Deposition of β-amyloid peptide (Aβ) and hyperphosphorylation of the τ protein are associated with neuronal dysfunction and cell death in Alzheimer's disease. Although the relationship between these two processes is not yet understood, studies have shown that both in vitro and in vivo exposure of neurons to Aβ leads to τ hyperphosphorylation and neuronal dystrophy. We previously reported that the microtubule-stabilizing drug paclitaxel (Taxol) protects primary neurons against toxicity induced by the Aβ25-35 peptide. The studies in this report were undertaken to characterize the actions of paclitaxel more fully, to assess the effectiveness of structurally diverse microtubulestabilizing agents in protecting neurons, and to determine the time course of the protective effects of the drugs. Primary neurons were exposed to Aβ in the presence or absence of several agents shown to interact with microtubules, and neuronal survival was monitored. Paclitaxel protected neurons against Aβ1-42 toxicity, and paclitaxel-treated cultures exposed to Aβ showed enhanced survival over Aβ-only cultures for several days. Neuronal apoptosis induced by Aβ was blocked by paclitaxel. Other taxanes and three structurally diverse microtubule-stabilizing compounds also significantly increased survival of Aβ-treated cultures. At concentrations below 100 nM, the drugs that protected the neurons did not produce detectable toxicity when added to the cultures alone. Although multiple mechanisms are likely to contribute to the neuronal cell death induced by oligomeric or fibrillar forms of Aβ, low concentrations of drugs that preserve the integrity of the cytoskeletal network may help neurons survive the toxic cascades initiated by these peptides.
The role of Aβ peptides in the pathogenesis of Alzheimer's disease (AD) is supported by the association of genetic mutations in the amyloid precursor protein and enhanced Aβ peptide formation in individuals with mutations in the presenilin genes (Selkoe, 2001). Furthermore, both in vivo and in vitro studies have shown that accumulation of Aβ fibrils can trigger neurodegenerative changes (Busciglio et al., 1995; Mattson, 1997; Geula et al., 1998). The neurotoxicity of Aβ involves dramatic morphological alterations, including formation of dystrophic neurites associated with the collapse of the cytoskeleton (Yankner et al., 1990; Pike et al., 1992; Grace et al., 2002). Cytoskeletal disruptions may result from aggregation of the microtubule (MT)-associated protein τ and loss of its activity in stabilizing the MT network. When τ is hyperphosphorylated, it self-assembles into polymers that form paired helical filaments constituting the neurofibrillary tangles (NFTs) in AD brain (Grundke-Iqbal et al., 1986). Highly phosphorylated τ no longer stabilizes MTs, and intracellular aggregates of the τ protein are believed to cause disturbances in axonal transport of molecules and organelles in affected neurons (Lee, 1995). Recent studies have suggested that deposition of Aβ enhances τ phosphorylation in vitro in neuronal cell cultures (Busciglio et al., 1995; Ferreira et al., 1997; Grace et al., 2002; Li et al., 2003) and in vivo when Aβ is deposited in the brain (Sigurdsson et al., 1997; Geula et al., 1998).
The discovery of human τ gene mutations in familial dementias, the “tauopathies”, provided the first demonstration that abnormalities in τ could lead to neurodegeneration (Lee et al., 2001). Transgenic mice expressing both mutant human τ and amyloid precursor protein developed not only Aβ plaques but also NFT-like lesions and severe neuronal death (Lewis et al., 2001). Furthermore, injecting Aβ into the brains of mutant τ mice led to significant increases in neuronal death above that occurring with mutant τ alone (Gotz, 2001). These observations taken together suggest a potential link between the two neuropathological lesions characteristic of AD.
If Aβ deposition does lead to excessive phosphorylation of τ and a disruption in MT integrity, drugs that can stabilize the MT network might help protect neurons against Aβ toxicity, as hypothesized by Lee et al. (1994). We previously reported that the MT-stabilizing agent paclitaxel (Taxol) significantly enhanced the survival of primary neurons in culture following exposure to Aβ25-35 peptides, reduced activation of the apoptotic protease caspase 3, and blocked Aβ-induced increases in abnormal τ phosphorylation (Michaelis et al., 1998; Li et al., 2003).
Paclitaxel promotes MT stability, and its protective actions may involve preservation of the structure of the MT network in the presence of a destabilizing activity of Aβ. However, emerging literature on the role of MTs in cell signaling suggests that very complex cellular actions may be involved in protective actions of paclitaxel in neurons (Gundersen and Cook, 1999), and our observation that paclitaxel prevented the activation of τ phosphorylation in the presence of Aβ suggests that the presumed interaction of the drug with MTs may have multiple signaling consequences in neurons. Because of the potential complexity underlying the pharmacological action of paclitaxel, it was of interest to determine whether protection against Aβ toxicity was unique to this agent or was shared by other MT-stabilizing compounds, some with quite diverse structures.
The goal of the present study was to characterize more fully the activity of MT-stabilizing agents in protecting neurons against Aβ-induced toxicity. Studies were designed to determine whether paclitaxel protected neurons against toxicity induced by both the Aβ25-35 peptide and the Aβ1-42 peptide found in plaques in AD brain, whether structurally diverse MT-stabilizing agents demonstrated protective effects against Aβ toxicity, and the time course of the effects of Aβ in the presence or absence of MT-stabilizing drugs. Results showed that both Aβ25-35 and Aβ1-42 led to a 40% rate of cell death in 48 h, and pretreatment with paclitaxel significantly reduced the rate of cell death with both peptides. Nanomolar concentrations of diverse MT-stabilizing drugs also blocked Aβ toxicity, and drug-treated cultures exposed to Aβ survived longer than Aβ-only cultures. Pharmacological analyses indicated that the agents typically were partial agonists and led to a maximal survival rate that was ∼90% of that in control cultures never exposed to Aβ. Taxanes were slightly less potent than the other agents, suggesting there may be some selectivity in the targets of the drugs in neurons or that subtle mechanisms not shared equally by the diverse agents are required to slow the toxic cascades initiated by Aβ.
Materials and Methods
Neuronal Cell Cultures. Dissociated cortical cell cultures were established from embryonic day 18 rat fetuses recovered from pregnant Sprague-Dawley rats (Harlan, Indianapolis, IN) as described previously (Michaelis et al., 1994). Pups were delivered by cesarean section while the dam was anesthetized with pentobarbital (140 mg/kg i.p.), and the brains were recovered according to National Institutes of Health-approved protocols. After the final precipitation step, neurons were suspended in fresh Dulbecco's modified Eagle's medium/F-12 (Sigma-Aldrich, St. Louis, MO) with 10% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA) and plated at a density of 2.5 × 105 cells in 35-mm glass bottom dishes (MatTek Co., Ashland, MA), coated with poly-d-lysine. Serum-containing medium was removed after 24 h, and the cells were maintained in serum-free Dulbecco's modified Eagle's medium/F-12 containing the N2 supplements. Cultures were grown at 37°C in 5% CO2 and 97% humidity as described (Michaelis et al., 1994).
Aβ Peptide and Drug Treatments. After 5 days in culture, the primary neurons were exposed to either Aβ25-35 or Aβ1-42 in the presence or absence of the indicated MT-stabilizing agents. The Aβ25-35 was synthesized and purified in the Biochemical Research Services Lab, University of Kansas. The reverse-sequence peptides Aβ35-25 and Aβ1-42 were purchased from Bachem California (Torrance, CA). Prior to adding the peptides to the cultures, the Aβ peptide stocks (1.3 mg/ml in double-distilled H2O) were diluted into 10 mM Tris/Cl, pH 7.4 and maintained for 24 h at 37°C. Each batch of Aβ peptides was analyzed for β-sheet formation by circular dichroism, but no effort was made to separate oligomers from fibrils. The peptides were added directly to the culture medium, usually at 10 μM final concentration.
Paclitaxel and 10-deacetylbaccatin III were obtained from Dabur India, Ltd. (Uttar Pradesh, India), and docetaxel from Aventis (Strasbourg, France). UK-100 (Klar et al., 1998) was a gift from Schering AG, Berlin, Germany, and discodermolide was a gift from Novartis Pharmaceutical Co., East Hanover, NJ. The succinylated paclitaxel analog Tx-67 was prepared by parallel solution phase synthesis as previously described (Liu et al., 2002). A mixture of epothilones A and B (∼1:2) was extracted from the myxobacterium Sorangium cellulosum and separated into A and B as described (Bollag et al., 1995), and epothilone A (Epo A) was used in the neuronal cultures. All MT-stabilizing agents were prepared as 2.5 mM stocks in dimethyl sulfoxide (DMSO) and maintained at -20°C. The final concentrations of the MT-stabilizing agents added to the cultures in the various experiments are indicated in the respective figure legends. Control cultures received the DMSO vehicle alone, and the final concentration of DMSO never exceeded 0.04%. Unless indicated otherwise, assays were carried out 48 h following Aβ peptide addition.
Measurement of Cell Viability. The effects of the Aβ peptides and the MT-stabilizing drugs were primarily determined by monitoring neuronal cell survival using the Live/Dead assay as previously described (Michaelis et al., 1998). Following the indicated periods of exposure to the peptides and/or the drugs, cells were labeled with 20 μM propidium iodide (PI) and 150 nM calcein acetoxy-methylester (Molecular Probes, Eugene, OR) for 30 min at 37°C. After incubation with the dyes, the dishes were rinsed with phosphate-buffered saline (PBS) and placed on the stage of a Nikon inverted microscope (Nikon Eclipse TE200; Nikon, Tokyo, Japan) with filters for fluorescein isothiocyanate and Texas Red. Digital images were captured with a Dage camera (Dage-MTI, Michigan City, IN) and saved in Adobe Photoshop. The number of viable (green) and dead (red) neurons was determined by counting the cells, and this was done in 6 to 12 microscopic fields per culture dish in duplicate dishes for each treatment. All experimental treatments were carried out on at least two separate embryonic neuronal preparations. Thus, approximately 1500 neurons were scored under each treatment condition. To establish validity of the assay criteria, initially two investigators unaware of treatment conditions independently scored ∼20% of the cultures by counting the number of cells labeled with PI or calcein. Greater than 90% agreement on the percentage of viable cells was obtained between two independent observers unaware of the treatment conditions in randomly selected test fields. The fraction of viable cells in each field was calculated based on the total number of cells counted in each field. Raw data from each experiment were combined and the significance of differences between cultures exposed to various treatments was determined using Student's t test. Data in the doseresponse curves for the various MT-stabilizing agents were analyzed by nonlinear, least-squares fitting to a hyperbolic function (r2, 0.95-0.98) to obtain the estimated EC50 and Vmax values for each of the compounds. Neuronal survival in the untreated control samples was considered to represent maximal viability and that in the Aβ-only samples the minimal viability.
Induction of apoptosis during exposure of the neurons to Aβ was monitored using an in situ assay for caspase activation, CaspACE FITC-VAD-FMK (Promega, Madison, WI). Neurons plated on the glass cutout dishes were exposed to the Aβ peptide in the presence or absence of the drugs, rinsed with PBS, and incubated with 20 μM FITC-VAD-FMK for 20 min at 37°C in the dark. Plates were rinsed three times with PBS and the cells viewed immediately in a fluorescence microscope at 450 to 490 nm excitation/515 nm emission. The fluorescent caspase inhibitor is cell-permeable and irreversibly binds to activated caspases, allowing for the detection of the apoptotic cascade in intact cells. A phase-contrast image of each field of neurons was captured using differential interference contrast optics, and the same field was captured under fluorescence as described above. The two images were superimposed, and the percentage of the total cell population that was fluorescent, and thus, undergoing apoptosis, was determined by direct cell counting. Multiple fields were obtained for each treatment condition as described in the Live/Dead assay, and data were expressed as the percentage of apoptotic cells.
c-Jun Terminal Kinase (JNK) Assay. JNK activity was assayed using a GST-c-Jun fusion protein as the substrate. Primary cortical neurons at day 7 in vitro were treated with DMSO only, with paclitaxel, or with Epo A for 24 h. In some cultures the JNK inhibitor indirubin-3′-monoxime (I-3′-M) (Sigma-Aldrich) was added for 30 min, followed by DMSO vehicle or the JNK activator sodium arsenite for 4 h. Cells were washed twice with PBS, exposed to a lysis buffer [20 mM Tris-HCl, pH 7.4, 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 10 mM NaF, 0.1% (v/v) Nonidet P40, 1 mM EDTA, and 1 μl of Protease Inhibitor Cocktail Set III (Calbiochem, San Diego, CA) per 10 ml of buffer], and centrifuged at 10,000 rpm for 10 min. Protein concentrations of the supernatants were determined by the bicinchoninic acid assay (Pierce, Rockford, IL). Aliquots of the supernatant (200 μg of protein) were incubated with 5 μg of GST-c-Jun (1-79) fusion protein (BioVision, Inc., Mountain View, CA) for 1 h at 4°C. Glutathione-conjugated agarose beads (40 μl of a 50% slurry in PBS, 0.1% Triton X-100) were added and mixed at 4°C for 30 min. The mixture was centrifuged for 1 min at 10,000 rpm to pellet beads with GST-c-Jun-JNK bound. The beads were washed three times in kinase activity buffer (10 mM MgCl2, 50 mM Tris-HCl, 1 mM EGTA, and 50 μM ATP) and resuspended in 40 μl of kinase activity buffer containing 50 μM [γ-32P]ATP. Following incubation for 30 min at 37°C, the reaction was stopped by the addition of 10 μl of 5× electrophoresis sample buffer [(250 mM Tris-HCl, pH 6.8, 33% (v/v) glycerol, 13% (w/v) SDS, 2.5% (w/v), bromphenol blue, 250 mM dithiothreitol, and 650 mM iodoacetamide)]. Samples were boiled for 5 min and resolved by SDS-polyacrylamide gel electrophoresis on 12% gels. Images were obtained using a Bio-Rad phosphorimaging system (Hercules, CA), the data analyzed by densitometry, and results plotted as the percentage of the DMSO control samples.
Effects of Paclitaxel on Aβ1-42-Induced Toxicity. In our initial report on the protective action of paclitaxel in Aβ-treated neurons, we used the short toxic fragment Aβ25-35 to induce cell death. Neurons were also exposed to similar concentrations of the reverse-sequence peptide Aβ35-25 as a control, but no increase in cell death occurred with this peptide. Although the cascade of events leading to neuronal cell death with the Aβ25-35 has many similarities to that initiated by the Aβ1-42 peptide that accumulates in AD brain, we conducted experiments to determine whether paclitaxel also enhanced neuronal survival when this larger peptide was used as the toxic stimulus. Exposure to 5 μM Aβ1-42 for 48 h reduced the percentage of viable neurons from 77 to 46% (Fig. 1). When 100 nM paclitaxel was added to the cultures 2 h before the Aβ1-42, the neuronal survival was increased to 66%. Exposure to 10 μM Aβ1-42 alone reduced the survival rate to 39%, and pretreatment with paclitaxel resulted in 70% survival (Fig. 1). Very similar results were observed when Aβ25-35 was tested in the presence or absence of 100 nM paclitaxel (Fig. 1). Thus, even though some signaling events initiated by the Aβ1-42 may differ somewhat from those initiated by Aβ25-35, the presence of the MT-stabilizing agent appeared to afford a substantial level of protection, regardless of the Aβ peptide used to induce toxicity. As we reported previously, the presence of paclitaxel alone at a concentration of 100 nM led to no significant decrease in the viability of the neurons (Michaelis et al., 1998).
Time Course of the Actions of Paclitaxel. The protective effects of paclitaxel against Aβ toxicity were initially observed either 24 or 48 h after addition of the peptide, with paclitaxel added 2 h before Aβ. Although some neurons are lost over time in untreated primary cultures, it was of interest to determine whether the effects of paclitaxel against Aβ persisted beyond the 48-h period and to determine the effects of treatment with paclitaxel alone on longer-term cell survival. Figure 2 shows the time course of neuronal survival following the one-time addition of Aβ in the presence or absence of Taxol on day 0. Viability of the cultures was monitored for 6 days after initiation of the treatments. Because of the loss of neurons across time in the control cultures, data from cells exposed to various treatments have been normalized to the controls defined as 100% for each time point. As shown in Fig. 2, the survival rate for 100 nM paclitaxel-only cultures was ∼90% of control values after 4 days and 80% after 6 days. In cultures exposed only to Aβ, the percentage of surviving cells was reduced to ∼40% by day 4 and to less than 30% by day 6. When paclitaxel and Aβ were both present, cell survival was increased to ∼65% on day 4 and to 50% on day 5. In these studies, the paclitaxel and Aβ were added only once at day 0, although 25% of the culture medium was removed and replaced with fresh medium on day 4. At all time points, cultures pretreated with paclitaxel showed better survival than those exposed to the Aβ peptide only.
Paclitaxel Blockade of Aβ-Induced Apoptosis in Neurons. Some studies have shown evidence of apoptosis in AD brain (Lassmann et al., 1995), and Aβ addition in vitro in neuronal cell culture models leads to activation of apoptotic cascades (Loo et al., 1993; Ivins et al., 1999). As a further characterization of the activity of paclitaxel, we conducted experiments to determine whether pretreatment of the neurons with the drug reduced the percentage of apoptotic cells observed following incubation with the Aβ25-35 peptide. An in situ fluorescence assay for apoptotic neurons was used to evaluate the effects of Aβ with or without paclitaxel pretreatment. Staurosporine, an established inducer of apoptosis, was used as a positive control. As shown in Fig. 3, the primary cultures typically had ∼7 to 8% apoptotic neurons after 6 days in culture, and the addition of 100 nM paclitaxel alone for 24 h produced no significant increase in this percentage. Aβ25-35 increased the population of apoptotic cells to ∼30% after 24 h. Cultures to which paclitaxel was added 2 h before Aβ contained about ∼17% apoptotic cells, a statistically significant reduction compared with cultures exposed to the peptide only. When staurosporine (1 μM) was added to the cultures for 24 h, ∼22% of the neurons were labeled with the fluorescent caspase substrate. Cultures pretreated with paclitaxel, followed by staurosporine, showed only 10% labeled cells, a somewhat greater reduction than was observed in the Aβ plus paclitaxel-treated cultures. Although the fluorescent caspase substrate used for the in situ determination of apoptotic cells is not specific for a given caspase enzyme, the increase in the percentage of apoptotic cells in the presence of Aβ was completely blocked by addition of the cellpermeable caspase-8 inhibitor IEDT-FMK (data not shown). Thus, initiation of the apoptotic cascade may be an early event in Aβ toxicity as suggested by others (Ivins et al., 1999).
Effects of Structurally Diverse MT-Stabilizing Agents on Aβ-Treated Cultures. The results obtained with the well characterized MT-stabilizing drug paclitaxel suggested that disruption of the cytoskeleton may contribute to the cell death cascade initiated in the presence of Aβ in the culture medium. If this is the case, one might predict that other drugs known to stabilize MTs would also be protective against Aβ toxicity, even if they differ from paclitaxel in terms of chemical structure. Before assessing the neuroprotective properties of several synthesized or isolated compounds known to interact with tubulin, we determined the MT-stabilizing activity of each agent relative to that of paclitaxel in an in vitro tubulin assembly assay as described previously (Liu et al., 2002). The structures of the agents are shown in Fig. 4. All of the compounds tested except 10-deacetyl baccatin III were found to promote in vitro MT assembly with EC50 values very similar to that for paclitaxel, i.e., concentrations ranging from 0.5 to 2.0 μM. The 10-deacetyl baccatin III, on the other hand, was ∼100-fold less active than all of the other compounds, although it shares many structural features with the taxanes. The initial sets of agents tested are taxanes and thus structurally closely related to paclitaxel, whereas the other agents are structurally quite diverse. The semisynthetic taxoid molecule docetaxel was twice as potent as paclitaxel in the in vitro tubulin assembly, whereas the succinylated analog Tx-67 was slightly less effective. Epo A, a macrolide isolated from Sorangium cellulosum, also exhibited MT-stabilizing activity comparable to paclitaxel as has been reported by others (Bollag et al., 1995). A synthetic compound designated as “UK-100” is a nontaxoid borneol derivative with MT-stabilizing activity comparable to paclitaxel but markedly reduced cytotoxicity in cancer cell lines (Klar et al., 1998), and discodermolide, a marine sponge product, was also quite similar to paclitaxel in MT-assembly, although it is structurally unrelated (Fig. 4).
Initially, the taxanes were evaluated for relative potencies in protecting neurons against Aβ toxicity as shown in the dose-response curves in Fig. 5. In all of the experiments, exposure to Aβ alone typically reduced the percentage of live cells from ∼75 to 80% to less than 40%. The three taxanes with demonstrated MT-stabilizing activity enhanced neuronal survival in a dose-dependent, statistically significant manner, beginning at concentrations in the 20 to 40 nM range (Fig. 5, A, B, and C). The estimated EC50 values for paclitaxel, docetaxel, and Tx-67 were 13, 32, and 34 nM, respectively (P values from analysis of variance were <0.02 for all curves). The maximal viability calculated from each curve indicated that only Tx-67 provided full protection, with paclitaxel and docetaxel exhibiting only partial agonist activity at ∼75 and 90% respectively. Higher concentrations either showed no further increase in cell survival or some loss in viability. The 10-deacetyl baccatin III, on the other hand, provided no statistically significant protection against Aβ toxicity at either low or high concentrations (Fig. 5D).
When each of the active taxane drugs was tested alone at the highest concentration used in combination with Aβ, none of the three taxanes produced statistically significant toxicity. Docetaxel, which is the most potent of the taxanes in the in vitro tubulin assembly assay, did show significant toxicity at concentrations above 60 nM. The succinate derivative Tx-67 did not produce significant toxicity, even when the drug alone was tested at a concentration of 150 nM.
When three structurally diverse MT-stabilizing agents, Epo A, discodermolide, and UK-100, were tested under similar conditions, all were associated with a statistically significant increase in neuronal survival at a concentration of 20 nM (Fig. 6). The nontaxane agents seemed to have somewhat greater potency than the taxanes, with the estimated EC50 values of 3.3, 1.7, and 4.2 nM for Epo A, discodermolide, and UK-100, respectively (p values from analysis of variance were <0.002 for all curves). Nevertheless, all of these agents were partial agonists in that the maximal survival in the presence of Aβ was ∼89% of non-Aβ-exposed controls with Epo A and UK-100 and 75% with discodermolide. Again, adding higher concentrations of the drugs showed no further protection against Aβ toxicity. Although we found that discodermolide alone seemed to show a small amount of toxicity at 100 nM, neither UK-100 nor Epo A reduced cell viability even when added alone at concentrations up to 200 nM (data not shown). When the effects of pretreating the neurons with either UK-100, Epo A, or discodermolide before Aβ were monitored for several days, all three agents significantly enhanced neuronal survival compared with cultures that were exposed to Aβ only (Fig. 7).
To confirm that at least a part of the protective effect of paclitaxel might be due to prevention of MT depolymerization, we used the known MT-depolymerizing agent nocodazole to induce cytoskeletal disruption and neuronal cell death. Exposure of the neurons to 5 nM nocodazole led to a 32% decrease in viable neurons in 2 h, whereas a 2-h pretreatment with 100 nM paclitaxel before addition of 5 nM nocodazole led to only a 7% loss of neurons (data not shown). Similar protective effects were observed when cell viability was assessed after a 4-h exposure to nocodazole. These results show that the neuronal cell death initiated directly by disruption of the cytoskeletal network can be significantly moderated by a drug known to have MT-stabilizing properties.
MT-Stabilizing Drugs and JNK Activity in Primary Cortical Neurons. Consistent with our earlier observations and those of several others (Furukawa et al., 2003; Sponne et al., 2003), the MT-stabilizing drugs did not lead to cell death at concentrations of ∼80 nM or lower. Given that MT-stabilizing drugs are used as anticancer agents because of cytotoxicity produced in proliferating cells, the effects of these agents on some known components of the cell death signaling cascade were evaluated in our neuronal cultures. Although paclitaxel has been reported to induce apoptosis in cancer cells by causing phosphorylation of the antiapoptotic protein Bcl-2 (Blagosklonny et al., 1997), others have not found such an increase in Bcl-2 phosphorylation in postmitotic primary neurons (Figueroa-Masot et al., 2001). However, these authors reported that paclitaxel induces apoptosis in cortical neurons through a pathway that requires JNK activation but is independent of Bcl-2 phosphorylation. Although we observed no drug-induced increase in apoptotic neurons under our culture conditions, we tested our neurons to determine whether an increase in JNK activity occurred with the drug treatment alone. When JNK activity was assessed following a 24-h exposure to either paclitaxel or Epo A, no drug-induced increase in the activity of the kinase was observed (Fig. 8). The known JNK activator sodium arsenite did produce a statistically significant increase in activity that was totally blocked by the JNK inhibitor indirubin-3′-monoxime. As expected, the inhibitor alone reduced the basic cellular activity in the cells. Thus, phosphorylation of the GST-c-Jun fusion protein substrate does appear to be an adequate reporter of the JNK activity under the various conditions to which the neurons were exposed. We also carried out assays to determine whether the drugs increased the amount of phosphorylated Bcl-2 in our cultures but, consistent with a previous report (Figueroa-Masot et al., 2001), no difference was detected. The lack of these phosphorylation events in our neurons in response to the drug treatments is consistent with the absence of caspase activation (Fig. 3).
The present studies were conducted to characterize more fully our previous observation that nanomolar concentrations of the MT-stabilizing drug paclitaxel protected primary cortical neurons against Aβ toxicity and blocked Aβ-induced hyperphosphorylation of τ (Michaelis et al., 1998; Li et al., 2003). Our initial demonstration of protection against Aβ was carried out with paclitaxel, a prototype MT-stabilizing agent that is quite insoluble and also fails to cross the bloodbrain barrier (Cordon-Cardo et al., 1989). These limitations raised the question of whether MT-stabilizing compounds with different physicochemical properties would also exhibit neuroprotective activity in the nanomolar range so they might ultimately be tested in vivo. The recent identification of natural products and new synthetic agents that promote tubulin assembly in vitro enabled us to evaluate several structurally diverse compounds for effects in the neuronal cell cultures. Data demonstrating the protective effects of three taxanes with comparable potency in promoting tubulin assembly (Fig. 5) are representative of more than 30 active taxoids tested on neuronal cultures. The failure of the inactive 10-deacetyl baccatin III to protect the neurons suggests that some effects on cytoskeletal structure may be involved in the pathways through which the drugs reduced toxic signaling cascades initiated by Aβ. Analysis of the pharmacological activities of the taxanes indicated that, with the exception of Tx-67, none of the compounds led to 100% protection of the Aβ-treated cultures, although the rate of neuronal survival in the presence of the drugs typically was close to 90% of that in control cultures. Interestingly, the three nontaxane compounds with in vitro tubulin assembly-promoting activity were somewhat more potent than taxanes in protecting the neurons (Fig. 6). The structural diversity of protective agents indicates that it may be possible to design new compounds that protect the cytoskeletal network and have requisite properties to permit the in vivo testing of a link between Aβ toxicity and neurofibrillary pathology in mouse models of AD.
The finding that drugs used clinically for their cytotoxic properties enhance the survival of postmitotic neurons exposed to agents that initiate cell death cascades seems counterintuitive. Although paclitaxel-induced activation of JNK in cortical neurons has been correlated with apoptosis in some studies (Figueroa-Masot et al., 2001), we did not observe increased apoptosis in our cultures exposed to 100 nM paclitaxel (Fig. 3) or direct activation of JNK in the presence of either paclitaxel or Epo A (Fig. 8). Differences in the neuronal culture conditions might contribute to these divergent observations. The data in Figs. 5 and 6 indicate that the MT-stabilizing drugs did not produce toxicity when added alone at concentrations that protected against Aβ, although most taxanes did lead to cell loss at drug concentrations greater than 100 nM. This certainly suggests that the use of such agents in neurons would have a narrow therapeutic concentration range for helping to protect brain neurons against Aβ toxicity.
Although Aβ-containing plaques and NFTs composed of hyperphosphorylated τ are the hallmarks of AD, a mechanistic link between the two lesions has not yet been proven. Nevertheless, many experimental studies in both cell culture systems and animals strongly suggest that τ hyperphosphorylation and neuronal dystrophy develop in the presence of Aβ oligomers and/or fibrils, leading ultimately to the death of the neurons (e.g., Geula et al., 1998; Grace et al., 2002). Expression of τ was found to be critical for Aβ-induced toxicity in hippocampal neurons in culture, providing additional support for a link between the two lesions (Rapoport et al., 2002). Neurons from τ-deficient mice showed no neuritic degeneration with a 24-h exposure to 20 μM Aβ1-40, although the τ null neurons did exhibit sensitivity to the peptide after a 6-h pretreatment with 1 to 10 μM paclitaxel. Analysis of the MT proteins in neurons from the τ-knockout mice indicated a trend toward less stabilized forms of tubulin and altered expression of other MT-associated proteins, leading the authors to suggest that the presence of more dynamic MTs might confer resistance to Aβ. However, the τ-null neurons also cannot undergo the τ hyperphosphorylation and aggregation process that may be a primary event leading to cell death in normal neurons exposed to Aβ. We have previously reported that exposure of neurons to nanomolar paclitaxel concentrations also blocks the Aβ-induced increase in τ phosphorylation through pathways that seem to involve regulation of intracellular Ca2+ (Li et al., 2003). The susceptibility to Aβ toxicity in the τ null neurons with addition of paclitaxel concentrations an order of magnitude greater than that used in our studies (Rapoport et al., 2002) could be due to effects of the drug on other processes initiated by excessive stabilization of the cytoskeleton and loss of dynamic instability independently of τ. Both the conditions under which we found MT-stabilizing drugs to protect neurons against Aβ and the cytoskeletal composition of the neurons are markedly different between the two experimental paradigms, making it difficult to assess the implications of each study for the effects of the drugs on neurofibrillary pathology in AD. Nevertheless, both studies point to the pivotal role that the cytoskeletal proteins, particularly τ, play in the toxic actions of Aβ.
Although the results of our experiments taken together suggest that some interaction with MTs is involved in the protective effects of the drugs, the cellular mechanisms underlying enhanced neuronal survival have yet to be delineated, primarily due to the fact that the molecular basis for Aβ toxicity has still not been fully elucidated. There may be differences in the molecular mechanisms through which the full-length Aβ1-40/42 and the Aβ25-35 we used here initiate cell death cascades, although Grace et al. (2002) have clearly demonstrated that both the short and longer peptides induce marked neuronal dystrophy and synaptic loss. Furthermore, it has been reported that Aβ25-35, like the longer peptide, does form toxic assemblies that may create channels in the plasma membrane and induce inward currents that allow for Ca2+ influx into neurons (Arispe et al., 1993; Lin and Kagan, 2002). Our finding that paclitaxel protected against both peptides indicates that signaling pathways initiated by the two peptides involve common elements that can be blocked by drugs that share the property of MT stabilization, although MTs may not be the only target for these compounds. Despite possible differences in early initiation steps, disruption of the regulation of intracellular Ca2+ distribution may be a final common pathway underlying the toxicity of both peptides. Paclitaxel has been shown to protect neurons against several insults leading to abnormal elevations in intraneuronal Ca2+ (Burke et al., 1994; Furukawa and Mattson, 1995; Furukawa et al., 2003). Altered Ca2+ regulation may be the first step in a sequence of events in which Aβ leads to signaling pathways that produce cytoskeletal rearrangement, possibly through activation of focal adhesion proteins (Cotman et al., 1998). The focal adhesion pathway leads to increased activity of two proline-directed kinases implicated in τ hyperphosphorylation and formation of NFTs, cyclin-dependent kinase 5 (cdk 5) (Li et al., 2000) and glycogen synthase kinase 3β (Frame and Cohen, 2001). Elevations in cytosolic free Ca2+ have also been implicated in the hyperphosphorylation of τ as a result of calpain-mediated cleavage of the p35 protein that normally regulates the activity of cdk 5 (Dhavan and Tsai, 2001). We previously reported that paclitaxel blocked Aβ-induced phosphorylation of τ, apparently by preventing the cleavage of p35 by calpain, although the drug did not directly inhibit the activity of either calpain or cdk 5 (Li et al., 2003). Given the extensive evidence implicating dysregulation of intracellular Ca2+ in Aβ toxicity, the MT-stabilizing drugs may ultimately help maintain Ca2+ homeostasis in the presence of an array of stressful stimuli, including Aβ.
Several laboratories have now reported that nanomolar concentrations of paclitaxel protect neurons against various toxic insults. These include the Ca2+ ionophore A23187 (Burke et al., 1994), glutamate-induced excitotoxicity (Furukawa and Mattson, 1995), the induction of reactive oxygen species by nonfibrillar Aβ (Sponne et al., 2003), and the neurotoxic action of a τ mutation associated with frontotemporal dementia (Furukawa et al., 2003). Trushina et al. (2003) reported that MT destabilization is a very early step in the toxicity leading to Huntington's disease, and paclitaxel treatment markedly enhanced survival of primary striatal neurons expressing a toxic mutant huntingtin protein. These authors observed no toxicity with addition of 5 to 100 nM paclitaxel in neurons monitored for 50 h, indicating that toxicity in the presence of nanomolar concentrations of MT-stabilizing drugs is not inevitable. Furthermore, drugs that block early MT depolymerizing signals associated with various insults appear to permit cells to engage cascades that enhance survival. These reports, in addition to our observations here with diverse MT-stabilizing drugs, suggest that the state of cytoskeletal proteins may serve as a surveillance system that transduces stress signals leading either to survival or cell death. New experimental strategies designed to delineate, at the molecular level, the participation of specific populations of cytoskeletal proteins in signaling cascades may well provide opportunities for innovative therapeutic interventions that capitalize on these targets.
We thank Jennifer Bean and Misty Bechtel for the preparation of primary neuronal cultures and Yanbin Liu and Brandon Turunen for synthesis of Tx-67.
- Received July 24, 2004.
- Accepted September 16, 2004.
This work was supported by the Alzheimer's Disease and Related Disorders Association, the Institute for the Study of Aging, and National Institutes of Health Grants HD 02528, CA 82801, and CA 79641. E.R.R. was supported by National Institutes of Health Training Grant GM 07775 and DAMD 17-001-0303. E.R.R. and K.I.S. received support from the American Foundation for Pharmaceutical Education.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: Aβ, β-amyloid; AD, Alzheimer's disease; MT, microtubule; NFT(s), neurofibrillary tangle(s); Epo A, epothilone A; DMSO, dimethyl sulfoxide; PI, propidium iodide; JNK, c-Jun terminal kinase; PBS, phosphate-buffered saline; GST, glutathione S-transferase; cdk 5, cyclin-dependent kinase 5; UK-100, isoborneol derivative (C29H34N2O6).
- The American Society for Pharmacology and Experimental Therapeutics