Transcriptional Regulation of Amyloid Precursor Protein During ...

12 downloads 2868 Views 1MB Size Report
(APP) during development of the nervous system sug- in vitro culture models .... RNase-free DNase (35 U; Promega) for 25 mm at 37°Cand. NG1O8-15 cells ...
Journal of Neurochemistry Lippincott—Raven Publishers, Philadelphia © 1997 International Society for Neurochemistry

Transcriptional Regulation of Amyloid Precursor Protein During Dibutyryl Cyclic AMP-Induced Differentiation of NG1O8-15 Cells Masoud Shekarabi, Martin Bourbonnière, André Dagenais, and Josephine Nalbantoglu Department of Neurology and Neurosurgery, McGill Centerfor Studies in Aging, McGill University, Montreal, Quebec, Canada

Abstract: Early expression of amyloid precursor protein (APP) during development of the nervous system suggests that this protein may play an important role first in axogenesis and later in synaptogenesis. To study regulation of APP mRNA expression in neuronal cells, NG1O815 neuroblastoma x glioma cells were induced to differentiate in the presence of dibutyryl cyclic AMP. Steadystate levels of APP mRNA and APP isoforms increased gradually, concomitantly with the appearance of differentiated phenotype. Northern blot analysis showed a threefold increase in APP expression at day 6 of dibutyryl cyclic AMP treatment. Nuclear run-on assays and transient transfections performed using APP promoter/reporter constructs confirmed a twofold increase in the rate of APP gene transcription. The stability of the mRNA was unchanged, with differentiated and nondifferentiated cells having the same half-life of about 21 h. These results strongly suggest that APP mRNA induction in the differentiated NG1 08-15 cells is due to an increase in the rate of transcription of the gene. Key Words: Amyloid precursor protein—Expression—Transcription—Differentiation. J. Neurochem. 68, 970—978 (1997).

though the function of APP is not known, studies with in vitro culture models point to a role in neuronal survival and differentiation (Araki et a!., 1991; Mattson et a!., 1993) and in promoting neuritic outgrowth and branching (Milward et al., 1992; Koo et al., 1993; Jin et al., 1994; Allinquant et al., 1995). The axonal transport of APP (Koo et al., 1990; Ferreira et al., 1993) and its immunolocalization to the axon shaft and growth cones of cultured hippocampal embryonic neurons (Mattson, 1994) also support its involvement in synaptic plasticity. Further evidence for this postulate is provided by the developmental expression profile of APP. During embryogenesis, APP expression is restricted initially to developing central and peripheral nervous systems (Salbaum and Ruddle, 1994; Lorent et al., 1995). In the mouse, APP is first detected in 9.5-day embryos in motor neurons of the hind brain and spinal cord (Salbaum and Ruddle, 1994); APP may be involved in axogenesis, because it is expressed only on differentiated neurons at a time when they grow neurites and generate axons. In the rat, levels of total APP mRNA are highest in the second postnatal week (Sherman and Higgins, 1992; Ohta et al., 1993), the time of brain maturation, suggesting that, at this later period, APP expression may be associated with synaptogenesis. The presence of APP at points of neuronal contact in adult rat brain also implicates the protein as functioning in cell—cell contact (Shivers et al., 1988; Masliah et al., 1992). This pattern of expression implies that regulation

Amyloid precursor protein (APP) was first identifled through the isolation of cDNA clones encoding /3-amyloid protein (Goldgaber et al., 1987; Kang et al., 1987; Robakis et al., 1987; Tanzi et al., 1987). This 39—43 amino acid peptide is the major constituent of senile plaques, which accumulate in brain parenchyma in Alzheimer’s disease, Down’s syndrome, and, to a lesser extent, normal aging (Glenner and Wong, 1984; Masters et al., 1985). APP is a family of extensively processed proteins that are generated by tissuespecific alternative splicing of a single copy gene on human chromosome 21 (for review, see Selkoe, 1994). Analysis of expression by in situ hybridization and immunocytochemistry has shown neurons to be the main site of APP mRNA and protein synthesis in nervous tissue (Bahmanyar et al., 1987; Palmert et al., 1988; Schmechel et al., 1988; Arai et al., 1991). Al-

July 24,October 1996; revised manuscript received September 24,Received 1996; accepted 21, 1996. Address correspondence and reprint requests to Dr. J. Nalbantoglu at Montreal Neurological Institute, 3801 University Street, Montreal, Quebec, Canada H3A 2B4. Abbreviations used: APP, amyloid precursor protein; CAT, chloramphenicol acetyltransferase; CRE,GAPDH, cyclic AMP-regulated elements; dbcAMP, dibutyryl cyclic AMP; glyceraldehyde-3-phosphate dehydrogenase; SDS, sodium dodecyl sulfate; SSC, saline— sodium citrate buffer.

970

APP EXPRESSION DURING NEURONAL DIFFERENTIATION

of the APP gene is both developmental and cell-type specific. To date, no neuron-specific regulatory Sequence elements have been identified in the promoter of the APP gene (Salbaum et al., 1988). In addition,

the transcription factors and environmental stimuli that govern its early expression in the developing nervous system are not known. Hence, we undertook a study of the regulation of the APP gene in neuronal cells

undergoing differentiation. As clonal lines of neuroblastoma and somatic hybrids derived from neuroblastoma have provided useful models of neuronal development, we chose the mouse neuroblastoma x rat glioma hybrid cells NG1O8-15, which were generated

by cell fusion of the mouse neuroblastoma cell line N18TG2 and rat glioma cell line C6-BU-1 (Nelson et al., 1976). After treatments that increase the intracellular cyclic AMP pool, NG1O8-15 cells extend numerous

971

1992). After 16 h of exposure to the precipitate at 37°Cin a 3% CO 2 incubator, each culture was washed once with 5 ml of phosphate-buffered saline, and cells were divided

equally in two dbcAMP 60-mm dishes. Cells at were allowed to attach for h before was added a final concentration of 14mM to one of the cultures. Transfections were allowed to proceed for 48 h at 37°Cin a 5% CO2 incubator. The transfected plasmids consisted of the following. A

3.8-kb BamHI fragment containing the APP promoter (Salbaum et al., 1988) was introduced in the BamHI site of pBLCAT6, a promoterless chloramphenicol acetyltransferase (CAT)-expression vector (Boshart et al., 1992) creating pAPPCAT-3699. pAPPCAT-96 was engineered from pAPPCAT-3699 by making an internal deletion using the restriction enzymes Sail and NarI. After digestion, the ends were filled with the Klenow fragment of DNA polymerase I and

ligated. CAT assays were performed as described previously (Bourbonnière and Nalbantoglu, 1996) and were quantified

neurite-like processes that can form functional synaptic contacts with cultured myotubes (Nelson et al., 1978), and intercellular synapse formation can be increased after transfection with synapsin Ill, (Han et al., 1991).

by exposing the thin-layer chromatography plates to a Phosphorlmager screen (Molecular Dynamics) and analyzed using the ImageQuant software.

These cells also express many other neuronal characteristics, such as excitability, synthesis, and release of

Northern blot analysis Total RNA was extracted with TRIZOL Reagent (Gibc0BRL). RNA (10 ,ug) was electrophoresed on a 1% agar-

acetylcholine (Hamprecht, 1977; Furuya and Furuya, 1983; Han et al., 1991), and expression of muscarinic

ose/formaldehyde gelmembrane (Sambrook(Amersham et al., 1989) and transferred to Hybond-N International)

acetylcholine receptors, a2-adrenergic receptors, and receptors for a variety of neuropeptides (Nirenberg et

by capillary blotting overnight in 20 x SSC [1 X SSC is 0.15 M NaC1, 0.015 M sodium citrate]. RNA was UV cross-

al., 1983).

linked in a Stratalinker (Stratagene) set at the automatic

To begin a dissection of gene regulatory sequences

setting, and the blots were prehybridized and hybridized as described (Sambrook et al., 1989) with radiolabeled eDNA

that might be important in governing expression levels of APP during neuronal development, we studied APP

probes prepared by the random primer labeling technique

expression during treatment of NG1 08-15 cells with dibutyryl cyclic AMP (dbcAMP). We observed an upregulation of APP expression during the differentiation period. We investigated the mechanism of the increase

(GibcoBRL). After overnight incubation at 42°C, filters were washed at increasing stringencies, with the last wash consisting of 0.1 X SSC, 0.1% sodium dodecyl sulfate (SDS) at 42°Cfor The probes that were used in this work consist of15themm. following: a 620-bp human APP

by performing nuclear run-on assays, transient transfections with APP promoter/reporter constructs, and determination of APP mRNA stability. We report that the induction in APP mRNA occurs through transcrip-

eDNA probe from 3’ coding and noncoding sequences (nucleotides 1,795—2,415; numbering according to Kang et al., 1987); a chicken cDNA probe for /3-actin (Kost et al., 1983); a eDNA probe for rat glyceraldehyde-3-phosphate

tional mechanisms.

dehydrogenase (GAPDH; Tso et al., 1985); a cDNA probe of ribosomal protein L32 (Bozzoni et al., 1981); and l8S

MATERIALS AND METHODS Cell culture and differentiation NG1O8-15 cells were maintained in Dulbecco’s modified

Eagle’s medium (GibcoBRL) supplemented with 10% fetal calf serum, 2 mM glutamine, 50 U/ml penicillin, and 50 ~~g/ ml x l0~) streptomycin were cultured (GibcoBRL). 4 days in Dulbecco’ For differentiation, s modified cells Eagle’s (1 medium, 5% (vol/vol) fetal calf serum with 1 mM dbcAMP (Sigma). The medium was then changed with the addition of fresh dbcAMP and 1% fetal calf serum every 2 days. Analysis of RNA degradation rate was carried out in the presence of actinomycin D (20 ~.tg/ml;Boehringer-Mannheim); in dbcAMP-treated cells, actinomycin D was added at day 6 of differentiation.

Transfection and promoter/reporter assays NG1O8-15 cells were plated at 5 X l0~cells in 60-mm dishes. Cells were transfected 24 h later by a modification of the standard calcium phosphate method (Ausubel et al.,

(Grummt, 1981) to normalize loading variations. Autoradiographic bands were quantified by scanning densitometry using a Phosphorlmager (Molecular Dynamics) and ImageQuant software.

Nuclei preparation and nuclear were grown run-on in theassay presence or 7 cells absence A totalofofdbcAMP 8 X i0 for 6 days, washed and harvested in cold 1 mM EDTA in phosphate-buffered saline, and resuspended in lysis buffer as described (Schibler et al., 1983). Nuclear pellets were resuspended in the presence of storage buffer (Schibler et al., 1983) containing 100 U/ml RNasin and kept at —70°C. For each assay, 100 btl (2 X iO~nuclei) of the nuclei preparation was used in the elongation reaction as described (Dufort et al., 1993). The reaction was then digested with RNase-free DNase (35 U; Promega) for 25 mm at 37°Cand with 0.2 jig/al of proteinase K solution for 30 mm at 42°C to eliminate DNA and protein contaminations, respectively. RNA extraction was done by TRIZOL Reagent as described J. Neurochem., Vol. 68, No. 3, 1997

972

M. SHEKARABI ET AL.

(O’Conner and Wade, 1992). The RNA pellet was then resuspended in 20 mM HEPES (pH 7.5), 5 mM EDTA, passed through a Sephadex G-50 spun column, and treated with 0.62 M NaOH for 13 mm; the reaction was quenched with 1 MHEPES free acid (pH 5.5), and RNA was precipitated by addition of 0.3 M sodium acetate (pH 5.6) and ethanol for 3 h. The precipitate was then dissolved in hybridization solution and counted in a /3 counter. Plasmid DNA (10 ~.tg)for each of the probes (APP, GAPDH, /3-actin, and L 32 cDNAs) was linearized, denatured, and immobilized onto 0.45-~.tmnylon membranes (GibcoBRL) by using a dot blot apparatus (Bio-Rad) as described (Kafatos et al., 1979). The strips were prehybridized in individual tubes for 4 h at 42°Cin a mix containing 50 mM HEPES (pH 7), 0.75 M NaC1, 50% formamide, 0.5% SDS, 2 mM EDTA, 10 x Denhardt’s, 500 ~.tg/ml salmon sperm DNA, and polyribo-ADP (10 ~ig/ml). Hybridization was performed in the same mix plus 3 x 106 cpm of the RNA probe for 70 h at 42°C.The strips were then washed in 2 X SSC for 30 mm at 42°C,incubated in fresh 2 x SSC and 10 ~g/ml of RNase A for 30 mm at 37°C,and washed in 2 x SSC for 30 mm at 37°C. To quantify the results, densitometry was performed (as described above). A slot containing plasmids with the inserted fragments and a slot containing the plasmid DNAs (no insert) were hybridized in a separate experiment. No hybridization ofthe vectors to the labeled RNAs was observed, indicating that the signals detected are specific to the inserts (eDNA).

SDS—polyacrylamide gel electrophoresis and immunoblotting Immunoblot analysis of APP isoforms was carried out essentially as described previously (Piccardo et al., 1993). For detection of APP isoforms secreted into the medium, NG1O8-15 cells were differentiated for the appropriate time and the medium replaced for 18 h with OPTI-MEM I (GibcoBRL), a serum-free synthetic medium that was supplemented with 1 mM dbcAMP. The medium was centrifuged at 3,500 g for 15 mm, and the supernatant was filtered on a 0.22-tim Millipore membrane, dialyzed with several changes against distilled water, and then lyophilized. A fraction of the medium corresponding to the equivalent amount of cells to obtain 100 ~.tgof protein was denatured in sample buffer prior to SDS—polyacrylamide gel electrophoresis on 7% resolving gels. Immunoblots were incubated with a monoclonal antibody directed against the N-terminal domain of APP, mAb22Cl 1 (Boehringer-Mannheim) at a concentration of 0.25 ,ug/ml. After extensive washing, the immunoblots were reacted with anti-mouse IgG linked to biotin (1:350; Amersham), followed by a further incubation with streptavidin linked to horseradish peroxidase (1:400; Amersham). The immunocomplexes were revealed with 4-chloro-

FIG. 1. Effect of dbcAMP treatment on the morphology of the

NG1O8-15 cells. A: Control cells after 6 days in culture. B: Cells treated for 6 days with 1 mM dbcAMP. Although NGIO8-15 cells express APP isoforms at low levels constitutively, dbcAMP-treated cells have increased levels of cell-associated APP and secreted APP upon differentiation (Fig. 2). Immunoblot analysis with a monoclonal antibody (mAb22Cl 1) against an N-terminal epitope of APP revealed a complex profile of protein bands: four major bands of 105, 112, 118, and 125 kDa were resolved in the 100—130 kDa range; additional diffuse bands were detected around 140 kDa, at 139 and 144 kDa. The secreted extracellular domain of APP was highly enriched in the medium and was detected by this antibody as several isoforms ranging in size from 97 to 120 kDa (Fig. 2, lane 4), which increased steadily in concentration during differentiation (data not shown). Thus, this cell line

l-naphthol (Sigma) as chromogen in 50 mM Tris-HC1 (pH

seems to express the full complement of posttransla-

7.5), containing 0.01% (vol/vol) H2O2 (Sigma).

tionally modified APP isoforms in a pattern that is similar to what has been reported in homogenates of

RESULTS

brain tissue and neuronal cell lines, such as PCI2 (Selkoe et al., 1988; Buxbaum et al., 1990).

APP is up-regulated in the presence of dbcAMP NG1O8-15 cells cultured with dbcAMP began to round up, slow down proliferation, and extend thick long processes on the third day of treatment. After 6

Preliminary results indicated that APP mRNA is also increased during treatment of NGIO8-15 cells with 1 mM dbcAMP. To determine the mechanism of upregulation and to establish the time course of the in-

days, the cells (in 1% serum) showed well-differentiated features similar to normal cultured neurons, such as sympathetic and dorsal root ganglion cells (Fig. 1).

crease, we performed northern blot analysis to quantify steady-state levels of APP mRNA (Fig. 3). The signals on the blot were normalized using l8S RNA as a stan-

.1. Neurochem., Vol. 68, No. 3, 1997

APP EXPRESSION DURING NEURONAL DIFFERENTIATION

973

hybridize a maximum amount of newly transcribed radioactive mRNA, and to reduce the influence of the nonlabeled pre-mRNA that was already synthesized

before isolation of the nuclei that might compete with the radioactive mRNA for the same eDNA, a large amount of plasmid (10 ,ug) was blotted onto the membrane. Results of the nuclear run-on assays showed an increase in the rate of transcription of 2.0 ±0.2-fold in APP, with no changes in L 32 transcription in comparison with the control (Fig. 4). The results of the nuclear run-on assays for the /3-actin and GAPDH tranFIG. 2. Immunoblot analysis of APP isoforms expressed by

NG1O8-15 cells treated with dbcAMP. Cell lysates of untreated NG1O8-15 cultures (lane 1) and cultures treated with dbcAMP for 6 days (lane 2) and 8 days (lane 3) were electrophoresed with samples of conditioned medium (lane 4) obtained after 8 days of dbcAMP treatment as described in Materials and Meth-

ods. Molecular mass markers are in kDa.

dard for lane loading and compared with the same time point in the control untreated cells. As shown in Fig. 3B, the gradual increase of APP mRNA in dbcAMPtreated cells was observed at early days (1.3-fold) and continued up to day 6 of differentiation (3.1 ±0.2-

scripts were comparable to those from the northern blot analysis (Fig. 4). To verify that the effect of dbcAMP is mediated through 5’ upstream regulatory sequences of the APP gene, a reporter construct in which the bacterial CAT

gene is under the control of different portions of the APP promoter (—96 to + 100, —3,699 to + 100) was transfected into NG1O8-15 cells, which were treated subsequently with dbcAMP (Fig. 5). These transient transfection experiments revealed that dbcAMP acti-

yates transcription from pAPPCAT-3699 and increases

the reporter CAT levels twofold (Fig. 5). Taken together, these data strongly suggest that the observed

fold). This level of expression was maintained until day 8, and APP mRNA levels decreased starting at day 9 (data not shown). Similar results were obtained with a probe derived from the 3’ noncoding region, unique to APP eDNA, indicating that the signal detected in Fig. 3A is not due to cross-hybridization with mRNAs such as APLP I and APLP2 (amyloid precursor-like proteins) (Wasco et al., 1992, 1993). The same northern blot was stripped and hybridized

with a eDNA probe for the ribosomal protein L32, which was chosen to control for signal variations in nuclear run-on assays (see below). L32 mRNA pattern

showed a slight increase of 1.75-fold on the first day and reached a plateau from day 3 to 6, when there was no increase in comparison with the control (Fig. 3). The results obtained with L32 indicated that there is no likely up-regulation of rRNA in differentiated cells. During the course of differentiation, we observed changes in the expression pattern of /3-actin and

GAPDH, which are widely used as standards for normalizing signals on northern blots. In comparison with control at day 6, /3-actin had increased 1.5 ±0.2fold, whereas GAPDH increased during the first day of treatment by 1.4 ± 0.1-fold, followed by a slight decrease by day 6 (0.9 ±0.1-fold; data not shown). The increase of APP expression is transcriptional The steady-state level of mRNA is affected by both its rate of synthesis and its rate of degradation. To examine whether the observed increase of APP expression in dbcAMP-treated cells at day 6 is due to an increase in the rate of APP transcription, run-on assays were performed on nuclei isolated from treated and nontreated NG1O8-l5 cells at day 6 of differentiation

to quantify levels of nuclear transcripts. In order to

FIG. 3. Time course of dbcAMP-induced APP gene expression

in NG1Q8-15 cells. A: Total RNA was isolated from control cells or cells treated with Northernthe analysis was performed using1 mM 10 ~gdbcAMP of RNA for from6 days. each sample; same blot was probed with eDNA specific for APP and L.32 and a probe for 1 8S. B: Quantification of northern blot signals by Phosphorlmager scanning. Loading variations were corrected using 1 8S probe as a standard. Fold increase was calculated by comparison ± with the represent same timethree pointdifferent in the control untreated Means SEM experiments for cells. APP probe. The increases observed at days 5 and 6 were significant by paired t test (day 5, p = 0.03; day 6, p = 0.01).

.1. Neurochem., Vol. 68, No. 3, 1997

974

M. SHEKARABI ET AL.

The up-regulation of APP does not seem to be an early event in the differentiation process, but is rather

correlated with extension of neurites from days 3 to 6. In NG 108-15 cells, APP induction occurs after that of the cytoskeletal marker neurofilament L, but precedes

the up-regulation of GAP-43, a protein that is enriched in growth cones (M. Shekarabi and J. Nalbantoglu, unpublished observations). In previously studied neuronal cell culture models, acquisition of the neuronal

FIG. 4. Effect ofdbcAMP on the rate of APP transcription. Nuclei

were isolated from NG1O8-15 cultures untreated or treated for 6 days with 1 m M dbcAMP. Run-on transcription was allowed 2P]UTP. Labeled RNA to forand 30 mm in the presence of [° wasproceed isolated, equal amounts of radiolabeled material were hybridized to plasmid DNA immobilized on nylon paper. The extent of hybridization for APP, GAPDH, and l3-actin was quanti-

fied by Phosphorlmager scanning, and fold increase was calculated. Means ± SEM represent three different assays. The increase in APP hybridization detected in the dbcAMP-treated cells was significant by the paired t test (p = 0.01).

phenotype has been shown to be accompanied by induction of APP transcripts or changes in the splicing of APP transcripts. In P19 teratocarcinoma cells differentiated with retinoic acid, there is de novo synthesis of APP transcripts and protein, predominantly APP695

(Fukuchi et al., 1992),cell whereas treatment the human NT2 teratocarcinoma line with retinoicofacid results not in up-regulation of expression, but in a switch from APP75O/771 to APP695 (Wertkin et al., 1993). In PCI2 cells treated with nerve growth factor, results

have been contradictory. There have been reports of changes occurring only in the differential splicing pat-

tern and not in overall levels of APP transcript (Smith

APP mRNA induction at day 6 in northern blot experiments is due to an increase in the rate of transcription

of APP mRNA. The half-life of APP mRNA is not affected by the dbcAMP In order to determine if changes in mRNA stability contribute at all to the increase in APP mRNA steadystate levels, we measured the half-life of APP mRNA

in the presence and absence of 1 mM dbcAMP in NG1O8-15 cells. The rate of degradation of APP mRNA was estimated by culturing NG1O8-15 cells in the presence of actinomycin D, an inhibitor of de novo transcription. As shown in Fig. 6, no changes were observed in the APP mRNA half-life (=~2 I h) between

differentiated and nondifferentiated NG 108-15 cells. The relatively long half-life of APP mRNA in these cells is consistent with the presence of APP RNA binding proteins identified as nucleolin and heterogeneous

nuclear ribonucleoprotein C by Zaidi and Malter (1995) (data not shown). DISCUSSION We have shown that long-term treatment of NGIO815 cells with dbcAMP causes a gradual increase in the

expression of APP mRNA along with cell differentiation. This is accompanied by a concomitant increase in the levels of APP isoforms expressed by the cells.

FIG. 5. Activation of human APP promoter by dbcAMP treat-

The highest levels of mRNA were observed from days 6 to 8 of treatment, with a decrease from day 9. To investigate the mechanism of up-regulation of expres-

ment. A: CAT assay ofconstructs NG1O8-15pAPPCAT-96 cells transfected withRepresentative the APP promoter/reporter (basal promoter) and pAPPCAT-3699 without dbcAMP (lanes 1 and 3) and following 48 h of dbcAMP treatment (lanes 2 and 4). B: Relative CAT activity of the two promoter/CAT constructs

sion, the present studies were carried out at day 6 of treatment, a time point at which the cells showed well-

after dbcAMP treatment. Transfections The wereerror performed using two independent plasmid preparations. bars indicate SEM (n 4). The transcriptional increase observed after treat-

differentiated features and had acquired the phenotype

ment with dbcAMP was significant by the paired t test (p