Apoptosis signal-regulating kinase 1 (ASK1) induces neuronal differentiation and survival of PC12 cells.

Apoptosis signal-regulating kinase 1 (ASK1) is a ubiquitously expressed mitogen-activated protein kinase kinase kinase that activates the c-Jun N-terminal kinase (JNK) and p38 mitogen-activated protein kinase signaling cascades. We report here that expression of constitutively active ASK1 (ASK1DeltaN) induces neurite outgrowth in the rat pheochromocytoma cell line PC12. We found that p38 and to a lesser extent JNK, but not ERK, were activated by the expression of ASK1DeltaN in PC12 cells. ASK1DeltaN-induced neurite outgrowth was strongly inhibited by treatment with the p38 inhibitor SB203580 but not with the MEK inhibitors, suggesting that activation of p38, rather than of ERK, is required for the neurite-inducing activity of ASK1 in PC12 cells. We also observed that ASK1DeltaN induced expression of several neuron-specific proteins and phosphorylation of neurofilament proteins, confirming that PC12 cells differentiated into mature neuronal cells by ASK1. Moreover, ASK1DeltaN-expressing PC12 cells survived in serum-starved condition. ASK1 thus appears to mediate signals leading to both differentiation and survival of PC12 cells. Together with previous reports indicating that ASK1 functions as a pro-apoptotic signaling intermediate, these results suggest that ASK1 has a broad range of biological activities depending on cell types and/or cellular context.

The rat pheochromocytoma cell line PC12 is a useful model of neuronal differentiation and death (1). On treatment with nerve growth factor (NGF), 1 PC12 cells differentiate with sympathetic neuron-like characteristics including neurite outgrowth. Upon binding of NGF, the cell surface receptor tyrosine kinase is activated, and the activated kinase in turn activates the small GTP-binding protein Ras followed by sequential phosphorylation and activation of three members of the mitogen-activated protein kinase (MAPK) superfamily, MAPK kinase kinase (MAPKKK)/Raf, MAPK kinase (MAPKK)/MEK, and MAPK/extracellular signal-regulated kinase (ERK) (2)(3)(4). Activation of the ERK pathway is thought to play roles during NGF-induced neuronal differentiation, since constitutively active MEK induced neurite outgrowth and selective blockade of the ERK pathway using dominant-negative MEK or the MEK inhibitor PD98059 resulted in inhibition of NGF-induced neurite outgrowth (5)(6)(7). It has also been suggested that persistence of ERK activation is critical for differentiation of PC12 cells, since transient activation of ERK induced by agents such as EGF is insufficient to induce neurite outgrowth (8 -10).
However, a contradictory finding was recently reported, in which an interfering mutant of the small G protein Rap1 which blocks the sustained phase of ERK activation did not inhibit NGF-induced neurite outgrowth, suggesting that sustained activation of ERK is not necessarily required for NGF-induced neuronal differentiation (11). In another study, NGF-induced neurite outgrowth of PC12D cells, a subline of PC12 cells, was shown to be resistant to the addition of PD98059 (12). Furthermore, neuronal differentiation of PC12 cells by treatment with bone morphogenetic protein (BMP)-2 was induced in the absence of ERK activation (13). These studies strongly suggest that signaling pathways other than the ERK cascade also contribute to neuronal differentiation of PC12 cells.
Candidates for signaling pathways regulating neuronal cell death and differentiation include two different recently identified MAPK cascades that converge on c-Jun N-terminal kinase (JNK; also known as SAPK, stress-activated protein kinase) and p38 MAP kinase. Whereas the ERK signaling cascade is generally involved in the control of cell proliferation and differentiation, JNK and p38 are preferentially activated by cytotoxic stressors such as UV radiation, x-rays, heat shock and osmotic shock, and by proinflammatory cytokines such as tumor necrosis factor (TNF) and interleukin-1 (14 -16). Importantly, activation of the JNK and/or p38 pathway has been observed in response to deprivation of trophic factors in differentiated PC12 cells and other neuronal cells, suggesting the possible involvement of JNK and p38 in neuronal death (17)(18)(19)(20). When the JNK pathway was inhibited using a dominantinterfering c-Jun mutant, neuronal death elicited by trophic factor withdrawal was decreased, implying that the JNK pathway is a necessary component of neuronal death induced by trophic factor deprivation. Moreover, mice lacking the JNK3 gene, a member of the JNK family (21), and JunAA mice, in which endogenous Jun is replaced by a dominant-negative mutant Jun allele (22), were reported to exhibit marked reduction in excitotoxicity-induced apoptosis of hippocampal neu-rons, clearly demonstrating the requirement of JNK activity for neuronal death. More recently, compound mutant mice lacking the JNK1 and JNK2 genes were shown to be embryonic lethal and have severe dysregulation of apoptosis in brain, suggesting that JNK1 and JNK2 regulate region-specific apoptosis during early brain development (23).
On the other hand, several lines of evidence have suggested that the JNK and p38 pathways are involved in neuronal differentiation. Differentiation of PC12 cells induced by expression of GTPase-deficient, constitutively active forms of the heterotrimeric G q family members, G␣ q and G␣ 16 , was accompanied by persistent activation of JNK but not of ERK (24). Expression of constitutively activated c-Jun, which partly mimicked the phosphorylated form of the protein, or retrovirusmediated overexpression of c-Jun induced neuronal differentiation of PC12 cells independently of upstream signals (25). Staurosporine, a protein kinase inhibitor and promoter of neurite outgrowth in PC12 cells, specifically induced prolonged activation of a novel JNK isoform (26). In addition, p38 was recently shown to be activated in response to NGF and to be required for NGF-induced differentiation of PC12 cells (27,28). More recently, it was reported that p38 was activated by treatment with BMP-2 in PC12 cells and that activation of p38 might be sufficient to induce neuronal differentiation of PC12 cells (29). These observations suggested that the JNK and p38 pathways mediate important biological signals not only for neuronal cell death but also for neuronal differentiation.
Apoptosis signal-regulating kinase 1 (ASK1) is a ubiquitously expressed MAPKKK that activates the SEK1-JNK and MKK3/MKK6-p38 signaling cascades (30). Overexpression of ASK1 in epithelial cells in low serum condition induced apoptosis, and in ovarian cancer cells expression of a kinaseinactive mutant of ASK1 inhibited microtubule-interfering agent-induced apoptosis, suggesting that ASK1 plays a role in the mechanism of stress-induced apoptosis (30 -32). We recently found that overexpression of ASK1 induced death of NGF-differentiated PC12 cells and primary rat sympathetic neurons (SCGs). Moreover, dominant-negative ASK1 reduced the neuronal death induced by NGF withdrawal from these cells (33). ASK1 was activated upon treatment with TNF-␣ or agonistic anti-Fas antibody, and a kinase-inactive mutant of ASK1 reduced TNF-␣-and Fas-induced JNK activation and apoptosis, suggesting that ASK1 is a pivotal component in cytokine-induced apoptosis as well (30,34,35). Considering the involvement of JNK and p38 in neuronal death and differentiation as described above, it is of great interest whether ASK1 mediates signals leading to death or differentiation of undifferentiated PC12 cells.
We report here that transient or stable expression of constitutively active ASK1 (ASK1⌬N) induces neurite outgrowth in PC12 cells. We found that p38 and to a lesser extent JNK, but not ERK, were activated by expression of ASK1⌬N in PC12 cells. Experiments using MEK inhibitors and a p38 inhibitor suggested a significant contribution of activation of p38, rather than of ERK, to the neurite-inducing activity of ASK1 in PC12 cells. We also found that ASK1⌬N induced expression of several neuron-specific proteins and phosphorylation of neurofilament proteins, confirming that PC12 cells differentiated into mature neuronal cells by expression of ASK1⌬N. Moreover, we found that ASK1⌬N-expressing PC12 cells survived in serumstarved condition. Taken together, these findings suggest that ASK1 may mediate signals leading to both differentiation and survival of undifferentiated PC12 cells.
Adenovirus Vectors-Recombinant adenoviruses encoding HAtagged ASK1 mutants and ␤-galactosidase were constructed as described (31). Nearly 100% infection of recombinant adenoviruses to PC12 cells can be achieved at a multiplicity of infection (m.o.i.) of 100 as determined by ␤-galactosidase staining (data not shown).
Neurite Outgrowth Assay-Approximately 1 ϫ 10 5 WT-PC12 cells per well were plated in 6-well cell culture plates (Becton Dickinson Labware) coated with bovine type IV collagen (Cellmatrix, Nitta Gelatin) and allowed to grow in DMEM supplemented with 10% FCS and 10% HS for overnight. Cells were then washed twice with phosphatebuffered saline (PBS) and refed with DMEM containing 1% HS and a recombinant adenovirus encoding HA-ASK1⌬N or ␤-galactosidase. Twenty-four hours after infection, cells were washed with PBS and further cultured in DMEM containing 1% HS for an additional 48 h. In the case of PC12-ASK1⌬N cells, approximately 1 ϫ 10 5 cells per well were plated in type IV collagen-coated 6-well cell culture plates and allowed to grow in DMEM supplemented with 10% FCS, 10% HS, and 500 ng/ml Tet overnight. Cells were then washed twice with PBS and refed with DMEM containing 1% HS and given concentrations of Tet.
To determine the percentage of neurite-bearing cells, the number of cells with a process longer than one cell diameter was determined and compared with the total number of cells counted.
Western Blot Analysis-Cells were lysed in a lysis buffer containing 150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS, and cell extracts were clarified by centrifugation. To detect the expression of neuronal marker proteins, cells were lysed in the same lysis buffer except with the concentration of SDS increased to 1% and sonicated briefly before clarification by centrifugation. The protein concentration of the supernatant was measured using a DC Protein Assay (Bio-Rad), and the same amounts of protein were resolved on SDS-polyacrylamide gel electrophoresis and electroblotted onto polyvinylidene difluoride membranes. After blocking with 5% skim milk in TBS-T (150 mM NaCl, 50 mM Tris-HCl, pH 8.0, and 0.05% Tween 20) for 1 h, the membranes were probed with antibodies. The antibody-antigen complexes were detected using the ECL system (Amersham Pharmacia Biotech).
Survival Assay-Approximately 5 ϫ 10 4 PC12-ASK1⌬N cells per well were plated in type IV collagen-coated 24-well cell culture plates and allowed to grow in DMEM supplemented with 10% FCS, 10% HS, and 500 ng/ml Tet overnight. The cells were washed twice with PBS and refed with serum-free DMEM containing given concentrations of Tet (day 0). Using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay-based Cell Counting Kit-8 (Dojindo), relative cell numbers were determined in triplicate by estimating the value of day 0 as 1. To examine further the roles of ASK1 in PC12 cells, we established a stable cell line expressing HA-tagged ASK1⌬N (PC12-ASK1⌬N cells), in which the expression of ASK1⌬N is under the control of a Tet-repressible promoter (see "Experimental Procedures"). After the complete removal of Tet from the culture medium, expression of ASK1⌬N was turned on and first detectable as early as 3 h later by Western blot analysis (data not shown). Twenty-four h after the reduction of Tet in the culture medium, no leaky expression of ASK1⌬N was detected in the presence of Tet (500 ng/ml), and the level of ASK1⌬N expression was tightly controlled by the concentration of Tet ( Fig. 2A). Fig. 2B shows representative morphology of PC12-ASK1⌬N cells expressing (Tet 0 ng/ml) or not expressing (Tet 500 ng/ml) ASK1⌬N 24 h after the reduction of Tet. The ASK1⌬N-expressing cells exhibited marked extension of neurites, similar to WT-PC12 cells infected with Ad-ASK1⌬N.

Constitutively Active ASK1 Induces Neurite Outgrowth in
We next examined the time and concentration effects of Tet on the induction of neurite outgrowth in PC12-ASK1⌬N cells. Twenty-four h after the reduction of Tet, approximately 15% of cells extended neurites with 50 ng/ml Tet (Fig. 2C). With 0 or 10 ng/ml of Tet, the proportion of neurite-bearing cells increased to approximately 45-50% in parallel with the expression of ASK1⌬N protein detected by Western blot (Fig. 2, A and  C). Forty eight h after the reduction of Tet, the proportion of neurite-bearing cells further increased in a time-dependent manner with 50 or 10 ng/ml Tet. However, when Tet was completely removed from the culture medium (Tet 0 ng/ml), the proportion of neurite-bearing cells peaked at 24 h but was decreased at 48 and 72 h. With a longer period of culture (up to 2 weeks), ASK1⌬N-induced neurite outgrowth was found to be most prominent at 10 ng/ml Tet (data not shown).
ASK1⌬N Preferentially Activates the p38 MAP Kinase Pathway in PC12 Cells-ASK1 activates the JNK and p38 MAP kinase pathways in COS cells (30). To explore the signaling pathway(s) leading to the induction of neurite outgrowth by ASK1, we examined which MAP kinase cascade is activated in PC12 cells by the expression of ASK1⌬N. To measure the activation status of endogenous ERK, JNK, and p38, we performed Western blot analysis using polyclonal antibodies that specifically recognize the dually phosphorylated active forms of these enzymes. We tested the dose-dependent effect of ASK1⌬N expression on activation of MAPKs in PC12-ASK1⌬N cells 24 h after the reduction of Tet (Fig. 3A). Consistent with our previous report for COS cells (30), an intense activation of p38 was readily detected with expression of ASK1⌬N. When the same membrane was reprobed with anti-phospho-JNK antibody, intense signals corresponding to phosphorylated p54 and p46 isoforms of JNK were detected in the lysate of sorbitoltreated cells, indicating the presence of intact signaling components leading to JNK in this cell line. Unexpectedly, in contrast to p38, JNK was barely activated by ASK1⌬N; only marginal activation of JNK was detected when an excess amount of ASK1⌬N was expressed at 10 or 0 ng/ml Tet. We detected no clear increase in ERK activation induced by ASK1⌬N on reprobing the same membrane with the anti-phospho-ERK antibody (data not shown; see below).
We next examined the time course effects of ASK1⌬N on JNK and p38 in PC12-ASK1⌬N cells (Fig. 3B). In this experiment, we used the same sets of Tet concentration and time points as those in Fig. 2C, in order to determine the correlation between the ASK1⌬N-induced neurite outgrowth and the activation of downstream MAPKs. With 50 ng/ml Tet, expression of ASK1⌬N increased time-dependently and reached a plateau at 48 h, and constant activation of p38 was maintained from 24 to 72 h. On the other hand, activation of JNK was first detectable 72 h after reduction of Tet. Neither p38 nor JNK was activated in the cells maintained in the presence of Tet (500 ng/ml) throughout the time course of observation. At 10 or 0 ng/ml Tet, activation of p38 peaked at 24 h and decreased thereafter in accordance with the expression level of ASK1⌬N, whereas clear activation of JNK was first detected 48 h after reduction of Tet. The preferential activation of p38 by ASK1⌬N was further confirmed in WT-PC12 cells, which we infected with Ad-ASK1⌬N or Ad-ASK1KM and tested for activation of MAPKs 24 h after infection. Although there was a quantitative difference in the adenovirus-mediated expression of ASK1⌬N and ASK1KM, expression of ASK1⌬N, but not of ASK1KM, clearly activated p38 but not JNK or ERK (Fig. 3C and data not  shown). These observations strongly suggest that ASK1⌬N preferentially activates the p38 pathway in PC12 cells. Furthermore, together with the observation that ASK1⌬N-induced neurite outgrowth was already found within 24 h after the induction of ASK1⌬N, it is likely that p38, rather than JNK or ERK, plays a primary role in ASK1⌬N-induced neurite outgrowth in PC12 cells.
Basal Activity of ERK Only Minimally Contributes to ASK1⌬N-induced Neurite Outgrowth-It has been suggested that activation of the ERK pathway plays crucial roles in NGF-induced neuronal differentiation. Since we did not detect ASK1⌬N-dependent activation of ERK in either WT-PC12 or PC12-ASK1⌬N cells, we examined whether basal activity of ERK is required for ASK1⌬N-induced neurite outgrowth (Fig.  4). In PC12-ASK1⌬N cells, a relatively high level of basal phosphorylation of ERK was observed in the presence of 500 ng/ml Tet, but no additional phosphorylation was induced by expression of ASK1⌬N. In the presence of 10 mM each of the MEK inhibitors PD98059 and U0126, the phosphorylation of ERK was partially and completely inhibited by PD98059 and U0126, respectively. When the percentage of neurite-bearing cells was determined 48 h after the induction of ASK1⌬N in the presence or absence of the MEK inhibitors, PD98059 and U0126 were found to only slightly decrease the number of neurite-bearing cells. Importantly, despite the obvious difference in ERK phosphorylation status, no significant difference was detected between PD98059 and U0126 in ability to modulate ASK1⌬N-induced neurite outgrowth. These results suggest that the ERK pathway makes little contribution, if any, to ASK1⌬N-induced neurite outgrowth in PC12 cells.
SB203580 Inhibits ASK1⌬N-induced Neurite Outgrowth-To assess the requirement of p38 for ASK1⌬N-induced neurite outgrowth, we examined the effect of SB203580, a specific inhibitor of p38, on the neurite-inducing activity of ASK1⌬N. When PC12-ASK1⌬N cells were induced to extend neurites in the presence or absence of SB203580, ASK1⌬Ninduced neurite outgrowth was inhibited in the presence of only 200 nM SB203580 (Fig. 5A). We confirmed that this concentration of SB203580 had no cytotoxic effect on PC12-ASK1⌬N cells (data not shown). Fig. 5B shows that the neuriteinducing activity of ASK1⌬N was inhibited in a dose-dependent manner by addition of SB203580. Since the presence of higher concentrations above 200 nM SB203580 in the culture medium decreased the expression of ASK1⌬N, we could not determine whether higher concentrations of SB203580 completely inhibited ASK1⌬N-induced neurite outgrowth. However, these findings strongly suggest that p38 is required for the neurite outgrowth induced by ASK1⌬N.
Characterization of ASK1-induced Differentiation of PC12 Cells-To confirm that the neurite outgrowth induced by ASK1⌬N was accompanied by neuronal differentiation of PC12 cells, we examined the expression of several neuron-specific proteins as markers of neuronal differentiation. For compari- son with NGF-induced differentiation, PC12-ASK1⌬N cells were cultured in parallel in the presence of 50 ng/ml NGF and 500 ng/ml Tet. PC12-tTA cells, the parental cell line of PC12-ASK1⌬N cells that stably expresses only the gene for a tetracycline transactivator (see "Experimental Procedures"), were also cultured under the same conditions. After culture for 10 days with or without induction of ASK1⌬N, cells were lysed and subjected to Western blot analysis using a variety of antibodies to neuron-specific markers (Fig. 6A). ␤-Tubulin isoform III (tubulin-␤III) is synthesized exclusively by neurons of higher vertebrates and increases in conjunction with the rate of neuronal differentiation (38). By using a monoclonal antibody to tubulin-␤III, we detected a marked increase in the expression of tubulin-␤III following treatment with NGF in both PC12-ASK1⌬N and PC12-tTA cells. PC12-ASK1⌬N cells but not PC12-tTA cells cultured in the reduced concentration of Tet also exhibited a dramatic increase in the expression of tubulin-␤III, indicating that the up-regulation of tubulin-␤III was induced by the expression of ASK1⌬N and was not a nonspecific effect of Tet reduction from the culture medium. Similar results were achieved using a monoclonal antibody to Tau, which is a group of microtubule-associated proteins (MAPs), and is known to be selectively expressed in unique subsets of neurons (39). Consistent with a previous report (40), expression of Tau was up-regulated in both cells by treatment with NGF. Expression of ASK1⌬N also increased Tau protein in PC12-ASK1⌬N cells. These results indicated that ASK1⌬N-induced morphological change of PC12-ASK1⌬N cells was accompanied by up-regula- FIG. 3. Effect of ASK1⌬N expression on p38 and JNK in PC12 cells. A, preferential activation of p38 by ASK1. PC12-ASK1⌬N cells were cultured in DMEM containing 1% HS and the indicated concentration of Tet for 24 h. As positive controls for activation of JNK and p38, cells were exposed to 0.5 M sorbitol for 5 min in the presence of 500 ng/ml Tet. Cells were then lysed and immunoblotted with specific antibodies to HA, phosphorylated p38, total p38, phosphorylated JNK, and total JNK. B, time course of activation of p38 and JNK with different concentrations of Tet in PC12-ASK1⌬N cells. PC12-ASK1⌬N cells were cultured in DMEM containing 1% HS and the indicated concentration of Tet for 24, 48, and 72 h. Cells were lysed at each time point and immunoblotted with specific antibodies as indicated. As a positive control, the lysate of the cells exposed to 0.5 M sorbitol for 5 min in the presence of 500 ng/ml Tet was used (indicated as S). C, activation of p38 in WT-PC12 cells by infection with Ad-ASK1⌬N. WT-PC12 cells were infected with an m.o.i. 50 (ϩ) or 200 (ϩϩ) of Ad-ASK1⌬N or an m.o.i. 100 (ϩ) or 400 (ϩϩ) of Ad-ASK1KM for 24 h. As a positive control, cells were exposed to 0.5 M sorbitol for 5 min. Cells were then lysed and immunoblotted with specific antibodies to HA, phosphorylated p38, and total p38. tion of neuron-specific components of microtubules, as was the case with NGF-induced differentiation.
We next examined the expression of neurofilament proteins, which are major elements of the neuronal cytoskeleton. Neurofilaments are composed of three intermediate filament proteins, neurofilament light (NF-L), neurofilament medium (NF-M), and neurofilament heavy (NF-H). Of these three subunits, NF-M and NF-H have extensive C-terminal domains including major phosphorylation sites (41). In untreated PC12-ASK1⌬N and PC12-tTA cells, a polyclonal antibody to NF-M detected endogenous NF-M as a doublet band. Since this antibody recognizes NF-M independently of its phosphorylation state, the observed slower and faster migrated bands were thought to correspond to phosphorylated and non-phosphorylated forms of NF-M, respectively. In both cells, treatment with NGF significantly increased both the phosphorylated and non-phosphorylated forms of NF-M. Interestingly, expression of ASK1⌬N in PC12-ASK1⌬N cells induced a band shift of NF-M without a change in total amount, suggesting that ASK1⌬N induces phosphorylation of NF-M in vivo. A monoclonal antibody that recognizes NF-H independently of its phosphorylation state detected endogenous NF-H as a single band in untreated PC12-ASK1⌬N and PC12-tTA cells. When the cells were treated with NGF, expression of NF-H was detected as a broad band, suggesting that NGF up-regulates both protein level and phosphorylation of NF-H. Notably, expression of NF-H in the ASK1⌬Nexpressing cells was also detected as a broad band with extremely slowly migrating species, suggesting that expression of ASK1⌬N strongly induces phosphorylation of NF-H. To con-firm the ASK1⌬N-induced phosphorylation of NF-H, we next used a monoclonal antibody that specifically recognizes highly phosphorylated forms of NF-H (36,37). The anti-phospho-NF-H antibody clearly detected the highly phosphorylated NF-H in the lysate of the ASK1⌬N-expressing cells, suggesting again that ASK1⌬N and/or its downstream kinase(s) phosphorylates neurofilaments.
To examine whether the adenovirus-mediated expression of ASK1⌬N also induces phosphorylation of NF-H, we infected WT-PC12 cells with Ad-ASK1⌬N or Ad-␤-galactosidase and lysed them 72 h after infection. The cell lysate was subjected to Western blot analysis using anti-NF-H and anti-phospho-NF-H antibodies (Fig. 6B). The anti-NF-H antibody, but not the antiphospho-NF-H antibody, detected basal expression of NF-H as a single band in untreated Ad-␤-galactosidase-infected cells. When the Ad-␤-galactosidase-infected cells were treated with NGF, expression of NF-H was detected as a slightly broad band, suggesting that 72 h treatment with NGF induces phosphorylation of NF-H to some extent in WT-PC12 cells. Consistent with the results for PC12-ASK1⌬N cells, NF-H expressed in Ad-ASK1⌬N-infected cells was detected as a doublet band, the upper band of which was recognized by the anti-phospho-NF-H antibody, demonstrating that adenovirus-mediated expression of ASK1⌬N strongly induced phosphorylation of NF-H in WT-PC12 cells.
To investigate further the kinetics of phosphorylation of NF-H induced by ASK1, we examined the ASK1⌬N-induced phosphorylation of NF-H over a short time course, i.e. within 24 h. In PC12-ASK1⌬N cells, phosphorylation of NF-H was first detected 12 h after the removal of Tet, following the expression of ASK1⌬N and the activation of p38, which were readily detected 8 h after Tet removal (Fig. 6C). Since upregulation of other neuron-specific proteins required several days of culture after the induction of ASK1⌬N (data not shown), phosphorylation of NF-H was a nearly immediate event caused by expression of ASK1⌬N. These results suggest that p38 activated by ASK1 or ASK1 itself directly phosphorylates NF-H.
Taken together, these findings indicate that ASK1⌬N-induced morphological change of PC12 cells was accompanied by the expression of neuron-specific proteins and intense phosphorylation of neurofilament proteins. These findings strongly suggest that ASK1⌬N-induced neurite outgrowth results from neuronal differentiation of PC12 cells.
ASK1⌬N-expressing PC12 Cells Survive in Serum-starved Condition-NGF induces not only differentiation but also survival of neurons. To test the possibility that ASK1 also mediates survival signals in PC12 cells, we tested the effect of ASK1⌬N expression on serum-starved PC12-ASK1⌬N cells. PC12-ASK1⌬N cells were cultured in different concentrations of Tet in serum-free medium for up to 7 days. Relative number of surviving cells was determined at different time points after the reduction of serum and Tet (Fig. 7A). In contrast to the gradual decrease in the number of ASK1⌬N-non-expressing cells (Tet 500 ng/ml), the number of ASK1⌬N-expressing cells (50 or 0 ng/ml of Tet) was maintained or even increased in the serum-free medium for 7 days, suggesting that ASK1⌬N-expressing cells were viable in serum-starved condition. It also appeared that moderate expression of ASK1⌬N is required to obtain the optimal survival effect, since this effect was more prominent in 50 ng/ml than 0 ng/ml Tet. Fig. 7B shows the morphology of the cells expressing (Tet 50 ng/ml) or not expressing (Tet 500 ng/ml) ASK1⌬N cultured in the serum-free medium for 5 days. ASK1⌬N-expressing cells in the serum-free medium appeared similar to those in the medium containing 1% HS (Fig. 2B), whereas the cells not expressing ASK1⌬N had an apoptotic appearance and detached from the culture plate. We also found that PC12-tTA and WT-PC12 cells infected with Ad-ASK1⌬N, but not with Ad-ASK1KM or Ad-␤-galactosidase, survived in the serum-free medium (data not shown). These observations suggest that moderate activation of ASK1 is capable of mediating survival signal in PC12 cells. ASK1 thus appeared to mediate signals leading to both differentiation and survival of PC12 cells.
SB203580 Inhibits ASK1⌬N-mediated Survival Effect-We next assessed the requirement of p38 for the ASK1⌬N-mediated survival effect by treatment with SB203580. To avoid evaluating the inhibitory effect of SB203580 on ASK1⌬N-induced differentiation, we used PC12-ASK1⌬N cells differentiated by expressing ASK1⌬N under optimal culture conditions (in DMEM containing 10 ng/ml of Tet and 1% HS) for 1 week. The differentiated PC12-ASK1⌬N cells were then cultured in the presence (50 or 200 nM) or absence of SB203580 in serumfree medium containing 50 ng/ml Tet, the concentration optimal for ASK1⌬N-mediated survival effect (Fig. 7A). At different time points after the deprivation of serum, the relative number of surviving cells was determined (Fig. 8). In the absence of SB203580, the same number of differentiated PC12-ASK1⌬N cells was maintained in the serum-free medium for 3 days, confirming that ASK1⌬N-expressing cells were viable in serum-starved condition. In the presence of SB203580, however, the number of surviving cells was decreased, indicating that SB203580 inhibited the survival effect mediated by ASK1⌬N. These findings suggest that activation of p38 is required not only for ASK1⌬N-induced neurite outgrowth but also for ASK1⌬N-mediated survival effect and that p38 is an important downstream mediator of ASK1 signaling in PC12 cells.

DISCUSSION
In the present study, we demonstrated that expression of ASK1⌬N induced neuronal differentiation of undifferentiated PC12 cells, which was characterized by neurite outgrowth, induction of neuron-specific proteins, and phosphorylation of neurofilament proteins. In addition, ASK1 appeared to transduce survival signal in PC12 cells.
Concomitant with the marked neurite outgrowth, rapid and intense activation of p38 was induced by expression of ASK1⌬N (Fig. 3). In contrast, no additional activation of ERK was induced by expression of ASK1⌬N, and inhibition of basal activ-ity of ERK did not markedly decrease the number of ASK1⌬Ninduced neurite-bearing cells (Fig. 4), suggesting that ASK1⌬N-induced neurite outgrowth does not require the activation of ERK. Given that NGF-induced differentiation of PC12 cells requires activation of the ERK pathway and is sensitive to MEK inhibitors (7, 10), we suggest that a major set of signaling pathways mediating the neurite-inducing activity of ASK1 differs from that activated in NGF-induced differentiation of PC12 cells. Expression of a dominant-negative mutant form of ASK1 (ASK1KM) consistently did not inhibit the NGF-induced neurite extension in PC12 cells (data not shown). Importantly, a p38 inhibitor inhibited ASK1⌬N-induced neurite outgrowth (Fig. 5), suggesting that the activation of p38 is required for this morphological event. Recently, p38 was reported to be required for NGF-and BMP-2-induced differentiation of PC12 cells (28,29). Although it is still unknown whether activation of p38 is sufficient to induce neurite outgrowth and/or neuronspecific markers, p38 appears to be a generally required element in neuronal differentiation of PC12 cells.
We previously reported that ASK1 activates both the JNK and p38 pathways in various cell types including COS cells, SCGs, and NGF-differentiated PC12 cells (30,33). Unexpectedly, ASK1-induced activation of JNK was relatively weak in undifferentiated PC12 cells. Whereas intense activation of p38 was easily detected within 24 h after the induction of ASK1⌬N, activation of JNK was first detected only after 48 h (Fig. 3B). Since the neurite outgrowth was induced within 24 h after the induction of ASK1⌬N, it is unlikely that the activation of JNK triggers neuronal differentiation of PC12 cells. JNK and its substrate c-Jun have been suggested to participate in death signaling of neurons, especially trophic factor deprivation-induced death. Interestingly, it was recently reported that expression of constitutively active MEKK1 (MEKK1⌬) in PC12 cells induced activation of both JNK and p38 to a similar extent, and that these cells underwent apoptosis (20), and that a dominant-interfering c-Jun mutant could inhibit MEKK1⌬-induced apoptosis, suggesting the possible involvement of JNK in death signaling in undifferentiated PC12 cells. Our observation that expression of ASK1⌬N induced differentiation, but not death, in PC12 cells may be related to the absence of intense activation of JNK. In addition, expression of ASK1⌬N in PC12-ASK1⌬N cells with the complete absence of Tet (0 ng/ml), but not with the moderate reduction of Tet (50 or 10 ng/ml), resulted in a decrease in the number of neurite-bearing cells after a longer period of culture (Fig. 2C). The differential effect of ASK1 may be explained by the relative increase in JNK activity in the complete absence of Tet (Fig. 3B). Although the extent of JNK activation induced by ASK1⌬N is insufficient to elicit death of PC12 cells, the weak activation of JNK by ASK1 might antagonize the neurite-inducing activity of the ASK1-p38 axis. It may also be possible that optimal balance between activation of the JNK and p38 pathways is required for ASK1⌬N-induced neurite outgrowth in PC12 cells.
We found that expression of ASK1⌬N up-regulated neuronspecific components of microtubules and strongly induced phosphorylation of neurofilament proteins (Fig. 6), suggesting that neurite outgrowth induced by expression of ASK1⌬N indeed resulted from neuronal differentiation of PC12 cells. Surprisingly, the extent of phosphorylation of NF-H induced by expression of ASK1⌬N was much more prominent than that induced by treatment with NGF. Phosphorylation of NF-H is thus a major characteristic of ASK1⌬N-induced differentiation of PC12 cells. NF-M and NF-H are among the most highly phosphorylated proteins in the nervous system. Although the role of their phosphorylation is not fully understood, phosphorylated NF-M and NF-H are major components of the cytoskeleton in many types of mature neurons and are particularly abundant in myelinated axons (41). Most of the phosphorylation sites in NF-M and NF-H are located in the C-terminal tail region containing multiple copies of Lys-Ser-Pro (KSP) motif (42,43). To date, several kinases including cyclin-dependent kinase 5 (Cdk5) (44), glycogen synthase kinase-3 (GSK-3) (45), ERK1 and -2 (46,47), and JNK1 (48,49) have been implicated in the phosphorylation of NF-M and NF-H. Relevant to the present study, it is of interest that proline-directed MAP kinases ERK and JNK are reported to phosphorylate directly NF-H (47,48), since p38, another proline-directed kinase, may phosphorylate NF-H in ASK1⌬N-expressing PC12 cells. Together with the observation that the phosphorylation of NF-H was detectable soon after the activation of p38 (Fig. 6C), it is possible that p38, rather than JNK, activated by ASK1 directly phosphorylates NF-H. Alternatively, other unidentified kinase(s) activated by ASK1 or ASK1 itself may phosphorylate NF-H.
The observation that ASK1⌬N-expressing cells could survive in serum-starved condition suggests another unexpected function of ASK1 (Fig. 7). We found that expression of ASK1⌬N induced apoptosis in NGF-differentiated PC12 cells and SCGs (33). Although these findings appear to contradict those of the present study, we suggest that the expression level of ASK1⌬N could determine the cell fate; moderate expression of ASK1⌬N induces differentiation and survival, whereas excessive ASK1 activity induces apoptosis in PC12 cells. In fact, we observed that excess overexpression of ASK1⌬N induced apoptosis even in undifferentiated PC12 cells (data not shown). On the other hand, it is difficult to compare the precise expression level of transfected ASK1⌬N in undifferentiated PC12 cells with that in NGF-differentiated PC12 cells or SCGs; we also consider the possibility that ASK1 may preferentially mediate apoptosis signal in NGF-differentiated PC12 cells and SCGs due to the different cellular context. We are currently investigating the difference in the ASK1-induced activation of downstream signaling pathways between undifferentiated and NGF-differentiated PC12 cells.
Similar to the ASK1⌬N-induced neurite outgrowth, ASK1⌬N-mediated survival effect was inhibited by a p38 inhibitor (Fig. 8), suggesting that the activation of p38 plays crucial roles in both differentiation and survival induced by ASK1. However, in the absence of data in primary neurons, we cannot exclude a possibility that the differentiation/pro-survival effects of ASK1 might be a PC12-specific phenomenon. Future experiments using undifferentiated neuronal precursor cells may enable understanding of diverse roles of ASK1 in neuronal physiology.