Distinct Mechanistic Roles of Calpain and Caspase Activation in Neurodegeneration as Revealed in Mice Overexpressing Their Specific Inhibitors*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 15, Issue of April 15, pp , by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Distinct Mechanistic Roles of Calpain and Caspase Activation in Neurodegeneration as Revealed in Mice Overexpressing Their Specific Inhibitors* Received for publication, January 25, 2005, and in revised form, February 4, 2005 Published, JBC Papers in Press, February 7, 2005, DOI /jbc.M Makoto Higuchi, Masanori Tomioka, Jiro Takano, Keiro Shirotani, Nobuhisa Iwata, Hajime Masumoto, Masatoshi Maki, Shigeyoshi Itohara**, and Takaomi C. Saido From the Laboratory for Proteolytic Neuroscience, RIKEN Brain Science Institute, Wako, Saitama , Laboratory of Molecular and Cellular Regulation, Department of Applied Biological Sciences, School of Agricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya , and **Laboratory for Behavioral Genetics, RIKEN Brain Science Institute, Wako, Saitama , Japan Enzymatic proteolysis has been implicated in diverse neuropathological conditions, including acute/subacute ischemic brain injuries and chronic neurodegeneration such as Alzheimer disease and Parkinson disease. Calcium-dependent proteases, calpains, have been intensively analyzed in relation to these pathological conditions, but in vivo experiments have been hampered by the lack of appropriate experimental systems for a selective regulation of the calpain activity in animals. Here we have generated transgenic (Tg) mice that overexpress human calpastatin, a specific and the only natural inhibitor of calpains. In order to clarify the distinct roles of these cell death-associated cysteine proteases, we dissected neurodegenerative changes in these mice together with Tg mice overexpressing a viral inhibitor of caspases after intrahippocampal injection of kainic acid (KA), an inducer of neuronal excitotoxicity. Immunohistochemical analyses using endo-specific antibodies against calpain- and caspase-cleaved cytoskeletal components revealed that preclusion of KA-induced calpain activation can rescue the hippocampal neurons from disruption of the neuritic cytoskeletons, whereas caspase suppression has no overt effect on the neuritic pathologies. In addition, progressive neuronal loss between the acute and subacute phases of KA-induced injury was largely halted only in human calpastatin Tg mice. The animal models and experimental paradigm employed here unequivocally demonstrate their usefulness for clarifying the distinct contribution of calpain and caspase systems to molecular mechanisms governing neurodegeneration in adult brains, and our results indicate the potentials of specific calpain inhibitors in ameliorating excitotoxic neuronal damages. Diverse proteolytic enzymes have been indicated to mediate molecular processes of neurodegeneration (1) because axonal, dendritic, and synaptic integrity is targeted by protease activities provoked by different types of insults (2 4). A family of * This work was supported by a research grant from RIKEN BSI and from the Ministry of Education, Culture, Sports, Science, and Technology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Both authors contributed equally to this work. To whom correspondence may be addressed. Tel.: ; Fax: ; mhiguchi@brain.riken.jp. To whom correspondence may be addressed. Tel.: ; Fax: ; saido@brain.riken.jp. This paper is available on line at nonlysosomal calcium-activated neutral cysteine proteases, referred to as calpain family ( - and m-calpain isoforms in the present study), has been mechanistically implicated in the regulation of various cellular functions (5, 6). Because multiple lines of evidence indicate that neuronal cytoskeletal constituents, including microtubule-associated proteins (MAPs), 1 neurofilaments, spectrin, and actin, are preferred substrates for calpain (7 11), calpain is likely to play essential roles in the pathophysiological derangement of cytoskeletons. Acute and subacute brain insults such as traumatic injury and hypoxia/ ischemia involve prominent calpain activation evoked by calcium dysregulation in neurons. Moreover, several independent studies have highlighted the elevation of calpain activity in chronic neurodegenerative disorders including Alzheimer disease (12, 13) and Parkinson disease (14). However, it remains unclear whether the relationship between calpain activation and the disorganization of the neuronal cytoskeletons in these pathologies is causal or coincidental. To our knowledge, there has been a relatively small amount of in vivo information for the molecular consequences of calpain activation in neurons. Although attempts to intervene in calpain activation in experimental animal models of head trauma (15), ischemia (16), and several other brain pathologies have been made, there have been the following limitations in modulating calpain activity (1): knock-out mice lacking the 30-kDa small regulatory subunit of calpain suffer embryonic lethality (17, 18), indicating difficulty in direct reverse-genetic suppression of calpain expression (2); and all synthetic peptidic, peptide-mimetic, and nonpeptidic calpain inhibitors currently available have problems in terms of specificity, metabolic stability, water-solubility, or penetration through the blood-brain barrier (19, 20). Calpastatin (CAST), the only natural inhibitor of both - and m-calpains (21, 22), is a potentially valuable tool to investigate the significance of calpain activation in living animals. A recent study, utilizing adenovirus-mediated overexpression of CAST for suppressing the calpain activity in an experimental mouse model of Parkinson disease, demonstrated protective effects of locally overexpressed CAST on striatal neurons (14). However, generation of genetically engineered mice carrying the CAST transgene is more advantageous in accomplishing a stable and persistent 1 The abbreviations used are: MAPs, microtubule-associated proteins; CAST, calpastatin; hcast, human CAST; KA, kainic acid; mcast: murine CAST; LSD, least squares difference; MTs, microtubules; NF, neurofilament; ntg, nontransgenic; Tg, transgenic; VIC, viral inhibitor of caspases; TUNEL, transferase-mediated dutp nickend labeling; ANOVA, analysis of variance.

2 15230 Calpain and Caspase Activation in Excitotoxicity overexpression of CAST in broad areas of the mouse brain. Such transgenic (Tg) mice could also be cross-bred with numerous models of neurodegenerative disorders to examine the contribution of calpain activation to various neuropathologies. We recently generated Tg mice overexpressing a viral inhibitor of caspases (VIC), and we assessed the effects of caspase inhibition on neuronal loss induced by a local injection of kainic acid (KA), an agonist for excitatory amino acid receptors (23). A similar experimental paradigm may be applicable to investigating the pathological role of the calpain-cast system. In the present study, we have generated Tg mice overexpressing human CAST (hcast) and raised antibodies to calpain- and caspase-cleaved cytoskeletal constituents, which can be employed to immunohistochemically visualize the activities of these proteases using both frozen and paraffin sections. The hcast and VIC Tg mice and these antibodies permitted us to establish an experimental paradigm that is ideal for distinctively analyzing the mechanistic roles of calpain and caspase activation. Whereas both the VIC and hcast transgenes were driven by the same neuron-specific calcium/calmodulindependent protein kinase-ii subunit promoter (24), only the hcast Tg mice showed a significant preservation of the neuritic cytoskeletons and an attenuation of progressive neuronal loss between the acute and subacute phases following the KA challenge. In contrast, there was little difference in the pathological consequences of KA injection between nontransgenic (ntg) and VIC Tg mice. These findings support the view that the calpains-cast system is critically involved in the molecular events taking place in excitotoxicity-induced neuropathology in adult brains, and also demonstrate the usefulness of our experimental model for obtaining insights into the mechanisms governing a wide range of neurodegenerative disorders. MATERIALS AND METHODS Tg Mice Overexpressing hcast and VIC The hcast cdna (25, 26) was cloned into pnn265, from which the NotI fragment was subcloned into pmm403 containing the calcium/calmodulin-dependent protein kinase-ii promoter (27). The SfiI-linearized DNA construct was microinjected into the pronucleus of C57BL/6Cr zygotes, which were transferred into foster females to create hcast Tg founders. The genotype was determined by PCR with primers 5 -CATGAACCACAGACAGCT- TGGTTGAC-3 and 5 -GGAGGATTTGATATTCACCTGGCCCG-3 that yield a 350-bp product. The VIC Tg mice were described previously (23). We used the line of mice with the highest expression of VIC for the present study. Intrahippocampal Injection of KA Administration of KA (1 nmol) into the hippocampal CA1 region of 6-month-old ntg, VIC Tg (23), and hcast Tg mice was performed as described previously (23). Transcardial perfusion of the mice for biochemical and neuropathological studies was performed 4 and 24 h and 7 days after the KA injection. All the animal experiments were performed in accordance with our institutional guidelines. Western Blot Analysis of hcast Expression in the Brain Levels of Tg hcast protein in different brain regions of untreated ntg and Tg mice were examined by means of immunoblot analysis. The brain tissue samples taken from the mice were homogenized in 10 ml/g tissue ice-cold lysis buffer (50 mm Tris-HCl (ph 8.0) containing 150 mm NaCl, 50 mm EDTA, 1% Triton X-100, and protease/phosphatase inhibitor mixture) and centrifuged at 14,000 g for 40 min at 4 C. Equal amounts (10 g of protein) of the resultant supernatants were employed for immunoblot analysis with the monoclonal anti-hcast antibody. Histochemical and Immunohistochemical Studies Mice under deep anesthesia were transcardially perfused with 15 ml of ice-cold PBS, followed by 20 ml of 4% paraformaldehyde in phosphate buffer (23). 4- m-thick paraffin sections and 20- m-thick frozen sections of the brains were immunostained based on either a standard protocol for immunofluorescence staining (23) or the tyramide signal amplification method with a TSA-Direct kit (PerkinElmer Life Sciences) using antibodies described below. We employed the staining techniques based on the method described by Wang et al. (28). for double or triple fluorescence labeling. Viability of the hippocampal neurons was studied by NeuN immunostaining; cell morphology and DNA fragmentation were monitored by cresyl violet staining and transferase-mediated dutp nick-end labeling (TUNEL) assay (Apoptosis Detection System; Promega, Madison, WI), respectively. Antibodies and Dilutions Rabbit polyclonal antibody against calpain-generated C-terminal fragment of -spectrin, the 150-kDa fragment (1:200 dilution) (29), was used for monitoring the magnitude of the calpain activity. In addition, we developed rabbit polyclonal antibodies to the calpain-generated N-terminal fragment of -spectrin, the 136- kda fragment (1:2,000 dilution), and the caspase-generated actin fragment, clact32 (1:2,000 dilution), using the synthetic peptides CQQQEVY and CYELPD, respectively (30, 31), as described previously (29). We also raised a polyclonal antibody against mcast using the synthetic peptide CKKTEEVSKPKAKEDARHS. We used a mouse monoclonal antibody to hcast (CSL1 5; 1:1,000 dilution; Takara Biomed). Mouse monoclonal antibodies to tau used here include T49 (1:500 dilution; murine tau-specific, phosphorylation-independent) (32) and AT8 (1:2,000 dilution; phosphorylated tau-specific; Innogenetics) (33). The following antibodies to neurofilament (NF) subunits were employed for immunostaining: mouse monoclonal RMdO9 to nonphosphorylated NFH (1:500 dilution) (34); RMO24 to phosphorylated NFH (1:500 dilution) (34); RMO189 to phosphorylated and nonphosphorylated NFM (1:500 dilution) (34); and rabbit antisera against NFL (anfl; 1:1,000 dilution) (35). Other antibodies used in the present study are as follows: mouse monoclonal antibodies to MAP2 (AP20; 1:10 dilution; Roche Diagnostics), -tubulin (DM1A; 1:1,000 dilution; Sigma), synaptophysin (1:50 dilution; Progen Biotechnik), and NeuN (1:2,000 dilution; Chemicon); and rabbit polyclonal antibodies to SV2A (1:100 dilution; Calbiochem-Novabiochem), GluR1 (1:1,000 dilution; Chemicon), mglur2/3 (1:1,000 dilution; Chemicon), and Cdk5 activator p25/p35 (C19; 1:1,000 dilution; Santa Cruz Biotechnology). Statistical Analysis The significance of differences between groups was examined by means of either t statistics or multiple comparisons by Fisher s LSD test following ANOVA. RESULTS The hcast Tg mice robustly expressed the transgene-derived protein in the forebrain regions. One of the founders that stably expressed hcast showed the highest copy number of transgenes and the highest level of hcast protein expression (Fig. 1A). This line of mice (L7) was used for the following experiments. The regional distribution of the overexpressed hcast was analyzed by immunohistochemistry with antihcast antibody (Fig. 1, B and C). The brain calpastatin activity in biochemical terms, defined as the activity to inhibit purified bovine m-calpain activity (36), was 3.05 times greater in the Tg mice than in the ntg controls. The transgene-derived hcast was highly expressed in neurons of the hippocampal formation and neocortex (Fig. 1C). Similar to the endogenous mcast (Fig. 1D), hcast was present mainly in the neuronal cell bodies, dendrites, and synaptic terminals (Fig. 1E). Double immunolabeling clearly demonstrates the high level expression of hcast in the presynaptic terminals (Fig. 1F) and dendrites (Fig. 1G). None of the hcast Tg mice exhibited marked motor or behavioral impairments at any age up to 15 months. In addition, there was no significant difference in body weight or longevity between the Tg and ntg mice. Immunohistochemical analyses for cytoskeletal components and synaptic markers (see Materials and Methods ) also revealed no pronounced abnormalities in the amounts and distributions of these molecules in the brain of Tg mice at any age. KA-induced calpain activation was prominently suppressed in the hippocampus of hcast Tg mice (Fig. 2). Immunohistochemistry using antibodies to the 150-kDa fragment (data not shown) and the 136-kDa fragment (Fig. 2, A and B) of calpaincleaved -spectrin (fodrin) revealed calpain activation in the hippocampal CA1 neurons 4 h after the KA injection. The 136-kDa fragment immunoreactivity in the ipsilateral hippocampus of the hcast mice was significantly decreased by 63% in comparison with that of the ntg mice (Fig. 2E). Caspase activation visualized by means of immunostaining for a 32-kDa fragment of caspase-cleaved actin, fractin

3 Calpain and Caspase Activation in Excitotoxicity FIG. 1.Transgenic expression of hcast in mice. A, Western blot analyses with anti-hcast (upper panel) and anti-murine CAST (mcast) (lower panel) antibodies for the protein samples extracted from the brains of ntg and four different lines of hcast Tg mice. No remarkable difference in the level of endogenous mcast was found among all the mice analyzed. B and C, immunofluorescence observations of ntg (B) and hcast Tg (C) mouse brain sections with an anti-hcast antibody. High levels of hcast expression were observed in the hippocampus and neocortex of Tg mice. D and E, localization of endogenous mcast (D) and transgene-derived hcast (E) in the dentate gyrus. Immunofluorescence signals of hcast were more concentrated in soma than processes as compared with those of mcast. The mcast immunoreactivity was particularly intense in the middle molecular layer, whereas hcast immunoreactivity was more prominent in the outer molecular layer. F and G, confocal images of double immunofluorescence staining with a combination of antibodies against hcast and presynaptic marker SV2 (F) or dendritic marker MAP2 (G) for sections from the CA3 (F) and CA1 (G) sectors of hcast Tg. The transgene-derived hcast was abundantly present in the presynaptic terminals (F) and dendrites (G). Scale bars: 1000 m (B and C) and 100 m (D G). (clact32), was observed primarily in the CA3 sector 4 h after the KA injection (Fig. 2B) and spread to the CA1 region by 24 h (Fig. 2, G and H). The clact32 immunoreactivity represents activation of the caspase cascade (30). The calpain activation preceded the caspase activation in the CA1 region, and the hcast Tg mice exhibited diminution of not only the 136-kDa fragment signals but also clact32 staining (Fig. 2, F and I), implying a requirement of calpain activation for the caspase activation in the CA1 neurons. The VIC Tg mice showed a marked suppression of caspase activity, but not calpain activity at all, in the hippocampus at every time point when compared with the ntg mice (Fig. 2, D F and J). Overexpression of hcast also halted excitotoxicity-induced progressive loss of hippocampal neurons. The majority of the hippocampal CA1 and CA3 neurons exhibited apparent atrophy and pyknosis as shown by cresyl violet staining 24 h after the KA administration (Fig. 3, A C). No significant difference in the number of pyknotic neurons was observed among the ntg, hcast Tg, and VIC Tg mice at this time point, although shrinkage of the neuronal cytoplasm was less severe in the hcast Tg mice than in the ntg and VIC mice. Additionally, there was no significant difference in the number of TUNEL-positive neurons among the three groups of mice at any time point after the KA treatment (data not shown). The CA1 region of the three groups of mice also showed similar reductions of NeuN immunoreactivity (Fig. 3, D F and J) 24 h after the KA injection. In contrast, FIG. 2. Calpain- and caspase-catalyzed proteolysis in ntg, hcast Tg, and VIC Tg mice in early stages after intrahippocampal KA injection. A D, double immunofluorescence staining with the anti-136-kda fragment (green) and anti-clact32 (red) antibodies for sections of ntg (A and B), hcast Tg (C), and VIC Tg (D) mouse hippocampal formations 4 h after injection of PBS (A) orka(b D). The CA1 sector of ntg mice exhibited marked calpain-catalyzed proteolysis indicated by the 136-kDa fragment immunoreactivity, whereas caspase-catalyzed fractin formation was localized primarily to the CA3 region (B). Relative to ntg mice, calpain activation was attenuated in hcast Tg mice (C); VIC Tg mice showed suppression of caspase activation (D). E, quantification of the intensities of the 136-kDa fragment (green columns) and clact32 (red columns) immunofluorescence signals in the whole hippocampal formations of ntg, hcast Tg, and VIC Tg mice 4 h after the KA challenge (n 3 in each group). Data were normalized against the mean value given by KA-challenged ntg mice. The quantification was performed within a linear range. F, intensities of clact32 immunostaining signals in the CA1 region 4 (pink columns) and 24 h (red columns) after KA injection (n 3 in each group). Data were normalized against the mean value given by ntg mice at 4 h. G J, confocal images of double immunolabeling with the anti-136-kda fragment (green) and anti-clact32 (red) antibodies for the CA1 region of ntg (G and H), hcast Tg (I), and VIC Tg (J)mice4(G) and 24 h (H J) after KA injection. Calpain activation preceded caspase activation (G), and thereafter caspase activation emerged, partially colocalizing in the soma and dendrites in ntg mice (H). Reduced signals of both the 136-kDa fragment and clact32 were observed in hcast Tg mice (I). In VIC Tg mice, calpain was activated in a manner similar to that in ntg mice, whereas caspase activation was prominently attenuated (J). Vertical line represents S.E. Scale bars, 300 m (A D) and 75 m (G J). *, p 0.05; **, p 0.01 versus ntg mice by ANOVA/LSD. there was a striking difference in the amount of NeuN-positive nuclei 7 days after the KA treatment (Fig. 3, G I and J). Notably, the hcast Tg mice did not show any progression of the NeuN decline from 24 h to 7 days, unlike the ntg and VIC Tg mice. These results indicate that inhibition of calpains, but not caspases, protected hippocampal neurons against KA-triggered subacute death, whereas the suppression of calpain or caspase activities did not modulate acute excitotoxic neuronal degeneration in the present experimental paradigm. Dendrites and axons in hcast Tg mice retained their integrities in the course of KA-induced neurodegeneration in a manner much better than those in ntg and VIC Tg mice (Fig. 4).

4 15232 Calpain and Caspase Activation in Excitotoxicity FIG. 3.Protective effect of overexpressed hcast on KA-induced subacute death of hippocampal neurons. A C, cresyl violet staining of the CA1 neurons in ntg (A and B) and hcast Tg (C) mice 24 h after PBS (A) orka(b and C) treatment. The majority of neurons in KA-treated ntg mice exhibited apparent atrophy and pyknosis (B). Although the neurons in hcast Tg mice also became atrophic, shrinkage of soma was less severe than in ntg mice (C). D I, NeuN immunostaining of the CA1 sector (D and E) and dentate gyros (G I) in PBS-treated (D and G) and KA-treated (E and H) ntg and KA-treated hcast Tg (F and I) mice 24 h (D F) and 7 days (G I) after injection. J, quantification of NeuN immunoreactivity in CA1 neurons from ntg, hcast Tg, and VIC Tg mice 24 h (black columns) and 7 days (gray columns) after KA treatment (n 3 in each group). Vertical bars represent S.E. Scale bars, 50 m (A C) and 250 m (D I). **, p 0.01 versus ntg mice by ANOVA/LSD.

5 Calpain and Caspase Activation in Excitotoxicity FIG. 4. Attenuation of KA-induced dendritic and axonal degenerations in hcast Tg mice. A L, immunofluorescence staining for MAP2 (antibody AP12) in the ipsilateral hippocampus (left panels) and for total tau (antibody T49) in the ipsilateral (middle panels) and contralateral (right panels) hippocampi from ntg (A F), hcast Tg (G I), and VIC Tg (J L) mice 24 h (left and middle panels) and 7 days (right panels) after injection of PBS (A C) or KA(D L). M O, quantitative analyses of MAP2 (M) and tau (N and O) immunoreactivities in the ipsilateral (M and N) and contralateral (O) hippocampal formations of ntg, hcast Tg, and VIC Tg mice 24 h (black columns) and 7 days (gray columns) after KA injection (n 5 in each group). Data were normalized against the mean values given by PBS-treated ntg mice. The vertical bars represent S.E. Scale bar, 700 m (A L). *, p 0.05; **, p 0.01 versus KA-treated ntg mice by ANOVA/LSD.

6 15234 Calpain and Caspase Activation in Excitotoxicity FIG. 5. Dual impacts of calpain activation on biochemical properties of tau. A, Western blot analyses for the 136-kDa fragment, MAP2a/b (AP20), total tau (T49), phosphorylated tau (AT8), and Cdk5 activator p25/p35 using protein samples extracted from ipsilateral (i) and contralateral (c) hippocampal formations of ntg and KA-treated hcast Tg mice 24 h after treatment with PBS or KA. B and C, ratio of AT8 to T49 (B) and ratio of p25 to p35 (C) in the ipsilateral (closed columns) and contralateral (open columns) hippocampi based on densitometric quantification of immunoblot signals (n 3 in each group). The ratios indicate the relative incidence of phosphorylation and limited proteolysis of tau and p35 proteins per molecule at this time point, respectively. The quantification was performed within a linear range. Vertical bars represent S.E. Data for tau phosphorylation are normalized against the mean value given by the PBS-treated mice. D and E, AT8 immunofluorescence staining in the CA1 neurons of ntg mice 24 h after PBS (D)orKA(E) treatment. KA-induced calpain activation not only reduced the total amount of tau but also promoted phosphorylation of remaining tau. Scale bar, 50 m (D and E). *, p 0.05; **, p 0.01 versus PBS-treated ntg mice by ANOVA/LSD. Although the dendritic marker, MAP2, showed a pronounced reduction in the entire hippocampal formation, except for the dentate gyrus, of the ntg and VIC Tg mice 24 h after the KA injection, the reduction of MAP2 immunoreactivity in the hcast Tg mouse hippocampus was significantly smaller (left panels in Fig. 4). The retention of the MAP2 immunoreactivity in the hcast Tg mice was also observed 7 days after the treatment. Although the tau immunoreactivity in the hippocampal axons on the injection side decreased to a lesser extent than the MAP2 signals (middle panels in Fig. 4), the entire hippocampal formation of the ntg and VIC Tg mice exhibited a remarkable and progressive decrease in the tau immunoreactivity after the KA treatment. In contrast to these two groups of mice, no marked reduction of the tau immunostaining was found in the hcast Tg mice. None of the mice showed marked alterations in the tau immunoreactivity in the contralateral hippocampus 24 h after the KA administration, whereas the intensity of the tau staining was significantly attenuated 7 days after the treatment (right panels in Fig. 4). Taking these observations together, suppression of the KAinduced calpain activation is likely to have rescued neurons of the entire hippocampal formation from both acute and subacute cytoskeletal degeneration of the dendrites and axons. However, subacute loss of axonal tau in the contralateral hippocampus, presumably caused by Wallerian degeneration, was not overtly inhibited by attenuation of calpain activity. We next quantified the levels of the cytoskeletal components 24 h after the KA injection by Western blotting (Fig. 5A). The 136-kDa fragment immunoreactivity was markedly elevated in the ipsilateral hippocampal samples from the ntg and hcast Tg mice, and the hcast Tg mice showed 32% reduction in the 136-kDa fragment signal intensity as compared with the ntg mice, consistent with the immunohistochemical observations. The appearance of the 136-kDa fragment signal in the contralateral hippocampus was also significantly lower in the hcast Tg mice than in the ntg mice. In accordance with the immunohistochemical data, the reduction of the MAP2 signal in the ipsilateral hippocampus was suppressed in the hcast Tg mice. The T49 level, representing total tau, in the ipsilateral hippocampus also remained preserved in the hcast Tg mice in comparison with the ntg mice. The levels of phosphorylated tau proteins assessed by AT8 immunostaining were elevated in both of the ntg and hcast Tg mice, and the ratio of phosphorylated tau to total tau (AT8/T49) also showed a marked increase in the ntg mice (Fig. 5B). Consistently, the increase in the AT8/T49 ratio was 30% smaller in the hcast Tg than in the ntg mice. We then investigated proteolytic cleavage of p35, the modulator of cyclin-dependent kinase 5 (Cdk5), to p25, since Cdk5 is one of the kinases capable of phosphorylating tau (37, 38). The p25 signal was increased in the KA-treated mice (Fig. 5A), and the ratio of p25 to p35 was significantly elevated in the ipsilateral hippocampus (Fig. 5C). Notably, the p25/p35 ratio of the ntg mice was 2-fold greater than that in the hcast Tg mice. Hence, the conversion of p35 to p25 may be promoted by the KA-induced calpain activation, giving rise to the enhanced tau

7 phosphorylation by activated Cdk5. Consistently, cytoplasmic AT8 immunoreactivity, representing tau phosphorylation, increased in the CA1 and CA3 neurons of the KA-treated mice (Fig. 5, D and E), implying that the phosphorylated tau remained unbound to MTs and may consequently accumulate in the neuronal soma. Double immunolabeling also revealed apparent dendritic colocalization of calpain activity with MAP2 in contrast to its limited colocalization with tau only in a small subset of axons (Fig. 6), providing a plausible explanation for the observation that progression of axonal cytoskeletal disorganization was much slower than that of dendritic disruption (see Fig. 7). Furthermore, the ntg mice exhibited a remarkable KA-induced reduction of endogenous mcast in the CA1 region, where a drastic elevation of the calpain activity and substantial loss of MAP2 took place (Fig. 6). In the hcast Tg mice, calpain activation as well as dendritic pathology was largely suppressed in the areas where hcast was abundantly expressed. Hence, the balance between calpain and CAST activities seems to primarily regulate cytoskeletal derangements during excitotoxic neuronal insults. Calpain and Caspase Activation in Excitotoxicity DISCUSSION The central nervous system is highly vulnerable to both nonapoptotic and apoptotic insults mediated by calpains and caspases, because various protein components of neuronal processes and synapses are substrates upon which these proteolytic enzymes act (39). Disruption of calcium homeostasis has been implicated in neuronal injuries, and thus roles of calpains, strictly regulated by intracellular calcium concentrations, have attracted particular research interest. In fact, vast arrays of experiments have revealed participation of calpain activation in apoptotic (1, 40, 41) as well as nonapoptotic (41 44) neuronal death. However, precise molecular mechanisms, primarily governed by calpain activation, in neurodegenerative processes remain elusive. In this study, the construction of an experimental paradigm using genetically engineered mice and proteolytic product-specific antibodies to monitor calpain and caspase activation enabled us to obtain direct evidence for distinct mechanistic roles of these enzymes in excitotoxicity-induced neuronal injury in vivo. It is noteworthy that calpain-mediated cytoskeletal disruption in neurites led to subacute cell death. In addition, stable overexpression of hcast and VIC in mice without emergence of unfavorable phenotypes indicates the potential usefulness of the hcast and VIC Tg mice and the endo-specific antibodies for unraveling involvement of calpain and caspase systems in a broad range of pathological circumstances. In our observations, the cytoskeletal disorganization was most prominent in the dendrites, where calpain was activated most extensively. Furthermore, this dendritic degeneration was markedly suppressed by the overexpressed hcast, which was also abundantly present in the dendrites, indicating a direct contribution of the calpain-cast system to the disruption of the dendritic structural components. The decrease of MAP2 immunoreactivity exceeded neuronal loss assessed by NeuN staining in the ntg hippocampus at 24 h (96 versus 72% of total CA1 neurons). This finding indicates that the loss of dendritic cytoskeleton did not arise simply as a consequence of the neuronal loss but rather preceded apoptotic events, thereafter exerting profound effects on the survival of the cells in a subacute stage. There also was a significant diminution of the axonal damage by the hcast overexpression, suggesting that the calpain-cast system also functions as a mediator of the axonal derangement. Most interestingly, excitotoxic treatment resulted in both an increase in axonal tau phosphorylation and a decrease in total tau quantity (Fig. 5). Conceivably, both of these changes accelerated disorganization of the axonal MTs FIG. 6.Inverse correlation between KA-induced calpain activation and hcast overexpression. A C, immunofluorescence staining showing the 136-kDa fragment (136-kf) (A), MAP2 (B), and merged (C) images in the hippocampus of ntg mouse 24 h after KA treatment. Calpain activation, indicated by the specific limited proteolysis of -spectrin, was spatially associated with severe loss of MAP2 immunoreactivity. D F, confocal images of the 136-kDa fragment/map2 (D and E) and 136-kDa fragment/tau (F) double immunofluorescence observations in ntg mice 4 (D and F) and 24 h (E) after the KA challenge. At 4 h, the 136-kDa fragment signals (red) colocalized well with MAP2 (green in D) but only partially with tau (green in F). Prominent loss of MAP2 immunostaining was observed 24 h after KA treatment (green in E), whereas the 136-kDa fragment staining remained intense (red in E). G L, confocal photomicrographs of triple immunolabeling for the 136- kda fragment (G and J), MAP2 (H and K), and either of mcast (I) or hcast (L) in the CA1 neurons of ntg (G I) and hcast Tg (J L) mice 24 h after KA injection. Prominent loss of MAP2 signals in ntg mice (H) was closely associated with calpain activation (G) and mcast reduction (I) as compared with the PBS-treated ntg control (inset in I). In contrast, hcast Tg mice exhibited retained hcast signals (L) closely associated with suppressed calpain activation (J) and preserved MAP2 immunoreactivity (K). Scale bars, 300 m (A C) and 50 m (D L). because tau phosphorylation lowers its affinity to MTs and thus destabilizes the MT assembly and dynamics (4). This phosphorylation is likely to be facilitated by an elevated level of Cdk5 activator p25 that is produced by calpain-mediated cleavage of p35 (37, 38). Mobilization of the calpain-p35/p25-cdk5 cascade suggested here was implicated previously in the phosphorylation of tau in excitotoxic neuronal injury (45). As our

8 15236 Calpain and Caspase Activation in Excitotoxicity FIG. 7.A proposed scheme for calpain-mediated neurodegenerative processes focusing on cytoskeletal disorganization. Excitotoxic insults induce calpain and caspase activation, whereas CAST and VIC can suppress activation of these proteases in a specific manner. Reduction of CAST may arise as a consequence of the calpain activation, driving a vicious cycle that further deteriorates the balance of the calpain-calpastatin system. Calpain is activated primarily in the somato-dendritic compartments of neurons and consequently degrades dendritic MAP2 and cytoplasmic MT-unbound tau in a preferential manner. Tau in the neuronal soma also undergoes hyper-phosphorylation catalyzed by Cdk5, which can be activated by a Cdk5 activator fragment (p25) converted from its intact form (p35) as a consequence of calpain-mediated proteolysis. Dendritic cytoskeletons become deranged as well in these locations by the calpain activation, giving rise to pronounced degeneration of dendrites in an acute phase. In contrast, calpain activation is smaller and slower in the axon, and consequently axonal cytoskeletons degenerate more slowly than those in the dendrites, resulting from degradation and hyper-phosphorylation of tau in the soma and subsequent depletion of axonal tau supplied from the soma. Finally, cytoskeletal disorganization in the neuritic processes substantially deteriorates neuronal functions, leading to neuronal death in a subacute phase. observations indicate that the somato-dendritic areas show the highest level of calpain activation relative to other cellular compartments, proteolytic degradation and hyper-phosphorylation of tau may take place primarily in the soma, resulting in depletion of axonal tau that should be constantly supplied from the soma. This can explain why the axonal cytoskeletal disruption progressed more slowly than the dendritic degeneration (Fig. 4). Incidentally, administration of KA at the dosage employed in the present study (1 nmol) into CAST-knock-out mouse brains (46) resulted in acute death of the mice, whereas the wild-type (ntg) and Tg mice remained alive, allowing the subsequent histochemical and biochemical analyses of the brain after the treatment. In the present study, use of this relatively high dose of KA was necessary to induce cytoskeletal disorganization particularly in axons. We therefore did not employ CASTknock-out mice here. The present experimental protocol may in part mimic the chronic axonal degeneration that arises in a number of neurological diseases accompanying tauopathy because we were able to observe delayed neurodegeneration and tau hyperphosphorylation (Figs. 3 and 4). The cytoskeletal disruption arose within 24 h after the KA treatment and progressed during the next 6 days in the ntg mice, whereas the hcast Tg mice showed relatively well sustained dendritic and axonal integrity. Additionally, we observed inexorable subacute loss of the CA1 neurons from 24 h to 7 days after the treatment in the ntg and VIC Tg mice but not in the hcast Tg mice. Taken together with the observation that there was no difference in the number of TUNEL-positive neurons among the ntg, hcast Tg, and VIC Tg mice after the KA challenge, these results demonstrate that subacute neuronal death involves calpain-mediated cytoskeletal disorganization independently of DNA fragmentation. The caspase activation in the soma of CA1 neurons appeared to arise partly as a consequence of the calpain activation because the clact32 immunoreactivity partially colocalizing with the 136-kDa fragment signals was reduced by the hcast overexpression. This is indicative of the cross-talk(s) between calpain and caspase systems in the Ca 2 -induced cellular injury as demonstrated recently by using primary neurons (45, 47). Notably, the suppression of the caspase activation in the VIC Tg mice did not rescue the CA1 neurons from subacute death; the pathological significance of the caspase activation in this region remains uncertain. Moreover, the cytoskeletal components in the VIC Tg mice were disrupted in a manner similar to those in the ntg mice, indicating that caspases do not play major roles in the disorganization of cytoskeletal networks following excitotoxic insults despite the previous knowledge that the dendritic and axonal constituents such as tau have been considered as preferred substrates for caspases (48, 49). In the experimental paradigm employed here, the roles of calpain and caspase systems in the machinery of apoptotic neuronal death remain unsettled. The TUNEL-stained cells seen 24 h after the KA challenge accounted for 45% of the total neurons in the hippocampus (Fig. 3). Therefore, the acute insult, induced by a relatively high dose of KA, was presumably too extensive to be modulated by suppression of calpain or caspase activity. We may also need to consider the potential physiological functions of calpain under excitotoxic conditions; calpain may play a protective role in the early phase of the excitotoxicity-induced neuronal damage at least in vitro (50, 51). Inhibition of calpain activation by the overexpressing CAST may thus not necessarily behave protectively in acute neuronal injury. This possibility should be further addressed by more intensive investigations using in vivo paradigms. The present findings also indicate that the balance between the levels of calpain and CAST activities determines the fate of neurons following excitotoxic challenges. If the amount of CAST is not sufficient to suppress the KA-induced calpain activity, CAST may undergo proteolysis as a consequence of calpain activation, driving a vicious cycle to promote the imbalance between calpain and CAST activities (Fig. 6). Although

9 Calpain and Caspase Activation in Excitotoxicity CAST is a possible substrate for calpain (52), caspases may also proteolyze CAST during apoptosis (53, 54). However, involvement of the caspase system appears to be negligible in the adult brain due to the apparent absence of caspase-3 (46). To conclude, the use of Tg mice overexpressing hcast and VIC has clarified the preeminent role played by the calpaincalpastatin system in the neuron-specific molecular mechanisms underlying excitotoxicity-induced cytoskeletal degeneration (schematically summarized in Fig. 7). The modulation of the balance between calpain and CAST activities demonstrated in the present study also offers a target for potential therapeutic interventions to treat neuritic pathologies. Exploitation of pharmacological means to solely regulate calpain activity through, for instance, determining three-dimensional structures of the calcium-bound forms of calpains may lead to establishment of an ideal medicinal approach to reverse or decelerate neuropathologies in diverse central nervous system disorders. Acknowledgments We thank Misaki Sekiguchi, Kaori Watanabe and Yukiko Dohzono for technical assistance. We also thank Mark Mayford, the Scripps Research Institute, for kindly providing the plasmids pnn265 and pmm403. We are grateful to John Q. Trojanowski and Virginia M.-Y. Lee, University of Pennsylvania School of Medicine, for providing anti-tau antibody T49 and antibodies against neurofilaments. REFERENCES 1. Rami, A. (2003) Neurobiol. Dis. 13, Chan, S. L., and Mattson, M. P. (1999) J. Neurosci. Res. 58, Coleman, M. P., and Perry, V. H. (2002) Trends Neurosci. 25, Higuchi, M., Lee, V. M.-Y., and Trojanowski, J. Q. (2002) Neuromolecular Med. 2, Melloni, E., and Pontremoli, S. (1989) Trends Neurosci. 12, Goll, D. E., Thompson, V. F., Li, H., Wei, W., and Cong, J. (2003) Physiol. Rev. 83, Johnson, G. V., Jope, R. S., and Binder, L. I. (1989) Biochem. Biophys. Res. Commun. 163, Johnson, G. V., Litersky, J. M., and Jope, R. S. (1991) J. Neurochem. 56, Fischer, I., Romano-Clarke, G., and Grynspan, F. (1991) Neurochem. Res. 16, Banik, N. L., Matzelle, D. C., Gantt-Wilford, G., Osborne, A., and Hogan, E. L. (1997) Brain Res. 752, Potter, D. A., Tirnauer, J. S., Janssen, R., Croall, D. E., Hughes, C. N., Fiacco, K. A., Mier, J. W., Maki, M., and Herman, I. M. (1998) J. Cell Biol. 141, Saito, K., Elce, J. S., Hamos, J. E., and Nixon, R. A. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Nixon, R. A., Saito, K. I., Grynspan, F., Griffin, W. R., Katayama, S., Honda, T., Mohan, P. S., Shea, T. B., and Beermann, M. (1994) Ann. N. Y. Acad. Sci. 747, Crocker, S. J., Smith, P. D., Jackson-Lewis, V., Lamba, W. R., Hayley, S. P., Grimm, E., Callaghan, S. M., Slack, R. S., Melloni, E., Przedborski, S., Robertson, G. S., Anisman, H., Merali, Z., and Park, D. S. (2003) J. Neurosci. 23, McIntosh, T. K., Saatman, K. E., Raghupathi, R., Graham, D. I., Smith, D. H., Lee, V. M.-Y., and Trojanowski, J. Q. (1998) Neuropathol. Appl. Neurobiol. 24, Neumar, R. W., Meng, F. H., Mills, A. M., Xu, Y. A., Zhang, C., Welsh, F. A., and Siman, R. (2001) Exp. Neurol. 170, Arthur, J. S., Elce, J. S., Hegadorn, C., Williams, K., and Greer, P. A. (2000) Mol. Cell. Biol. 20, Zimmerman, U. J., Boring, L., Pak, J. H., Mukerjee, N., and Wang, K. K. (2000) IUBMB Life 50, Wang, K. K. W., and Yuen, P. W. (1999) in CALPAIN: Pharmacology and Toxicology of Calcium-dependent Protease (Wang, K. K. W., and Yuen, P. W., eds) pp , Taylor & Francis, Philadelphia 20. Lubisch, W., Hofmann, H. P., Treiber, H. J., and Möller, A. (2000) Bioorg. Med. Chem. Lett. 10, Emori, Y., Kawasaki, H., Imajoh, S., Imahori, K., and Suzuki, K. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, Wang, K. K., and Yuen, P. W. (1994) Trends Pharmacol. Sci. 15, Tomioka, M., Shirotani, K., Iwata, N., Lee, H. J., Yang, F., Cole, G. M., Seyama, Y., and Saido, T. C. (2002) Brain Res. Mol. Brain Res. 108, Sahyoun, N., LeVine, H., III, Burgess, S. K., Blanchard, S., Chang, K. J., and Cuatrecasas, P. (1985) Biochem. Biophys. Res. Commun. 132, Asada, K., Ishino, Y., Shimada, M., Shimojo, T., Endo, M., Kimizuka, F., Kato, I., Maki, M., Hatanaka, M., and Murachi, T. (1989) J. Enzyme Inhib. 3, Hitomi, K., Yokoyama, A., and Maki, M. (1998) Biosci. Biotechnol. Biochem. 62, Mayford, M., Wang, J., Kandel, E. R., and O Dell, T. J. (1995) Cell 81, Wang, G., Achim, C. L., Hamilton, R. L., Wiley, C. A., and Soontornniyomkij, V. (1999) Methods 18, Saido, T. C., Yokota, M., Nagao, S., Yamaura, I., Tani, E., Tsuchiya, T., Suzuki, K., and Kawashima, S. (1993) J. Biol. Chem. 268, Yang, F., Sun, X., Beech, W., Teter, B., Wu, S., Sigel, J., Vinters, H. V., Frautschy, S. A., and Cole, G. M. (1998) Am. J. Pathol. 152, Manya, H., Inomata, M., Fujimori, T., Dohmae, N., Sato, Y., Takio, K., Nabeshima, Y., and Endo, T. (2002) J. Biol. Chem. 277, Mawal-Dewan, M., Henley, J., Van de Voorde, A., Trojanowski, J. Q., and Lee, V. M.-Y. (1994) J. Biol. Chem. 269, Goedert, M., Jakes, R., Crowther, R. A., Six, J., Lüke, U., Vandermeeren, M., Cras, P., Trojanowski, J. Q., and Lee, V. M.-Y. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Carden, M. J., Trojanowski, J. Q., Schlaepfer, W. W., and Lee, V. M.-Y. (1987) J. Neurosci. 7, Tu, P. H., Elder, G., Lazzarini, R. A., Nelson, D., Trojanowski, J. Q., and Lee, V. M.-Y. (1995) J. Cell Biol. 129, Nagao, S., Saido, T. C., Akita, Y., Tsuchiya, T., Suzuki, K., and Kawashima, S. (1994) J. Biochem. (Tokyo) 115, Kusakawa, G., Saito, T., Onuki, R., Ishiguro, K., Kishimoto, T., and Hisanaga, S. (2000) J. Biol. Chem. 275, Lee, M. S., Kwon, Y. T., Li, M., Peng, J., Friedlander, R. M., and Tsai, L. H. (2000) Nature 405, Saido, T. C., Sorimachi, H., and Suzuki, K. (1994) FASEB J. 8, Mattson, M. P. (2000) Nat. Rev. Mol. Cell. Biol. 1, Yamashima, T. (2000) Prog. Neurobiol. 62, Nath, R., Raser, K. J., Stafford, D., Hajimohammadreza, I., Posner, A., Allen, H., Talanian, R. V., Yuen, P., Gilbertsen, R. B., and Wang, K. K. (1996) Biochem. J. 319, Syntichaki, P., Xu, K., Driscoll, M., and Tavernarakis, N. (2002) Nature 419, Pang, Z., Bondada, V., Sengoku, T., Siman, R., and Geddes, J. W. (2003) J. Neuropathol. Exp. Neurol. 62, Nath, R., Davis, M., Probert, A. W., Kupina, N. C., Ren, X., Schielke, G. P., and Wang, K. K. (2000) Biochem. Biophys. Res. Commun. 274, Takano, J., Tomioka, M., Tsubuki, S., Higuchi, M., Iwata, M., Itohara, S., Maki, M., and Saido, T. C. (February 2, 2005) J. Biol. Chem /jbc.M Waterhouse, N. J., Finucane, D. M., Green, D. R., Elce, J. S., Kumar, S., Alnemri, E. S., Litwack, G., Khanna, K., Lavin, M. F., and Watters, D. J. (1998) Cell Death Differ. 5, Nakagawa, T., and Yuan, J. (2000) J. Cell Biol. 150, Canu, N., Dus, L., Barbato, C., Ciotti, M. T., Brancolini, C., Rinaldi, A. M., Novak, M., Cattaneo, A., Bradbury, A., and Calissano, P. (1998) J. Neurosci. 18, Gamblin, T. C., Chen, F., Zambrano, A., Abraha, A., Lagalwar, S., Guillozet, A. L., Lu, M., Fu, Y., Garcia-Sierra, F., LaPointe, N., Miller, R., Berry, R. W., Binder, L. I., and Cryns, V. L. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, Rami, A., and Krieglstein, J. (1993) Brain Res. 609, Lankiewicz, S., Luetjens, M. C., Truc Bui, N., Krohn, A. J., Poppe, M., Cole, G. M., Saido, T. C., and Prehn, J. H. (2000) J. Biol. Chem. 275, Wang, K. K., Posmantur, R., Nadimpalli, R., Nath, R., Mohan, P., Nixon, R. A., Talanian, R. V., Keegan, M., Herzog, L., and Allen, H. (1998) Arch. Biochem. Biophys. 356, Porn-Ares, M. I., Samali, A., and Orrenius, S. (1998) Cell Death Differ. 5,

10 Distinct Mechanistic Roles of Calpain and Caspase Activation in Neurodegeneration as Revealed in Mice Overexpressing Their Specific Inhibitors Makoto Higuchi, Masanori Tomioka, Jiro Takano, Keiro Shirotani, Nobuhisa Iwata, Hajime Masumoto, Masatoshi Maki, Shigeyoshi Itohara and Takaomi C. Saido J. Biol. Chem. 2005, 280: doi: /jbc.M originally published online February 7, 2005 Access the most updated version of this article at doi: /jbc.M Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts This article cites 53 references, 18 of which can be accessed free at