Journal of NeurochemistryLippincott—Raven Publishers, Philadelphia© 1997 International Society for Neurochemistry

Metabotropic Glutamate Receptor Subtype mGluRlaStimulates the Secretion of the Amyloid

ß-Protein Precursor Ectodomain

Roger M. Nitsch, *Amy Deng, tRichard J. Wurtman, and *John H. Growdon

Centerfor Molecular Neurobiology, University of Hamburg, Hamburg, Germany, *Department of Neurology,Massachusetts General Hospital, Boston; and tDepartment of Brain and Cognitive Sciences,

Massachusetts Institute of Technology, Cambridge, Massachusetts, U.S.A.

Abstract: To examine the effects of glutamatergic neuro-transmission on amyloid processing, we stably expressedthe metabotropic glutamate receptor subtype la (mGlu-Ria) in HEK 293 cells. Both glutamate and the selectivemetabotropic agonist 1-amino-i ‚3-cyclopentanedicar-boxylic acid (ACPD) rapidly increased phosphatidylinosi-toI (PI) turnover four- to fivefold compared with controlcells that were transfected with the expression vectoralone. Increased Pl turnover was effectively blocked bythe metabotropic antagonist a-methyl-4-carbophenylgly-cine (MCPG), indicating that heterologous expression ofmGluRla resulted in efficient coupling of the receptorsto G protein and phospholipase C activation. Stimulationof mGluRla with glutamate, quisqualate, orACPD rapidlyincreased secretion of the APP ectodomain (APPs);these effects were blocked by MCPG. The metabotropicreceptors were coupled to APP processing by proteinkinases and by phospholipase A2 (PLA2), and melittin, apeptide that stimulates PLA2, potently increased APPssecretion. These data indicate that mGluRlcs can be in-volved in the regulation of APP processing. Together withprevious findings that muscarinic and serotonergic recep-tor subtypes can increase the secretion of the APP ecto-domain, these observations support the concept that pro-teolytic processing of APP is under the control of severalmajor neurotransmitters. Key Words: Amyloid —Alzhei-mer‘s disease— Metabotropic glutamate receptor—a-Secretase processing.J. Neurochem. 69, 704—712 (1997).

ants fall into three main groups that differ in signalingmodes and in pharmacological characteristics (for re-view, see Nakanishi, 1994; Pin and Duvoisin, 1995).Group I receptors include mGluRl and mGluR5,which activate phospholipases, phosphatidylinositol(PI) hydrolysis, and adenylyl cyclase, whereas groupsII and III receptors inhibit adenylyl cyclase activity(Boss and Conn, 1992) but differ in agonist specifici-ties. mGluRl are widely distributed in the mammalianCNS and are expressed predominantly in hippocampalneurons, in cerebellar Purkinje cells, as well as in mi-tral and in tufted cells of the olfactory bulb (Shigemotoet al., 1992).

mGluR subtypes modulate both excitatory and in-hibitory neurotransmission (McBain et al., 1994; Ger-eau and Conn, 1995), and they play a role in synapticplasticity in as much as they promote long-term depres-sion (LTD) both in cerebellum and in hippocampus(Bolshakov and Siegelbaum, 1994; Shigemoto et al.,1994; Kobayashi et al., 1996; Yokoi et al., 1996).Furthermore, activation of mGluR can induce long-term potentiation (LTP) in hippocampus (Bashir etal., 1993; Bortolotto et al., 1994). Both LTD and LTPare believed to be synaptic representations of memoryprocesses (Bliss and Collingridge, 1993; Linden,

Glutamate is one of the majör neurotransmitters inmammalian brain and modulates both pre- and post-synaptic neuronal activities through a variety of iono-tropic and metabotropic receptors (Sugiyama et al.,1989; Nakanishi, 1992). Metabotropic glutamate re-ceptors (mGluRs) are a family of G protein—coupledreceptors with seven transmembrane domain topologybut with no known sequence homology to other Gprotein—coupled receptors (Sugiyama et al., 1987;Houamed et al., 1991; Masu et al., 1991; Tanabe etal., 1992). Eight mGluR subtypes and five splice van-

Received December 20, 1996; revised manuscript received April4, 1997; accepted April 4, 1997.

Address correspondence and reprint requests to Dr. R. M. Nitschat Center for Molecular Neurobiology, University of Hamburg, Mar-tinistr. 52, 20246 Hamburg, Germany.

Abbreviations used: Aß, amyloid ß-protein; ACPD, l-amino-l,3-cyclopentanedicarboxylic acid; AD, Alzheimer‘s disease; APLP2,APP-like protein 2; APP, amyloid ß-protein precursor; APPs, se-creted APP ectodomain; DEDA, 7,7-dimethyleicosadienoic acid;DMEM/F12, Dulbecco‘s modified Eagle medium/F 12 nutrientmix-ture; ECL, enhanced chemiluminescence; HBSS, Hanks‘ bufferedsaline solution; LTD, long-term depression; LTP, long-term potenti-ation; MCPG, a-methyl-4-carbophen~‘lglycine;mGluRla, metabo-tropic glutamate receptor subtype la; OPC, oleyloxyethyl phospho-ryicholine; PI, phosphatidylinositol; PKC, protein kinase C; PLA2,phospholipase A2 PMA, phorbol myristate acetate.

704

mGluRla REGULATION OF APP PROCESSING 705

1994), and they may be the electrophysiological basisfor the observation that mGluR agonists improve mem-ory-related behaviors in rats (Straubli et al., 1994). Inaddition to these physiological functions, mGluR maybe involved in neuropathological mechanisms; phar-macological activation of mGluR can induce seizuresand can also aggravate NMDA-induced excitotoxicdamage by facilitating NMDA receptor responses toglutamate (McDonald and Schoepp, 1992; McDonaldet al., 1993).

Reductions in glutamatergic neurotransmission arebelieved to contribute significantly to the cognitive im-pairment in Alzheimer‘s disease (AD) (Francis et al.,1993), a common neurodegenerative disorder that ischaracterized by the formation in brain of neurofibril-lary tangles and amyloid plaques. Amyloid plaques arecomposed of amyloid ß-protein (Aß), a hydrophobic39—43-amino acid peptide that is derived by proteo-lytic processing from the larger amyloid ß-proteinpre-cursor (APP) (Kang et al., 1987). Proteolytic pro-cessing of APP follows several intracellular routes thatinvolve cleavage events at diverse sites clusteredaround its Aß domain (for review, see Selkoe, 1994).These proteolytic mechanisms are referred to as secre-tase processing events. a-Secretase processing in-volves cleavage within the Aß domain (Sisodia et al.,1990), generating the secreted ectodomain (APPs)along with an ‘-~10-kDaC-terminal fragment that isfurther degraded by y-secretase processing at the C-terminus of the Aß domain to produce p3 and ankDa C-terminus (Haass et al., 1993). In contrast, APPcan be processed by ß-secretase at the N-terminus,generating a truncated APPs molecule (Seubert et al.,1993), as well as a 100-amino acid C-terminal frag-ment that is the immediate precursor to generate theintact Aß peptide by y-secretase cleavage (Haass etaI., 1993). a-Secretase processingoccurs mostly in thetrans-Golgi network as well as in secretory vesicles,and at the cell surface, whereas ß-secretase processinginvolves recycling of APP from the cell surface to lateendosomes or lysosomes. There are some indicationsthat ß-secretase processing can also occur in the secre-tory pathway (De Strooper et al., 1993; Thinakaranet al., 1996). Both a- and ß-secretase processing areregulated by internal and external signals including Gprotein—coupled cell surface receptors. These includemuscarinic ml and m3, and serotonergic 5~HT2aand5-HT2, receptors. (Buxbaum et al., 1992; Nitsch et al.,1992, 1996). Furthermore, a-secretase processing ismodulated in an activity-dependent manner by actionpotentials in rat brain slice preparations (Nitsch et al.,1993; Farber et al., 1995). Evidence to date indicatesthat protein kinases and phospholipases are criticalcomponents in the signaling pathway that couples re-ceptor activation to APP processing. To test whethermGluR subtypes that are coupled to this signaling path-way can also increase a-secretase processing of APP,we stably overexpressed mGluRla in HEK 293 cells,treated them with metabotropic receptor ligands, and

measured the amount of APPs released into the culturemedia. We show that secretion of the APP ectodomainwas increased within minutes by stimulation ofmGluR 1 and that increased secretion was mediated byreceptor-coupled activation of both protein kinase C(PKC) and phospholipase A2 (PLA2). These studiesconfirm and extend our prior report that glutamate in-fluenced APP processing in primary cells (Lee et al.,1995).

MATERIALS AND METHODS

ConstructsThe cDNA encoding the human mGluR subtype la (Masu

et al., 1991) was generously provided by S. Nakanishi. ANotI/SalI restriction fragment comprising the completereading frame including the signal peptide was ligated intopBK-CMV (Stratagene). The intact reading frame of theresulting eDNA was confirmed by DNA sequencing. pBK-CMV without insert was used for control transfections.

Cell culture and stable transfectionSubconfluent monolayers of 293 cells were transfected

with 10 ~tgplasmid DNA, precipitated with calcium phos-phate, followed by 15% glycerin shock according to Sam-brook et al. (1989). Stably transfected cells were selectedwith 500 ~tg/mlGeneticin (G418, GIBCO), and clonai lineswere produced by two rounds of low density plating andcollection of individual colonies with cloning cylinders. Indi-vidual clonai lines were screened for expression of function-ally intact receptors by analyzingglutamate-induced stimula-tion of FI turnover; 53% of the screened clones effectivelystimulated Pl turnover. Experiments were repeated in threeindividual positive cell lines. Cells were maintained in Dul-becco‘s modified Eagle mediumlFl2 nutrient mixture(DMEM/Fl2) supplemented with 10% fetal calfserum, andwith 500 ~.tg/mlG418. To reduce glutamate concentrationsin the culture media before the experiments, the growth andselection medium was replaced with serum-free DMEM/Fl2without glutamine and without histidine. Experiments werealso performed with this medium. Because G418 can inter-fere with G-protein coupling, it was removed from the mediaat least 2 days before the experiments.

Metabolic labeling and P! turnover analysisCells were labeled metabolically overnight with 1.25 1iCi/

dish of myo-[2-3H]inositol (20.5 Ci/mmol; New England

Nuclear) in inositol-deficient, serum- and glutamate-freeDMEM/F12 medium, washed twice with Hanks‘ balancedsalt solution (HBSS), and treated for 10 min with 10 mMlithium chloride in HBSS. Drugs were added in the presenceof 10 mM lithium for 5—60 min at 37°C.Cells were lysedwith ice-cold methanol, and lipids were removed by chloro-form/methanol/water (2:2:1; by volume) extraction. La-beled water-soluble inositol phosphates were separated fromfree [3H]inositol by ion-exchange chromatography, usingAG 1-X8 columns (Bio-Rad), and 1 M ammonium formateand 0.1 M formic acid as eluent. Radioactivity was quanti-tated by liquid scintillation spectrometry.

Drugs -

Glutamate (Sigma, St. Louis, MO, U.S.A.) was preparedfreshly before each experiment and was used at 0—500

1.tM. trans- (±)- I-Amino- 1 ‚3,-cyclopentanedicarboxylic acid

J. Neurochem., Vol. 69, No. 2, /997

706 R. M. NITSCH ET AL.

(ACPD; 10—500 ‚uM), a-methyl-4-carboxyphenylglycine(MCPG; 400 ‚uM), quisqualate (50—500 ‚uM), chelerythrinechloride (I ‚uM), staurosporine (1

1.tM), melittin (0.1—lmM), and thapsigargin (20 nM) were purchased from Re-search Biochemicals International (Natick, MA, U.S.A.).Phorbol myristate acetate (PMA; Sigma) was used at 100nM to 1 ‚uM for both stimulation and down-regulation ofPKC. Manoalide (3.2—10 ‚uM), 7,7-dimethyleicosadienoicacid (DEDA; 20—50 ‚uM), and oleyloxyethyl phosphoryl-choline (OPA; 0.1—1 mM) were purchased from Biomol(Plymouth Meeting, PA, U.S.A.).

Antibodies, western blotting, and densitometryCells were washed twice with histidine- and glutamine-

free serum, DMEM/F12 medium, and were incubated at37°Cin the presence or absence of test substances freshlydissolved in thesame medium. Antagonist andblockers wereadded 10 min before the coincubation with the agonist. Aftertime intervals of 5—60 min. conditioned media were col-lected, cooled to 4°C,centrifuged at 300 g, and supernatantfluids were desalted by gel-filtration chromatography usingG25 Sephadex (Pharmacia) columns, with water as eluent.Desalted proteins were dried by vacuum centrifugation, re-constituted in water followed by 2>< Laemmli gel loadingbuffer, and boiled for 3 min. Total cell protein was extractedwith 1% sodium dodecyl sulfate (SDS) in Tris-buffered sa-line and was quantitated by the bicinchoninic acid assay(Pierce). Equal volumes of reconstituted secretory proteinsolutions normalized to total cell protein were separated bySDS —polyacrylamide gel electrophoresis, electroblottedonto polyvinylidene difluoride (Immobilon P, Waters) mem-branes, and blocked with 5% nonfat dry milk (Carnation)in Tris-buffered saline containing 0.05% Tween 80. Mem-branes were probed with antibodies directed against variousdomains of APP and APP-like protein 2 (APLP2). Specifi-cally, the monoclonal antibodies 22C1 1 (Boehringer Mann-heim) against the APP ectodomain and 6ElO (Senetek)against residues 1—17 of the Aß domain, as well as thepolyclonal antiserum D2-I raised against full-length APLP2expressed in a baculovirus system (Slunt et al., 1994), wereused. Secondary antibodies were visualized on preflashed x-ray films (Kodak) by enhanced chemiluminescence (ECL;Amersham). Immunoreactive bands were quantitated by la-ser scanning densitometry with an LKB Ultroscan densitom-eter set to 40-~.tmvertical interval sizeand 2.4-mm horizontalslit width. Areas under the optical density curves were ex-pressed as arbitrary units (AU) and were normalized to areasgeneratedby immunoreactive proteins secreted under controlconditions determined on the same blot. Measurements wereperformed in the linear range of the ECL reaction as deter-mined from serial dilution curves of secreted proteins. Con-trol and stimulated conditions were always handled in paral-lel, processed identically, and run in parallel lanes on thesame blot. All experiments werç done in triplicate dishesand were repeated at least three o~rfour times in three indi-vidual clonal lines. Statistical analysis was performed byANOVA, using treatments as the independent variable.

RESULTS

After stable transfection and two rounds of cloning,we screened individual clones for intact coupling ofmGluRla to G-protein activation and stimulation ofPl turnover. In 10 of 19 clones, both glutamate (Fig.

FIG. 1. Stimulation of Pl turnover in 293 cells stably transfectedwith mGluRla (O), and nontransfected 293 cells (U) by metabo-tropic receptor agonists. Glutamate (A); ACPD, a selectivemGluRagonist (B). Both glutamate and ACPD increased Pl turn-over four- to fivefold in transfected cells but did not affect wild-type cells. Adding the mGluRl antagonist MCPG (400 1.tM; A)effectively blocked the ACPD-induced accumulation of inositolphosphates (IF). Data are means from triplicate culture dishes.

lA) and ACPD stimulated Pl turnover four- to fivefold(Fig. lB). In contrast, nontransfected parent cells andcells expressing the transfection vector alone, as wellas nine of 19 clones stably expressing mGluRla, failedto increase Pl turnover in response to mGluR la stimu-lation with either glutamate or ACPD. In the cell linesthat expressed functionally intact receptors, agonist-induced Pl was completely blocked by coincubationwith MCPG, an mGluR antagonist (Fig. lB). Similarresults were obtained with two additional clonal 293lines stably expressing mGluRla. Three of the clonallines that expressed functionally coupled mGluRlawere used for the further experiments. Under basal,unstimulated conditions, Pl turnover in the lines thatexpressed functionally coupled receptor was consis-tently nine times higher compared with the wild-type,nontransfected cells (data not shown). This higherbasal Pl turnover may reflect receptor stimulation un-der basal conditions induced by glutamate, which maybe derived from cellular synthesis, although both gluta-mine and histidine were absent in the serum-free label-ing and chase media. Collectively, the data indicatethat stable transfection of mGluRla in 293 cells re-sulted in the expression of functionally active receptorsthat are coupled to, and activate, Pl turnover.

To test whether agonist-induced increase in Pl turn-over in these cells was associated with changes in se-cretory APP processing, we analyzed conditioned me-dia by western blotting and by using antibodies di-rected against the ectodomain of APP. Both thenonselective agonists glutamate and quisqualate, aswell as the selective agonist ACPD, significantly in-creased the release of APPs by the cells that stablyexpressed mGluRla (Fig. 2A). As a positive control,direct activation of PKC by PMA also increased APPsrelease. The increased secretion induced by ACPD wasblocked by the metabotropic receptor antagonistMCPG (Fig. 2B). To show that APPs secreted in re-sponse to receptor activation, we used the monoclonal

J. Neurochem., Vol. 69, No. 2, 1997

mGluRJa REGULATION OF APP PROCESSING 707

FIG. 2. Western blots of APPs secreted into the culture media. A: Both the nonselective mGluRl agonists glutamate (Glu) andquisqualate (QA), as well asthe selective agonist ACPD increased APP5 secretion (detected by 22C1 1) from 293 cells stably transfectedwith mGluRla. APPs was also stimulated by direct activation of PKC with PMA. B: MCPG (400

1iM) blocked ACPD-induced increasein APP5 release (detected by 22C11) in 293 cells stably transfected with mGluRla. C: Increased APP5 secreted in response tomGluRlastimulation was derived from a-secretase processing as indicated by the monoclonal antibody 6E1 O directed against residues1—17 of the human Aß domain. D: In control experiments with 293 cells stably transfected with the vector only, glutamatergic agonistsfailed to increase APPs release. APP5 was detected with the monoclonal antibody 22C1 1; concentrations are indicated in micromolarunits and molecular masses in kilodaltons.

antibody 6E10 that is directed against the 16 N-termi-nal residues of the Aß domain and, thus, selectivelydetects APPs derived from a-secretase cleavage. Stim-ulation of the mGluR clearly increased 6E10 immuno-reactivity in western blots of secreted proteins (Fig.2C). Control cell lines that expressed the vector onlyfailed to respond to glutamate, to quisqualate, or toACPD with changes in secretory APP processing.None of these treatments changed basal rates of APPsrelease (Fig. 2D). Time course experiments revealedthat the major mGluRla-induced increase in APPs se-cretion occurred between 30 and 60 min within recep-tor stimulation (Fig. 3), and amaximum response wasattained after 120 min. Coincubation experiments withthe protein synthesis inhibitor cycloheximide showedessentially similar time courses of mGluR la-inducedincreases in APPs release. We quantitated these re-sponses by densitometry and found that concentrationsof APPs in stimulated cells expressing mGluRla werethree- to fourfold higher than the unstimulated, basallevels (Fig. 4). Dose—response analyses showed thatglutamate concentrations of 50—1,000 ‚uM effectivelyincreased APPs release (Fig. 4A). With the exception

of the 1 mM concentration, this dose—response wasequivalent to that of glutamate-induced stimulation ofPl turnover. At 1 mM, the response to glutamate waslower than that at 500 ‚uM. It is currently unknown

FIG. 3. Glutamate stimulated APPs release from 293 cells stablyexpressing mGluRla within 60 min. O, cells stimulated with 500~M glutamate; LI, unstimulated, basal release of APPs by thesame cell line. APP5 in culture media was quantitated by densi-tometry after western blotting and immunoreaction with themonoclonal antibody 22C1 1. Data are mean values of triplicateculture dishes from a representative experiment. AU, arbitraryunits.

J. Neurochem., Vol. 69, No. 2, 1997

708 R. M. NITSCH ET AL.

FIG. 4. Dose—response of APPs release induced by mGluRlagonists applied for 60 min to 293 cells stably transfected withmGluRla. The cells were treated with glutamate at concentra-tions of 10 ‚uM to 1 mM (A), or with ACPD at concentrations of10—500 ‚uM (B) were significantly different from nontreated cells(p 0.01 by ANOVA). The increase in APPs secretion inducedby ACPD was blocked with 400 ‚uM of the mGluRl antagonistMCPG. Filled columns represent ACPD treatment alone; shadedcolumns represent coadministration of ACPD and MCPG. Dataare mean ±SEM values of n = 3—5 experiments.

whether this ~f‘fect is related to potential cytotoxic ef-fects of high glutamate concentrations. Neither gluta-mate nor ACPD changed basal APPs secretion in wild-type 293 cells. In the presence of the mGluR antagonistMCPG, ACPD at various concentrations failed to in-crease basal APPs concentrations in stably transfectedcells, indicating that ACPD increased APPs secretionspecifically through the stimulation of the transfectedmGluRs (Fig. 4B).

To identify the component steps in the signalingpathways that couple mGluRla to APP-processingpathways, we reduced PKC activity bothby pharmaco-logical inhibitors and by down-regulation with PMA.Both the nonselective kinase inhibitor staurosporineand the more PKC-specific inhibitor chelerythrinechloride effectively inhibited the increase by glutamatein APPs release (Fig. SA, lanes 1 and 3). Down-regu-lation of PKC by overnight incubation with PMA abol-ished the increase in APPs secretion by PMA and glu-tamate (Fig. 6A, lanes 6 and 7). Most, if not all, ofthe APPs released in response to stimulation ofmGluRla was processed by a-secretase cleavage, as

indicated by the monoclonal antibody 6E10 that spe-cifically detects residues 1—16 of the Aß domain. Themonoclonal antibody 22C 11 recognizes bothAPPs andits homologue APLP2 (Slunt et al., 1994). To deter-mine whether mGluRla can also stimulate secretoryprocessing of APLP2, we probed western blots withD2- 1, a polyclonal antiserum that specifically recog-nizes APLP2 (Slunt et al., 1994). Glutamate stimula-tion clearly increased secretion of the ectodomain ofAPLP2 from cells stably transfected with mGluRla(Fig. 6B).

To determine whether PLA2 is involved in couplingmGluRla to APP processing, we coadministered inseparate experiments the PLA2 inhibitors manoalideand DEDA along with glutamate to cells stably ex-pressing mGluRla. The PLA2 inhibitors blunted gluta-mate-induced increase in APPs release (Fig. 6A). Toinvestigate the possibility that the blocking effects ofthe inhibitors were due to nonspecific interaction withPI turnover, we stimulated Pl turnover with variousconcentrations of glutamate in the presence and theabsence of manoalide, DEDA, and OPC. We foundthat manoalide and OPC, but not DEDA, inhibitedglutamate-induced Pl turnover (Table 1). Thus, re-duced PI turnover may explain in part the inhibitoryeffect of manoalide on APPs secretion. However, thatDEDA did not block Pl turnover but did block stimu-lated APPs secretion shows clearly that PLA2 can beinvolved directly in coupling mGluRla to increasedAPPs secretion. Furthermore, melittin, a peptide thatstimulates PLA2 activity, increased consistently APPsrelease more than twofold over the effect of glutamateor ACPD (Fig. 6B). Melittin‘ s effect on APPs releasewas greater in 293 cells stably expressing mGluRlathan in wild-type 293 cells (Fig. 6B).

DISCUSSION

The results of this study show that glutamate, quis-qualate, and the metabotropic agonist ACPD readilystimulate a-secretase processing of APP in culturedcells that stably express the metabotropic receptor sub-type mGluRla. This finding, along with the prior ob-servation that glutamate increases the secretion of the

FIG. 5. PKC coupled mGluRla to APPs se-cretion (A). Glutamate (Glu)-induced APP5secretion (60-min drug incubation) wasblocked by the kinase inhibitor staurosporine(ST; 1 ‚uM), as well as by the more specificPKC inhibitor chelerythrine chloride (CC; 1‚uM). Direct activation of PKC with PMA (1‚uM) also increased APPs release. After down-regulation of PKC by 15-h pretreatment withPMA (+)‘ both glutamate and PMA failed toincrease APP5 release. Glutamate increasedthe release of the ectodomain of the APLP2in 293 cells stably transfected with mGluRla(B). Western blot of media proteins from cellsstably expressing mGluRla were probed withthe APLP2-specific antibody D2-1.

J. Neurochem., Vol. 69, No. 2, 1997

mGluRIa REGULATION OF APP PROCESSING 709

FIG. 6. The PLA2 inhibitors manoalide (ML) and DEDA (DE) ef-fectively blocked glutamate (Glu)-induced APPs release (A).Thapsigargin (1G), a compound that releases calcium from intra-cellular stores, failed to change APPs release in these cell lines(A). Stimulation of PLA2 with melittin, a peptide that activatesPLA2, stimulated APP5 release in 293 cells that stably overex-press mGluRla (solid columns) as well as in wild-type 293 cells(shaded columns) (B). At 1 mMconcentrations, APPs secretionwas significantly greater in the transfected cells compared withthe wild-type cells (p < 0.05, by ANOVA). Basal, unstimulatedAPPs secretion was also significantly higher in the transfectedcells than in the wild-type cells (p < 0.01, by ANOVA).

APP ectodomain in primary neurons (Lee et al., 1995)indicates that glutamate receptors can regulate secre-tory APP processing. Together with the observationthat other G protein—coupled receptors including mus-carinic and serotonergic receptor subtypes (Nitsch etal., 1992, 1996), as well as the formation of actionpotentials in brain slices, can also increase a-secretaseprocessing (Nitsch et al., 1993; Farber et al., 1995),the data reported here support the concept that severalmajor neurotransmitters can regulate secretory APPprocessing.

Glutamate caused a rapid and dose-dependent in-crease in PI turnover in 293 cells stably transfectedwith mGluRla, as evidenced by the accumulation ofradiolabeled inositol phosphates in the presence of lith-ium after metabolic labeling of PI. This result demon-strates that our transfection protocol generated func-tionally intact and stably expressed surface receptorsthat couple to the expected signal transduction systems.Expression of functionally intact receptors was notconsistently observed in all clones, because we foundthat ‘—47% of the stable clones failed to stimulate P!turnover despite the expression of mGluRla message.In functionally positive clones, basal Pl turnover was

eightfold higher in cells transfected with mGluRla.Similar increases of basal Pl turnover by mGluRlawere previously obtained in transiently transfectedcells (Gabellini et al., 1994; Lee et al., 1995), and weattribute such increases toconstant, ligand-independentstimulation of the second messenger systems, or toendogenous glutamate synthesized and secreted by thecells. Despite the transfection-related increase in basalPl turnover, the heterologously expressed receptors re-tained the ability to respond to agonists with furtherincreases in Pl turnover; addition of mGluRl receptoragonists including glutamate, ACPD, and quisqualateincreased Pl turnover ‘—37 times higher than basal Plturnover in untransfected, wild-type 293 cells.

Western-blotting analyses of secreted proteins re-vealed that metabotropic receptor agonists caused adose-dependent increase in APPs secretion. The in-crease in APPs secretion was rapid, with the steepestrate of increase between 30 and 60 min after stimula-tion. These data are similar to the kinetics of serotonin-and carbachol-induced stimulation of APPs secretion(Nitsch et al., 1992, 1996) and support the view thatpreexisting APP holoprotein is the chief substrate forthe increases in APPs in response to receptor stimula-tion. The effects reported here are specific to the me-tabotropic receptor, because the agonist-induced in-creases of APPs secretion were effectively blockedby the metabotropic receptor antagonist MCPG. Thatagonists failed to change basal rates of APPs secretionin cells transfected with the vector alone supports thespecificity of the effect. These two lines of evidenceindicate that increased APPs secretion was indeedcaused by mGluRla stimulation. Most of APPs se-creted in response to receptor activation was derivedfrom a-secretase processing, because it contained the16 N-terminal residues of the Aß domain that are rec-ognized by the monoclonal antibody 6E 10.

To dissect the second messenger signaling pathwaysthat couple mGluR 1 a to increased a-secretase pro-cessing, we examined the independent effects of PKCand PLA2. Both the kinase inhibitor staurosporine andthe more selective PKC inhibitor chelerythrine chlo-ride inhibited glutamate-induced increases in APPs se-cretion. In addition, down-regulation of PKC bychronic, 15-h incubation with PMA blunted the ex-pected glutamate-induced increased in APPs secretion.These results demonstrate that PKC is critical in cou-

TABLE 1. Effects of PLA2 inhibitors on glutamate-induced Pl turnoverin 293 cells stably overexpressing mGluRla

Glutamate Control ML DEDA OPC

0 2.54 ±0.29 1.97 ±0.06 2.07 ±0.10 1.96 ±0.11250 ‚uM 10.8 ±0.50 6.34 ±0.45 11.7 ±1.72 11.1 ±0.61500 ‚uM 27.6 ±3.45 9.33 ±1.24 26.9 ±4.38 19.8 ±1.30

Data are expressed as cpm X l0~and are mean ±SD values from three independent assays. ML,manoalide.

J. Neurochem., Vol. 69, No. 2, 1997

710 R. M. NITSCH ET AL.

pling mGluRla to secretory APP processing. PLA2can also influence APP processing, as illustrated byincreased APPs secretion in response to the PLA2-stimulating agentmelittin. Inhibiting PLA2 activity hadthe expected opposite effect; i.e., the PLA2 inhibitorsmanoalide, DEDA, and OPC blocked glutamate-in-duced APPs secretion. To control for potential nonspe-cific effects on other signaling pathways, we examinedwhether PLA2 inhibitors interferewith agonist-inducedFI turnover. Whereas one (manoalide) inhibited gluta-mate-induced FI turnover, two others (DEDA andOPC) did not, indicating that their effects on blockingagonist-induced increase of APPs secretion were spe-cific to PLA2 and not related to nonspecific interactionswith Pl turnover. Similar effects of PLA2-mediatedcoupling of surface receptors to APP processing werereported for muscarinic receptors (Emmerling et al.,1993) and for serotonin receptors (Nitsch et al., 1996).Simply increasing intracellular concentrations of freecalcium was not sufficient to change secretory APPprocessing in the cells we used; i.e., thapsigargin,which releases calcium from internal stores, failed tochange basal increases in APPs secretion. Thus, thepreviously reported effects of thapsigargin on APPssecretion may be cell-type specific (Buxbaum et al.,1994; Querfurth and Selkoe, 1995).

APP is a member of a larger family of APP-likeproteins (APLP). APLP2 is homologous to APP inboth N- and C-terminal regions but lacks the Aß do-main. Proteolytic processing of APLP2 is similar tothat of APP, and we have previously shown that stimu-lation of 5HT2a and 5-HT2. serotonin receptors in-creases the secretion of the ectodomains of APLP2.Here, we demonstrate that mGluRla can also in-crease secretory processing of APLP2 and acceleratethe rate of secretion of its ectodomain. These findingsstrengthen the view that APP and APLP2 may be com-pete for similar processing pathways.

The cellular mechanism whereby external and inter-nal signals regulate rates of APP and APLP2 pro-cessing are not completely understood. It was pre-viously reported that PKC can accelerate the buddingof secretory vesicles from the trans-Golgi network (Xuet al., 1995). Our data emphasizing the central impor-tance of PKC in APP processing support the possibilitythat kinase activation influences the formation of APP-containing secretory vesicles and accelerates theirtransport to the cell surface. Another possibility is thatreceptor stimulation and the resultant second messen-ger signaling cascade stimulates directly the proteasesinvolved in a-secretase processing. Testing this hy-pothesis must await the results of the search to identifythese proteases.

The physiological relevance of having APP pro-cessing under neurotransmitter control is unclear be-cause the function of APP remains uncertain. Manyexperimental data point to a role of APP as a neuro-trophic and neuroprotective factor, possibly as an adhe-sion molecule with the ability to bind such extracellular

matrix proteins as laminin, and the ability to promoteneunte outgrowth (Milward et al., 1992). In contrast,deleting the APP gene in mice did not cause obviousbrain or behavioral abnormalities (Zheng et al., 1995).Thus, the precise role of APP in mammalian brainremains to be determined. APP is involved, however,in the pathophysiology of AD. APP is the immediateprecursor of Aß peptides that can aggregate to formamyloid plaques, a major pathological hallmark of AD.The most direct evidence for the central importance ofAPP in causing AD comes from molecular geneticstudies; i.e., mutations of the APP gene that causemisprocessing of APP cause familial forms of the dis-ease. In a Swedish kindred, a double mutation at theN-terminus of the Aß domain causes increases in Aßlevels; in other families, single-point mutations at co-don 717 of the APP gene lead to the formation oflonger Aß molecules that may aggregate more readily.Furthermore, mutations of the presenilin 1 gene thatcause some forms of early-onset familial AD are asso-ciated with increased formation of longer Aß peptidesin blood, fibroblasts (Scheuner et al., 1996), and brain(Mann et al., 1996). Thus, a unifying theme in early-onset autosomal dominantly inherited AD is an abnor-mality in APP processing that leads to increasedamounts of Aß1 -42 in blood and in brain. WhetherAß deposition causes late-onset sporadic AD remainscontroversial, but amyloid plaque formation is an earlyand prominent feature of AD histopathology.

A rational strategy to treat AD would be to reducethe formation of potentially amyloidogenic APP deriv-atives (Nitsch and Growdon, 1994). Our findings sug-gest that one way to accomplish this goal would be toincrease a-secretase cleavage of Aß through mGluRlareceptor stimulation. However, increased a-secretaseactivity raises a theoretical concern because, in addi-tion to APPs, the p3 fragment of APP may also in-crease. The p3 fragment is a hydrophobic APP deriva-tive that may have the potential of aggregating intoamyloid-like structures. Several lines of evidence,however, challenge a role for p3 in amyloid formationand neurodegeneration in AD brain; p3 is not a majorconstituent of neuritic plaques in AD brain. Amyloidplaques are stained with antibodies against the N-ter-minus of Aß, and sequencing of peptides derived fromsenile plaques consistently identified the N-terminusof Aß. In contrast, p3 was purified biochemically fromAD brain with massive diffuse amyloid deposits(Gowing et al., 1994).

We were unable to measure either p3 or Aß in the293 cells used in this study. Because 293 cells secretevery small amounts of these derivatives, incubationtimes of 24 h and longer would have been necessaryto detect measurable levels with immunoabsorbentassays or metabolic labeling and immunoprecipitation.Such long incubation times, however, preclude thestudy of receptor-coupled phenomena that are detect-able within time periods of minutes, and that attainmaximum responses within 1 or 2 h. Receptor-stimula-

J. Neuroche,n., Vol. 69, No. 2, 1997

mGluRla REGULATION OF APP PROCESSING 711

tion paradigms that are longer than these time intervalscause internalization and down-regulation of the sur-face receptors and result in loss of the response. Ingeneral, there is a reciprocalrelation inAPP processingbetween APPs and Aß; i.e., when APPs secretion in-creases, there is a concomitant decrease in Aß (Bux-baum et al., 1993; Gabuzda et al., 1993; Hung et al.,1993; Wolf et al., 1995). Direct proof that this ruleholds for mGluR la receptor stimulation remains a goalfor future research.

Acknowledgment: We thank Dennis Selkoe and SangramSisodia for antibodies and Shigetada Nakanishi for cDNAs.

REFERENCES

Bashir Z. I., Bortolotto Z. A., Davies C. H., Berratta N., Irving A. J.,Seal A. J., Henley J. M., Jane D. E., Watkins J. C., and Colling-ridge G. L. (1993) Induction of LTP in the hippocampus needssynaptic activation of glutamate metabotropic receptors. Nature363, 347—350.

Bliss T. V. P. and Collingridge G. L. (1993) A synaptic model ofmemory: long-term potentiation in the hippocampus. Nature361, 31—39.

Bolshakov V. Y. and Siegelbaum S. A. (1994) Postsynaptic induc-tion and presynaptic expression of hippocampal long-term de-pression. Science 264, 1148—1152.

Bortolotto Z. A., Bashir Z. I., Davies C. H., and Collingridge G. L.(1994) A molecular switch activated by metabotropic glutamatereceptors regulates induction of long-term potentiation. Nature368, 740—743.

Boss V. and Conn P. J. (1992) Metabotropic excitatory receptoractivation stimulates phospholipase D in hippocampal slices. J.Neurochem. 59, 2340—2343.

Buxbaum J., Ruefli A., Parker C., Cypess A., and Greengard P.(1994) Calcium regulates processing of the Alzheimer amyloidprotein precursor in a protein kinase C-independent manner.Proc. Nati. Acad. Sei. USA 91, 4489—4493.

Buxbaum J. D., Oishi M., Chen H. I., Pinkas-Kramarski R., JaffeE. A., Gandy S. E., and Greengard P. (1992) Cholinergic ago-nists and interleukin 1 regulate processing and secretion of theAlzheimer ß/A4 amyloid protein precursor. Proc. Nail. Acad.Sci. USA 89, 10075—10078.

Buxbaum J. D., Koo E. H., and Greengard P. (1993) Protein phos-phorylation inhibits production of Alzheimer amyloid ß/A4peptide. Proc. Nail. Acad. Sei. USA 90, 9195—9198.

Caporaso G. L., Gandy S. E., Buxbaum J. D., Ramabhadran T. V.,and Greengard P. (1992) Protein phosphorylation regulates se-cretion of Alzheimer ß/A4 amyloid precursor protein. Proc.Natl. Acad. Sei. USA 89, 3055—3059.

De Strooper B., Umans L., Vanleuven F., and Vandenberghe H.(1993) Study of the synthesis and secretion of normal andartificial mutants of murine amyloid precursor protein (APP) —

cleavage of APP occurs in a late compartment of the defaultsecretion pathway. J. Cell Biol. 121, 295—304.

Emmerling M. R., Moore C. J., Doyle P. D., Carroll R. T., and DavisR. E. (1993) Phospholipase A2 activation influences the pro-cessing and secretion of the amyloid precursorprotein. Biochem.Biophys. Res. Commun. 197, 292—297.

Farber S. A., Nitsch R. M., Schulz J. G., and Wurtman R. J. (1995)Regulated secretion of ß-amyloid precursor protein in rat brain.J. Neurosci. 15, 7442—7450.

Francis P. T., Sims N. R., Procter A. W., and Bowen D. M. (1993)Cortical pyramidal neurone loss may cause glutamatergic hypo-activity and cognitive impairment in Alzheimer‘s disease: in-vestigative and therapeutic perspectives. J. Neurochem. 60,1589—1604.

Gabellini N., Manev R. M., and Manev H. (1994) Is the heterolo-gous expression of metabotropic glutamate receptors (mGluRs)

an appropriate method to study the mGluR function? Experiencewith human embryonic kidney 293 cells transfected withmGluRl. Neurochem. mt. 24, 533—536.

Gabuzda D., Busciglio J., and Yankner B. A. (1993) Inhibition ofß-amyloid production by activation of protein kinase C. J.Neurochem. 61, 2326—2329.

Gereau R. W. and Conn P. J. (1995) Multiple presynaptic metabo-tropic glutamate receptors modulate excitatory and inhibitorysynaptic transmission in hippocampal area CAl. J. Neurosci.15, 6879—6889.

Gowing E., Rober A. E., Woods A. S., Cotter R. J., Chaney M.,Little S. P., and Ball M. J. (1994) Chemical characterization ofA beta 17-42 peptide, a component of diffuse amyloid depositsof Alzheimer‘s disease. J. Biol. Chem. 269, 10987—10990.

Haass C., Hung A. Y., Schlossmacher M. G., Teplow D. B., andSelkoe D. J. (1993) ß-Amyloid peptide and a 3-kDa fragmentare derived by distinct cellular mechanisms. J. Biol. Chem. 268,3021—3024.

Houamed K. M., Kuijper J. L., Gilbert T. L., Haldeman B. A.,O‘Hara P. J., Mulvihill E. R., Almers W., and Hagen F. S.(1991) Cloning, expression, and gene structure of a G protein-coupled glutamate receptor from the rat brain. Science 252,1318—1321.

Hung A. Y., Haass C., Nitsch R. M., Qiao Qiu W., Citron M., Wurt-man R. J., Growdon J. H., and Selkoe D. J. (1993) Activationof protein kinase C inhibits cellular production of the amyloidß-protein. J. Biol. Chem. 268, 22959—22962.

Kang J., Lemaire H-G., Unterbeck A., Salbaum J. M., Masters C. L.,Grzeschik K. H., Multhaup G., Beyreuther K., and Miller-HillB. (1987) The precursor of Alzheimer‘s disease amyloid A4protein resembles a cell-surface receptor. Nature 325,733—736.

Kobayashi K., Manabe T., and Takahashi T. (1996) Presynapticlong-term depression at the hippocampal mossy flber-CA3 syn-apse. Science 273, 648—650.

Lee R. K. K., Wurtman R. J., Slack B. E., Cox A. J., and NitschR. M. (1995) Amyloid precursor protein processing is stimu-lated by metabotropic glutamate receptors. Proc. NatI. Acad.Sei. USA 92, 8083—8087.

Linden D. J. (1994) Long-term depression in the mammalian brain.Neuron 12, 457—472.

Mann D. M., Iwatsubo T., Cairns N. J., Lantos P. L., Nochlin D.,Sumi S. M., Bird T. D., Poorkaj P., Hardy J., Hutton M., CrookR., Rossor M. N., and Haltia M. (1996) Amyloid ß protein(Aß) deposition in chromosome 14-linked Alzheimer‘s disease:predominance of Aß42(43). Ann. Neurol. 40, 149—156.

Masu M., Tanabe Y., Shigemoto R., and Nakanishi S. (1991) Se-quence and expression of a metabotropic glutamate receptor.Nature 349, 760—765.

McBain C. J., DiChiara T. J., and Kauer J. A. (1994) Activationof metabotropic glutamate receptors differentially affects twoclasses of hippocampal interneurons and potentiates excitatorysynaptic transmission. J. Neurosci. 14, 4433—4445.

McDonald J. W. and Schoepp D. D. (1992) The metabotropic excit-atory amino acid receptor agonists I S,3R-ACPD selectivelypotentiates N-methyl-D-aspartate-induced brain injury. Eur. J.Pharmacol. 215, 353—354.

McDonald J. W., Fix A. S., Tizzano J. P., and Schoepp D. D. (1993)Seizures and brain injury in neonatal rats induced by 1S,3R-ACPD, a metabotropic glutamate receptor agonist. J. Neurosci.13, 4445—4455.

Milward E. A., Papadopoulos R., Fuller S. J., Moir R. D., Small D.,Beyreuther K., and Masters C. L. (1992) The amyloid proteinprecursor of Alzheimer‘s disease is a mediator of the effects ofnerve growth factor on neunte outgrowth. Neuron 9, 129—137.

Nakanishi S. (1992) Molecular diversity of glutamate receptors andimplications for brain function. Science 258, 597—603.

Nakanishi 5. (1994) Metabotropic glutamate receptors: synaptictransmission, modulation, and plasticity. Neuron 13, 1031—1037. -

Nitseh R. M. and Growdon J. H. (1994) Role of neurotransmissionin the regulation of amyloid ß-protein precursor processing.Biochem. Pharmacol. 47, 1275—1284.

J. Neurochem., Vol. 69, No. 2, 1997

712 R. M. NITSCH ET AL.

Nitsch R. M., Slack B. E., Wurtman R. J., and Growdon J. H. (1992)Release of Alzheimer amyloid precursor derivatives stimulatedby activation of muscarinic acetylcholine receptors. Science258, 304—307.

Nitsch R. M., Farber S. A., Growdon J. H., and Wurtman R. J.(1993) Release of amyloid ß-protein precursor derivatives fromhippocampal slices by electrical depolarization. Proc. Nail.Acad. Sei. USA 90, 5191—5193.

Nitsch R. M., Deng M., Growdon J. H., and Wurtman R. J. (1996)Serotonin 5-HT

2, and 5-HT2, receptors stimulate amyloid pre-cursor protein ectodomain secretion. J. Biol. Chem. 271, 4188—4194.

Pin J. P. and Duvoisin R. (1995) The metabotropic glutamate recep-tors: structure and functions. Neuropharmacology 34, 1 —26.

Querfurth H. W. and Selkoe D. J. (1995) Calcium ionophore in-creases amyloid beta peptide production by cultured cells. Bio-chemistry 33, 4550—4561.

Sambrook J., Fritsch E. F., and Maniatis T. (1989) Molecular Clon-ing. A Laboratory Manual. Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, New York.

Scheuner D., Eckman C., Jensen M., Song X., Citron M., SuzukiN., Bird T. D., Hardy J., Hutton M., Kukull W., Larson E.,Levy-Lahad E., Viitanan M., Peskind E., Poorkaj P., Schellen-berg G., Tanzi R., Wasco W., Lannfelt L., Selkoe D., andYounkin S. (1996) Secreted amyloid beta-protein similar tothat in the senile plaques of Alzheimer‘s disease is increasedin vivo by the presenilin 1 and 2 and APP mutations linked tofamilial Alzheimer‘s disease. Nat. Med. 2, 850—852.

Selkoe D. J. (1994) Cell biology of the amyloid ß-protein precursorand the mechanism of Alzheimer‘s disease. Annu. Rev. CellBiol. 10, 373—403.

Seubert P., Oltersdorf T., Lee M., Barbour R., Blomquist C., DavisD., Bryant K., Fritz L., Galasko D., Thai L., Lieberburg I., andSchenk D. (1993) Secretion of ß-amyloid precursor proteincleaved at the amino terminus of the ß-amyloid peptide. Nature361, 260—263.

Shigemoto R., Nakanishi S., and Mizuno N. (1992) Distribution ofthe mRNA for a metabotropic glutamate receptor (mGluRl) inthe central nervous system: an in situ hybridization study inadult and developing rat. J. Comp. Neurol. 322, 121—135.

Shigemoto R., Abe T., Nomura S., Nakanishi S., and Hirano T.(1994) Antibodies inactivating mGluRl metabotropic gluta-mate receptor block long-term depression in cultured Purkinjecells. Neuron 12, 1245—1255.

Sisodia S. S., Koo E. H., Beyreuther K., Unterbeck A., and PriceD. L. (1990) Evidence that ß-amyloid protein of Alzheimer‘s

disease is not derived by normai processing. Science 248, 492—495.

Slack B. E., Nitsch R. M., Livneh E., Kunz G. M. Jr., Breu J., EldarH., and Wurtman R. J. (1993) Regulation by phorbol esters ofamyloid precursor protein release from Swiss 3T3 fibroblastsoverexpressing protein kinase Ca. J. Biol. Chem. 268, 21097—21101.

Siunt H. H., Thinakaran G., Von Koch C., Lo A. C. Y., Tanzi R. E.,and Sisodia S. 5. (1994) Expression of a ubiquitous, cross-reactive homologue of the mouse beta-amyloid precursor pro-tein (APP). J. Biol. Chem. 269, 2637—2644.

Straubli U., Rogers G., and Lynch G. (1994) Facilitation of metabo-tropic receptors enhances memory. Proc. Nail. Acad. Sei. USA91, 777—781.

Sugiyama H., Ito I., and Hirono C. (1987) A new type of glutamatereceptor linked to inositol phospholipid metabolism. Nature325, 531—533.

Sugiyama H., Ito 1., and Watanabe M. (1989) Glutamate receptorsubtypes may be classified into two major categories: a studyon Xenopus oocytes injected with rat brain mRNA. Neuron 3,129—132.

Tanabe Y., Masu M., Ishii T., Shigemoto R., and Nakanishi S.(1992) A family of metabotropic glutamate receptors. Neuron8, 169—179.

Thinakaran G., Teplow D. B., Siman R., Greenberg B., and SisodiaS. 5. (1996) Metabolism of the “Swedish“ amyloid precursorprotein variant in Neuro2a (N2a) cells—evidence that cleavageat the “beta-secretase“ site occurs in the Golgi apparatus. J.Biol. Chem. 271, 9390—9397.

Wolf B. A., Wertkin A. M., Jolly Y. C., Yasuda R. P., Wolfe B. B.,Konrad R. J., Manning D., Ravi S., Williamson J. R., and LeeV. M. Y. (1995) Muscarinic regulation of Alzheimer‘s diseaseamyloid precursor protein secretion and amyloid beta-proteinproduction in human neuronal NT2N cells. J. Biol. Chem. 270,4916—4922.

Xu H., Greengard P., and Gandy S. (1995) Regulated formationof Golgi secretory vesicles containing Alzheimer beta-amyloidprecursor protein. J. Biol. Chem. 270, 23243—23245.

Yokoi M., Kobayashi K., Manabe T., Takahashi T., Sakaguchi I.,Katsuura G., Shigemoto R., Ohishi H., Nomura S., NakamuraK., Nakao K., Katsuki M., and Nakanishi 5. (1996) Impairmentof hippocampal mossy fiber LTD in mice lacking mGluR2.Science 273, 645—647.

Zheng H., Jiang M. H., Trumbauer M. E., Sirinathsinghji D. J. S.,Hopkins R., Smith D. W., Heavens R. P., Dawson G. R., BoyceS., Conner M. W., Stevens K. A., Slunt H. H., Sisodia S. S.,Chen H. Y., and Van der Ploeg L. H. T. (1995) Beta-amyloidprecursor protein-deficient mice show reactive gliosis and de-creased locomotor activity. Cell 81, 525—531.

J. Neurochem., Vol. 69, No. 2, 1997