Cyclic AMP induces transactivation of the receptors for epidermal growth factor and nerve growth factor thereby modulating activation of MAP kinase, Akt and neurite outgrowth in PC12 cells

In PC12 cells, a well studied model for neuronal differentiation, an elevation in the intracellular cAMP level increases cell survival, stimulates neurite outgrowth, and causes activation of extracellular signal-regulated protein kinase 1 and 2 (ERK1/2). Here we show that an increase in the intracellular cAMP concentration induces tyrosine phosphorylation of two receptor tyrosine kinases, i.e. the epidermal growth factor (EGF) receptor and the high affinity receptor for nerve growth factor (NGF), also termed Trk(A). cAMP-induced tyrosine phosphorylation of the EGF receptor is rapid and correlates with ERK1/2 activation. It occurs also in Panc-1, but not in human mesangial cells. cAMP-induced tyrosine phosphorylation of the NGF receptor is slower and correlates with Akt activation. Inhibition of EGF receptor tyrosine phosphorylation, but not of the NGF receptor, reduces cAMP-induced neurite outgrowth. Expression of dominant-negative Akt does not abolish cAMP-induced survival in serum-free media, but increases cAMP-induced ERK1/2 activation and neurite outgrowth. Together, our results demonstrate that cAMP induces dual signaling in PC12 cells: transactivation of the EGF receptor triggering the ERK1/2 pathway and neurite outgrowth; and transactivation of the NGF receptor promoting Akt activation and thereby modulating ERK1/2 activation and neurite outgrowth.


INTRODUCTION
Detection of Neurite Outgrowth   Cells grown in 24-well dishes were exposed to forskolin, CPT-cAMP, EGF or NGF or vehicle for 24 h in serum-containing DMEM. Cells were visualized by phase-contrast microscopy and representative cells were photographed with a CCD camera. Images were prepared using Adobe PhotoshopTM 6.0 software. Immunoprecipitation of the EGFR   For the experiments 80% confluent serum-starved cells were used. About 3 × 10 6 cells grown in culture flasks were incubated with indicated agents at 37°C. At specified times, the incubation was stopped by the addition of lysis buffer (50 mM Hepes, pH 7.0, 100 mM NaCl, 0.2 mM MgSO 4 , 0.5 mM Na 3 VO 4 , 0.4 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, 10 µg/ml leupeptin, 10 µg/ml aprotinin).

Cell Death Assay/Detection of Apoptosis
The EGFR was immunoprecipitated by the addition of anti-EGFR antibody. After an incubation for 2 h at 4°C with gentle agitation, protein G-Sepharose was added and the incubation was continued for 2 h. Immunoprecipitates were washed three times in lysis buffer, resuspended in 2 × SDS sample buffer, boiled for 4 min, and separated on SDS polyacrylamide gels under reducing conditions.
Immunoblotting   Gel-resolved proteins were electrotransferred to polyvinylidene difluoride sheets, and immunoblotting was performed as recently described (48,49). Antigenantibody complexes were visualized using horseradish peroxidase-conjugated antibodies and the enhanced chemiluminescence system. X-ray films were scanned and processed by Adobe PhotoshopTM 6.0 software.
Reproducibility of the results   Results are representative of at least three experiments on different occasions giving similar results.

RESULTS
To investigate if the EGFR participates in cAMP-induced signaling, we examined the effects of forskolin, a direct activator of adenylyl cyclase, and of membrane-permeable cAMP analogs (8-Br-cAMP, CPT-cAMP) on tyrosine phosphorylation of the EGFR by antiphosphotyrosine immunoblotting of EGFR immunoprecipitates. As illustrated in Fig. 1A and 1B, forskolin and 8-Br-cAMP caused rapid and transient tyrosine phosphorylation of the EGFR with a maximum after 3-5 min. A similar result was obtained in EGFR overexpressing PC12 cells, when cellular lysates were analyzed by immunoblotting with an antibody that recognizes specifically the tyrosine phosphorylated EGFR (Fig. 1C).
The effect of cAMP on tyrosine phosphorylation of the NGFR was analyzed by immunoblotting of cell lysates with antibodies recognizing specifically phosphorylated forms of the NGFR. Forskolin induced tyrosine phosphorylation of the NGFR as revealed by immunoblotting of cellular lysates with an antibody recognizing specifically pY-674/675-NGFR in the activation loop (Fig. 1D). A similar result was obtained when immunoblotting was performed with an antibody recognizing pY-490-NGFR, the phosphorylation of which is crucial for NGF-induced ERK1/2 activation, differentiation and activation of PI3-kinase (50, 51). Compared with its effect on EGFR tyrosine phosphorylation, the kinetics of forskolininduced tyrosine phosphorylation of NGFR was less rapid and more sustained (maximum after 60 min). Thus, our data indicate that cAMP induces tyrosine phosphorylation of the receptors for EGF and NGF in PC12 cells.
Analysis of total cellular lysates from PC12 cells stimulated with forskolin for 3 to 5 min by anti-phosphotyrosine immunoblotting shows that forskolin induced rapid increase in tyrosine phosphorylation of several protein bands (Fig. 1E). Major forskolin-responsive bands migrated at 170, 130/140, and 100 kDa. pp170 comigrated with the EGFR. Longer periods of stimulation of the cells with forskolin did not result in detectable increase in protein tyrosine phosphorylation. These data indicate that forskolin induces rapid tyrosine phosphorylation of several proteins in addition to the EGFR and NGFR.
To assess if cAMP-mediated EGFR tyrosine phosphorylation is a general phenomenon or whether it is confined to PC12 cells, we studied the effect of forskolin on EGFR tyrosine phosphorylation in Panc-1 cells, a pancreatic carcinoma cell line with moderate EGFR expression. As illustrated in Fig. 2, incubation of Panc-1 cells with forskolin also caused tyrosine phosphorylation of the EGFR. In human mesangial cells, however, we did not detect forskolin-induced EGFR tyrosine phosphorylation (data not shown). Thus, cAMP-induced EGFR tyrosine phosphorylation appears to occur in some, but not all EGFR-expressing cell types.
RTK activation involves complex formation of the EGFR with the adaptor proteins Grb2, Shc and Gab1, and tyrosine phosphorylation of SH2-domain-containing substrates such as Shc (7). Immunoprecipitation of Shc and analysis of the immunoprecipitates with antiphosphotyrosine showed that forskolin caused tyrosine phosphorylation of Shc (Fig. 3A). The specific EGFR tyrosine kinase inhibitor AG1478 abolished forskolin-induced tyrosine phosphorylation of Shc. To investigate if cAMP-induced tyrosine phosphorylation of the EGFR is accompanied by recruitment of adaptor proteins to the EGFR, the cells were stimulated with forskolin and EGFR immunoprecipitates were analyzed by antiphosphotyrosine, anti-Shc, anti-Grb2, and anti-Gab1. As shown in Fig. 3B, forskolin increased the amount of Shc, Grb2 and Gab1 coprecipitating with the EGFR, indicating that activation of adenylyl cyclase induces complex formation of the EGFR with Grb2, Shc, and Gab1. In contrast, our attempts to detect adaptor proteins in NGFR immunoprecipitates were unsuccessful because the immunoprecipitation of the NGFR was insufficient.
To investigate the possible involvement of the EGFR or NGFR in cAMP-induced ERK1/2 activation as well as neurite outgrowth, we tested the effects of forskolin in parental and EGFR overexpressing PC12 cells. ERK1/2 activity was detected by immunoblotting of cellular lysates with an antibody recognizing the dually phosphorylated active form of ERK1/2. Forskolin and CPT-cAMP caused robust neurite outgrowth in EGFR overexpressing cells (Fig. 4B), whereas their effect on morphology of parental PC12 cells was minor (Fig.   4A). Forskolin-and CPT-cAMP-induced ERK1/2 activation were more pronounced and sustained in cells overexpressing the EGFR ( Fig. 4A and 4B), and its kinetics correlated well with EGFR tyrosine phosphorylation ( Fig. 1A-1C). EGFR tyrosine kinase inhibition by AG1478 strongly reduced ERK1/2 activation as well as neurite outgrowth in response to forskolin or membrane-permeable cAMP analogs ( Fig. 4A and 4B), whereas this manipulation had no effect on NGF-induced responses (data not shown). Similarly, PD165393, another EGFR tyrosine kinase inhibitor, abolished forskolin and CPT-cAMPinduced ERK1/2 phosphorylation (Fig. 4B). This indicates that activation of the EGFR is involved in cAMP-induced ERK1/2 activation and neurite outgrowth. However, the inhibitory effect of AG1478 was less pronounced on cAMP-driven responses compared to its effect on the EGF response (Fig. 4B), indicating that the effects of cAMP are not entirely dependent on activation and neurite outgrowth. To investigate if the NGFR participates in cAMP-induced ERK1/2 activation and neurite outgrowth, we studied the effect of forskolin on these responses in a NGFR-defective PC12 cell line. As illustrated in Fig. 4C, the ability of forskolin to induce ERK1/2 activation and neurite outgrowth was not impaired in NGFRdefective cells as compared to parental cells, indicating that the NGFR is not essential for cAMP-induced ERK1/2 activation and neurite outgrowth. In support of this assumption, cAMP-induced ERK1/2 activation and neurite outgrowth were not inhibited by the NGFR tyrosine kinase inhibitor K252a (data not shown).
Previous studies have provided evidence that G i/q -coupled receptors can induce EGFR transactivation by proteolytic cleavage of the EGF-like transmembrane precursor pro-HB-EGF by metalloproteinase activity (52, 53). Moreover, a recent study shows that the proforms of NGF and of brain derived neurotrophic factor are secreted and cleaved extracellularly by the proteases and can thereby activate neurotrophin receptors (54). To investigate if cAMPinduced tyrosine phosphorylation of the EGFR involves ligand-dependent mechanisms, we studied the effect of neutralizing anti-HB-EGF antibody on forskolin-induced EGFR tyrosine phosphorylation and ERK1/2 activation. Moreover, we tested the effect of anti-NGF antibody on forskolin-induced phosphorylation of the NGFR and Akt. All neutralizing antibodies had virtually no effect on forskolin responses, suggesting that cAMP-induced EGFR and NGFR activation are HB-EGF-and NGF-independent.
In PC12 cells the mechanism of ERK1/2 activation by a rise in [cAMP] i has been claimed to be Ras-dependent (21, 28) or Ras-independent, but Rap1-dependent (24). We inhibited forskolin-induced ERK1/2 phosphorylation to a similar extent as the EGF response.
These data indicate the involvement of Ras in cAMP-induced ERK1/2 activation.
Another response elicited by activated EGFR and NGFR is stimulation of Akt, which mediates the pro-survival effects of RTK, including that of NGF in neuronal cells by activation of Akt (12). As cAMP induces tyrosine phosphorylation of both the EGFR and NGFR, we investigated if cAMP-induced tyrosine phosphorylation of these RTKs is coupled to activation of Akt. Akt activation was detected by immunoblotting of cellular lysates with an antibody that specifically recognizes pS-473-Akt. Forskolin or CPT-cAMP increased Akt phosphorylation ( Fig. 6A and 6B). The kinetics of forskolin/CPT-cAMP-induced phosphorylation of Akt was considerably slower than that for Akt activation by EGF and NGF and followed the kinetics of cAMP-induced phosphorylation of the NGFR on tyrosine-490 and tyrosines 674/675 (Fig. 6A). It did not correlate with cAMP-induced ERK1/2 activation ( Fig. 6A) and EGFR tyrosine phosphorylation ( Fig. 1A-1C). The NGFR tyrosine kinase inhibitor K252a blocked forskolin-induced Akt phosphorylation without influencing ERK1/2 phosphorylation, whereas AG1478 inhibited forskolin-induced ERK1/2 phosphorylation without influencing phosphorylation of Akt in response to forskolin ( Fig. 6A and 6B). In NGFR-defective cells, forskolin had no effect on Akt phosphorylation, whereas EGF caused rapid and strong activation of Akt (Fig. 6C). Thus, cAMP-induced activation of Akt depends critically on NGFR activation and is EGFR-independent, although activation of the EGFR by EGF causes Akt activation. These data suggest that the response patterns of EGFR and NGFR stimulated by cAMP only partially overlap with those induced by their cognate ligands, i. e.
the cAMP-activated EGFR is coupled to ERK1/2 activation but not to activation of Akt, whereas the opposite holds for NGFR.
To investigate if the restricted response pattern of cAMP-activated EGFR and NGFR are due to differential efficacies in their coupling to the ERK1/2 and Akt cascade, we examined the dose-response curves for EGF and NGF to elicit activation of ERK1/2 and Akt.
As can be infered from Fig. 7, the dose-response curves of the growth factors to induce ERK1/2 and Akt phosphorylation were almost identical. Thus, the partial responses of cAMPactivated EGFR and NGFR may not be due to differential efficacies in their coupling to the ERK1/2 and Akt cascades. Expression of K179M-Akt did not reduce forskolin-induced pro-survival effect in both parental as well as NGFR-defective PC12 cells, whereas the effects of EGF and NGF were strongly reduced. Thus, our data indicate that the pro-survival effect of cAMP is independent of NGFR-dependent activation of Akt. AG1478 did not reduce forskolin-induced pro-survival effect (data not shown), indicating that the EGFR does not participate in cAMP-induced trophic effect as well.
Examining the effect of Akt on forskolin/CPT-cAMP-induced ERK1/2 activation and neurite outgrowth in PC12 cells, we found that expression of K179M-Akt significantly increased these responses (Fig. 9A and 9B). This suggests that Akt imposes an inhibitory The present study also shows that forskolin-induced tyrosine phosphorylation of the EGFR is accompanied by rapid tyrosine phosphorylation of the adaptor protein Shc and recruitment of Shc, Grb2, and Gab1 to the EGFR. These events are well known to occur upon EGF-induced activation of the EGFR and thus support the assumption that cAMP induces activation of the EGFR in PC12 cells. Furthermore, these signaling intermediates might be critical for cAMP-mediated signal transmission and the recruitment of these proteins to the EGFR could be involved in cAMP-induced activation of ERK1/2 and other responses such as differentiation, induction of neurite outgrowth and survival in PC12 cells (63).
The mechanisms of cAMP-induced tyrosine phosphorylation of the EGFR and the NGFR remains to be established. One possibility is that cAMP-induced tyrosine phosphorylation of the RTKs involves ligand-dependent mechanisms. Previous studies have provided evidence that G i/q -induced receptors can induce EGFR transactivation by proteolytic cleavage of EGF-like transmembrane precursor heparin-binding EGF-like growth factor (HB-EGF) by metalloproteinase activity (52, 53). Moreover, a recent study shows that the proforms of NGF and of brain derived neurotrophic factor are secreted and cleaved extracellularly by proteases and can thereby activate neurotrophin receptors (54). However, we did not observe any effect of neutralizing anti-HB-EGF antibody on forskolin-induced EGFR tyrosine phosphorylation or ERK1/2 activation. Likewise, immunoneutralizing anti-NGF antibody did not ablate forskolin-induced tyrosine phosphorylation of the NGFR and Akt. Thus, cAMPinduced EGFR and NGFR activation appears to be HB-EGF and NGF-independent. Similar to our observations, tyrosine phosphorylation of the NGFR in response to pituitary adenylyl cyclase activating peptide (PACAP), which is coupled to activation of adenylyl cyclase is not inhibited by anti-NGF antibody (61).

cAMP-induced ERK1/2 activation in PC12 cells has been reported to occur through a
Ras-dependent pathway (21) or a Ras-independent, Rap1-dependent pathway (24  Whether these mechanisms play a role in cAMP-induced survival in PC12 cells remains to be established. Whereas cAMP-induced Akt activation appears to be of minor importance in mediating survival response, the present study suggests that one role of cAMP-induced NGFR-dependent activation of Akt is suppression of ERK1/2 activation and neurite outgrowth. This is concluded from our finding that dominant-negative Akt strongly increases cAMP-induced ERK1/2 activation and neurite outgrowth. In agreement with an inhibitory role of Akt in PC12 cell differentiation, expression of dominant-negative Akt has recently been shown to enhance NGF-induced differentiation (76). As Akt has been reported to inhibit activation of c-Raf-1 and B-Raf (77, 78), Akt may inhibit cAMP-induced ERK1/2 activation and neurite outgrowth at the level of B-Raf, the major Raf isoform expressed in PC12 cells.
In summary, the present study demonstrates that activation of the cAMP pathway involves stimulation of two receptor tyrosine kinases, i. e. the EGFR triggering activation of the ERK pathway and cellular differentiation, and the NGFR mediating activation of Akt. Serum-starved Panc-1 cells were exposed to forskolin (20 µM) or EGF (10 ng/ml) for 3 min at 37°C in serum-free DMEM. The incubation was terminated by replacement of the medium with lysis buffer. Cell lysates were immunoprecipitated with anti-EGFR antibody followed by analysis of the immunoprecipitates by anti-phosphotyrosine immunoblotting. Antigenantibody complexes were visualized by horseradish peroxidase-conjugated antibodies and the enhanced chemiluminescence system.