Fibroblast Growth Factor Receptors Participate in the Control of Mitogen-activated Protein Kinase Activity during Nerve Growth Factor-induced Neuronal Differentiation of PC12 Cells*

The current paradigm for the role of nerve growth factor (NGF) or FGF-2 in the differentiation of neuronal cells implies their binding to specific receptors and activation of kinase cascades leading to the expression of differentiation specific genes. We examined herein the hypothesis that FGF receptors (FGFRs) are involved in NGF-induced neuritogenesis of pheochromocytoma-derived PC12 cells. We demonstrate that in PC12 cells, FGFR expression and activity are modulated upon NGF treatment and that a dominant negative FGFR-2 reduces NGF-induced neuritogenesis. Moreover, FGF-2 expression is modulated by NGF, and FGF-2 is detected at the cell surface. Oligonucleotides that specifically inhibit FGF-2 binding to its receptors are able to significantly reduce NGF-induced neurite outgrowth. Finally, the duration of mitogen-activated protein kinase (MAPK) activity upon FGF or NGF stimulation is shortened in FGFR-2 dominant negative cells through inactivation of signaling from the receptor to the Ras/MAPK pathway. In conclusion, these results demonstrate that FGFR activation is involved in neuritogenesis induced by NGF where it contributes to a sustained MAPK activity in response to NGF.

Growth factors participate in axon growth, neuron survival in the nervous system during embryonic development, and in regeneration of peripheral nerves of vertebrate organisms (for reviews see Refs. [1][2][3]. Several studies depicted the primordial role of NGF, 1 FGF-1, and FGF-2 in the differentiation and survival of neuronal cells in vivo and ex vivo (2, 4 -6). Other studies suggested that the NGF/NGFR and the FGF/FGFR transduction pathways are interdependent. For example, in the early stages of embryonic chicken development, FGFR mRNAs are expressed, and the decline of their presence is accompanied by NGFR mRNA expression and, ultimately, by a new round of de novo FGFR transcription (7). Moreover, FGF-2 stimulates NGFR gene promoter activity in a human neuroblastoma-derived cell line (8) and acts in synergy with NGF in neuronal stem cell differentiation and proliferation (9). Taken together, these observations imply that in the nervous system NGF and FGFs intervene alternatively and sequentially in neuronal differentiation and are, to some extent, interdependent and co-regulated.
The PC12 rat adrenal pheochromocytoma-derived cell line differentiates either into sympathetic neuron-like cells or into chromaffin-like cells (10). NGF, FGF-1, or FGF-2 differentiate PC12 cells into cells morphologically and biochemically resembling sympathetic neurons (11,12). NGF signal transduction is mediated through the activation of tyrosine kinase cell surface receptors (13). The NGF transduction pathways proceeded through p21 ras (14) and B-Raf (15) activation, leading to the activation of MAPK kinase (MEK), which is sufficient for PC12 differentiation (16). The activation of the MAPK by NGF is insufficient, however, to mediate differentiation of PC12 cells (17), suggesting that other transduction pathways involving Shc, phospholipase C-␥, or yet unidentified MAPK kinase-dependent pathways (18,19,16) may intervene. FGFs also signal through the activation of tyrosine kinase cell surface receptors. Four different structurally related FGF receptor sub-families (FGFR-1 to FGFR-4) have been identified, but little is known about the intracellular downstream signaling. However, it seems to involve Grb2/Sos and the MAPK pathway (20). phospholipase C-␥ does not play a significant role in PC12 cell differentiation because FGF-2 stimulates neuronal differentiation in cells that express a mutated FGFR-1 that does not bind phospholipase C-␥ (21)(22)(23). Recently, a 90-kDa tyrosine-phosphorylated protein (named SNT, SLP, or FRS-2) was involved in FGF signaling (24 -27) by participating in the sustained activation of the MAPK pathway through Grb2 and SHP-2 (27,28).
In this study, we show that FGFR-2 is involved in NGF signaling leading to PC12 neuronal differentiation. We report that FGFRs expression and activity are modulated by NGF. Furthermore, we show that endogenous FGF-2 expression is induced by NGF and participates in FGFR activation. Finally, we demonstrate that the NGF-induced transduction pathways involving FGFRs depend upon SLP/FRS activation and sustained MAPK activity. The hypothesis that in the nervous system, FGFR and FGF-2 are part of NGF signaling is discussed.

EXPERIMENTAL PROCEDURES
Cell Culture and DNA Electroporation-PC12 cells (ATCC CRL-1721) were grown on gelatin-coated Petri dishes in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 5% horse serum, and 4.5 g/liter glucose at 37°C in a 5% CO 2 atmosphere. NGF or FGF-2 treatments were performed once at the beginning of the experiment. The plasmid RK5 containing a 1.3-kb insert encoding a tyrosine kinase activity-deficient human FGF receptor-2 exon IIIc (DNFGFR-2 (29)) was kindly provided by Dr. J. Schlessinger. PC12 cells (8 ϫ 10 5 ) were electroporated with pRK5 (0.9 g) and pSVneo (0.1 g) at 170 V and 1800 microfarads. Isolated G-418-resistant clones were tested for the presence of high affinity FGF-2 binding sites as described by Moscatelli (30) and by cross-linking to 125 I-FGF-2.
125 I-FGF-2 and 125 I-NGF Binding, Cross-linking, and Scatchard Analysis-Human recombinant FGF-2 (Synergen, Boulder, CO; 10 g) was iodinated with 1 mCi of 125 I (ICN, Costa Mesa, CA) by the Iodo-Gen method (Pierce) to a specific activity of about 7.5 ϫ 10 4 cpm/ng. Murine 2.5 S NGF (Promega, Madison, WI; 10 g) was iodinated by the same method to a specific activity of about 1.85 ϫ 10 5 cpm/ng. After treatment with 30 ng/ml NGF in Dulbecco's modified Eagle's medium containing 1% fetal bovine serum and 0.5% horse serum (maintained throughout this work) for the indicated periods of time, cells were treated as described previously (29). The determination of the number of FGF-2 binding sites and the dissociation constants was analyzed according to Scatchard (33). The same procedure was used for 125 I-NGF binding except that cells were lysed in 1 M NaOH. Cross-linking experiments were performed and analyzed as described (29). The quantity of protein used in each experiment was normalized to and corresponded to that of 10 6 cells.
RNA Extraction and Northern Blots-Total RNA was extracted from 5 ϫ 10 6 cells. RNA/10 6 cells along the first 96 h of differentiation was quantified and used as a standard (8.5 g up to 15 g in PC12 or PCN1 cells). RNA was then subjected to Northern blot experiments with the indicated [␣-32 P]dCTP random primed-labeled probes.
Detection of Extracellular Membrane-bound FGF-2 by Cell Surface Biotinylation-NHS-SS-Biotin (Pierce) was used as recommended by the manufacturer to label 8 ϫ 10 6 cells either before or after a 5-min treatment by NGF. Cells were washed three times with cold phosphatebuffered saline and lysed in 200 l of 20 mM Tris-HCl (pH 7.4), 300 mM NaCl, 2% sodium deoxycholate, 2% Triton X-100, 0.2% SDS, and 0.2% bovine serum albumin. Lysates were incubated with anti-FGF-2 antibodies for 16 -20 h at 4°C followed by 1 h with protein A-Sepharose at 4°C. Immuno complexes were resolved by SDS-PAGE and transferred onto ImmobilonP, then biotinylated proteins were revealed with the Vector Biotin/Avidin System (Vector Laboratories, Burlingame, CA) and detected by chemiluminescence.
Kinase Assays-Phosphorylation of the myelin basic protein by ERK-1 or ERK-2 immunocomplexes was done as described elsewhere (35).

NGF Modulates FGF-2 Binding on PC12 Cell Surface-Pre-
vious studies have shown that FGFRs are present before the initiation of differentiation of PC12 cells (31). To monitor the evolution of the FGFRs in the early stages of NGF-induced differentiation, we carried out binding experiments with 125 I-FGF-2 on PC12 cells treated for 0, 2, 6, 12, 24, 48, and 96 h with NGF. No competition between FGF-2 and NGF was observed (data not shown). After 2 h of NGF treatment, the level of FGF-2 binding to low affinity binding sites (proteo-heparan sulfates) increased by 7-fold and returned to the initial level after 24 h (Fig. 1A) as described previously (36). Binding of FGF-2 to its high affinity sites after NGF treatment resulted in a 3-4-fold increase in binding detected after 6 h, maybe because of a cooperation with low affinity binding sites and at 48 h (Fig. 1A). No change in FGF-2 binding was observed when cells were treated with EGF (not shown). To characterize the membrane-associated molecules that bind FGF-2, cross-linking experiments were carried out during NGF treatment as indicated above. Fig. 1B depicts a ϳ157-kDa cross-linked complex, and similar to the binding experiments, the abundance of this complex was highest at 6 and 48 h. Cross-linking of 125 I-FGF-2 was competed by a 100-fold excess of unlabeled FGF-2 (not shown). Additional bands of lower molecular weight were detected during NGF treatment (121 kDa more so at 48 h; 103 kDa at 2, 6, 12, and 48 h and, 83 kDa at 2 and 6 h; Fig. 1B). The number and dissociation constants (K d ) of 125 I-FGF-2 binding to FGF receptors were determined by Scatchard analysis (33) in either untreated or 48-h NGF-treated cells. Two classes of receptor binding sites (class 1: K d ϭ 134 Ϯ 27 pM, number of sites ϭ 1.6 Ϯ 0.7 fmol/10 6 cells; class 2: K d ϭ 877 Ϯ 125 pM, number of sites ϭ 5.12 Ϯ 1.1 fmol/10 6 cells) were detected in untreated PC12 cells. After NGF treatment (48 h), the following binding values were observed: class 1, K d ϭ 160 Ϯ 11 pM, number of sites ϭ 3.6 Ϯ 0.2 fmol/10 6 cells; class 2: K d ϭ 457 Ϯ 55 pM, number of sites ϭ 7.7 Ϯ 2 fmol/10 6 cells (Fig. 1C).
NGF Stimulates FGFR-1 and FGFR-2 Tyrosine Phosphorylation-125 I-FGF-2 cross-linked material from untreated or NGF-treated cells was immunoprecipitated with anti-FGFR antibodies. FGFR-1 (ϳ160 kDa) and FGFR-2 (ϳ157 kDa) were principally present in PC12 cells under NGF treatment at 48 and 6 h, respectively ( Fig. 2A). Northern blotting experiments with either one of the four specific probes for FGFR-1 to FGFR-4 was performed. The FGFR-1 and FGFR-2 probes hybridized to the blots ( Fig. 2B) but not the FGFR-3 or the FGFR-4 probes (not shown). One 4-kb FGFR-1 and two FGFR-2 (2.3 kb and 2.9 kb) transcripts were identified. FGFR-1 mRNA was present before NGF stimulation, and its quantity increased by 2-3-fold during the first 60 min of NGF treatment and subsequently decreased to the initial level. The 2.9-kb FGFR-2 mRNA was present after 10 min of NGF treatment and disappeared after 60 min (Fig. 2B). The 2.3-kb transcript was present in untreated cells, and its amount increased between 30 to 60 min and dramatically decreased between 2 and 6 h. A sudden increase of the 2.9-kb mRNA after 48 h was also observed, and its presence might be required for survival. Reverse transcription-polymerase chain reaction experiments revealed that FGFR-2 exon IIIc was present in PC12 cells before NGF induction, and both IIIb and IIIc exons were present after 72 h of NGF or FGF treatment (data not shown). Activation of FGFRs was assessed in PC12 cells at early times after NGF treatment. Interestingly FGFR-1 as well as FGFR-2 was tyrosine-phosphorylated after 10 min up to 2 h of NGF stimulation (Fig. 2C).
FGFR Are Involved in NGF-Induced Neuritogenesis-PC12 cells were electroporated with a cDNA encoding a dominant negative FGFR-2, which inhibits signaling not only of an identical FGFR but also of the other FGFRs (37). DNFGFR-2expressing clones were identified by 125 I-FGF-2 binding (not shown) and cross-linking (Fig. 3). Of 13 DN clones studied, 3 clones (DN21, DN62, and DN63) exhibited high levels of DN receptor (10 -15-fold when compared with PC12 or PCN control cells), 6 (DN23, DN24, DN61, DN66, DN67, and DN68) exhibited medium levels of DN receptor (4 -6-fold), and 4 (DN22, DN64, DN65, and DN69) exhibited low levels (1-2-fold). In addition to the truncated FGFR-2 monomer, a higher molecular weight band was detected by cross-linking experiments (Fig. 3, CL). This band corresponds to multimeric complexes (homodimers of DNFGFR-2 receptors; heterodimers DNF-GFR-2 and endogenous FGF receptors) and may also include endogenous FGF receptor monomers. Moreover, no change in NGFR expression was detected in DN-expressing cells by 125 I-NGF binding experiments (not shown). In Fig. 3, an example of each class of FGFR-DN-expressing cells (DN21, DN24, and DN64) and PCN1 control cells is shown. Experiments described below were done with cells from three clones of each DN receptor expression level (DN21, DN62, and DN63 for high; DN23, DN24, and DN61 for medium; DN22, DN64 and DN65 for low DN expression levels) and two clones harboring the G418resistance vectors (PCN1 and PCN2). FGFR-1 as well as FGFR-2 was phosphorylated similarly in PC12, and in PCN1 (as shown in Fig. 2C) cells after 10 min of NGF treatment, but in DN21 cells, no phosphorylation was detectable (Fig. 4A). Neurite outgrowth in the presence of NGF for up to 10 days was measured in PC12, PCN1 and 2, DN21, DN24, and DN64 cells. With all of the clones studied, the number of cells with neurites reached a plateau after 96 h (Figs. 3 and 4B). Parental PC12 cells, PCN1, and PCN2 (data not shown) exhibited the same neurite outgrowth pattern. The morphology of PCN1, DN21, DN24, and DN64 was compared (Fig. 3). Neurite outgrowth of DN21 was the most strongly inhibited (70% at 48 h) upon NGF treatment, DN64 was the most weakly inhibited (45%), and DN24 exhibited an intermediary outgrowth pattern (60%); the quantification of this effect was performed by neurite-counting and reported in Fig. 4B. The percentage of cells with neurites did not significantly change over a 10-day period from the percentage estimated at day 4 (data not shown). A total of 11 clones were studied in the neurite outgrowth assay as indicated (9 clones expressing DNFGFR-2, 2 control clones expressing the geneticin-resistance gene only), and the number of cells with neurites after 4 or 10 days of NGF treatment is reported in Table I. These data indicate that inhibition of neurite outgrowth was observed in all DNFGFR-2 cell clones and that the degree of inhibition was correlated with the expression of DN-FGFR-2. About 76 Ϯ 3.7% of PC12 cells exhibited neurites upon FGF-2 treatment. The neurite outgrowth results were very similar with PCN1 or PCN2 cells. Upon FGF-2 treatment, DN21 and DN24 cells did not differentiate (Ͻ1% cells with neurites; not shown), and DN64 exhibited some neurite outgrowth (5-10% of the control; not shown). As other phenotypic differentiation marker, we analyzed peripherin mRNA expression in RNAs isolated from PCN1 or DN21 cells after NGF treatment during indicated periods of time (Fig. 4C). PCN1 cells exhibited a 2-3-fold increase in 73 mRNA levels after 12 h (Fig. 4C) as described previously for PC12 cells (38). The increase in peripherin mRNA level in DN21 cells was moderate and somewhat retarded (24 -48 h; Fig. 4C).
NGF Modulates FGF-2 Expression-FGF-2 expression was studied by Western blotting of cell extracts and culture media of cells stimulated by NGF. Practically no FGF-2 was detectable in cell extracts of nonstimulated PC12 or PCN1 control cells. After 5 min of NGF treatment, an 18-kDa FGF-2 isoform was detected in cell extracts (Fig. 5A). FGF-2 was present for up to 96 h with peaks at 1-6 h and at 48 -96 h. In DN21 cell extracts, 18-kDa FGF-2 was present in the cell extracts before NGF treatment and disappeared 60 min after (Fig. 5A). In the PCN1-conditioned media, an 18-kDa FGF-2 species was detected from 6 -12 h only, with a peak at 24 h. A 15-kDa FGF-2-related protein with apparition kinetics identical to FGF-2 was also revealed (Fig. 5A). Higher molecular mass bands were detected on Western blots by either anti-18-kDa or anti-high molecular weight FGF-2 antibodies at constant levels in PCN1 and DN21 cellular extracts (not shown). Because FGF-2 was detectable in the medium only after 6 -12 h of NGF treatment, we assessed the presence of FGF-2 on the cell surface. PC12 cells, treated or not by NGF, were cell surface-biotinylated, and cell extracts were immunoprecipitated with anti-FGF-2 antibodies. NGF induced in 10 min the appearance of an 18-kDa FGF-2 and a 15-kDa FGF-2-related protein on the cell surface (Fig. 5B), quantities of which declined from 30 min on.
Signaling Events Are Differentially Regulated in PC12 and PC12-FGFRDN Cells-Intracellular signaling pathways involved in the response to NGF or FGF-2 were then investigated. The membrane fraction of PC12 cells treated either by NGF or FGF-2 harbored tyrosine-phosphorylated proteins of 75 to 85 kDa that were associated with FGFR-1 and FGFR-2 (Fig.  7A). These proteins designated as SLP/FRS proteins (26,27) bind to p13 suc-1 -agarose beads. When plasma membranes from NGF-or FGF-2-treated PC12 or DN cells were solubilized and incubated with p13 suc-1 -agarose beads, tyrosine-phosphorylated FRS proteins were detected only in FGF-2 and to a lesser extent in NGF-treated PC12 cells. No tyrosine phosphorylation was detected in DN cells (Fig. 7B, Ib:␣PY). In both cases the amount of FRS-2 was comparable (Fig. 7B, Ib:␣FRS-2). The FRS-2 molecular mass shift observed in the immunoblotting experiments corresponded to tyrosine phosphorylation of these proteins. FRS proteins were also known to bind Grb-2 and to constitute a link between FGFR and MAPK activation (27). The association of Grb-2 and FRS proteins at the plasma membrane in response to NGF or FGF-2 was studied. As shown in Fig. 7C  (upper and middle panels), Grb-2 and FRS proteins were associated in response to FGF-2 and NGF stimulation with no major change in Grb-2 expression (Fig. 7C, lower panel). More-   6. Effect of anti-FGF-2 RNA oligonucleotides on NGF-induced neurite outgrowth. PCN1 cells treated with NGF (30 ng/ml (‚)) were incubated with either 100 g/ml heparin (E), 200 g/ml anti-FGF-2 antibodies (᭛), 100 nM NX-286 FGF-2 binding-modified RNA oligonucleotide (ƒ), or 100 nM NX-213 vascular endothelial growth factor-specific binding-modified RNA oligonucleotide (Ⅺ) that were added 10 min before growth factor addition. Squares depict neurite outgrowth assays in which only NGF was added to the medium, as a positive control. over, Shc was tyrosine-phosphorylated in PC12 cells under NGF or FGF-2 treatment, but in PC12DN cells, Shc was found tyrosine-phosphorylated only under NGF stimulation (Fig. 7D, upper panel) with no change in Shc expression (Fig. 7D, lower  panel). This result demonstrates that NGFRs are active, whereas endogenous FGFRs are inactivated by the dominant negative FGFR-2.
FGFRs Participate in the Net Sustained ERK Activity Induced by NGF-ERK-1 and ERK-2 activities were determined by phosphorylation of myelin basic protein (MBP). Under NGF stimulation, ERK-1 immunoprecipitates from either PCN1 or DN21 cell extracts presented a similar MBP phosphorylation activity pattern. ERK-1 activity was maximal at 5-10 min and stayed elevated (at about 50% of maximum) for up to 4 h (Fig.  8A). ERK-1 activity in DN21 cell extracts was somewhat higher than in PCN1 cells between 10 and 120 min. ERK-2 activity, however, was 2-3-fold lower in DN21 than in PCN1 cells at 5 min and 7-fold lower 120 min after the beginning of NGF stimulation, and no significant activity was detected after 4 h. ERK-1 and ERK-2 activation were similar in PCN1 cells under NGF or FGF-2 treatment, although the stimulation with the latter was 2-fold lower (Fig. 8A). Furthermore, both ERK-1 and ERK-2 activities were strongly inhibited in DN21 cells following FGF-2 stimulation when compared with control cells (Fig.  8A). Finally, ERK2 activities were similarly low in FGF-2 and NGF-treated DN21 cells. As shown Fig. 8B, the kinetics of MBP phosphorylation by ERK-2 immunoprecipitates as well as ERK-2 phosphorylation were significantly decreased in intensity and length in DN21 cells compared with PCN1, although the amount of protein immunoprecipitated was similar.

DISCUSSION
In this study, we show that in PC12 cells, FGFRs, and FGF-2 participate in NGF-induced neuritogenesis. This is based on the following observations. 1) NGF modulates FGFR expression and induces FGFR activation, 2) NGF modulates FGF-2 expression, 3) neurite outgrowth and FGFR activation are reduced in PC12 cells expressing dominant negative FGFR-2, 4) specific anti-FGF-2-modified RNA oligonucleotides inhibit NGF-induced neuritogenesis, and 5) specific FGFR transduction pathways play a crucial role in the maintenance of NGFinduced sustained MAPK activity.
The data presented here revealed the induction of two waves of FGFR expression under NGF treatment. In addition, one FGFR-1 and two FGFR-2 transcripts were identified during NGF stimulation. Numerous FGFR splicing variants have been identified (3,39,40). These include variants with either exon IIIb or IIIc usage, two immunoglobulin-like forms, and truncated or soluble FGFRs. Several molecular weight species were present in cross-linking studies, but only a unique complex was detected in either anti-FGFR-1 or anti-FGFR-2 immunoprecipitations. This discrepancy could be explained in part by the existence of FGFRs variants truncated at their carboxyl-terminal sequences (41,42) that are not immunoreactive with the antibodies used. Beside these effects of NGF on FGFR expression, we show here that NGF was able to promote FGFR activation. The functional significance of the FGF receptor activation upon NGF treatment was investigated using the dominant negative FGF receptor approach. Overexpression of dominant negative FGFR-2 reduces the NGF-induced neuritogenesis and reduces the expression of peripherin. Moreover, the key role of SLP/FRS protein in this mechanism was previously reported (24 -28, 43). The role of FRS-2 has been clearly demonstrated to participate in the sustained MAPK activity through its interaction with SHP-2 (28). Our data demonstrate also that in PC12DN the MAPK activity is reduced both in intensity and in duration possibly causing the inhibition of differentiation. Because the Shc pathway was affected in DN cells only after FGF-2 treatment, we may conclude that SLP/ FRS pathway is probably stimulated only after FGFR activation itself promoted by NGF treatment. The maximal neuritogenesis inhibition we observed was of about 70%, the remaining 30% possibly caused by NGFR activation itself.
The expression of dominant-negative FGFR-1 was recently reported to inhibit neuronal differentiation of PC12 cells induced by FGF-2, L1, Neural cell adhesion molecule (N-CAM) or, N-cadherin but not by NGF (44). The apparent discrepancy between these data and ours may simply reside in the types of dominant negative FGFR used. Indeed, the transdominant effect of dominant negative FGFRs is not absolute, and inhibition by DNFGFR-2 was demonstrated to be more efficient than that of DNFGFR-1 as a result of increased dimer stability (45). Moreover, when DNFGFR-1 or DNFGFR-2 are targeted to photoreceptors by using the rhodopsin promoter, only DNF-GFR-2 will induce focal photoreceptor degeneration (46), indicating that DNFGFR-2 is more potent than DNFGFR-1 in deregulating FGF signaling. Finally, the neurite outgrowth measured in our work reflected the number of cells with neurites and not the mean neurite length as described by Saffell et al. (44). Therefore, our observations may just be the result of an increased inhibition potential of DNFGFR-2.
The expression of the FGF-2 was stimulated by NGF in PC12 cells and found both in cell extracts and on the cell surface. We demonstrate here that newly produced FGF-2 mainly stays on the cell surface to act on its receptors. The feeble inhibitory effect of FGF-2-neutralizing antibodies has been already noted in previous studies (47) possibly because of their size which does not allow access to membrane-or extracellular matrixbound FGF-2. However, when using anti-FGF-2 RNA oligonucleotides that specifically neutralize extracellular FGF-2 (31), neuronal differentiation induced by NGF was inhibited at about 50%. This implies the existence of an NGF-driven FGF-2 autocrine loop. The activation of the FGFRs by endogenous FGF-2 is not incompatible with the participation of other FGFR ligands in neuritogenesis. For example, FGF-1 is expressed during differentiation of PC12 cells and possibly participates in FGFR-1 activation (48) or the adhesion molecule L1 signals through FGFRs and induces neuronal differentiation of neu- rons in culture (49). FGF-2 knock-out mice show some abnormalities in the cytoarchitecture of the neocortex and significant reduction in neuronal density in most layers of the motor cortex (50). Moreover, neuronal defects are present in the hippocampal commissure, and neuronal deficiencies are observed in the cervical spinal cord. Furthermore, FGF-2 knock-out mice present an impaired neural regulation of blood pressure by the baroreceptor reflex, suggesting a role for the sympathetic system (51). Therefore, the interpretation of our results in that context suggests that FGF-2 may be important for neuronal differentiation especially in the case of sympathetic neurones, modeled by PC12 cells.
The results described in the present work indicate that FGFRs are involved in neuronal differentiation of PC12 cells induced by NGF and that MAPK-sustained activation is dependent on functional FGFR system. Furthermore, autocrine FGF-2 participates in NGF-mediated FGFR activation. It is possible that the mechanism of NGF action described herein is specific for a population of sympathetic neurons and regulate their function. But also, this mechanism could have a general significance and operate in different neuronal cell populations, thus placing FGFs/FGFRs in or alongside the NGF signal transduction pathway.