Nerve growth factor modulates the expression and secretion of β-amyloid precursor protein through different mechanisms in PC12 cells

Authors


Address correspondence and reprint requests to Dr Angel Pascual, Instituto de Investigaciones Biomedicas (CSIC), Arturo Duperier 4, 28029 Madrid, Spain. E-mail: apascual@iib.uam.es

Abstract

The β-amyloid protein, component of the senile plaques found in Alzheimer brains is proteolytically derived from the β-amyloid precursor protein (APP), a larger membrane-associated protein that is expressed in both neural and non-neural cells. Overexpression of APP might be one of the mechanisms that more directly contributes to the development of Alzheimer's disease. The APP gene expression is regulated by a number of cellular mediators including nerve growth factor (NGF) and other ligands of tyrosine kinase receptors. We have previously described that NGF increases APP mRNA levels in PC12 cells. However, the molecular mechanisms and the precise signalling pathways that mediate its regulation are not yet well understood. In the present study we present evidence that NGF, and to a lesser extent fibroblast growth factor and epidermal growth factor, stimulate APP promoter activity in PC12 cells. This induction is mediated by DNA sequences located between the nucleotides − 307 and − 15, and involves activation of the Ras–MAP kinase signalling pathway. In contrast, we have also found that NGF-induced secretion of soluble fragments of APP into the culture medium is mediated by a Ras independent mechanism.

Abbreviations used
APP

β-amyloid precursor protein

CAT

chloramphenicol acetyl transferase

DMEM

Dulbecco's modified Eagle's medium

EGF

epidermal growth factor

FGF

fibroblast growth factor

NGF

nerve growth factor

PKC

protein kinase C.

One of the characteristic features of Alzheimer's disease is the presence of senile plaques in which the β-amyloid protein, a 40–42 amino-acid peptide, is the major component. This 4-kDa peptide is derived by proteolytic cleavage from a set of alternatively spliced β-amyloid precursor proteins (APP), which are encoded by a single gene located on chromosome 21 (Selkoe 1994). The APP gene is ubiquitously expressed in mammalian tissues, and gives rise to three major APP messenger RNAs that encode for the isoforms APP695, APP751 and APP770. The precursor protein APP plays a central role in Alzheimer's disease, and significant alterations of APP expression might contribute to its development (Zhong et al. 1994). The appearance of an Alzheimer-like pathology in Down's syndrome, probably caused by the trisomy 21-associated duplication of the APP gene (Neve et al. 1988), the degeneration of neurones overexpressing APP (Yoshikawa et al. 1992) or the appearance of β-amyloid-immunoreactive deposits in transgenic mice carrying the human APP cDNA (Quon et al. 1991; Games et al. 1995), strongly suggest a positive correlation between the overexpression of APP and the formation of amyloid deposits.

Overexpression of the APP gene might result in an aberrant processing of the amyloid precursor, which leads to a higher concentration of amyloidogenic fragments, and is associated with the formation of β-amyloid deposits in the brain (Fukuchi et al. 1992). The APP gene expression is regulated in different cell types by a variety of stimuli, including phorbol esters (Yoshikai et al. 1990; Trejo et al. 1994), thyroid hormones (Belandia et al. 1998; Latasa et al. 1998) or retinoic acid (Konig et al. 1990). In addition, and as NGF is considered to be of benefit in Alzheimer's and other neurodegenerative diseases, it is of interest to analyse the effects induced by NGF and other neurotrophins on APP. Nerve growth factor (NGF), the best characterized neurotrophic factor, has been found to increase APP mRNA in SH-SY5Y human neuroblastoma cells (Konig et al. 1990), in mouse brain primary cultures (Ohyagi and Tabira 1993) and in developing hamster brain (Mobley et al. 1988). In PC12 cells, a rat pheochromocytoma cell line that constitutively expresses APP, NGF has been reported to influence transcript levels (Cosgaya et al. 1996), splicing of APP mRNA isoforms (Smith et al. 1991; Fukuyama et al. 1993), localization (Fukuyama et al. 1993), catabolism and secretory processing of APP (Refolo et al. 1989; Rossner et al. 1998). Moreover, it has also been suggested that APP might act to mediate the effects induced by NGF on neurite outgrowth (Milward et al. 1992). However, so far the precise mechanisms involved in this regulation remain largely unclear.

The trophic effects of NGF are mediated by the specific tyrosine kinase receptor TrkA, and also by the low-affinity neurotrophin receptor p75NTR (Barbacid 1994; Segal and Greenberg 1996). Nerve growth factor binds to the receptor, and activates various intracellular signalling pathways that mediate the phosphorylation of specific transcription factors, the activation of target genes and finally the biological responses induced by the neurotrophin. Accumulating evidence indicates that the binding of NGF to its receptors, p75NTR and TrkA, mediates the responses of the neurotrophin on APP mRNA expression, splicing, and protein secretion in PC12 cells (Rossner et al. 1998). In addition, TrkA receptor gene expression is decreased in nucleus basalis (Higgins and Mufson 1989), and parietal cortex (Hock et al. 1998) of patients with Alzheimer's disease, suggesting that this receptor could play a role in the neurodegenerative process associated with this pathology. We have previously reported that in PC12 cells NGF, as well as epidermal (EGF) and basic fibroblast (bFGF) growth factors, increase APP mRNA levels by a mechanism that very likely involves activation of Ras (Cosgaya et al. 1996), and is probably mediated by specific sequences of the APP promoter (Lahiri and Nall 1995; Lahiri et al. 1999). However, the exact mechanisms, and the precise signalling pathways that mediate these effects of NGF remain elusive.

In this work we have examined the role of the ras oncogene in the response induced by NGF on APP in PC12 cells, and also in two different subclones of PC12 stably transfected with a dominant negative mutant of the ras oncogene (M-M17-26), or with an activated ras oncogene under control of the MMTV promoter (UR61). We have determined the promoter activity induced by NGF or dexamethasone in those cell lines, and partially analysed the sequences of DNA that mediate the response of the APP promoter to NGF. We have also examined whether the mitogen-activated protein kinase cascade is involved in this response by using several dominant negative mutants, and finally analysed the specific pattern of intracellular and secreted isoforms of APP induced by NGF in these cells. Nerve growth factor induces the APP promoter activity, and, according to our results, this activation is mediated by Ras throughout the MAP kinase signalling pathway. Nerve growth factor-induced activation of the promoter causes a specific increase of the intracellular content of APP. In contrast, the release of APP isoforms to the culture medium appeared to be quite Ras-independent. On the other hand, the response to NGF was significantly reduced in a small fragment (− 15/+102) of the promoter that, however, maintains a residual response likely to involve direct effects of NGF on different components of the basal transcriptional machinery.

Materials and methods

Cell cultures

PC12 cells were cultured in RPMI-1640 medium containing 10% donor horse serum and 5% fetal calf serum (GIBCO, Life Technologies Ltd, Paisley, UK) in collagen treated plates. The subclones of PC12 cells, UR61 and M-M17-26 cells, were grown in a similar medium containing 10% horse – 5% fetal calf serum. UR61 cells were derived from PC12 cells following transfection with a plasmid containing the transformant mouse N-ras oncogene under the transcriptional control of the dexamethasone-inducible mouse mammary tumour virus promoter (Guerrero et al. 1988). The PC12 subline M-M17-26 was obtained by Szeberenyi et al (1990) after transfection with the dominant negative mutant Ha-ras (Asn-17) gene transcribed from the promoter of the mouse methallothionein-1 gene. This subclone constitutively express high levels of mutant Ras (Val-12) protein, which could not be further induced by zinc. In our culture conditions both subclones, as well as parental PC12 cells, exhibit more than 95% viability as determined by trypan blue exclusion. In addition, the proliferative ability of M-M71-26 cells is not affected by constitutive expression of RasAsn17 (Szeberenyi et al. 1990). In UR-61 cells, RasVal12 induces differentiation; however, differentiation is not accompanied by a decrease in the protein content per culture. The same occurs in parental PC12 cells treated with NGF, whereas proliferative ability of M-M17-26 cells is not significantly affected by the neurotrophin.

The PC12 subclones were incubated with the different factors at the concentrations and for the times indicated in the figures. NGF and dexamethasone were obtained from Sigma (St Louis, MO, USA), and bFGF and EGF were obtained from PeproTech EC Ltd. (London, UK).

Plasmids

The chloramphenicol acetyl transferase (CAT) reporter plasmid containing the − 1099/+105 fragment of the human APP gene has been previously described (Belandia et al. 1998). Progressive 5′ deletions to − 487, − 307 and − 15 bp were prepared by polymerase chain reaction from the original − 1099/+105 bp fragment, kindly provided by Dr Lahiri's labouratory (Lahiri and Robakis 1991), and subcloned into the BamH1 site of pBL-CAT8, a plasmid that lacks the AP-1 binding site present in the pUC backbone. Expression vectors for the dominant negative mutants of Ras, Raf and MAPK have been previously described. These vectors contain the inhibitory Ha-rasAsn17 mutant (Feig and Cooper 1988), a dominant negative Raf (Raf-C4) (Bruder et al. 1992), or the Chinese hamster p44MAPK mutated in the kinase domain (Meloche et al. 1992).

DNA transfection and determination of CAT activity

The PC12 cells were transfected in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum. The RPMI growing culture medium was replaced by DMEM 4–6 h before transfection, and cells were transfected by the calcium phosphate coprecipitation method with 1 µg of reporter plasmids and carrier DNA. One hundred nanograms of a luciferase reference vector was simultaneously used as an internal control of the transfection efficiency. In cotransfection experiments, 1 µg of reporter plasmid and 10 µg of the corresponding dominant negative expression vector were used. After 16 h of incubation in the presence of calcium phosphate, the medium was discarded and washed with 5 mL of phosphate-buffered saline. A new RPMI medium containing 0.5% serum was added and the cells were then incubated for an additional period of 48 h in the presence or absence of NGF (50 ng/mL). Each treatment was performed in duplicate cultures that normally showed less than 5–15% variation in CAT activity, which was determined by incubation of [14C]-chloramphenicol with the same amount of cell lysate protein. After autoradiography, the non-acetylated and acetylated [14C]-chloramphenicol were quantified. Each experiment was repeated at least 2 or 3 times with similar relative differences in regulated expression. All data are expressed as the mean ± standard deviation.

Western blot analysis

Cellular proteins were extracted by lysis with a buffer (150 mm NaCl, 50 mm Tris pH 8, 2 mm EDTA, 1% Triton, 0.1% SDS) containing the protease inhibitors PMSF (1 mm) and leupeptin (10 µg/mL). The protein content of cells was determined by using the BCA assay (Pierce, Illinois, USA), according to the manufacturer's instructions. Identical amounts, 40 µg, of cell extracts were then electrophoresed in an 8% SDS-polyacrylamide gel and transferred to an immobilon polyvinylidine difluoride membrane. Non-specific binding was blocked with 5% non-fat dried milk in TBS-T (Tris buffered saline, 0.1% Tween 20) for 2–3 h at room temperature and the cellular APP was detected with a 1/1500 dilution of the rabbit polyclonal antibody 369 A, raised against the carboxi-terminal domain of human APP. After 1 h incubation at room temperature (25°C)the membrane was washed and incubated with a second biotinylated anti-rabbit antibody (1/2000) for an additional hour, washed again and finally incubated for 1 h with 1/2000 peroxidase-conjugated streptavidine. All incubations took place at room temperature, and detection by enhanced chemiluminiscence (ECL, Amersham International plc., UK) was carried out according to the manufacturer's indications.

Secreted full-length APP isoforms were detected by the same method from 50 µL (1 : 100 from total) of conditioned medium using the monoclonal antibody 22C11 at a final concentration of 10 µg/mL. The apparent molecular mass (kDa) of the detected bands was always determined by using a wide range protein standard (Mark 12 from Novex, San Diego, CA, USA).

Results

Induction of APP promoter activity by NGF and other neurotrophins

We have previously reported that NGF, as well as EGF or bFGF, effectively increases APP-mRNA levels in PC12 cells by a mechanism that requires activation of Ras (Cosgaya et al. 1996). To examine whether the specific changes induced by these growth factors on the APP mRNA are induced at transcriptional levels, we analysed the effect of NGF, EGF and bFGF on the APP promoter activity. As shown in Fig. 1(a), treatment with 50 ng/mLNGF for 48 h increased by approximately four-fold the CAT activity in PC12 cells transiently transfected with a chimeric plasmid containing the APP promoter linked to the CAT reporter gene. Epidermal growth factor and bFGF also stimulated APP promoter activity, although their effect was weaker than that found with NGF. Because the ras oncogene is involved in different actions of growth factor ligands of tyrosine kinase receptors in PC12 cells, we examined the effect of activated Ras on APP promoter activity in the PC12 subline UR61, which contains a transfected N-ras oncogene (Rasval12) under control of the glucocorticoid-inducible MMTV promoter. Figure 1(b) shows that promoter activity was significantly induced in UR61 cells treated for 48 h with 100 nm dexamethasone. Because, as also shown in Fig. 1(a), dexamethasone treatment of PC12 cells did not increase promoter activity, these results indicate that activated ras is responsible for promoter stimulation by the glucocorticoid in UR61 cells. The induction of APP promoter activity by dexamethasone in UR61 was comparable to that produced by NGF in the parental PC12 cells. In contrast, NGF caused a weaker increase in UR61 cells, in which the neurotrophin does not elicit neurite outgrowth (Guerrero et al. 1988; Cosgaya et al. 1997).

Figure 1.

Regulation of APP promoter activity by NGF and other neurotrophins in PC12 cells. Promoter activity was determined in three different subclones of PC12 cells, the native cell line (a), the M-M17-26 cell line that expresses a dominant negative ras mutant (c), and UR61 cells which express the N-ras (val-12) oncogene in response to dexamethasone (b). The cells were transfected with a CAT reporter gene containing the − 1099/+105 fragment of the APP promoter, and CAT activity was measured after a 48 h period of incubation in the presence or absence of 100 nm dexamethasone (Dex), 50 ng/mL NGF, 17 ng/mL bFGF or 60 ng/mL EGF. CAT activity was normalized for total protein, and results are expressed as fold induction over the levels obtained in the corresponding control untreated cells. Data are the mean ± SD from three independent experiments.

To determine whether endogenous Ras is required for growth factor stimulation of the APP promoter in PC12 cells, we examined the ability of NGF, EGF and bFGF to increase APP promoter activity in the PC12 subclone M-M17-26, which as indicated above constitutively expresses the dominant inhibitory RasAsn17 mutant. As illustrated in Fig. 1(c), none of the three growth factors were able to stimulate APP promoter activity in the presence of the dominant negative ras mutant, showing the requirement of functional Ras for this response.

Characterization of the signalling pathway that mediates the effects of NGF on the APP promoter

The results obtained in M-M17-26 cells transfected in a stable manner with the dominant inhibitory ras mutant were confirmed in parental PC12 cells transiently transfected with RasAsn17. Figure 2 shows that transfection with this mutant totally abolished the response to NGF in PC12 cells. As Raf can act downstream of Ras in growth factor signal transduction, the influence of transient transfection with an expression vector for a dominant negative form of Raf was also analysed. As shown in Fig. 2, the mutant Raf had an effect identical to that caused by RasAsn17, blocking the response of the APP promoter to the neurotrophin in PC12 cells. The same was found with a MAP kinase mutant, which also very significantly reduced the promoter response to NGF. In addition, transfection of these mutants decreased basal CAT levels (data not shown), thus suggesting that even in the absence of NGF there exists a certain activation of the Ras/MAPK pathway that contributes to basal APP promoter activity.

Figure 2.

Stimulation of APP promoter activity by NGF is mediated by the MAP kinase signalling pathway in PC12 cells. The CAT plasmid containing the − 1099/+105 fragment of the APP promoter was transfected alone or in combination with dominant negative forms (dn) of Ras, Raf and MAPK. CAT activity was determined in untreated cells and in cells treated with 50 ng/mL NGF for 48 h. CAT activity is expressed in each case as fold induction above the control values found in NGF-untreated cells, and represent mean ± SD values of four determinations (duplicates from two independent experiments).

Identification of DNA regions mediating the regulation of APP transcriptional activity

To map the DNA sequences of the APP gene that mediate the NGF-induced response, progressively deleted fragments of the promoter (− 1099, − 487, − 307 and − 15) were linked to the CAT reporter gene and transfected into PC12 cells. As illustrated in Fig. 3, not only the basal activity, but also the response to NGF was affected along the different fragments of promoter studied. Deletion from nucleotides − 1099 to − 487 increased basal promoter activity without affecting significantly the response to NGF. Deletion to nucleotide − 307 caused a decrease of CAT levels similar to those found with the fragment extending to nucleotide − 1099. In addition, induction with NGF was similar, indicating that sequences between nucleotides − 307 and − 1099 do not significantly contribute to the response of the APP promoter to NGF. However, further deletion of sequences between − 307 and − 15 caused not only a strong decrease of basal promoter activity, but also a marked reduction of the response to NGF, as only a residual response to the neurotrophin was found in PC12 cells transfected with the plasmid containing the fragment from 15 to 102.

Figure 3.

Analysis of DNA regions that mediate the effect of NGF on the APP promoter. PC12 cells were transiently transfected with pBL-CAT plasmids containing progressive deletions of the APP promoter and CAT activity was determined in cell extracts after a 48 h incubation in the absence or presence of NGF. Data are expressed relative to CAT activity obtained with the construct containing the − 1099/+ 105 bp fragment and are the mean ± SD from two separate experiments performed with duplicates.

Effects of NGF on the cell-associated proteins and the accumulation of secreted forms into the culture medium

After 48 h of incubation in the presence or in the absence of NGF (PC12 and M-M17-26 cells), or dexamethasone (UR61 cells), the content of APP was analysed in cells and culture medium by western blot using the antibodies 369 A and 22C11, respectively. The analysis of the intracellular APP isoforms is illustrated in Fig. 4(a). Based on previous descriptions, the immunoreactive bands detected by western blot analysis in PC12 cells contain at least six protein species corresponding to immature and mature forms of APP (APP695, APP751 and APP770; Buxbaum et al. 1990; Weidemann et al. 1989). Incubation of PC12 cells with NGF (50 ng/mL) for 48 h, does not affect the general pattern of intracellular proteins (data not shown), but leads to a generalized increase of the intracellular APP content, that is not observed in the M-M17-26 cell line, which constitutively expresses the dominant negative Ras mutant. In contrast, incubation of UR61 cells with 100 nm dexamethasone, which induces the expression of the stably transfected ras oncogene, leads to a similar increase in the intracellular content of APP. In addition, dexamethasone did not affect the intracellular content of APP in the native PC12 cells or in M-M17-26 cells (data not shown). Figure 4(b) shows the effects of NGF and dexamethasone on the levels of APP soluble isoforms accumulated into the culture medium. As expected, NGF treatment leads to an increased sAPP content in the PC12 conditioned medium. However, and in contrast to that observed with the intracellular APP content, the amount of secreted isoforms was not modified by dexamethasone in UR61 cells. Furthermore, the amount of sAPP accumulated into the culture medium was increased in the NGF-treated M-M17-26 cells.

Figure 4.

Western blot analysis of cellular and secreted APP isoforms. After 48 h of incubation in the absence or presence of NGF (PC12 and M-M17-26 cells), or dexamethasone (UR61 cells), the intracellular and secreted isoforms of APP were analysed by western blot with the antibodies 369 A (a) and 22C11 (b), respectively. The figure includes blots obtained in a representative experiment. The apparent size (kDa) of the protein bands is indicated at the right by arrows. The relative changes induced by the different treatments are indicated by a number, at the bottom, representing the mean ± SD obtained from three independent experiments.

Discussion

It has been extensively documented that NGF can modulate the postranscriptional processing and secretion of APP in PC12 cells (Refolo et al. 1989; Rossner et al. 1998; Rossner et al. 1999). In addition, we have previously demonstrated that NGF, as well as the growth factors bFGF and EGF, increases APP-mRNA levels through a Ras-dependent mechanism (Cosgaya et al. 1996).

In this study we have examined the effects of NGF on APP expression in three variants of PC12 cells, the parental cell line and two different subclones, which represent a good model to investigate the role of Ras in NGF-induced effects in PC12 cells. UR61 cells express oncogenic N-ras and M-M17-26 cells contain a dominant inhibitory mutant of ras. Transient transfection studies demonstrate that NGF and to a lesser extent bFGF and EGF stimulate APP promoter activity in PC12 cells. It has been reported that NGF can stimulate APP promoter activity when these cells are treated with the growth factor for a duration of 4 days prior to, and 4 days after transfection with a plasmid containing the APP promoter (Lahiri and Nall 1995). We demonstrate here that 48 h of incubation with NGF after transfection are sufficient to activate the APP promoter. These results demonstrate that NGF increases APP gene expression at transcriptional level.

Activation of the protein tyrosine kinase receptors initiates a cascade of intracellular signalling events leading to regulation of specific genes and finally to regulation of a wide variety of cellular responses. Numerous proteins and several second messengers, among them the Ras-MAP kinase cascade (Muroya et al. 1992; Thomas et al. 1992; Wood et al. 1992), have been implicated in the signalling pathway that follows binding of neurotrophins to their Trk receptors. Evidence that activation of endogenous Ras mediates the effects of NGF on APP promoter activity comes from the experiments with the PC12 subclone M-M17-26, which constitutively expresses the dominant inhibitory mutant of ras Asn-17. This mutant has a reduced affinity for GTP and inhibition of endogenous Ras function by this mutant has been suggested to occur through competition with normal Ras for regulatory proteins that promote nucleotide exchange (Feig and Cooper 1988). Szerebenyi et al. (1990), have shown that the mutant blocks the neurite outgrowth induced by NGF or FGF, as well as the induction of different genes. Our present results demonstrate that expression of the Asn-17 mutant also blocks the stimulation of APP promoter activity induced by NGF in PC12 cells, thus showing the requirement of functional Ras for this response. The results obtained in dexamethasone-induced UR61 cells support this hypothesis further. In this subclone of PC12 cells the transfected ras proto-oncogene is under control of a dexamethasone inducible MMTV promoter. As dexamethasone does not affect the expression of APP in PC12 cells, it can be assumed that Ras is the signal transduction pathway used by dexamethasone to increase the intracellular content of APP in UR61 cells. Moreover, the results obtained with dominant negative mutants of Ras, Raf and MAP kinase, strongly suggest that NGF-induced stimulation of the APP promoter activity not only requires activation of Ras, but also the complete activation of the Ras-MAP kinase signalling pathway.

Many transcription factors that constitute final targets of specific transduction pathways (Hunter and Karin 1992) bind to specific response elements in the regulated genes, and mediate the effects induced by neurotrophins. The APP promoter has the typical structure of a housekeeping gene, lacking the TATA and CAAT elements. Its sequence is extremely well conserved between rat, mouse and human (Chernak 1993), and contains multiple positive and negative regulatory elements and consensus sequences for the binding of several transcription factors (Salbaum et al. 1988;La Fauci et al. 1989; Vostrov and Quitschke 1997; Querfurth et al. 1999; Wright et al. 1999). In agreement with previous descriptions (Lahiri et al. 2000), we found that progressive deletions of promoter sequences led to significant variations of the basal promoter activity. This activity was maximal in the construct containing the − 487 fragment of the APP promoter, suggesting the presence of a silencer element between this position and the nucleotide − 1099, and it was further decreased in successive deletions to nucleotides − 307 and − 15. Deletion to nucleotide − 307 caused a significant decrease of basal activity, which could be secondary to the loss of the ‘5′-GC element/AP1 site’ tandem described in the − 383/−358 region of the promoter (Querfurth et al. 1999). Further deletion of sequences to nucleotide − 15 caused an additional reduction which probably represents the loss of previously described response elements that are essential to maintain the basal promoter activity (Salbaum et al. 1988;La Fauci et al. 1989; Vostrov et al. 1997; Querfurth et al. 1999).

The response to NGF was maintained in the plasmid that contains the first 307 bp of the 5′ upstream sequence, but it was significantly reduced in the next deletion plasmid extending to nucleotide − 15. These results indicate that stimulation of the APP promoter activity by NGF is mainly mediated through sequences located between nucleotides − 307 and − 15. In addition the smallest promoter fragment used (− 15/+102) retains a residual response that probably involves direct effects of NGF on different components of the basal transcriptional machinery, or alternatively the existence of regulatory sequences located downstream of the main transcriptional initiation site.

Our results strongly suggest that transcriptional activation of APP gene expression is responsible for the NGF-induced increase in APP transcripts previously described by us in PC12 cells (Cosgaya et al. 1996). We have reported that NGF increases APP mRNA levels in PC12 cells. Furthermore, expression of activated ras in UR61 cells also leads to a significant increase in APP transcripts, whereas the expression of a dominant negative mutant of ras blocked the induction of APP gene expression by NGF in the M-M17-26 cell line. Stimulation of transcription should lead also to an increase in the cellular content of APP proteins. This was indeed observed in PC12 cells, where NGF treatment resulted in a significant increase in the intracellular levels of the different APP isoforms. Moreover, our results indicate that this effect was also largely dependent on Ras. As also occurred with promoter activity (Cosgaya et al. 1996), the NGF-induced increase of cellular APP observed in PC12 cells was not reproduced in the M-M17-26 subline, which expresses the inhibitory mutant of Ras, and was mimicked by dexamethasone-induced expression of Ras in UR61 cells. According to our results NGF appears to induce a generalized increase of the different APP species detected. However, and since a complete definition of bands has not been possible, we cannot exclude that, as previously described (Fukuyama et al. 1993), NGF only affected the synthesis of APP695, thus increasing specifically the levels of both the immature and mature forms of APP695 in PC12 cells.

In addition, it has been also reported that NGF, as well as other neurotrophins, regulates the secretion of sAPP in PC12 cells (Refolo et al. 1989; Fukuyama et al. 1993; Desdouits-Magnen et al. 1998). In agreement with those descriptions we have found that NGF increases the amount of sAPP accumulated into the cultured medium in the parental PC12 cells. In contrast, expression of activated Ras in the UR61 subline was unable to increase the extracellular sAPP content. In addition, the increased release of sAPP induced by NGF in PC12 cells was not completely abolished in the M-M17-26 cells that express a dominant negative mutant of Ras. Therefore, and contrarily to the effects induced by NGF on promoter activity, levels of mRNA and cellular content of APP, the regulation of APP secretion by NGF in PC12 cells appears to be mainly mediated by a Ras-independent mechanism.

These results are in apparent contradiction with the description that activation of the MAP kinase cascade in PC12 results in the rapid secretion of sAPP isoforms. Nevertheless, it has been also reported that although activation of MAP kinase promotes APP secretion, the inhibition of this kinase is not sufficient to reduce APP secretion (Rossner et al. 1999). Taken together, these results are compatible with the existence of a Ras-independent mechanism responsible for APP secretion in response to NGF in PC12 cells. A number of possible mechanisms could explain the specific MAP kinase-dependent/Ras-independent regulation of APP secretion by NGF. Other TrkA-interacting proteins, such as PLC-γor PI 3-Kinase (Greene and Kaplan 1995), or docking proteins such as FRS2, that are phosphorylated in response to NGF stimulation (Kouhara et al. 1997; Hadari et al. 1998) could be involved in the observed effect. In addition, activation of protein kinase C (PKC) downstream of PLC-γ could contribute to Ras-independent activation of MAP kinase through PKC phosphorylation of Raf (Sozeri et al. 1992).

Our results demonstrate that NGF activates transcription of the APP gene, thus increasing expression of APP. If this also occurs in humans in vivo, the use of NGF, proposed as a paliative treatment in Alzheimer's disease, could rather result in a risk factor for the disease. However, the neurotrophins can also increase the release of neurotrophic sAPP fragments, which very probably should be accompanied by a reduced generation and release of the β-amyloid peptide. If this is the case NGF could indeed be useful to define new therapeutic strategies for the disease. Therefore, it will be of interest to analyse the effect of NGF on APP gene expression and on the secretion of neurotrophic sAPP in humans.

Acknowledgements

This research was funded by grants from the Spanish Comisión Interministerial de Ciencia y Tecnología (SAF 97–0183) and Dirección General de Investigación de la Comunidad de Madrid (08.5/0036/1998. AV was the recipient of a fellowship from Consejo Superior de Investigaciones Científicas, Glaxo-Wellcome. We thank Drs G. M. Cooper and Dr A. Pellicer for the PC12 subclones, D K Lahiri and N K Robakis for providing the APP promoter and S. E. Gandy for the polyclonal antibody 369 A. We also thank Dr A. Aranda for critical reading of the manuscript.

Ancillary