm, 48 h) induced a pronounced decrease in the expression of all NF proteins, the older cultures (21 DIV) being affected more severely than the young ones (7 DIV; Holopainen et al., 2001).

Calpains are known to be directly involved in glutamate-induced hippocampal damage, and the activation of these proteases is suggested to be an early, requisite step in the cascade of events initiated by excitotoxic and hypoxic injuries which, through various intracellular signalling pathways, contributes to cytoskeletal degradation, and finally leads to nerve cell death both in vivo and in vitro (Siman et al., 1989; Minger et al., 1998; Lankiewicz et al., 2000; Stys & Jiang, 2002; Rami, 2003; Araujo et al., 2004; Wu et al., 2004). Moreover, spectrin breakdown products, indicators of the Ca²⁺-activated calpain activity, have been detected in the adult rat hippocampus 3 h after the intracerebroventricular injection of KA (Siman et al., 1989), and as little as 1 h after KA-induced seizures in the DG of the adult rat hippocampus (Bi et al., 1996). In addition, increased calpain activity occurs within 20 min of the ischaemic insult in the CA1 region (Rami, 2003). The fast NF degradation shown in our present study, together with earlier studies by other groups, suggests that the degradation of cytoskeleton might be an early step and not a late consequence in the process of neuronal death. Moreover, these studies further suggest that calpain activation could be of major importance in many pathological conditions affecting the human brain, including excitotoxic nerve cell damage.

AMPA/KA receptor antagonist and calpain inhibitors promote NF integrity

The marked KA-induced degradation of all three NF proteins in our OHCs was effectively inhibited with the calpain inhibitor MDL-28170 and with the potent AMPA/KA receptor antagonist CNQX, whereas the blockade of Ca²⁺ entrance with the l-type Ca²⁺-channel blocker nifedipine and the use of the NMDA receptor antagonist MK-801 had no protective effect. The observed beneficial effect of calpain inhibitors in preserving cytoskeletal integrity is in accordance with earlier studies, which have shown different calpain inhibitors to be effective in ameliorating excitotoxic and hypoxic injuries, and decreasing spectrin, calcineurin and NF breakdown (Li et al., 1998; Kunz et al., 2004; Wu et al., 2004). Moreover, calpain inhibition has been effective in protecting neurons from excitotoxic insults both in vitro (Chen et al., 1997; Lankiewicz et al., 2000; Araujo et al., 2004; Wu et al., 2004) and in vivo conditions (Wu et al., 2004; Higuchi et al., 2005). The fact that the calpain inhibitor was now added simultaneously with KA, and not before KA, mimics the clinical situation in a more relevant way, and further corroborates the idea that calpain activation at the early stage of the insult might contribute to NF degradation. This renders neurons more vulnerable to insults, which were effectively inhibited by the calpain inhibitor.

The low KA concentration (5 µm) used in our study is within the range (3–8 µm), which activates mainly KA receptors, although AMPA receptors are also activated to some extent (Kristensen et al., 2001), allowing calcium influx. In the presence of CNQX, calcium entry via AMPA/KA receptors is reduced, and results in lower calpain activation. This could explain the observation that MDL-28170 prevented NF degradation to the same extent as did CNQX. This speculation is in accordance with the recent study by Araujo et al. (2004), in which they showed that the specific AMPA receptor antagonist NBQX inhibited the cleavage of calpain substrate in cultured hippocampal neurons in a manner similar to that of the calpain inhibitor MDL-28170. Moreover, our hypothesis of the CNQX mechanism of action is in agreement with a recent study, in which AMPA receptor-positive modulators activated calpain proteases and resulted in a breakdown of spectrin, and this has effectively been inhibited by the calpain inhibitor (Jourdi et al., 2005).

Neuronal damage and neuroprotection

The current FJB and thionin staining results showed that the neuronal death process (FJB staining), and the ultimate death (thionin staining) in OHCs were effectively attenuated by CNQX, and attenuated in a concentration-dependent manner by MDL-28170. Moreover, these treatments not only protected NF proteins from degradation, but also enhanced their immunoreactivity in specific hippocampal subregions, i.e. in the CA3 and hilar regions and in the inner molecular layers of the DG. Although the molecular basis for this is currently unknown, it is tempting to speculate that it as a sign of an intracellular protection process, as recently suggested in a mouse model of amyotrophic lateral sclerosis (Lariviere & Julien, 2004). As caspase-3 and poly(ADP)ribose polymerase, the key mediators of the apoptotic pathway, are not activated in our experimental conditions (Holopainen et al., 2004), calpain activation could be the key mediator of NF degradation and contribute to neuronal death, and, as suggested earlier, might convert excitotoxic nerve cell death into a caspase-independent process (Lankiewicz et al., 2000).

In conclusion, our present study in KA-treated OHCs, as a model system for excitotoxic nerve cell damage, suggests that there exist a link between the degradation of NF proteins by calpain proteases and neuronal damage after the insult. As the NF proteins are crucial for many basic cellular functions, e.g. radial growth, dendritic arborization, maintenance of myelinated axons and axonal transport (Lee & Cleveland, 1996; Fuchs & Cleveland, 1998; Julien & Mushynski, 1998; Julien, 1999), their stability is essential for nerve cell integrity, intracellular architecture and survival. Consequently, the inhibition of their degradation could be an important step in a cascade of events promoting nerve cell viability after brain insults. However, the relationship between inhibition of calpains, cytoskeletal degradation and neuronal survival is complex, and further research is required to better understand the intracellular signalling pathways involved in calpain- and AMPA receptor-mediated effects on the cytoskeleton, and nerve cell survival in the immature hippocampus.