Induction of neurite outgrowth by interleukin-6 is accompanied by activation of Stat3 signaling pathway in a variant PC12 cell (E2) line.

PC12-E2 cells, a stable variant subcloned from native cell populations, produce neurites in a rapid, transcription-independent manner upon exposure to nerve growth factor (NGF) or basic fibroblast growth factor (bFGF). They also give a similar morphological response to interleukin-6 (IL-6), which is, however, transcription-dependent and with a slower onset, a phenomenon basically not observed in native PC12 cells. The response profile of PC12-E2 cells to NGF and bFGF is similar to that observed for native PC12 cells pre-exposed (primed) to NGF, and such cells also respond to IL-6 in a fashion indistinguishable from PC12-E2 cells. Mechanistically, NGF and bFGF induce a sustained phosphorylation and activation of ERK1 and ERK2 in both cells, while IL-6 produces only a transient and weak tyrosine phosphorylation. However, it does stimulate a prolonged and biphasic tyrosine phosphorylation and nuclear translocation of Stat3 (signal transducers and activators of transcription 3; at least 24 h) and, to a lesser extent, Stat1. Gel shift and supershift analyses confirm that IL-6 predominantly activates Stat3 (and some Stat1) and stimulates sis-inducible element binding activity. Other members of the same cytokine subfamily, including ciliary neurotrophic factor and leukemia inhibitory factor, also cause a transient initial phase of tyrosine phosphorylation and activation of Stat1 and Stat3 (up to 1 h) but fail to stimulate a second phase of response and do not produce significant neurites. These results suggest that sustained signaling of either STAT or ERK pathways in PC12-E2 cells leads to induction of neuronal differentiation. However, only the latter is effective in native PC12 cells as the activation of Stat3 and Stat1 in native PC12 cells by IL-6 fails to induce neuronal differentiation. Thus, the response of PC12-E2 cells to IL-6 suggests the constitutive expression of a required factor(s) for differentiation, that is induced in native PC12 cells by NGF or bFGF (possibly by ERK activation), but not by IL-6 via Janus kinase/STAT activation. This factor(s), which has a sufficient half-life to allow primed cells to remain responsive to IL-6 for several days, is necessary but not sufficient for differentiation (as measured by neurite proliferation) to occur.

PC12-E2 cells, a stable variant subcloned from native cell populations, produce neurites in a rapid, transcription-independent manner upon exposure to nerve growth factor (NGF) or basic fibroblast growth factor (bFGF). They also give a similar morphological response to interleukin-6 (IL-6), which is, however, transcriptiondependent and with a slower onset, a phenomenon basically not observed in native PC12 cells. The response profile of PC12-E2 cells to NGF and bFGF is similar to that observed for native PC12 cells pre-exposed (primed) to NGF, and such cells also respond to IL- 6

in a fashion indistinguishable from PC12-E2 cells. Mechanistically, NGF and bFGF induce a sustained phosphorylation and activation of ERK1 and ERK2 in both cells, while IL-6 produces only a transient and weak tyrosine phosphorylation. However, it does stimulate a prolonged and biphasic tyrosine phosphorylation and nuclear translocation of Stat3 (signal transducers and activators of transcription 3; at least 24 h) and
, to a lesser extent, Stat1. Gel shift and supershift analyses confirm that IL-6 predominantly activates Stat3 (and some Stat1) and stimulates sis-inducible element binding activity. Other members of the same cytokine subfamily, including ciliary neurotrophic factor and leukemia inhibitory factor, also cause a transient initial phase of tyrosine phosphorylation and activation of Stat1 and Stat3 (up to 1 h) but fail to stimulate a second phase of response and do not produce significant neurites. These results suggest that sustained signaling of either STAT or ERK pathways in PC12-E2 cells leads to induction of neuronal differentiation. However, only the latter is effective in native PC12 cells as the activation of Stat3 and Stat1 in native PC12 cells by IL-6 fails to induce neuronal differentiation. Thus, the response of PC12-E2 cells to IL-6 suggests the constitutive expression of a required factor(s) for differentiation, that is induced in native PC12 cells by NGF or bFGF (possibly by ERK activation), but not by IL-6 via Janus kinase/STAT activation. This factor(s), which has a sufficient half-life to allow primed cells to remain responsive to IL-6 for several days, is necessary but not sufficient for differentiation (as measured by neurite proliferation) to occur.
PC12 cells are derived from a rat pheochromocytoma and respond to external stimuli in different ways (1,2). Some growth factors, such as nerve growth factor (NGF) 1 or bFGF, induce a reversible neuronal-like morphology that is preceded by a cycle of cell division. Other factors, such as EGF, insulinlike growth factor-1, or IL-6, show little or no differentiative activity with these cells, although EGF and insulin-like growth factor-1 do enhance mitotic responses to some degree.
It has been well established that NGF and FGF function by activating their receptor tyrosine kinases, triggering several signaling pathways, including the RAS-dependent cascade, that leads to the initiation of transcription of immediate early genes and the subsequent coordinated expression of delayed response genes (3)(4)(5)(6)(7). The 24 -36 h required for the onset of the differentiative state underscores the transcriptional requirements for this response in PC12 cells and the cell division observed may represent the "default" activity of these cells that is initially induced by the factor until the delayed response gene products can achieve the cessation of division. EGF, which also functions via a tyrosine kinase-containing receptor, seems to stimulate this "basal" response to some degree but fails to cause neurite proliferation in native PC12 cells, apparently because the duration of the signals generated at the plasma membrane are insufficient to induce the necessary level of response (8 -11).
In contrast to the neurotrophic receptors, the receptors for IL-6 belong to the family of cytokine receptors that do not possess intrinsic tyrosine kinase activity. Instead, the biological response of IL-6 is mediated by a receptor complex that is formed by a dimer of the ligand binding ␣ subunit, and a dimer of the signal transducing component gp130 (12,13). It has been shown that the JAK family of tyrosine kinases (Jak1, Jak2, and Tyk2) is associated with the dimerized gp130 (14 -19). Activation of the JAKs by IL-6 induces the tyrosine phosphorylation and subsequent nuclear translocation of the STAT proteins including Stat1 and Stat3 (20,21). Both Stat3 homodimers and Stat3-Stat1 heterodimers bind to IL-6 response elements identified in the promoter regions of various IL-6-induced acute phase proteins, e.g. ␣ 2 -macroglobulin, and immediate-early genes, such as interferon regulatory factor 1, intercellular adhesion molecule 1, c-fos and junB, that stimulate transcription (22)(23)(24). A similar signaling cascade is also induced by other members of the cytokine subfamily sharing the same signal transducer gp130 (␤-signal transducing receptor component), including LIF, oncostatin M, interleukin-11, and CNTF; STAT proteins can also be affected by polypeptide growth factors, such as platelet-derived growth factor and EGF (15,18).
Increasing evidence suggests that the actions of cytokines are not limited to immune and inflammatory responses, and IL-6 has been specifically connected to the activation of osteoclasts and bone resorption (25,26) and various pathophysiological conditions in the nervous system (27)(28)(29). It also prevents cell death and promotes survival in some cerebral neurons (30 -32). In transgenic mice, overexpression of IL-6 in the central nervous system causes neuronal degeneration and other neuronal and glial abnormalities (33,34). Moreover, it has been shown that IL-6 is induced in axotomized sensory neurons, suggesting a function in neuronal regeneration (35).
Recently, a PC12 cell variant (designated E2) that is characterized by very rapid responses to NGF and FGF has been described (36). These cells, in contrast to native PC12 cells, also respond differentially to IL-6 in a manner that is morphologically indistinguishable from that produced by NGF or FGF, albeit that onset is somewhat slower. It is demonstrated in the present study that the morphological differentiation of PC12-E2 cells by IL-6 is not mediated by activation of a RASdependent signaling pathway. Instead, the stimulation of tyrosine phosphorylation and nuclear translocation of Stat3 and Stat1 proteins suggests that the JAK-STAT pathway is involved thus demonstrating that that form of signaling can also lead to neurite proliferation under appropriate conditions in a PC12 type cell. However, activation of the STAT pathway alone in native PC12 cells is not sufficient for the induction of neurite outgrowth indicating an unknown factor(s), constitutively expressed in E2 cells and presumably induced by NGF or FGF in native PC12 cells, is required for the differentiative action of IL-6. The robust differentiative response to IL-6 of native PC12 cells, which have been previously exposed to NGF ("primed"), supports this view.
Cell Culture-PC12 cells were obtained from Dr. D. Schubert (Salk Institute, San Diego, CA). The variant E2 cells were isolated from the parental line and have been characterized (36). Cells were maintained as monolayer cultures in 162-cm 2 tissue culture flasks (Costar, Cambridge, MA) in DMEM containing 10% heat-inactivated horse serum (Life Technologies, Inc.), 5% fetal calf serum (Life Technologies, Inc.) and 1% penicillin-streptomycin solution (complete medium) in a 5% CO 2 humidified atmosphere. Cells were subcultured once a week by shaking the flask and replating in a 1:4 to 1:6 ratio. The medium was changed every 3 days. The HepG2 cells were obtained from ATCC (HB8065) and maintained in DMEM and Ham's modified F-12 medium (1:1) containing 10% fetal calf serum and 1% penicillin-streptomycin. Cells were subcultured once a week by trypsinization and replating in a 1:4 ratio.
Neurite Outgrowth-The neurite outgrowth assay was performed as described previously (39). For the time course and dose response of IL-6-induced neurite outgrowth, cells were grown in complete media for 6 h and changed to low serum media (DMEM containing 1% horse serum and 1% penicillin-streptomycin) with the addition of growth factor(s) for the specified duration. For neurite regeneration experiments, cells were primed in complete media containing 100 ng/ml NGF for 7-9 days. Neurites were sheared from cell bodies by vigorously shaking the flask followed by three brief centrifugations and resuspension. Cells were then replated as described above for neurite outgrowth.
Immunoprecipitation and Immunoblot Analyses-Cells were grown in 100-or 150-mm collagen-coated dishes for 2 days in complete media and continued in culture for 1 additional day in low serum media. Growth factor was added directly to the media during the last period of cell culture for the indicated duration. For preparation of total cell extracts, cells were rinsed twice with ice-cold phosphate-buffered saline containing 200 M orthovanadate and 0.5 mM PMSF and solubilized in 100 or 150 l 2x lysis buffer (pH 7.4) containing 1% Nonidet P-40, 50 mM Hepes, 10% glycerol, 1 mM EDTA, 30 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 0.5 mM PMSF, 10 g/ml leupeptin, and 10 g/ml aprotinin. Cell lysates were cleared by centrifugation at 10,000 g for 15 min. Crude nuclei were isolated by sucrose gradient centrifugation as described by Ausubel et al. (40) with the addition of proteases and phosphatases inhibitors (1 mM sodium orthovanadate, 0.5 mM PMSF, 10 g/ml leupeptin, and 10 g/ml aprotinin) to the lysis buffer and the sucrose gradient. The nuclei were then extracted with buffer containing 10 mM Tris-HCl, pH 8.0, 420 mM NaCl, 20% glycerol, 0.2 mM sodium orthovanadate, 0.5 mM PMSF, 10 g/ml leupeptin, and 10 g/ml aprotinin for 30 min on ice, followed by centrifugation at 16,000 ϫ g for 20 min. The supernatant was used as nuclear extract. Protein concentration was determined by the Bradford protein assay using bovine serum albumin as standard (41).
For determining the time course of stimulation of tyrosine phosphorylation, 120 g of cellular proteins was separated by 7.5% SDS-PAGE, transferred to an Immobilon-P membrane (Millipore Corp., Bedford, MA), and probed with anti-phosphotyrosine antibody (4G10) and 125 Ilabeled goat anti-mouse IgG. For STAT protein immunoprecipitations, 5 mg of cellular protein was incubated for 2 h at 4°C with 3 g/ml anti-STAT antibody. The resulting immune complexes were collected by the addition of 20 l of 50% protein A-Sepharose (Pharmacia Biotech Inc.) and incubated at 4°C for additional 2 h. After incubation, the protein A-Sepharose immune complexes were subjected to three washes with lysis buffer and two washes with 20 mM Hepes buffer (pH 7.4).

FIG. 1. Induction of neurite outgrowth by IL-6, CNTF, and LIF in PC12-E2 cells.
Cells were grown in complete media for 6 h and changed to low serum media in the presence of 10 ng/ml IL-6, 30 ng/ml CNTF, 30 ng/ml LIF, or in the absence of factors (control) for 1 or 2 days. Bar, 100 m.
Proteins were eluted from Sepharose beads with 30 l of 2 ϫ Laemmli's sample buffer and boiled for 3 min. For immunoprecipitating ERK, 2 mg of cellular protein was incubated for 2 h at 4°C with 5 l of anti-ERK agarose (SC93AC). After incubation, the immune complexes were treated the same as described above. The immunoprecipitate was separated by a 7.5% (STAT proteins) or 12.5% (ERK) SDS-PAGE, transferred to an Immobilon-P membrane, and probed with anti-phosphotyrosine antibody (4G10) and 125 I-labeled goat anti-mouse IgG.

RESULTS
Neurite Outgrowth in PC12-E2-Although a slow and modest induction of neurites in one PC12 cell line after 1 week of stimulation with IL-6 has been reported by Satoh et al. (43), no other PC12 line (including the native PC12 cells used in this study) has been reported to produce neurites following treatment by IL-6. However, in the recently isolated PC12 variant E2 cells, IL-6 induces a rapid and robust neurite outgrowth (36) (Fig. 1). The response is dose-dependent (0.3-30 ng/ml) and time-dependent (Fig. 2); neurites extending longer than 2 cell bodies appear within a day, and maximal response is reached in 2 days. The long branching neurites that form a network-like pattern induced by IL-6 after 2-3 days are indistinguishable from those produced by NGF. However, the NGF-induced neurite outgrowth in these cells is mediated by a transcriptionindependent mechanism and the morphological changes are observed within 1-2 h after the addition of the growth factor (36). In contrast, the effect of IL-6 is abolished by actinomycin D treatment and thus appears to be transcription-dependent (data not shown). In general, the IL-6-induced morphological changes are seen in 6 -8 h after addition of the factor. The differential sensitivity to actinomycin D inhibition in E2 cells suggests that the responses of IL-6 and NGF are most likely mediated by two independent mechanisms.
Although not as potent as IL-6, other members of the IL-6 cytokine subfamily, including LIF and CNTF (1-100 ng/ml), are also able to promote morphological changes in E2 cells (Fig.  1). The majority of cells extend short processes (less than 2 cell bodies) after more than 3 days of treatment; however, only FIG. 3. Induction of neurite outgrowth by IL-6 in PC12 cells primed by previous exposure to NGF. PC12 cells were primed in complete media in the presence of NGF (100 ng/ml) for 7 days, replated in low serum media, and then treated with or without IL-6 (30 ng/ml) or NGF (100 ng/ml) for 1 or 2 days. Native PC12 cells were cultured in complete media alone and replated in low serum media in the presence or absence of IL-6 (30 ng/ml) or NGF (100 ng/ml) for 2 days. Bar, 100 m.
30 -40% of cells bear neurites longer than 2 cell bodies at the highest concentration tested (100 ng/ml).
IL-6 Stimulates Neurite Outgrowth in Primed PC12 Cells-The time course of NGF and bFGF responses by PC12-E2 cells resembles that in native PC12 cells previously exposed ("primed") to NGF (36,44). One possibility is that an unknown factor(s) that is induced in primed cells by NGF and bFGF is constitutively expressed in PC12-E2 cells and thus allows the transcription-independent, rapid responsiveness of these cells to growth factors. If such a factor(s) allows the IL-6 response of PC12-E2 cells, then IL-6 should stimulate primed cells in a similar fashion. Indeed, IL-6 is able to induce a dose-dependent neurite outgrowth in primed PC12 cells with a similar EC 50 as in PC12-E2 cells (Fig. 3). The response reaches a maximum after 2 days of IL-6 treatment, and, despite the continued presence of IL-6, neurites eventually degenerated within 6 -7 days. As in E2 cells, a latency of 6 -8 h is required for the morphological changes induced by IL-6 as opposed to 1-2 h for the NGF-induced neurite regeneration. The effect of NGF in primed cells is transcription-independent (44); the effect of IL-6 is abolished in the presence of actinomycin D (1 M). To rule out the possibility that the effect of IL-6 is due to the synergistic interaction of IL-6 with the residual NGF left over from priming, the effect of NGF antiserum on these cells was also tested. Affinity-purified NGF antibody (120 g/ml), which inhibits NGF (100 ng/ml)-induced neurite outgrowth in primed cells, failed to affect the IL-6-induced neurite regeneration (data not shown). These data suggest that IL-6 is capable of promoting neurite regeneration in PC12 cells in the absence of NGF provided that the expression of an unknown factor(s) has been induced by previous exposure to NGF.
Kinetics of Protein Tyrosine Phosphorylation-Increased protein tyrosine phosphorylation following stimulation by IL-6, CNTF, oncostatin M, or LIF in many cell types is associated with their physiological functions. These factors have been shown to stimulate tyrosine phosphorylation of many intracellular substrates including phospholipase C␥, phosphoinositol 3-kinase, phosphotyrosine phosphatase (PTP1D), pp120, Shc, Stat1, Stat3, ERK1, and ERK2 (45). Several proteins of similar molecular size are also tyrosine-phosphorylated by IL-6 in PC12 or PC12-E2 cells (Figs. 4 and 5). Among these proteins, it has been found that one of the most strongly phosphorylated protein, p89, co-migrated with the newly identified Stat3. This protein is modified within 5 min and reaches a maximum level of tyrosine phosphorylation within 15 min. The mobility of Stat3 is reduced after IL-6 stimulation, which results in a doublet on some blots, and also reaches a maximum in 15 min. Tyrosine phosphorylation and the mobility shift returns to near basal level within approximately 1 h. (The pattern can be more easily seen with additional time points around 1 h.) However, a prolonged second phase of tyrosine phosphorylation occurs after 1 h and is sustained for at least 24 h. The mobility shift of p89 does not correlate with the level of tyrosine phosphorylation and is most likely the result of serine phosphorylation FIG. 4. Kinetics of protein tyrosine phosphorylation after stimulation by IL-6. PC12 and PC12-E2 cells were grown in complete media for 2 days and continued in culture in low serum media for 1 d before the treatment. 120 g of total cellular protein from cells treated with IL-6 (30 ng/ml) for different durations was separated by a 7.5% SDS-PAGE. Protein tyrosine phosphorylation was detected by immunoblotting probed with anti-phosphotyrosine antibody (4G10) and 125 I-goat anti-mouse IgG. The phosphotyrosine blots were stripped and reprobed with anti-Stat3 polyclonal antibody and detected by ECL. One protein band, Stat3, was detected by this antibody and co-migrates with the p89 protein band shown in the anti-P-Tyr blot. The anti-Stat3 blots were again stripped and reprobed with anti-Stat1 polyclonal antibody and 125 I-protein A. Two protein bands, Stat1␣ and Stat1␤, were detected by this antibody, and Stat1␣ co-migrates with the p91 protein band shown in the anti-phosphotyrosine blot. ERK1 and ERK2 co-migrate with two protein bands, p44 and p42 (blots not shown). (46 -49). The phosphotyrosine blot was stripped and reprobed with the anti-Stat3 antibody to reveal the protein level, which shows a notably increased expression of Stat3 after 2 h of IL-6 treatment, and it is up-regulated by severalfold after 24 h of IL-6 treatment. The Stat1 protein has also been implicated in the function of these cytokines. However, the phosphorylation of Stat1␣, shown by the Stat1 protein blot to migrate at the position of p91, is not as pronounced as p89 and is not detectable on phosphotyrosine blots. In addition, unlike the expression of Stat3, IL-6 treatment (within 24 h) does not affect the expression of Stat1.
The phosphotyrosine blot shows that the pattern, the intensity, and the kinetics of phosphorylation induced by IL-6 in both PC12 and E2 cells are similar (Fig. 4). Expression of some key signaling molecules including Stat1, Stat3, ERK1, and ERK2 are not significantly different (Figs. 6 and 7). These data indicate that the lack of differentiative response of IL-6 in native PC12 cells cannot be simply explained by some missing components of the signaling machinery.
As with IL-6, CNTF and LIF induced a rapid and biphasic tyrosine phosphorylation of p89; however, the effect is not as prolonged as with IL-6 (Fig. 5). The phosphorylation of p89 was greatly reduced after 4 h for either CNTF or LIF treatments. No detectable tyrosine phosphorylation of p89 and only a slightly increased expression of Stat3 were observed after 24 h of treatment. In addition, as with IL-6, p42 and p44 (co-migrate with ERK1 and ERK2) are only weakly and transiently tyrosine-phosphorylated. CNTF and LIF are likely to stimulate the serine phosphorylation of Stat3, as indicated by a timedependent (5-60 min) mobility shift of Stat3 seen on the immunoblot.
Tyrosine Phosphorylation of ERK-Consistent with other re-

FIG. 5. Kinetics of protein tyrosine phosphorylation after stimulation by LIF and CNTF in PC12-E2.
Cells were treated with LIF (30 ng/ml), CNTF (30 ng/ml), and IL-6 (10 ng/ml) for different durations. Other conditions were the same as in Fig. 4. C represents the cell extract prepared from untreated cells at 24 h. FIG. 6. NGF but not IL-6 stimulates sustained protein tyrosine phosphorylation of ERK1 and ERK2. Immunoprecipitates of ERK1 and ERK2 from cells stimulated with 10 ng/ml IL-6 or 100 ng/ml NGF were separated by 12.5% SDS-PAGE and immunoblotted with antiphosphotyrosine antibody. The phosphotyrosine blots were stripped and reprobed with affinity purified anti-ERK polyclonal antibody and 125 I-protein A as shown on the lower panels.
ports that the activation of the RAS/ ERK pathway is a major component of the differentiative effect of NGF (50, 51), a strong and prolonged activation of ERK1 and ERK2 by NGF in PC12-E2 cells was observed (36). To further characterize the effect of IL-6 on tyrosine phosphorylation and activation of ERKs, phosphotyrosine blots of ERK immunoprecipitates from cells treated with either IL-6 or NGF were prepared. As compared with NGF, IL-6 stimulates only a weak and transient tyrosine phosphorylation of ERK2 in both cells (Fig. 6). In addition, the effect of IL-6 on ERK appears to be much weaker than that induced by EGF in both cells (62). The low level of ERK response by EGF has been suggested to account for its small neurite promoting effect (10). These data further support the view that IL-6 activates a distinct signaling pathway and activation of the ERK pathway is not responsible for the action of IL-6.
Tyrosine Phosphorylation of Stat1 and Stat3-To determine whether activation of STAT proteins could be an alternative pathway for the induction of differentiation by IL-6 in E2 cells, the regulation of tyrosine phosphorylation of Stat1 and Stat3 by IL-6 and NGF was examined. Immunoprecipitates of Stat1 and Stat3 were prepared from PC12, PC12-E2, or HepG2 (as a positive control) cell extracts stimulated with growth factors for different durations. IL-6 stimulated prolonged tyrosine phosphorylation of Stat3 in both cells (Fig. 7A). However, the effect on Stat1 reached a maximum within 15 min and mostly disappeared within 1 h (Fig. 7B). As reported for HepG2 cells, both Stat1 and Stat3 are equally and strongly phosphorylated following IL-6 stimulation for 15 min (16,20,21). When the responses in PC12 and PC12-E2 cells were compared to that in HepG2 cells, it was found that the Stat3 was at least equally phosphorylated in the PC12 samples as in the HepG2 cells; however, in contrast, the Stat1 appeared to be much less phosphorylated than in HepG2. Similar levels of phosphorylation of Stat1 and Stat3 were also observed following CNTF or LIF stimulation for 15 min (data not shown). NGF, on the other hand, had no effect on tyrosine phosphorylation of either Stat1 or Stat3 within 24 h ( Fig. 7 and data not shown). It was noted that 5-15 min of NGF treatment retarded the mobility of Stat3, which is likely to be a result of serine phosphorylation as seen after IL-6 treatment (Fig. 7A).
Nuclear Localization of Stat1 and Stat3-Immunoblot analyses using both phosphotyrosine and STAT protein antibodies were further performed to examine whether the STAT proteins are translocated to the nucleus after IL-6 stimulation. The tyrosine-phosphorylated Stat3 protein appears in nuclear extracts of IL-6-treated PC12, PC12-E2, or HepG2 cells, but not in the untreated controls (Fig. 8). The identity of this protein was further confirmed by a Stat3 blot of immunoprecipitates from IL-6-treated nuclear extracts. The phosphorylated Stat1 protein has also been detected in the nucleus of IL-6-treated HepG2 cells; however, it is present at greatly reduced levels in the nucleus of IL-6-treated PC12 or PC12-E2 cells (Fig. 8). Thus, these data show that the Stat3 protein and, to a much lesser extent, the Stat1 protein are not only tyrosine-phosphorylated but also translocated to the nucleus after IL-6 treatment.
SIE Binding Activity-To determine if IL-6 is capable of inducing SIE binding activity, EMSA analyses using a 32 Plabeled high affinity SIE element and nuclear extracts from cells treated with IL-6 for different durations were performed. Treatment of PC12-E2 cells with IL-6 leads to the formation of three protein-DNA complexes, which co-migrate with SIFA, -B, and -C in IL-6-treated HepG2 cells (Fig. 9, A and B). Consistent with previous reports (21), the slowest migrating SIFA contains the Stat3 homodimers, as indicated by the supershift analyses. It is the strongest complex, which is rapidly formed within 5 min of IL-6 stimulation, diminishes to near basal levels within 60 min and is followed by a prolonged second phase of activation, lasting for at least 24 h. In contrast to the strong and sustained activation of SIFA, SIFB, which contains the heterodimers of Stat1 and Stat3, is less intense and is only transiently induced within 60 min of IL-6 stimulation. SIFC, which contains the homodimers of Stat1, is very weakly and transiently induced by IL-6 in E2 cells. These data suggest that IL-6 preferentially activates Stat3 in PC12 and PC12-E2 cells.
EMSA and supershift analyses were also performed using nuclear extracts from LIF-treated E2 cells (Fig. 9, C and D). The pattern of STAT-DNA complexes and their maximum activation are similar to that of IL-6, i.e. SIFA is the strongest, followed by SIFB and SIFC. On the other hand, the kinetics of induction are quite different from those of IL-6. Using nuclear extracts from LIF-treated cells, both SIFA and SIFB displayed biphasic activation. As with the kinetics of LIF-induced protein tyrosine phosphorylation of Stat3, activation of SIFA is weaker and less prolonged than by IL-6. There is no detectable SIFA after 24 h of LIF treatment. In contrast, activation of SIFB is stronger and of longer duration than that induced by IL-6.
Activation of SIFC also appears to be stronger than by IL-6. These data suggest that LIF has a stronger effect on the activation of Stat1 than does IL-6, but that Stat3 is still preferentially activated.

DISCUSSION
In PC12-E2 cells, IL-6 induces a morphological differentiation that is indistinguishable from that of NGF, but the underlying mechanisms of action for these two factors are clearly different. In contrast to native PC12 cells, the neurite outgrowth induced by NGF is mediated by a transcription-independent mechanism (36); the neurite induction of E2 cells by IL-6 is transcription-dependent. This difference probably accounts for the shorter latency period for the appearance of neurites in E2 cells after NGF treatment than after IL-6 treatment. This conclusion is further supported by the fact that IL-6 does not have a significant effect on tyrosine phosphorylation and/or activation of the ERKs, which have been shown to be essential for the induction of neuronal differentiation by NGF or bFGF in PC12 cells (50,51), suggesting that this pathway is not likely to play a major role in the induction of neurite outgrowth by IL-6 in PC12-E2 cells. The evidence presented is consistent with the view that the alternative JAK/STAT signaling pathway, well characterized for a number of cytokines, is utilized instead.
IL-6 receptors do not possess intrinsic tyrosine kinase activity but instead recruit and activate JAK kinases, which, in turn, induce tyrosine phosphorylation and translocation of FIG. 8. Tyrosine phosphorylation and translocation of Stat1 and Stat3 to the nucleus after IL-6 stimulation. Nuclear extracts were prepared from PC12-E2 (100 g of protein), PC12 (50 g of protein), and HepG2 cells (50 g of protein) stimulated with IL-6 (30 ng/ml) or EGF (5 ng/ml) for different durations. Total cellular extracts (50 or 100 g of protein) were from untreated cells. Immunoprecipitates (IP) of Stat3 were from nuclear extracts of E2 cells stimulated with or without IL-6 for 15 min. These samples were separated by a 7.5% SDS-PAGE and immunoblotted with anti-phosphotyrosine antibody as shown on upper panels. The phosphotyrosine blots were stripped and reprobed first with anti-Stat3 polyclonal antibody. One protein band, Stat3, was detected in untreated total cell extracts, IL-6-treated nuclear extracts, and Stat3 immunoprecipitates of treated nuclear extracts. Stat3 co-migrates with the p89 protein band shown in anti-phosphotyrosine. The anti-Stat3 blots were stripped again and reprobed with anti-Stat1 antibody. Two protein bands, Stat1␣ and Stat1␤, were detected in untreated total cell extracts, and Stat1␣ was detected in IL-6-treated nuclear extracts. Stat1␣ co-migrates with the p91 protein band shown in the anti-phosphotyrosine blot.
FIG. 9. Stimulation of DNA binding activity by IL-6 and LIF. A, nuclear extracts were prepared from HepG2 or PC12-E2 cells treated with IL-6 (30 ng/ml) for 15 min. EMSA was performed with a 32 P-labeled SIE probe. Competing unlabeled oligonucleotides were present at a 100-fold STAT proteins to the nucleus. These proteins serve directly as transcriptional factors for the induction of specific genes. As in other cells, IL-6 induces tyrosine phosphorylation and activation of at least two STAT proteins, Stat3 and Stat1, in PC12 and PC12-E2 cells, with Stat3 being preferentially activated in both cases. In general, there is a more sustained and stronger stimulation of tyrosine phosphorylation of Stat3 than Stat1, the presence of a higher amount of Stat3 in the nucleus, and a prolonged and more intense formation of SIFA than SIFB and SIFC after IL-6 treatment. It is interesting that IL-6 treatment leads to a biphasic activation of Stat3 but not Stat1 in PC12 cells in contrast with a strong but transient activation of both factors in HepG2 cells (20,52). Moreover, a gradual up-regulation in the expression of Stat3 was observed, which is in good temporal correlation with the delayed increase in the tyrosine phosphorylation of Stat3 and the formation of Stat3-DNA complexes. In addition to a severalfold increase in Stat3 expression, it is possible that the level of other components of the STAT pathway may also be altered. It has been reported that the IL-6 receptor and signal transducer gp130 are internalized and desensitized after 1 h of IL-6 treatment followed by a feedback regulation and rapid resynthesis (53,54). The reappearance and phosphorylation of gp130 in the continued presence of IL-6 may subsequently contribute to the second phase of tyrosine phosphorylation of JAK and Stat3. Similar biphasic responses to IL-6 have not been shown in other cell systems. Although an up-regulation of STAT proteins by interferon-␣ or interferon-␥ has been reported (55,56), the increased expression of Stat1 and Stat2 in interferon-treated cells does not lead to an increased transcriptional activity. Collectively, the data presented indicate that the response of PC12-E2 cells by IL-6 is most likely mediated by the Stat3 signaling pathway.
It is clear that the activation of the STAT signaling pathway by itself is not sufficient for the induction of differentiation in these cells. In every respect, the relative cellular concentration and activation of each STAT protein by IL-6 is similar in both PC12 and PC12-E2 cells. However, IL-6 is able to induce morphological differentiation only in the E2 cells. It has previously been suggested that an unidentified differentiation-related factor(s) may be constitutively expressed in PC12-E2 cells. Such a factor may be induced by neurotrophic factors such as NGF and bFGF but not IL-6 in native cells (36). The ability of these factors to cause a sustained activation of the ERKs (which IL-6 cannot) suggests that the synthesis of this factor(s) depends on this pathway. This possibility is further supported by the fact that IL-6 is able to promote neurite regeneration in NGFprimed PC12 cells. NGF is also able to induce a rapid transcription-independent neurite regeneration in these cells similar to that observed in PC12-E2 cells. The response of primed PC12 cells to IL-6 suggests a significant but finite half-life for this factor(s).
If activation of the Stat3 pathway is involved in the response of IL-6 in PC12-E2 cells, one would expect that other members of this subfamily of cytokines including CNTF and LIF might also produce a similar response. Thus, it was surprising to see that only modest induction of short processes was observed after prolonged treatment of E2 cells with these factors. The lack of response cannot be simply explained by the lack of these cytokine receptors because both high and low affinity CNTF binding sites are present in PC12 cells (57). Instead, as in the case of ERK activation, the distinct kinetics of STAT activation by different inducers may contribute to different biological responses. All three factors are able to stimulate a comparable rapid and transient increase in tyrosine phosphorylation of Stat3; however, only IL-6 is able to promote a strong and sustained delayed phase of activation. The STAT-DNA binding activities also display a similar distinct kinetics after IL-6 and LIF treatment. All these cytokines use a common signal transducer, gp130, but signal through somewhat different pathways (58). Binding of IL-6 to its receptor triggers the association and dimerization of gp130. In contrast, LIF binds to heterodimerized LIFR and gp130. The same heterodimer is induced after CNTF binds to its ␣ receptor component. Thus, different kinetics of tyrosine phosphorylation and activation of the STAT proteins by these cytokines may be due to different kinetics of internalization of receptor complexes as well as differential desensitization of their respective ␣-receptor components and the signaling transducing molecules gp130 and LIFR. Activation of ERK has been shown to induce phosphorylation of Ser-1044 of LIFR and potentially may play a role in regulation of LIFR activity (59). In addition, both gp130 and LIFR associate with and activate at least three members of the JAK family (Jak1, Jak2, and Tyk2) and other non-receptor tyrosine kinases (60). Different members of the IL-6 cytokine subfamily may activate distinct combinations of these JAKs and cause different patterns of tyrosine phosphorylation (16,19). On the other hand, these cytokine receptors may differentially associate with a specific protein-tyrosine phosphatase and result in differential dephosphorylation of STAT proteins. It is also possible that IL-6 can specifically activate an as yet unidentified transcription factor (or other STAT factor) that is responsible for its response in PC12-E2 cells.
In addition to the induction by cytokines, the STAT proteins can also be activated by some growth factors such as PDGF and EGF. Similar to the activation of ERK in PC12-E2 cells, EGF produces a weak and transient stimulation of tyrosine phosphorylation of Stat3 and Stat1 (data not shown), which did not result in any significant translocation of STAT proteins to the nucleus. Although EGF produces modest effects on both ERK and STAT pathways, it appears insufficient to produce significant neurite outgrowth in either cells because it sustains only a transient signal from the plasma membrane (10). This further suggests that the duration of activation of both pathways (RAS/ERK and JAK/STAT) is a crucial determinant of their biological responses.
It has been recently reported that the full transcriptional activity of Stat1 and Stat3 requires both tyrosine and serine phosphorylation (21,46,48). Consistent with reports in other cells, a transient stimulation of serine phosphorylation by IL-6 was observed in PC12 and PC12-E2 cells. The nature of the serine kinases involved is not clear, but it has been shown by others that a highly conserved sequence PMSP in the COOHterminal region of the STAT proteins is similar to the ERK recognition consensus sites and that the STAT proteins may be phosphorylated by ERK in vitro. Furthermore, it has been shown that ERK2 activity is required for Stat1-mediated gene expression by interferon ␣ and interferon ␤ (61). It is possible that the weak activation of ERKs by IL-6 contributes to some degree to the response of IL-6. Nonetheless, the results from this study provide evidence suggesting that activation of the molar excess. The STAT proteins present in the protein-DNA complexes were identified by polyclonal antibodies to Stat3, Stat1, or combination of both. The three protein-DNA complexes formed (A, B, and C) were separated from free probes by electrophoresis on a 4% polyacrylamide gel containing 0.5 ϫ TBE. B, kinetic analysis of the activation of Stat3 and Stat1 by IL-6 in PC12-E2 cells. Nuclear extracts were prepared from PC12-E2 cells treated with IL-6 for the indicated time. C represents the nuclear extract prepared from untreated PC12-E2 cells at 24 h. C, PC12-E2 cells were treated with LIF (30 ng/ml) for 15 min. Other conditions were the same as described above. D, kinetic analysis of the activation of Stat3 and Stat1 by LIF in PC12-E2 cells.