Increased Protein Tyrosine Phosphorylation in Apoptotic Neural Cell Death Due to Microtubule Perturbations

Abstract

The microtubule-perturbing drugs colchicine and taxol have been found to induce apoptosis in a CNS neuronal cell line. Apoptosis in drug-treated rat B103 neuroblastoma cells was evident in characteristic morphological changes, internucleosomal DNA fragmentation, and loss of nuclear content. Since colchicine and taxol have opposite actions on microtubule integrity, disruption of the active turnover of the microtubule network appears to be a crucial step for apoptosis to occur. It has been suggested that the basis for apoptosis by these drugs derives from their known block of the cell cycle at G₂/M, but this does not appear the sole reason as both colchicine and taxol were able to evoke high levels of apoptosis in cells differentiated by Bt₂cAMP or serum withdrawal. Further tests of cellular consequences of microtubule perturbation revealed a specific impact on signal transduction involving protein tyrosine phosphorylation. Immunoprecipitation with antibodies against tyrosine phosphorylated proteins showed a striking increase in the phosphorylation of a Triton-insoluble ~90 kDa protein, roughly concurrent with the onset of internucleosomal DNA fragmentation. Cycloheximide and genistein significantly reduced cell death and blocked appearance of the ~90 kDa tyrosine phosphorylated protein. Data suggest the hypothesis that signal transduction leading to apoptosis can be triggered by anomalous microtubule turnover and that the mechanism involves tyrosine phosphorylation of a ~90 kDa Triton-resistant protein.

INTRODUCTION

Microtubules have a central role in cell division, and their disruption is thought to imbalance cell cycle signals/checkpoints in a manner that triggers apoptosis (Sorger et al., 1997). This apoptosis has been investigated primarily in proliferating, non-neuronal cell types and can be induced by drugs with opposite effects. Colchicine and taxol, e.g., both kill proliferating cells by their impact on microtubules (Sorger et al., 1997), but colchicine causes microtubule depolymerization (Panda et al. 1995), while taxol induces microtubule stabilization (Yvon et al., 1999).

There are indications that these microtubule-directed toxins also may impact neuronal cells (Ceccatelli et al, 1997, Helson et al, 1993). Proliferating neuroblastoma and pheochromacytoma cells are killed by both colchicine and taxol (Nakagawa-Yagi, 1994, Lindenboim et al., 1995, Helson et al., 1993). However, in post-mitotic, untransformed cerebellar neurons, it has been reported that taxol opposes the apoptosis caused by colchicine (Bonfoco et al., 1995). Thus there may be differences in neuronal responses dependent upon the state of cellular metabolism and differentiation. How microtubule-perturbing drugs affect cells of neuronal origin thus requires further study.

The current paper addresses the conflicting information regarding the toxic nature of taxol versus colchicine in neuronal cells, examines the possible relevance of differentiation, and suggests a basis for the neurotoxic mechanism. Experiments have used rat B103 cells, a spontaneously neuritogenic neuronal cell line of CNS origin. Differentiating-promoting conditions were used to determine if the state of proliferation might have a critical impact on neurotoxicity of taxol and colchicine. With respect to mechanism, experiments focused on possible roles for altered protein tyrosine phosphorylation, which has been implicated previously in neuronal apoptosis (Muragaki et al., 1997). Data support the hypothesis that microtubule anomalies constitute a signal for neuronal apoptosis, which is propagated through protein tyrosine phosphorylation. This work is especially relevant to CNS microtubule anomalies, such as might occur in AD.

MATERIALS AND METHODS

All drugs and supplies from Sigma (St. Louis, MO), unless otherwise noted.

Cell culture

B103 rat CNS neuroblastoma cells (Schubert et al., 1974), were maintained in Dulbecco's modified Eagle medium (DMEM), 10% fetal calf serum (FCS), and antibiotics (streptomycin, penicillin, and fungizone)(all from Life Technologies, Gaithersburg, MD), in 6% CO₂. For all experiments, cells were plated at low density and grown to 70% confluency. To perturb microtubules, colchicine or taxol was added to 100 nM. To perturb actin filaments, cytochalasin D was added to 1 µM. To differentiate B103 cells, medium containing 0.5 mM dibutyrl cyclic AMP (Bt₂cAMP) or medium without serum, was added for 24 to 48 hours. In experiments with genistein, cells were pretreated with 5 or 50 µM genistein for 1 hour prior to addition of anti-microtubule drugs. For experiments with cycloheximide, cells were treated with 5–10 µg/ml cycloheximide added at the same time as the anti-microtubule drug.

Phase microscopy and LIVE/DEAD^® assay

The LIVE/DEAD^® assay was carried out essentially as described in Lambert et al. (1994) and by the manufacturer's protocol (Molecular Probes, Eugene, OR). Briefly, B103 cells were plated on a Permanox^® 8-well Lab-Tek^®(Nalge Nunc Int., Naperville, IL) culture slide at a dilution of 1 × 10⁴cells/well. The cells were grown for 2 days to reach 70% confluency. Cells were then treated with appropriate drugs for 24 hours. After drug treatment, cells were washed with Dulbecco's Phosphate Buffered Saline (D-PBS, Life Technologies) and were then treated with 250 µl of D-PBS containing 12 µM ethidium homodimer and 1 µM calcein-AM for 30 minutes at 37°C in the dark. The cells were examined using phase-contrast and fluorescent imaging on a Nikon Diaphot microscope and images were captured using a Hamamatsu 10-bit digital CCD camera with the MetaMorph^® image analysis software (Universal Imaging Corp., Westchester, PA). Quantitation was done by counting the fluorescent cells and percent death was achieved by dividing the number of dead cells by the total number of cells.

DNA agarose gel electrophoresis

Fragmented DNA was obtained and characterized essentially as described by Nakagawa-Yagi (1994). Samples were harvested with trypsin, centrifuged at ~500 × g (International Equipment Co., Needham Heights, MA), and resuspended in 460 µl of a hypotonic detergent buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 0.3% Triton X-100) on ice for at least 30 min. to lyse cells. The mixture was then centrifuged at 27,000 × g for 20 min. to separate fragmented DNA (supernatant) from intact chromatin DNA (pellet). The supernatants were incubated at 50°C (Precision water bath, Chicago, IL) in the presence of 0.5 mg/ml proteinase K for one hour, followed by addition of 0.4 mg/ml RNAse A for another hour, and then DNA was extracted with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:l)(United States Biochemical, Cleveland, Ohio). DNA was then extracted from the aqueous phase with 1/10 the volume of sodium acetate (0.3 M) and twice the volume of 100% ethanol at −80°C overnight. DNA was pelleted, washed with 70% ethanol, dried by speed vac (Savant, Holbrook, NY), and dissolved in 1 × TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). DNA fragments were separated by electrophoresis in 1.5% agarose using 40 mM Tris-acetate, 1 mM EDTA: pH 7.2 (1 × TAE buffer) with 5–10 µg of DNA. The DNA was visualized with 0.5 µg/ml ethidium bromide under an ultraviolet light. Total cellular DNA was obtained according to Renauld et al. (1995) using a SDS lysis buffer (0.15 mM NaCl, 10 mM Tris pH 8, 12.5 mM EDTA, 0.5% SDS, 0.5 mg/ml proteinase K). The samples were incubated overnight at 56°C. DNA was separated and visualized as above.

Flow cytometry

Cellular DNA was labeled and examined as previously described (Nicoletti et al., 1991, Lindenboim et al., 1995). Cells were harvested, pelleted at 500 × g, rinsed in Hank's buffer (Life Technologies), and resuspended with gentle trituration in a hypotonic detergent solution containing propidium iodide (PI) (0.1% sodium citrate, and 0.1% Triton-X 100, 50 µg/ml PI). The samples were incubated at 4°C overnight in the dark before analyzing on a FACScan^® flow cytometer (Becton Dickinson, Mountain View, CA, U.S.A.). The manual peak fit protocol was used within the CellFIT^® software program to obtain percentages of apoptotic cells within the population of each sample. DNA QC particles (Becton Dickinson) were used to standardize the flow cytometer and normalize cell cycle peaks. The G₀-G₁ cell population peak was arbitrarily set at approximately 250 in fluorescence area (FL2-A). The peak to the left of the G₀-G₁ peak was defined as the hypo-G₁peak and contains the apoptotic population (A₀). Cells in G₂ or M phase contain twice the nuclear content and are present in a peak at about twice the fluorescence area of the G₀-G₁ peak (G₂/M). Cells undergoing DNA synthesis are designated between the two peaks (S).

Immunoprecipitation

Cell lysis, immunoprecipitation, and immunoblotting were carried out essentially as described (Berg et al., 1997). Cells were seeded at 2.25 × 10⁶ cells on 150 mm plates, grown for two days, and then treated with fresh growth media and the appropriate drug for 24 hours. Cells were scraped off with a rubber policeman in 500 µl of ice-cold 2 × Triton buffer + a cocktail of protease inhibitors (2% Triton X-100, 2 mM EGTA, 100 mM Tris, pH 7.2, 4 µg/ml aprotinin, 1 µg/ml pepstatin A, 2 µg/ml leupeptin, 2 mM sodium vanadate, 50 mM sodium fluoride, and 1 mM phenylmethylsulfonyl fluoride (PMSF)). The samples were incubated for 10 min. on ice and Triton soluble proteins (supernatant) were separated by a 10 min. 12,000 × g spin at 4°C. The Triton-insoluble pellet was washed 2 × in 1 × Triton buffer + protease inhibitors. The resulting pellet was solubilized in extraction buffer with protease inhibitors (10 mM Na₂HPO₄, 10 mM Tris, 150 mM NaCl, 1 mM CaCl₂, 0.1% SDS, and 1% triton, pH 7.4) and clarified by centrifugation at 10K × g for 10 min. at 4°C. Protein was quantified using the BioRad protein assay (BioRad, Hercules, CA). Immunoprecipitation was carried out with 50 µg of protein, added to 200 µl of Protein-A-sepharose beads (1 mg/ml; Pierce, Rock-ford, IL) attached to an anti-phosphotyrosine primary antibody (PY20; Transduction Labs, San Diego, CA) and shaken overnight at 4°C. The beads were washed- twice with TES + protease inhibitors (25 mM Tris pH 7.5, 5 mM EDTA pH 8.0, 250 mM NaCl) and protein was separated using 8% SDS-PAGE and transferred to Hybond ECL Nitrocellulose (Amersham, Piscataway, NJ). Membranes were blocked in 1% B.S.A in TBS-T2 (10 mM Tris-HCl, pH 7.4, 0.9% NaCl, 0.05% MgCl, and 0.2% Tween 20) for 1 hour at room temperature. Anti-phosphotyrosine primary antibody (RC20H, Transduction Labs) conjugated to horseradish peroxidase (HRP) was added for 1.5 hours at room temperature. Visualization was obtained using ECL reagents (Amersham). Molecular weight standards (Novex, San Diego, CA) were visualized using HRP-conjugated streptavidin (Amersham).

RESULTS

Evoked cell death

In the first experiment we tested if either colchicine or taxol would induce cell death in the B103 CNS neuronal cell line. Actively growing cultures at 70% confluence were incubated with 100 nM colchicine or taxol for 24 hours. Viability was then quantified by the LIVE/DEAD® dual fluorescence assay. In this assay, ethidium homodimer intercalates into the DNA of cells with compromised membranes (dead or dying cells) and undergoes a 40-fold enhancement of fluorescence in the red range. Calcein-AM is cleaved into the fluorescent calcein product and fluoresces green only in metabolically active living cells. Despite having opposite effects on the microtubule system (reviewed in Sorger et al., 1997), each drug was found to evoke approximately 6 to 7-fold increases in the percentage of dead cells (figure 1). As a control for other cytoskeletal disrupting drugs, the anti-actin filament drug cytochalasin D was tested and found to have no effect on cell survival. The data shown in figure 1 indicate that perturbations in the microtubule network, but not in microfilaments, induce cell death in the B103 cells. Maintenance of appropriate microtubule dynamics (microtubule turnover) appears to be involved in the prevention of apoptosis.

Figure 1. Cell death is evoked in B103 cells by perturbing the microtubule network, but not by perturbation of actin filaments.

B103 cells were treated with colchicine (100 nM), taxol (100 nM), cytochalasin D (1 µM), or appropriate vehicle (ethanol (control) and DMSO, 0.1%) for 24 hours and then assayed for viability using the LIVE/DEAD^® assay from Molecular Probes (see Methods). Percent cell death is the number of ethidium homodimer-positive cells divided by the total number of counted cells times 100. At least three different fields were counted for each of four experiments. Error bars depict the standard deviation (SD). Colchicine and taxol are seen to induce a 6- to 7-fold increase in cell death, while cytochalasin D confers no significant increase in cell death

Cell morphology

The quantitation of the LIVE/DEAD^® fluorescence assay does not discriminate between apoptotic and necrotic cell death. To help make this distinction, morphological changes were observed following colchicine or taxol treatment. Examination of cells at two hpurs showed that cell bodies had begun to round up and neurites retract. At approximately 7 hours, membrane blebbing became obvious, as visualized by VEC-DIC microscopy (data not shown). By 24 hours, most cells showed morphological changes consistent with apoptotic cell death, including extensive membrane blebbing, picnotic nuclei, and condensed chromatin (figure 2, panels A and C). The chromatin evident at the edges of the nuclear membrane indicate nuclear fragmentation and DNA marginization, visualized here as a punctate pattern of ethidium homodimer fluorescence (figure 2, panels D and F). The ethidium homodimer positive cell (panel F) includes an apoptotic body. The morphological data strongly suggest that colchicine and taxol induce apoptosis in the B103 CNS neuronal cultures.

Figure 2. Cells treated with colchicine and taxol exhibit morphological features of apoptotic cell death.

This figure shows phase and fluorescent images of B103 cells treated with colchicine (A,B), ethanol (C,D), and taxol (E,F) for 24 hours. Membrane blebbing and neurite retraction at the cell periphery are seen in Panels A and E, but are absent from control cells (Panel C). Panel D shows calcein fluorescence (living cells). No ethidium homodimer fluorescence is present. Ethidium homodimer fluorescence in panels B and F indicates these cells have damaged plasma membranes and are dying or are already dead. Panels B and F also highlight the condensed chromatin of the dead cells treated with either drug, with the DNA present at the edges of the nucleus. This DNA marginization is typical of apoptotic cell death. Furthermore, the DNA is fragmented as indicated by the punctate rather than uniform fluorescence. An apoptotic body is also visible in panel F, where a membrane-bound piece of DNA has been released from the nucleus (see Color Plate I at the back of this issue)

DNA laddering

To complement the morphological data, DNA gel electrophoresis was carried out to verify apoptotic internucleosomal DNA fragmentation. Vehicle control samples showed no fragmentation, but fragmented DNA extracted from colchicine and taxol treated B103 cells exhibited distinct ladders (figure 3). DNA laddering also was observed in extracts of total cellular DNA obtained from colchicine and taxol treated cells, but not control cells (data not shown). Cells treated with cytochalasin D showed no evidence of DNA fragmentation (figure 3, lane 5). Thus, perturbations in F-actin fail to produce internucleosomal DNA fragmentation, while microtubule perturbations lead to DNA fragmentation in B103 cells.

Figure 3. Microtubule perturbations induce DNA fragmentation.

Fragmented DNA extracted from B103 cells treated as in figure 1, was run on a 1.5%. agarose gel and stained with ethidium bromide. “Ladders” of uniform fragments indicative of apoptotic cell death are seen only in Lanes 2 and 4, which are cultures treated with colchicine and taxol, respectively. No ladder is seen with vehicle-treated cells, Lanes 1 and 3, ethanol and DMSO, respectively, or with cytochalasin D-treated cells, Lane 5. These data suggests that loss of microtubule maintenance, but not loss of actin filaments, causes apoptosis in B103 cells

Flow cytometry

Flow cytometry previously has been used to discriminate between apoptotic and necrotic cell death in non-neuronal cells (Huschtscha et al., 1994) and to quantify the extent of apoptosis (Nicoletti et al., 1991). Flow cytometry therefore was done to quantify the extent of apoptotic death due to colchicine and taxol treatments of B103 cells. Figure 4 illustrates an experiment in which the increase in apoptotic nuclei can readily be seen. Results from at least three flow cytometry experiments showed that apoptosis was evident in 63.2 ± 15.93% of colchicine treated cells, 56.0 ± 12.0% of taxol treated cells, and 6.11 ± 4.95% of control cells. As expected, treatment with cytochalasin D showed no increase in apoptotic cells. The flow cytometric analysis was able to show that microtubule perturbations for 24 hours induce apoptosis in approximately 60% of B103 cells.

Figure 4. Quantitation of apoptotic cells by flow cytometry.

Flow cytometry was carried out on B103 cultures treated as above to further analyze the state of the cells. The area under each peak was quantified using CellFIT^® software. Histograms shown here represent one of four separate experiments. Vehicle control cultures in panels A and C (ethanol and DMSO, respectively) exhibit typical cell cycle characteristics, with approximately 3.5% hypo-G₁ (A₀) cells. Cultures treated with 100 nM colchicine and taxol (panels B and D, respectively) exhibit large increases in the A₀ population (up to 73%) and concomitant decreases in the other cell cycle populations following 24 hours. Cytochalasin D-treated cultures (1 µM, panel E) do not undergo a significant increase in their apoptotic population

Apoptosis independent from mitotic block

Both colchicine and taxol are known to confer a G₂/M transition block in proliferating cells and it has been suggested that this block is responsible for apoptosis (Jordan et al., 1992, Liebmann et al., 1994, reviewed in Sorger et al., 1997). To assess this hypothesis, B103 cells first were treated with Bt₂cAMP, which is known to induce morphological differentiation (Schubert, 1974). After 24 hours in Bt₂cAMP, cells were incubated with colchicine (figure 5, left panel) or taxol (middle panel) for another twenty four hours. Vehicle-treated cells showed no fragmentation, but anti-MT drug treatment induced typical DNA laddering. Bt₂cAMP treatment was only partially effective at reducing cells in S phase, however (decreased by ~25%; not shown). Therefore, B103 cells next were induced to differentiate by serum withdrawal. Again, the anti-MT drug-treated cells showed characteristic DNA fragmentation (figure 5, right panel). By flow cytometry, the decrease in S-phase cells was more than 65%, but the extent of microtubule drug-induced apoptosis still exceeded 50%, similar to that in proliferating populations (data not shown). Differentiated versus proliferating cells also were compared by looking at their extent of apoptosis at early time points. Proliferating cells required approximately 24 hours (figure 4). In differentiated cells, however, a large increase in the hypo-G₁ peak occurred within 2–4 hours of taxol exposure. At this time, cells treated with colchicine remained viable (figure 6), with significant apoptotic nuclei not evident until at least 12 hours (data not shown). Apoptosis thus occurs quicker in differentiated cells than in asynchronous cells, with effects of taxol more rapid than colchicine. The rapid induction of apoptosis in differentiated cells provides further evidence that the mitotic block is not essential for microtubule network perturbation induced apoptosis.

Figure 5. Differentiation of B103 cells does not inhibit apoptosis.

B103 cells were differentiated for 24 hours using Bt₂cAMP (left and center) and serum-starvation (right). Total (center) and fragmented (left and right) DNA was extracted and analyzed using agarose gel electrophoresis. Far left: 123 bp DNA standard. DNA from cultures treated with colchicine (+ COL, left and right) or taxol (TAX, center and right) show a “ladder” of DNA of roughly 180–200 bp, indicative of apoptosis, under both differentiation conditions. DNA fragments in controls (CON and DMSO) are substantially less than in experimental cultures. These data indicate that maintaining cells in G₀ does not prevent apoptosis caused by addition of microtubule-perturbing drugs

Figure 6. Rapid apoptotic cell death occurs following forced microtubule polymerization in differentiated B103 cells.

Panels A and C show cells treated with colchicine. Panels B and D show cells treated with taxol. Panels A and B are histograms of cultures treated with Bt₂cAMP for 24 hours, then treated with drug for 2 hours. Panels C and D are cultures that are serum-starved for 24 hours and then treated with drug for 4 hours. While colchicine treated cells gave apoptotic levels similar to controls (data not shown), taxol-treated cells show significant rapid apoptosis

Dependence on protein synthesis

The induced synthesis of cell death-related proteins is an integral step for many apoptotic signaling cascades (reviewed in Vaux, 1993), including following microtubule network perturbations (Nakagawa-Yagi, 1994, Liebmann et al., 1994). However, evidence suggests some examples where new protein production is not necessary following anomalous microtubule turnover (Takano et al., 1993, Gangemi et al., 1995). To assess the necessity of nascent protein synthesis in this system, some cultures treated with the anti-MT drug also were given cycloheximide. DNA isolated from cells treated with anti-MT drugs and cycloheximide were analyzed by DNA gel electrophoresis. Even in the presence of cycloheximide, DNA “ladders” resulting from apoptotic cell death were evident (data not shown). DNA from B103 cells that were treated with cycloheximide alone did not exhibit DNA fragmentation. To determine if the overall amount of apoptosis might nonetheless be reduced by cycloheximide, we used flow cytometric analysis. As shown in figure 7, cycloheximide significantly lowered the amount of apoptotic cell death. Cycloheximide lowered colchicine-induced apoptosis by 85% and taxol-induced apoptosis by 65%. Controls using cycloheximide alone showed no impact on apoptosis, as the hypo-G₁ populations showed only 4–5% of the total nuclei in these samples. The effects of cycloheximide thus show the importance of nascent protein production on the anti-MT drug-mediated apoptotic cascade.

Figure 7. Apoptotic cell death is sensitive to both protein production and protein tyrosine phosphorylation.

Differentiated cultures treated with either colchicine (A) or taxol (C) in the presence of cycloheximide show a large decrease in apoptotic cells (A₀ peak, compare to Figure 4B and D). Cycloheximide alone (E) elicits no apoptotic increase. Genistein pretreatment (1 hour) also greatly reduces apoptosis induced by colchicine (B) and taxol (D). (Again, compare to Figure 4B and D). The genistein only sample (F), which has an increased G₂/M peak at 500 FL2-A, shows no apoptosis. These data indicate that the apoptotic cell death initiated by microtubule-perturbing drugs is sensitive to protein synthesis and protein tyrosine phosphorylation

Requirement for protein tyrosine phosphorylation

Recent evidence has shown that apoptotic cell death may involve changes in protein tyrosine phosphorylation (reviewed in Lavin et al., 1996) and in some systems tyrosine phosphorylation has been shown to be increased by taxol treatment (Ding et al., 1993, Manthey et al., 1992). The tyrosine kinase inhibitor genistein, moreover, can be an effective apoptosis inhibitor (Yousefi et al., 1994, Liu et al., 1994). Therefore, we investigated how genistein affected colchicine- and taxol-induced apoptosis in B103 cells. Despite the presence of genistein, internucleosomal DNA cleavage was still detectable following colchicine and taxol treatment (data not shown). However, when apoptosis was quantified by flow cytometry, a reduction in apoptosis was clear (figure 7). Genistein lowered colchicine-and taxol-induced apoptosis by approximately 70%. Controls using genistein alone showed no impact on apoptosis, as the hypo-G₁ populations showed only 4–5% of the total nuclei in these samples. The impact of genistein in LIVE /DEAD^® assays, which measure membrane breakdown rather than nuclear DNA loss, was consistent with the flow cytometry results (not shown). This protection was evident at 50 µM genistein but not at 5 µM. The higher dose also inhibited neurite retraction caused by the colchicine and taxol treatments. The data establish that genistein was able to protect cells against death caused by colchicine and taxol.

Tyrosine phosphorylation of a 90 kDa protein

The genistein-mediated reduction in cell death indicated that tyrosine phosphorylation of specific proteins may play a role in the apoptotic signaling cascade caused by microtubule perturbations in B103 cells. Therefore, western immunoblotting was carried out on Triton soluble and insoluble fractions of tyrosine phosphorylated B103 proteins. As shown in figure 8, colchicine induced a striking increase in a cytoskeletal Triton-insoluble 90 kDa tyrosine phosphorylated protein. Genistein and cycloheximide were both able to prevent the increased expression of the 90 kDa tyrosine phosphorylated protein (lanes 3 and 4). Since cycloheximide and genistein were able to block the increase in the 90 kDa protein following colchicine, and also inhibit apoptosis, the 90 kDa protein may function in an apoptotic signaling cascade. Genistein was also able to block the increase in phosphorylation of the 90 kDa protein following taxol treatment (data not shown). As was the case for inhibition of cell death, 50 µM but not 5 µM genistein blocked tyrosine phosphorylation of the 90 kDa protein. The ability of genistein to completely block the tyrosine phosphorylation of the 90 kDa protein and inhibit apoptosis suggests that the 90 kDa protein may be functionally essential in an apoptotic cascade caused by microtubule network perturbations.

Figure 8. Striking increase in a ~90 kDa protein occurs following microtubule perturbations.

B103 cultures were treated with colchicine (100 nM), colchicine plus cycloheximide (10 µg/ml), and colchicine plus a 1-hour pretreatment with genistein (50 µM) for 24 hours. From these cultures, tyrosine phosphorylated proteins were immunoprecipitated using PY20 antibody. The immunoprecipitates were separated on a 4–20% polyacrylamide gel using SDS-PAGE and samples were transferred to nitrocellulose. Tyrosine phosphorylated proteins were visualized with RC20H antibody and ECL. A striking increase in a ~90 kDa tyrosine phosphorylated protein occurs after treatment with colchicine for 24 hours (COL), while no increase is seen in ethanol-treated control cells (CON). Both cotreatment with cycloheximide (C+C) and pretreatment with genistein (G+C) fully block the appearance of this protein

DISCUSSION

Data reported here show that microtubule perturbation in a CNS neuronal cell line causes death by apoptosis. Apoptosis occurred following taxol-induced microtubule stabilization and colchicine-induced microtubule breakdown. Survival does not appear to depend on actin filament integrity, as cytochalasin D treated cells did not apoptose. Apoptosis was observed in differentiated cells as well as proliferating cells, showing the mechanism does not require a G₂/M cell cycle block. The mechanism appears to involve specific changes in protein tyrosine phosphorylation, as induced apoptosis was accompanied by new appearance of tyrosine phosphorylation in a 90 kDa protein. Apoptosis and tyrosine phosphorylation of the 90 kDa protein both were reduced by genistein and by cycloheximide.

In the current study, apoptosis in CNS B103 nerve cells has been assessed through nuclear fragmentation, condensation, and marginization, as well as apoptotic body formation (figure 2), DNA laddering (figure 3), and nuclear content loss via flow cytometry (figure 4). Although there has been some debate about the essential criteria for apoptotic cell death (reviewed in Stewart, 1994), all morphological and biochemical data show that death caused by microtubule perturbations in B103 cells is by apoptosis. The presence of ethidium homodimer in the dying cells also presents the possibility that these cells were undergoing secondary necrosis following the apoptotic cascade.

It has been widely proposed that microtubule perturbations induce apoptosis by disrupting mitotic spindle formation at the G₂/M cell cycle transition (Jordan et al., 1992, reviewed in Sorger et al., 1997) and that cell cycle block at this transition point is essential for the induced apoptosis (Liebmann et al., 1994, Jordan et al., 1996). However, apoptosis still was evident in colchicine or taxol-treated B103 cells that had been induced to differentiate. This extends to a neuronal cell system the idea that cells at G₀ can apoptose in response microtubule to perturbations (Donaldson et al., 1994).

Experiments comparing apoptosis in differentiated cultures with asynchronously growing cultures provided further insight into the survival role of microtubule dynamics outside the G₂/M transition. Cells in G₀ were found to exhibit apoptosis much more quickly than cells proliferating in asynchronous cultures. This was especially prominent with taxol, which induced apoptosis within 4 hours in differentiated cells. Because these cells died quickly, before reaching a putative G₂/M checkpoint, the result further supports the conclusion that microtubule-induced apoptosis does not require the G₂/M block. The data indicate that some differentiated neuronal cells may be particularly sensitive to microtubule anomalies of the sort induced by taxol.

Data presented here suggest that the apoptotic cascade in B103 neuronal cells caused by microtubule perturbations involves signal transduction via specific changes in protein phosphorylation. A substantial body of evidence has implicated changes in serine/threonine phosphorylation in the cellular response to anti-microtubule drugs (Stewart et al., 1999, Wang et al., 1999). Here, we have found that after either colchicine or taxol treatment, there is a marked increase in a 90 kDa Triton-insoluble tyrosine phosphorylated protein. These results are consistent with reports that microtubule anomalies cause a net increase in tyrosine phosphorylation (Bershadsky et al., 1996, Wolfson et al., 1997), but decreases also have been reported (Roberge et al., 1993). More recently, microtubule perturbations have been shown to activate the mitogen-activated protein kinase (MAPK) (Extracellular signal-regulated kinase, ERK) pathway (Ding et al., 1996, Gibson et al., 1999, Yujiri et al, 1999), which suggests a possible mechanism for microtubule-associated apoptosis in non-neuronal cells. A panERK antibody (Transduction Laboratories), however, failed to bind the 90 kDa B103 protein (data not shown). Gycloheximide and genistein added to B103 cell cultures blocked upregulation of the tyrosine phosphorylated 90 kDa protein and greatly reduced apoptotic cell death. We suggest the hypothesis that microtubule perturbations in neurons cause apoptosis through a signal transduction pathway involving selective protein tyrosine phosphorylation. There are indications of similar mechanisms in fibroblasts (Bershadsky et al., 1996).

Triggered in neurons, microtubule-associated apoptosis could have a role in neurodegenerative diseases. Alzheimer's, Parkinson's, and Huntington's diseases all show evidence of both apoptosis and inappropriate microtubule function (reviewed in Vogel, 1998). Indeed, a defining feature of Alzheimer's pathology is neurofibrillary tangles made of hyperphosphorylated tau (reviewed in Goedert, 1998), a protein that normally functions to regulate microtubule turnover (Trojanowski and Lee, 1994, Panda et al., 1999). A link to Alzheimer's disease is especially intriguing. Alzheimer's pathology includes anomalies in tyrosine phosphorylation (Shapiro et al., 1991, Wood and Zinsmeister, 1991), and Alzheimer's-afflicted neurons (positive for tangles made of hyperphosphorylated tau) have highly elevated levels of Fyn (Shirazi and Wood, 1993), a protein tyrosine kinase linked to tau and hence to microtubules. Fyn binds tau and is implicated both in tau's tyrosine and serine phosphorylation (Lee et al., 1998, Lesort et al., 1999). Fyn also mediates toxic actions of the Alzheimer's-related Aβ peptide (Lambert et al., 1998, Zhang et al., 1996), which include altered tau phosphorylation (Lambert et al., 1994, Busciglio et al., 1995). Alzheimer's-type tau phosphorylation, moreover, is induced by changes in the state of microtubules (Mattson, 1992; Pope 1994). Moreover, Guise et al. (1999) have shown that cells undergoing taxol-induced apoptosis also show elevated levels of serine phosphorylated tau. These results suggest a feedback in which microtubule perturbations exacerbate tau abnormalities. Whether microtubule breakdown ultimately plays a role in activation of neuronal apoptosis in Alzheimer's, perhaps via the 90 kDa protein seen here, is a possibility for future investigation.

Abbreviations

MTs

microtubules

Bt₂cAMP

dibutyrl cyclic AMP

VEC-D1C

Video-enhanced-contrast-diffcrcntial-interference-contrast

DMEM

Dulbecco's modified Eagle media

DMSO

dimethylsulfoxide

D-PBS

Dulbecco's Phosphate buffered saline

HRP

horse radish peroxidase

PMSF

phenylmethylsuifonyl fluoride

PI

propidium iodide

NET

ncurofibrillary tangles

ADDLs

Amyloid-βeta derived diffusible ligands

AD

Alzheimer's disease