Ciliary Neurotrophic Factor Stimulates the Phosphorylation of Two Forms of STAT3 in Chick Ciliary Ganglion Neurons*

Ciliary neurotrophic factor (CNTF) is a neuropoietic cytokine that was identified, purified, and cloned based on its neurotrophic activity on cultured chick ciliary ganglion neurons. The molecular mechanisms by which CNTF elicits its effects on these neurons are unknown. We have previously identified functional receptors for CNTF on ciliary ganglion neurons and demonstrated the CNTF-specific tyrosine phosphorylation of an approximately 90-kDa protein. Here we show that CNTF induced the rapid tyrosine phosphorylation and nuclear accumulation of this protein and identify it as an avian form of the transcription factor, STAT3. Identification was confirmed by its recognition with two distinct anti-STAT3 antibodies and the lack of binding to antibodies against STAT1, -2, -4, -5, or -6. The phosphorylation was stable for up to 2 h but required the continued presence of CNTF. CNTF also induced the tyrosine phosphorylation of a similar protein in cultured chick dorsal root ganglion and retinal neurons. In addition, we identify a second, 100-kDa form of STAT3 that appears in response to CNTF. Unlike previous reports, utilizing mammalian cell lines that detected a slower migrating form of STAT3 resulting from H7-sensitive protein phosphorylation, H7 did not prevent the appearance of the 100-kDa form in ciliary neurons. Thus, the 100-kDa avian protein may represent a novel form of CNTF-inducible STAT3.

Ciliary neurotrophic factor (CNTF) 1 is a cytokine that shows activity toward a variety of cell types in the nervous system. However, the molecular mechanisms for these effects have not been clearly elucidated in defined neuronal populations. CNTF belongs to a family of neuropoietic cytokines, which display a high degree of redundancy in their biological activities, have structural similarity, share receptor subunits and signal transduction pathways, and yet are capable of generating unique cellular responses as well (1,2). The known receptor complexes for CNTF, leukemia inhibitory factor, oncostatin M, cardiotrophin-1, and interleukin-6 and -11, include gp130 as a ␤-receptor component, and all except interleukin-6 and -11 share the leukemia inhibitory factor ␤-receptor component (3)(4)(5)(6)(7). CNTF and interleukin-6 and -11 also utilize unique ligand-binding ␣ subunits (1,8,9). Receptor activation by CNTF-related cyto-kines results in dimerization of the ␤-receptor components and activation of receptor-associated tyrosine kinases of the Jak family. Jaks or other cytoplasmic kinases phosphorylate tyrosine on the cytokine receptor components that selectively bind cytoplasmic transcription factors of the STAT (signal transducers and activators of transcription) family (10). In this model, established primarily from cell lines, STAT proteins are directly tyrosine-phosphorylated by Jaks and then dimerize and translocate to the cell nucleus and regulate gene transcription (11)(12)(13). There are at least six members of the STAT family, and the selection and activation of a particular STAT is dependent on receptor substrate-specifying motifs (10). Thus, cytokines utilizing gp130 mainly activate STAT3, initially identified as an acute phase response factor (10). However, in certain cell types, STAT1 is also activated by the gp130-associated class of cytokines (11). It remains to be determined whether these selectivity patterns are retained in primary cells and neurons. It has recently been suggested that a cytoplasmic serine kinase may also phosphorylate STAT1 and STAT3 and be necessary for maximal gene activation (14,15).
Mammalian CNTF and the avian homologue growth-promoting activity (GPA) were identified, purified, and cloned based on their neurotrophic activity for embryonic chick ciliary ganglion neurons (16 -19). These neurons require CNTF for survival and express high levels of CNTF receptors, and yet the signal transduction pathway of CNTF action on ciliary ganglion neurons has not been defined (16,20). They represent a well characterized population of cholinergic neurons of which one-half of the neurons die during a period of cell death between the embryonic day 8 and 13 (21). This neuronal death is target-sensitive and coincides temporally with the formation of functional synapses within the target (21) with an increase in expression of GPA in the developing eye (22) and of CNTF receptor ␣ subunit mRNA in ciliary ganglia. 2 In culture all embryonic ciliary neurons are rescued by a select group of CNTF-like neurotrophic factors but not by members of the neurotrophin family (20,22). Also, CNTF has trophic activities for motor, sensory, sympathetic, and hippocampal neurons, regulates acetylcholine receptors on autonomic neurons, and regulates neurotransmitter expression in cultured sympathetic and central nervous system neurons (23)(24)(25)(26)(27)(28)(29).
We have identified a class of functional CNTF receptors on cultured ciliary ganglion neurons where 125 I-CNTF cross-linking studies detected a receptor component different in size from the previously described mammalian receptor subunits (20). These results also showed that CNTF produced the rapid induction of a tyrosine-phosphorylated protein with an apparent molecular mass of approximately 90 kDa. The induction of the 90-kDa phosphoprotein was specific for CNTF and avian GPA in that it was not observed in response to other CNTF-related cytokines or basic fibroblast growth factors (20). Tyrosine phos-phorylation of this 90-kDa protein, therefore, represents a potential signal transduction pathway activated specifically by mammalian CNTF and chick GPA. In this study we show for the first time in primary chicken neurons that the 90-kDa phosphoprotein is translocated to the nucleus and is immunologically identical with the transcription factor, STAT3. In addition, CNTF induces a modified 100-kDa form of STAT3 that appears distinct from previously described mammalian serinephosphorylated STAT3 (14).

EXPERIMENTAL PROCEDURES
Cell Culture-Ciliary ganglion neurons were dissociated from 8-dayold chick embryos and grown at 37°C in a 5% CO 2 , 95% air atmosphere as described previously (20). Cells for the tyrosine phosphorylation assays were plated on 35-mm culture dishes on a substratum of poly-D-lysine and mouse laminin. The culture medium consisted of Eagle's minimal essential medium (MEM) containing 10% (v/v) heat-inactivated horse serum, 50 units/ml penicillin, and 50 mg/ml streptomycin. The medium was adjusted to a final concentration of 25 mM KCl, which allows for complete survival of neurons and negligible survival of nonneuronal cells. Freshly dissociated retina from 11-day-old embryos were trypsin-digested (0.25%) for 30 min at 37°C and then processed as described for ciliary ganglion neurons except that cells were plated on tissue culture dishes coated with poly-D-ornithine (10 g/cm 2 ) at a final density of 5 ϫ 10 5 cells/cm 2 in MEM with 10% horse serum. Dorsal root ganglion neurons from 11-day-old embryos were cultured on a substratum of rat tail collagen in medium supplemented with 100 ng/ml nerve growth factor, as described previously (30).
Immunoblot Detection of Proteins-Ciliary ganglion neurons were grown in culture for 5 days, rinsed, and incubated in serum-free unsupplemented medium for 3 h; the neurons were then treated for the indicated periods with human CNTF (2 nM, 50 ng/ml unless otherwise indicated), washed in MEM containing 1 mM Na 3 VO 4 , and solubilized in Laemmli sample buffer (containing 5% 2-mercaptoethanol, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 50 mM NaF, 2 mM Na 3 VO 4 , 10 g/ml leupeptin, and 2 g/ml aprotinin) prior to immunoblotting as described previously (13,20). Where indicated, cultures were treated for 20 min with 0.1 or 1 M staurosporine, for 30 min with 200 M H7, or for 5 min with 20 mM sodium fluoride prior to the treatment with CNTF. The proteins were separated on a 7.5% SDS-polyacrylamide gel and then transferred to nitrocellulose. The membranes were blocked for 2 h in Tris-buffered saline (10 mM Tris-HCl, pH 7.5, 137 mM NaCl, and 0.05% Tween 20) with 1% ovalbumin (or 5% bovine serum albumin) and incubated for 4 h with anti-phosphotyrosine monoclonal antibody PY4G10 (1:1000) as described previously (20), with a monoclonal antibody to STAT3 (1:1000) (13), a polyclonal antibody directed against tyrosine-phosphorylated (Tyr-705) STAT3 (1:1000) (13), or other anti-STAT monoclonal antibodies as recommended by the supplier. Following incubation with either goat anti-mouse IgG or anti-rabbit IgG conjugated to horseradish peroxidase the reactive proteins were visualized using enhanced chemiluminescence (ECL) according to the manufacturer's instruction (Amersham Life Sciences, Inc.). The level of antibody binding was quantified by scanning densitometry (Ultrascan XL, Pharmacia Biotech Inc.). In some experiments the nitrocellulose membranes were stripped and analyzed using a different primary antibody as described in the ECL Manual (Amersham Life Sciences, Inc.).
Immunoprecipitation-Ciliary ganglion neurons (3.4 ϫ 10 5 cells; 50 ganglion eq) grown for 5 days in culture were serum-starved for 3 h at 37°C in MEM and then treated for 5 min with 2 nM CNTF. Cells were rinsed, solubilized with 30 l of boiling 1% SDS in 10 mM Tris-HCl, pH 7.4, and then diluted with 150 l of 2 ϫ cold immunoprecipitation buffer (1 ϫ buffer is 150 mM NaCl, 1% Triton X-100, 0.5% Nonidet P-40, 1 mM EGTA, 1 mM EDTA, 1 mM Na 3 VO 4 , 0.5 mM phenylmethylsulfonyl fluoride, 20 g/ml leupeptin, 20 g/ml aprotinin, 10 mM Tris base, pH 7.4) and 120 l of H 2 O. Anti-STAT3 monoclonal antibody or nonimmune mouse serum was added to cell lysates (1:100) and incubated for 1 h at 4°C. Rabbit anti-mouse IgG (Cappel) was then added for 30 min, and the antibody-bound proteins were captured with protein A, washed twice with immunoprecipitation buffer, and boiled in Laemmli sample buffer containing 2 mM Na 3 VO 4 . The soluble proteins after the immunoprecipitation with anti-STAT3 antibodies were precipitated with 6 volumes of ice-cold acetone, and the dried pellet was solubilized in boiling Laemmli sample buffer containing 2 mM Na 3 VO 4 .
Cell Fractionation-Crude nuclear and cytoplasmic extracts were prepared as described (32). Ciliary ganglion neuron cultures were washed and lysed by homogenization in hypotonic buffer (20 mM HEPES, pH 7.0, 10 mM KCl, 1 mM MgCl 2 , 0.1% Nonidet P-40, 10% glycerol, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 2 g/ml leupeptin). Nuclei were separated by centrifugation at 200 ϫ g, and the pellet containing the nuclei was washed twice in hypotonic buffer without Nonidet P-40 before immunoblot analysis as described above. The cytoplasmic fraction (the supernatant from the first centrifugation) was clarified by centrifugation at 14,000 ϫ g, and the proteins were precipitated with acetone and subjected to immunoblot analysis as above.
Materials-Embryonated chick eggs (mixed heavy breed, Rhode Island Red/Barred Rock/White Rock) were obtained from Gawlak Farms (Lawton, NY) and grown at 37°C in a humidified incubator. Cell culture reagents and molecular weight markers were obtained from Life Technologies, Inc. Laminin was from Collaborative Biomedical Products (Bedford, MA), staurosporine and H7 were from Calbiochem, and H7 was also obtained from Seikagaku America (Rockville, MD). Monoclonal antibody 4G10 was from Upstate Biotechnology (Lake Placid, NY), polyclonal anti-phospho-STAT3 antibody was from New England Biolabs (Beverly, MA), and monoclonal antibodies to all other STATs were from Transduction Laboratories (Lexington, KY). Recombinant human CNTF was generously provided by Regeneron Pharmaceuticals (Tarrytown, NY).

Nuclear Translocation of a 92-kDa CNTF-induced
Tyrosinephosphorylated Protein-CNTF induced the phosphorylation of an approximately 90-kDa protein in ciliary ganglion neurons that corresponded to the size described for members of the STAT family of transcription factors (M r ϭ 84,000 -113,000) (10,20). Thus we determined whether this CNTF-induced tyrosine-phosphorylated protein from ciliary ganglion neurons was translocated from the cytoplasm to the nucleus as required for transcriptional activity. A tyrosine-phosphorylated 92-kDa protein was detected at very low levels in the nucleus before CNTF stimulation; however, after 15 min of CNTF treatment the 92-kDa phosphoprotein had accumulated in the nuclear fraction ( Fig. 1, left panel). Sodium fluoride (20 mM), an inhibitor of nuclear translocation of activated STAT1 (33), inhibited 80% of the CNTF-induced nuclear translocation of the 92-kDa phosphoprotein ( Fig. 1, left panel). These results indicated that the CNTF-induced tyrosine-phosphorylated protein was translocated to the nucleus by a fluoride-sensitive mechanism.
Recognition of the 92-kDa Protein by Anti-STAT3 Antibodies-The STAT family of transcription factors includes six known members through which cytokines can differentially modulate gene transcription. The CNTF family of cytokines has been reported to activate both STAT1 and STAT3 (11). We examined cell lysates of cultured ciliary ganglion neurons using antibodies against STAT1, -2, -3, -4, -5, or -6 to determine if similar sized products were detectable. Only antibodies to STAT3 produced a visible reaction at 92 kDa (Ϯ1 kDa, n ϭ 6) in ciliary ganglion neurons ( Fig. 2A). Two products close in size, a doublet of about 112/119 and 105 kDa, were detected using anti-STAT5 and -6 antibodies, respectively. In control experiments to test for cross-reactivity of STAT1, -2, and -4 with chick proteins, antibodies to STAT1 detected a doublet of 93/95 kDa in lysates from chick embryo fibroblasts, and the STAT2 antibody detected a band of 90 kDa in chick embryo spleen (data not shown). Therefore, these antibodies, except for anti-STAT4, recognize chick proteins comparable in size with those seen in mammalian cells. To confirm the identity of the 92-kDa product as STAT3 as compared with STAT1 we used antibodies specific for STAT1 and STAT3 to precipitate proteins from extracts of CNTF-treated ciliary ganglion neurons. Monoclonal antibodies to STAT3 immunoprecipitated proteins of 92 kDa (Fig. 2B). In contrast, anti-STAT1 antibodies and nonimmune mouse serum failed to immunoprecipitate the 92-kDa protein (data not shown and Fig. 2C). Anti-STAT3 immunoprecipitates were further analyzed by phosphotyrosine immunoblotting. The antiphosphotyrosine antibody, PY4G10, specifically detected the STAT3-precipitated 92-kDa protein only from CNTF-treated cells (Fig. 2B). The majority of the CNTF-induced 92-kDa tyrosine-phosphorylated protein was depleted from cell lysates after immunoprecipitation with anti-STAT3 indicating that most, if not all, of this phosphoprotein was recognized by anti-STAT3 antibodies (Fig. 2B). Further, the 92-kDa phosphoprotein that appeared in the nuclear fraction following CNTF treatment was recognized by the anti-STAT3 antibody (Fig. 1,  right panel). A second band of approximately 80 kDa was often observed after probing blots of cell lysates with the anti-STAT3 antibody (Fig. 2C, left). This 80-kDa band may correspond to an alternatively spliced STAT3 product previously described (34); however, it was not tyrosine-phosphorylated following CNTF treatment. These results identify the CNTF-induced 92-kDa tyrosine-phosphorylated protein in ciliary ganglion neurons as immunologically related to mammalian STAT3.
Characterization of CNTF-induced Tyrosine Phosphorylation of STAT3-Protein tyrosine phosphorylation initiated by ligand-receptor interaction is usually described as a rapid and transient event typically occurring within 1-5 min and declining after 30 -60 min (14). Time course experiments of CNTFinduced tyrosine phosphorylation of the 92-kDa protein in cultured ciliary ganglion neurons showed unusual stability. Increased tyrosine phosphorylation of the 92-kDa protein was detectable within 3 min of CNTF exposure and increased up to 15 min and remained at elevated levels during 2 h of continuous exposure (Fig. 3A). To further confirm the identity of the 92-kDa phosphoprotein we used a recently developed antibody from New England Biolabs, Inc. that specifically recognizes tyrosine-phosphorylated (Tyr-705) STAT3. The time course of tyrosine phosphorylation was similar whether determined using the anti-phosphotyrosine (4G10) or the anti-phospho-STAT3 antibodies, and little change in total STAT3 protein was observed with continued CNTF treatment (Fig. 3A). These results provide strong evidence that the 92-kDa protein represents avian STAT3. A lighter intensity band of 100 kDa (Ϯ1 kDa, n ϭ 4) was visualized in material prepared from CNTFtreated cells when immunoblots probed with the anti-STAT3 antibody were re-exposed to film for longer periods (Fig. 3A,  inset).
CNTF has effects on a variety of chick neuronal types in both the peripheral and central nervous systems. The CNTF-mediated activation of STAT3 was not limited to parasympathetic neurons of ciliary ganglia as cells cultured from both dorsal root ganglia and neural retina showed a 2-fold increase in tyrosine phosphorylation of STAT3 after 10 min of CNTF treatment (Fig. 4). Thus, other neurons whose in vitro survival is supported by CNTF, dorsal root ganglion neurons (35), as well as neurons whose cholinergic development is regulated in vitro by CNTF, neural retina (36), also show tyrosine phosphorylation of STAT3.
The survival of cultured ciliary ganglion neurons requires
Relative mobility of the 92-kDa protein is indicated by arrows. In a replicate experiment (C), blots of anti-STAT3 immunoprecipitated material (lanes 1, 2, 4, and 5) as described in panel B or normal mouse serum immunoprecipitated material (lanes 3 and 6) were first probed with anti-STAT3 antibodies (lanes 1-3) and then stripped and reprobed with an anti-phospho-STAT3 antibody (lanes 4 -6) and exposed after each to film for an extended time (90 and 20 min, respectively) to reveal both 92-and 100-kDa (arrow) forms of STAT3.
the continued presence of trophic factor support (37). 3 The tyrosine phosphorylation of STAT3 also required the continued presence of CNTF. Ciliary ganglion neurons were treated with CNTF for 5 min; the neurons were then rinsed and incubated with CNTF-free medium. After 1 h of CNTF deprivation the level of tyrosine phosphorylation returned to near prestimulation levels (Fig. 3B). These results indicated that maintenance of the tyrosine phosphorylation of STAT3 depended on the continued presence of CNTF. In addition, chronic growth in CNTF did not eliminate the response. Neurons grown in either 25 mM KCl medium, 1 nM CNTF, or 10 nM basic fibroblast growth factor/heparin for 4 days, each of which promotes survival of ciliary neurons, followed by incubation of cells for 3 h in unsupplemented medium and stimulation with 5 nM CNTF induced tyrosine phosphorylation of STAT3 (data not shown).
Studies using inhibitors further suggested a relationship between tyrosine phosphorylation and cellular function. Incubation of neurons with a tyrosine kinase inhibitor, staurosporine (0.1-1.0 M), for 48 h prevented greater than 92% (n ϭ 4) of CNTF-mediated cell survival. Preincubation of ciliary ganglion neurons with staurosporine also inhibited the CNTFinduced stimulation of tyrosine phosphorylation of STAT3 (Fig. 5A).
Analysis of the 100-kDa STAT3-like Component-The 100-kDa protein cross-reacting with the STAT3 antibody underwent tyrosine phosphorylation in response to CNTF and was recognized by the anti-phospho-STAT3 antibody (Fig. 5B). The 100-kDa product was probably not seen in experiments utilizing PY4G10 due to masking by background tyrosine-phosphorylated proteins in the 94 -105-kDa range following prolonged exposures. The appearance of the tyrosine-phosphorylated 100-kDa protein was prevented by pretreatment of cells with staurosporine (data not shown) but was unaffected by treatment with the serine/threonine kinase inhibitor, H7 (Fig. 5B). H7 from two different sources were each without effect on the appearance of the 100-kDa component. Confirmation that the H7 preparations were active was provided by experiments showing that the same concentration of H7 blocked the cell survival effects of CNTF on ciliary ganglion neurons (data not shown). The appearance of the 100-kDa protein following treatment with CNTF was enhanced over a similar time course (Fig.  3A, inset) as tyrosine phosphorylation of the 92-kDa form. In addition, the appearance of the 100-kDa component and the tyrosine phosphorylation of the 92-and 100-kDa components were each induced over a similar range of CNTF concentrations (Fig. 6). Finally, immunoprecipitation with anti-STAT3 antibodies of lysates from CNTF-treated cells revealed a 100-kDa product recognized by either the anti-STAT3 or the anti-phospho-STAT3 antibodies (Fig. 2C). The 100-kDa band was not detected in immunoprecipitates from untreated cells. These results suggest that both the 100-and 92-kDa bands represent structurally related forms of avian STAT3.

Identification of STAT3 in Ciliary Ganglion Neurons-We
have previously shown that CNTF stimulation of ciliary ganglion neurons resulted in the rapid induction of a 92-kDa tyrosine-phosphorylated protein (20). The 92-kDa phosphoprotein was the only major phosphoprotein that consistently showed a specific increase in tyrosine phosphorylation in re-  1 and 2) and neural retina (lanes 3 and 4) cultures were plated as described under "Experimental Procedures" and treated for 10 min with 5 nM CNTF (lanes 2 and 4) or untreated (lanes 1 and 3). Cell lysates were subjected to immunoblot analysis with an anti-phospho-STAT3 antibody (top). Blots were subsequently stripped and reprobed with the anti-STAT3 antibody (bottom) for normalization of protein levels. The arrows indicate the position of a 92-kDa phosphoprotein recognized by the anti-phospho-STAT3 antibody and visualized by ECL. Scanning densitometry revealed a 2.0-fold increase in the intensity of the normalized signals following CNTF treatment of both dorsal root ganglion and retinal neurons. Similar results were obtained in three additional independent experiments. sponse to physiologically relevant concentrations of human CNTF and avian GPA but was not induced by a series of CNTF-related cytokines (19) (interleukin-6, interleukin-11, and oncostatin M). In this study we identified the 92-kDa tyrosine-phosphorylated protein in ciliary ganglion neurons as a member of the STAT family of transcription factors based on three criteria. First, the 92-kDa protein underwent tyrosine phosphorylation and translocation from the cytoplasm to the nucleus following CNTF stimulation, a process necessary for DNA binding and gene regulation by STATs. Second, the nuclear translocation of the 92-kDa phosphoprotein was inhibited by sodium fluoride, which also inhibits activated STAT nuclear translocation in response to interferons and epidermal growth factor (32,33). Finally, the 92-kDa tyrosine-phosphorylated protein was immunoprecipitated with a monoclonal antibody specific for the carboxyl terminus of mammalian STAT3 and was recognized by a polyclonal antibody specific for tyrosinephosphorylated (Tyr-705) STAT3. A second 100-kDa tyrosinephosphorylated protein was induced by CNTF and recognized by both STAT3 antibodies. There was no evidence for crossreactivity of either the 92-or 100-kDa proteins with antibodies to STAT1, -2, -4, -5, or -6. Although these antibodies were raised against mammalian forms of STAT proteins except for STAT4, control studies confirmed that all cross-react with chick homologues either in ciliary ganglion, chick spleen, or chick embryo fibroblasts.
Characterization of STAT3 Phosphorylation-Mammalian STAT3 shares 40 -50% amino acid identity to STAT1 and STAT2 and recognizes a DNA binding element that resembles the consensus sequence found for STAT1, -4, -5, and -6 and Drosophila STAT (10). However, while STAT1 is activated by a number of cytokines and growth factors, STAT3 is activated mainly by the CNTF family of cytokines and epidermal growth factor and is implicated in the regulation of transcription of acute phase protein genes in liver cells. The gene for vasoactive intestinal peptide in rat also contains a cytokine response element (10,12,38,39). The specificity of STAT3 activation, at least in part, is achieved by the tyrosine-based motifs in gp130 and LIFR␤, the signal transducer receptor components for the CNTF family of cytokines (40). Recently, two cytokine-induced tyrosine-phosphorylated forms of STAT3 have been characterized, STAT3f and STAT3s, differing in H7-sensitive secondary serine/threonine phosphorylation (14). The two forms differ in apparent molecular mass by 2 kDa, undergo nuclear translocation, and recognize the same DNA sequences in vitro but may act on distinct target genes in vivo, depending on the cell type (14). In ciliary ganglia we found that CNTF induced the appearance of an additional 100-kDa form of tyrosine-phosphorylated STAT3. We do not know if it was generated by a secondary modification of the 92-kDa STAT3. However, the 92-kDa STAT is the most likely source since it is the only other protein in cell lysates recognized by both the STAT3 and phospho-STAT3 antibodies. This CNTF-induced shift of apparent molecular mass of 8 kDa was significantly greater than the 2-kDa shift observed in mammalian cell lines (14,15). In cell lines, cytokines induce a time-dependent shift of STAT3f to STAT3s, whereas in ciliary ganglion neurons the 100-kDa form remained a small percentage of the total STAT3 immunoreactive pool even at higher CNTF concentrations and after extended treatment. It is yet to be determined if the newly generated 100-kDa pool is static or is actively turning over. A final difference between STAT3s and the chick 100-kDa STAT3 is the lack of sensitivity to H7 by the latter. Two different sources of H7 were unable to prevent the appearance of the 100-kDa STAT3, whereas H7 effectively prevented the conversion of STAT3f to STAT3s in CNTF-treated SH-SY5Y human neuroblastoma cells. 4  Immunoblots of cellular proteins were probed with anti-phospho-STAT3 antibodies (top) and then stripped and reprobed with anti-STAT3 antibodies (bottom) and visualized by ECL. The blots were exposed normally (20 and 5 s, respectively) to detect the 92-kDa STAT3 band (arrows) or overexposed (5 and 2 min, respectively) to detect the 100-kDa STAT3-immunoreactive product (arrowheads). Equivalent results were obtained in a replicate experiment.  3 and 4). Immunoblots of cell proteins were analyzed with anti-phospho-STAT3 antibodies and visualized by ECL. The blots were exposed normally during ECL to detect the 92-kDa STAT3 band (arrow, 15 s) or overexposed to detect the 100-kDa STAT3-immunoreactive product (arrowhead, 5 min). cells. 3 Therefore, the 100-kDa band appears to be a novel STAT3 induced in chick neurons with distinct properties from those of the serine-phosphorylated STAT3s previously identified in cell lines. The question remains whether it represents an active form or a form associated with signal termination.
The CNTF-induced tyrosine phosphorylation of STAT3 in ciliary ganglion neurons was unusually stable. Analysis of the time course of the tyrosine phosphorylation showed that the effect was induced rapidly after CNTF stimulation and that the level of phosphorylation remained elevated for up to 2 h in the continued presence of CNTF. Removal of CNTF, however, produced a decrease in the phosphorylation to control levels within 1 h indicating that the stability of the induced tyrosine phosphorylation of 92-kDa STAT3 depended on the continued presence of CNTF. Rapid but transient tyrosine phosphorylation of receptor components and signaling molecules, including STAT proteins, characterizes CNTF and related cytokine responses in a variety of cell lines (14). The functional importance of the stable tyrosine phosphorylation of STAT3 in ciliary ganglion neurons is not clear yet. Cultured embryonic ciliary ganglion neurons are rescued by CNTF, and it is possible that continuous activation of STAT3 is required for CNTF-mediated cell survival. Consistent with this, staurosporine treatment inhibited both CNTF-induced STAT3 tyrosine phosphorylation and CNTF-mediated cell survival.
Conclusion-There are at least three physiological effects of CNTF on cultured embryonic ciliary ganglion neurons: long term survival, increased neuronal growth, and down-regulation of ␣7-containing nicotinic acetylcholine receptors (20,24,41). Induction of STAT3 appears to be specific for the signal transduction pathway induced by CNTF in ciliary ganglion neurons. Identification of the CNTF-inducible 92-and 100-kDa tyrosine-phosphorylated proteins as immunologically identical and functionally similar to STAT3 in ciliary ganglion neurons, which display biologically relevant responses to CNTF/GPA, provides insight into the potential signal transduction mechanism(s) producing these diverse biological effects. The role of the newly identified 100-kDa form of STAT3 in CNTF signal transduction remains to be determined as does the composition of the activated STAT3 transcription complex and its target genes.