Calcium-dependent Cleavage of Endogenous Wild-type Huntingtin in Primary Cortical Neurons*

Huntington's disease (HD) is caused by a polyglutamine expansion in the amino-terminal region of huntingtin. Mutant huntingtin is proteolytically cleaved by caspases, generating amino-terminal aggregates that are toxic for cells. The addition of calpains to total brain homogenates also leads to cleavage of wild-type huntingtin, indicating that proteolysis of mutant and wild-type huntingtin may play a role in HD. Here we report that endogenous wild-type huntingtin is promptly cleaved by calpains in primary neurons. Exposure of primary neurons to glutamate or 3-nitropropionic acid increases intracellular calcium concentration, leading to loss of intact full-length wild-type huntingtin. This cleavage could be prevented by calcium chelators and calpain inhibitors. Degradation of wild-type huntingtin by calcium-dependent proteases thus occurs in HD neurons, leading to loss of wild-type huntingtin neuroprotective activity.

Huntingtin is a 348-kDa cytoplasmic protein that is important for cell survival (1)(2)(3)(4)(5). Interest in this protein stems from the fact that mutation into the encoding gene causes HD, a slowly progressing neurodegenerative disease characterized by the selective death of the striatal and cortical neurons (6).
In the pathology, the amino-terminal portion of huntingtin is characterized by an expanded polyglutamine stretch conferring a newly acquired toxic function to the protein (7). Several observations implicate mutant huntingtin proteolysis in the pathogenesis of HD 1 because amino-terminal huntingtin fragments aggregate into the nucleus and cytoplasm of human neurons and overexpression of these fragments in vitro causes cell death (8).
Despite the genetic and experimental evidence indicating a gain of function mechanism in HD, loss of normal huntingtin function may also be important (9). Normal huntingtin is antiapoptotic (1,5) and exerts a peculiar function for striatal neurons, given that it sustains the production of cortically derived BDNF, which they depend on for their activity (10). Proteolysis of wild-type huntingtin may cause loss of these physiological function(s). However, thus far, most of the research on the role of proteolysis in HD has been directed toward analysis of mutant huntingtin. Mutant huntingtin is cleaved by caspases at residues 513 and 552. Mutation at these sites impairs caspase-3 cleavage and shows reduced toxicity after transfection in neuronal cells (11). Furthermore, transfection of cortical neurons with mutant huntingtin in the presence of the broad spectrum caspase inhibitor z-VAD-fmk blocks cleavage of exogenous huntingtin and reduces cell toxicity (12). Moreover, in HD-transgenic mice, caspase inhibition results in delayed onset of symptoms, slowed progression, and prolonged survival (13,14).
Other proteases different from caspases may also be involved. Indeed, analyses of huntingtin proteolysis in HD postmortem brains have revealed fragments in the cortex and striatum not justified by caspase cleavage alone (15,16). More recently, the possibility that wild-type huntingtin may be cleaved by calpains has been brought to our attention by the work of Kim et al. (17). Calpains are calcium-dependent noncaspase cysteine proteases that are enriched in neuronal cells. m-calpain and -calpain, the two main isoforms of calpains in the brain (18), are activated in dendrites following excitotoxic and ischemic insults in cell culture and in vivo (19). It was found that adding m-calpain to lysates from total mouse brain causes cleavage of endogenous wild-type huntingtin and that a calpain inhibitor is able to block the cleavage of transfected wild-type huntingtin (17).
Here we aimed at evaluating whether endogenous wild-type huntingtin found in primary cortical neurons could be cleaved by stimuli that increase intracellular calcium levels, leading to calpain activation. We show that calcium influx, induced by a calcium-selective ionophore as well as by glutamate or 3-nitropropionic acid (3-NP) exposure, leads to cleavage of wild-type huntingtin in cortical neurons. Wild-type huntingtin proteolysis, without activating caspase-3, produces specific huntingtin fragments of about 75 and 60 kDa and is inhibited by calcium chelators or calpeptin. These data indicate that stimuli that trigger calcium-dependent calpain activation lead to proteolysis of wild-type huntingtin in primary brain neurons.
Sigma. z-VAD-fmk was from Promega. Antibody mAb 2166 was from Chemicon; anti-tubuline monoclonal antibody and anti-caspase-3 monoclonal antibody were from Santa Cruz Biotechnology.
Cell Culture-Primary neuronal cultures from cerebral cortex were obtained from E16-E18 Sprague-Dawley rats according to standard procedures (20). Pregnant animals were killed by cervical dislocation under CO 2 anesthesia, and the fetuses were removed and put into ice-cold Hanks' balanced salt solution. After dissection of the cortices, cells were dissociated, plated on poly-L-ornithine-treated 35-mm plates in DMEM plus 10% fetal calf serum and incubated at 37°C, 5% CO 2 . The following day the medium was replaced with differentiating medium (three-fourths: neurobasal medium from Euroclone, B27 and N2 from Invitrogen, Pen-Strep, and glutamine; one-fourth: DMEM F12, B27, and Pen-Strep).
18 days after plating, neurons were exposed to various treatments. In experiments where calpeptin, BAPTA-AM, and z-VAD-fmk were used, the drugs were added to the cells in Krebs solution (NaCl 118.5 mM, KCl 4.8 mM, CaCl 2 2.5 mM, KH 2 PO 4 1.2 mM, MgSO 4 1.2 mM, NaHCO 3 25 mM, Glucose 11 mM) 30 min before the addition of ionophore, 3-NP, or glutamic acid. Control cultures were maintained in Krebs solution.
For measurement of intracellular free calcium [Ca 2ϩ ] i changes after 3-NP, cortical neurons were plated on poly-L-ornithine-coated 12-mm glass coverslips and then incubated in 0.5 ml of differentiating medium. Analyses were conducted 18 days after plating.
Western Blot Analysis-Cells were washed with phosphate-buffered saline and lysed for 60 min on ice in a buffer containing Tris HCl 10 mM, NaCl 100 mM, Triton X-100 1%, EGTA 5 mM, EDTA 5 mM, and protease inhibitors. Samples were then centrifuged for 15 min at 14,000 rpm, 4°C, supernatants were recovered, and protein concentration was evaluated by the Bradford reagent assay (Bio-Rad).
Equal amounts of protein were separated by 7.5% SDS-PAGE. The blotted protein were exposed to mAb 2166 antibody (dilution 1:5000) or anti-caspase-3 (dilution 1:500). Bands were detected using a horseradish peroxidase secondary antibody and an enhanced chemiluminescence system (ECL, Amersham Biosciences).
Densitometric Analysis-The bands on the autoradiography corresponding to full-length huntingtin and ␤-tubulin were scanned and quantified using NIH Image software. The ratio between full-length huntingtin and tubulin in the untreated sample is considered 100% of basal huntingtin.
m-Calpain Biochemical Assay-Neuronal cells grown on a 100-mm plate were washed with phosphate-buffered saline, resuspended in 400 l of ice-cold hypotonic lysis buffer (Tris HCl 50 mM, KCl 10 mM, pH 7.4, ␤-mercaptoethanol 2%), and lysed for 40 min on ice. After a 30-s sonication, samples were centrifuged for 15 min at 14,000 rpm, 4°C. The supernatant was aliquoted. Specific treatments were applied to each sample. The reactions were conducted for 15 min at 30°C. After the addition of sample buffer, samples were boiled and loaded onto a 7.5% SDS-PAGE.
Intracellular Calcium Measurements-Cultured cells plated on glass coverslips were loaded for 60 min at 37°C with 2 M Fura-2 pentacetoxy methylester in Krebs-Ringer solution buffered with HEPES (150 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 2 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES/NaOH, pH 7.4), washed in the same solution to allow the deesterification of the dye, and transferred to the recording chamber of an inverted microscope (Axiovert 100, Zeiss) equipped with a calcium imaging unit. A modified CAM-230 dual wavelength microfluorimeter (Jasco, Tokyo) was used as a light source for the assays. The images of Fura-2 fluorescence at two excitation wavelengths (340 and 380 nm) were collected with a PCO Super VGA SensiCam (Axon Instruments, Forest City, CA) and analyzed with the Axon Imaging Workbench 2.2 software (Axon). Ratio values in discrete areas of interest were calculated from the sequences of ratio images to obtain temporal analyses. Images were acquired at 1-5 340/380 ratios/s. Experiments were performed in a static bath at room temperature (24 -25°C). Drugs were applied in 2 ml of solution. The various reagents were added by loading the appropriate volumes of 100ϫ concentrated solutions into a syringe connected to the incubation chamber via a small tube. Aspiration into the syringe of 1 ml of extracellular medium followed by the reintroduction of the mixture into the chamber ensured accurate and rapid delivery and mixing.

Proteolysis of Endogenous Wild-type Huntingtin in Primary
Neurons Follows Calcium Influx-We aimed at analyzing whether endogenous huntingtin expressed in cortical neurons undergoes proteolysis induced by calcium entry into the cells.
Eighteen days after plating, rat cortical neurons were exposed to the calcium selective ionophore (A23187) at a final concentration of 5 M. Lysates were collected after 30 min, 1, 2, or 3 h of treatment and processed for Western blotting analyses using mAb 2166 antibody recognizing residues 181-812 of huntingtin. A 348-kDa band corresponding to huntingtin was detected in control cultures (Fig. 1A, lane 0, top panel). Upon exposure of the cultures to the ionophore, wild-type huntingtin levels decreased in a time-dependent manner with an 80% reduction observed at 180 min (Fig. 1A). These data support the hypothesis that an increase in the intracellular calcium level leads to degradation of endogenous wild-type huntingtin in neuronal cells.
Importantly, Fig. 1A also shows increased levels of two mAb 2166-immunoreactive bands of molecular mass of 75 and 60 kDa in coincidence with the time-dependent decrease in intact endogenous huntingtin. These data indicate that normal huntingtin undergoes proteolytic cleavage in cortical neurons exposed to calcium ionophores, originating two amino-terminal mAb 2166-immunoreactive fragments. In the same blot, two additional bands of 68 and 70 kDa are present in all samples, likely representing cross-reactive proteins of unknown origin.
Huntingtin has been previously demonstrated to be cleaved by caspase-3 (21). We therefore investigated whether active caspase-3 could be measured under the experimental conditions indicated above that led to huntingtin proteolysis. We therefore evaluated caspase-3 activity by two means, i.e. by fluorimetric assays using specific fluorogenic substrate (Ac-DEVD-aminomethylcoumarin; not shown) and by Western blot analyses using an anti caspase-3 antibody able to recognize both the inactive and the two active p20 and p11 caspase-3 fragments. As shown (Fig. 1B, left panel), in the absence or presence of the calcium ionophore, a single band corresponding to inactive caspase-3 is detected in all samples. To validate further the specificity of the anti-caspase-3 antibody, cells were exposed to serum-deprived medium, a condition that activates caspase-3 in our cellular system, as shown by the appearance of the two immunoreactive bands corresponding to active caspase-3 (Fig. 1B, right panel). The same bands were not present in a parallel lysate obtained from cells where the caspase-3 inhibitor, z-VAD-fmk, was present.
We conclude that under our experimental conditions caspase-3 activation appears not to be a major requirement for wild-type huntingtin proteolysis. Furthermore, cells exposed to stimuli that evoke caspase-3 activation do not seem to produce fragments similar to those observed in calpain activated cells (not shown).
Proteolysis of Huntingtin in Primary Neurons Is Calpainmediated-Given that the addition of calpain induces proteolytic cleavage of huntingtin in total lysates from rodent brain (17) and that calpeptin, a specific inhibitor of -calpain and m-calpain, prevents proteolyses of transfected huntingtin in immortalized cells (17), we explored the possibility that a similar calpain-dependent cleavage of endogenous huntingtin could occur in primary cortical neurons. We exposed cortical neurons to different doses of calpeptin for 30 min before adding the calcium ionophore. As shown in Fig. 1C, cleavage of endogenous wild-type huntingtin could be prevented by the calpain inhibitor in a dose-dependent manner.
We then tested whether endogenous wild-type huntingtin could be cleaved by calpain in vitro. Since the working temperature of calpain in vitro is 30°C, we first checked whether simple exposure of the lysates to this temperature could evoke degradation of huntingtin. Lanes 1 and 2 in Fig. 2 show that similar amounts of endogenous wild-type huntingtin are present at 4 and 30°C. We next performed the assay at 30°C by incubating equal amounts of proteins with 0.1 unit/ml m-calpain in the presence or absence of CaCl 2 5 mM, as calpain activator. As shown in Fig. 2 Fig. 3 cortical neurons were exposed to the excitatory amino acid glutamate at a dose of 100 (Fig. 3A) or 500 M (Fig. 3B) or to 5 mM 3-NP (Fig. 3C), a mitochondrial toxin. [Ca 2ϩ ] i homeostasis under these conditions was analyzed by loading the cortical neurons with Fura-2 ( Fig. 4). Increased cytoplasmic concentrations of calcium was promptly measured in single neuronal cells after exposure to 100 M glutamate (Fig. 4, A and C). A 60-min latent period was observed after exposure to 3-NP after which [Ca 2ϩ ] i reached levels comparable with those evoked by glutamate. These data indicate that both glutamate and 3-NP were able to induce calcium movements under our experimental conditions. Under these same conditions we analyzed the levels of endogenous wild-type huntingtin by Western blotting. As shown in Fig. 3A, exposure of primary neurons to 100 M glutamate for 24 h caused a 56% decrease in the amount of endogenous intact huntingtin. Given that glutamate increases intracellular calcium levels in neurons (Fig. 4, A and C), we applied 50 M BAPTA-AM, an intracellular calcium chelator, and found that it could prevent the loss of intact wild-type huntingtin observed in the presence of glutamate. In a subsequent experiment, 500 M glutamate was applied for different time periods. As shown in Fig. 3B, the levels of endogenous wild-type huntingtin decrease starting at 2 h after glutamate addition and was maximal at 24 h, when fragments of huntingtin could be detected. Also under this condition, the addition of calpeptin was able to prevent cleavage. Finally and as expected, given the changes in [Ca 2ϩ ] i levels observed with 5 mM 3-NP, (Fig. 4, B and C) we found that exposure of the neuronal cultures to this toxin could also evoke changes in the levels of intact wild-type huntingtin and the appearance of huntingtin fragments, an effect that was prevented by calpeptin (Fig. 3C). Also, under this condition, proteolysis of intact wild-type huntingtin originated mAb 2166immunoreactive bands. The caspase inhibitor z-VAD-fmk was partially able to prevent depletion of endogenous wild-type huntingtin (Fig. 3C). DISCUSSION Here we show that changes in calcium homeostasis evoked by glutamate and 3-NP lead to calpain-dependent proteolysis of endogenous wild-type huntingtin in cortical neurons. Previous calpain inhibitor calpeptin before addition of the ionophore. Samples were then analyzed for the levels of intact wild-type huntingtin and cleavage products. Densitometric analysis was performed by normalizing full length huntingtin signal to ␤-tubulin. Data shown are from one of three different experiments producing the same results. FIG. 1. A, calcium-dependent cleavage of wild-type huntingtin in primary neurons. Cortical neurons were exposed to A23187 calcium ionophore at a concentration of 5 M and then lysed. A time-dependent decrease in endogenous wild-type huntingtin is observed (top panel). Incubation of the same membranes with mAb 2166 antibody gave rise to two immunoreactive bands of about 60 and 75 kDa, which increased in intensity upon exposure to the ionophore (middle panel). Graph, densitometric analysis performed by normalizing huntingtin signal to the ␤-tubulin band. Data shown are from one of three different experiments producing the same results. B, under our experimental conditions, bands corresponding to active caspase-3 were not detected in the blots (left panel). Right blot, positive control, active caspase-3 is detected in ST14A cells exposed to serum-deprived medium. The same fragments are not visible in z-VAD-fmk-pretreated cells. C, cleavage of wild-type huntingtin in primary neurons is calpain-mediated. Cortical neurons were pretreated for 30 min with increasing amounts of the data by Kim et al. (17) indicated that endogenous huntingtin from total brain homogenates is cleaved in vitro by calpains and that cleavage of transfected huntingtin could be prevented by a specific calpain inhibitor. We now report that stimuli that lead to changes in calcium levels in primary neurons cause proteolyses of endogenous wild-type huntingtin in the absence of caspase-3 activation (Fig. 1A). Under our conditions, exposure of neuronal cultures to 3-NP was able to evoke intracellular calcium influx comparable with that observed in glutamate treated neurons (Fig. 4). In both conditions, analyses of endogenous huntingtin revealed a similar decrease in protein levels, an effect that could be prevented by the addition of calpain inhibitors or calcium chelators, respectively.
We found that calpain-mediated cleavage of wild-type huntingtin results in the appearance of amino-terminal fragments of about 75 and 60 kDa. Fragments of a similar molecular mass were identified after in vitro calpain-mediated cleavage of huntingtin from brain homogenates (17).
The evidence of calpain-mediated cleavage of wild-type hun- tingtin in neuronal cells is important for several reasons. On one hand, it supports the hypothesis that an imbalance in calcium homeostasis may lead to huntingtin proteolysis and, on the other hand, it strengthens the idea that loss of full-length wild-type huntingtin protein and of its physiological activity may occur in HD and in other brain diseases characterized by alterations in calcium levels.
Calcium-mediated excitotoxicity and mitochondrial dysfunction have long been hypothesized to play a role in the pathogenesis of HD (22). In R6/2 HD transgenic mice and, to a lesser extent, in YAC72 mice expressing the full-length mutant gene, an increased intracellular [Ca 2ϩ ] i flux evoked by selective activation of NMDA receptors was demonstrated at both the presymptomatic and symptomatic stages. Parallel increases in NMDA-R1 receptors and decreases in NMDA-R2A/B subunit proteins were also reported (23)(24)(25). The latter modification is of particular interest given that calpains are able to cleave the R2A/B subunit of NMDA glutamate receptor (26). It is also worth noting that NMDA-R1 receptors (which are not calpain substrates) are enriched in the striatal neurons that colocalize somatostatin, neuropeptide Y, and NADPH-diaphorase (NADPH-d); these neurons are selectively spared in Huntington's disease and also resistant to experimentally induced excitotoxic cell death (27). Calbindin D28k level, a protein with a high capacity for buffering Ca 2ϩ , was also increased in the distal dendrites of spiny striatal neurons in HD postmortem striatum (28). Altogether, these and other new findings (29) indicate that calcium handling is impaired in Huntington's disease, therefore possibly influencing the activation of calciumdependent proteases.
Other data point to a striatum-specific activation of calpains. Indeed, the ratio of the active versus inactive calpains in different areas of brain from post-mortem HD patients revealed a significant increase in calpain activation in the putamen, whereas no changes were observed in the frontal cortex or cerebellum. On the contrary, in patients with Alzheimer's disease, all brain regions were similarly characterized by an increase in calpain activation with respect to controls (30).
Proteolysis of wild-type huntingtin is expected to impact cell survival and function (9). Indeed, normal huntingtin is antiapoptotic in brain cells (1,2,5,31) and is also found to protect from toxicity induced by the mutant protein (3,4) and to support the production of cortically derived BDNF (10).
Earlier data point to caspase-mediated polyglutamine-dependent cleavage of mutant huntingtin as one of the key events leading to cell toxicity (11). In this scenario, compounds able to inhibit caspase activation have been proven able to delay disease onset in experimental animal models of HD (13). The discovery of important physiological function(s) of wild-type huntingtin in brain cells (9) and the reports by Kim (17) and also in this paper that wild-type huntingtin is a target for calpain-mediated proteolyses in neurons support the idea that drugs able to inhibit calpains in humans may also lead to increased levels of wild-type huntingtin, therefore potentially restoring its neuroprotective activity.