The Journal of Neuroscience, August 1992, fZ(8): 2973-2981

NMDA and Non-NMDA Receptor-mediated Increase of c-fos mRNA in Dentate Gyrus Involves Calcium Influx via Different Routes

Leslie S. Lerea,’ Linda S. Butler,’ and James 0. McNamara1,2 ‘Department of Medicine, Division of Neurology, and 2Departments of and Neurobiology and Research Laboratory, Veterans Administration Medical Center, Duke University, Durham, North Carolina 27710

We examined the effects of selective of ionotropic minate in different programs of early gene expression and excitatory amino acid (EAA) receptor subtypes on induction correspondingly different patterns of target gene expres- of the immediate early gene c-fos. We used in situ hybrid- sion. ization to measure c-fos mRNA and fura- imaging to mea- sure intracellular calcium (Ca:+) in individual dentate gyrus The molecular mechanismsby which brief experiencesproduce neurons maintained in vitro. Activation of either NMDA or lasting modifications of nervous system function are unknown. non-NMDA receptor subtypes is sufficient to induce the rapid Long-term changescould be due to posttranslational modifi- and dramatic increase of c-fos mRNA. Activation of either cations ofexisting and/or alterations in geneexpression. NMDA or non-NMDA receptors also induces a rapid and dra- Immediate-early genes(IEGs) such as C-$X have been impli- matic increase of Ca:+, effects blocked by the removal or cated in the conversion of short-term stimuli to long-term al- chelation of extracellular calcium (Ca:+). c-fos mRNA induc- terations in cellular by regulating gene expression tion by either receptor subtype is Ca2+ dependent, since (Curran and Morgan, 1985; Goelet et al., 1986; Morgan and chelation of Ca:+ with EGTA prevents c-fos mRNA induction Curran, 1989). Activation of glutamate receptors is critical to by both NMDA and non-NMDA receptor agonists. The in- some long-lasting forms of neuronal plasticity. Transient acti- crease in Ca:+ induced by activating non-NMDA receptors vation of both the NMDA and non-NMDA subtypes of glu- is inhibited either by removal of extracellular sodium (Na:) tamate receptor have been linked to induction of long-term or by the voltage-sensitive calcium channel (VSCC) blocker potentiation (Collingridge et al., 1983; Harris et al., 1984;Anik- . By contrast, the increase of Ca:+ induced by ac- sztejn and Ben-Ari, 199 1). Transient activation of the NMDA tivating NMDA receptors is not inhibited by removal of Na: receptor also is necessaryfor induction of the long-lasting hy- or nifedipine. Consistent with these effects on Ca;+, nife- perexcitability in the kindling model of epilepsy (McNamara et dipine inhibits induction of c-fos mRNA by non-NMDA, but al., 1988; Stasheffet al., 1989). not by NMDA, receptor agonists. These findings indicate that The stimulus necessaryfor induction of kindling, a brief sei- Ca*+ serves as a second messenger coupling ionotropic zure, is sufficient to induce rapid and dramatic increasesof EAA receptors with transcriptional activation of c-fos mRNA. mRNA of multiple IEGs, including c-fis (Dragunow and Rob- The route of Ca*+ entry into dentate neurons, however, de- ertson, 1987; Shin et al., 1990). Activation of the NMDA sub- pends on the EAA receptor subtype stimulated. Non-NMDA type of during kindled seizuresis necessary receptor activation results in Ca*+ influx indirectly via VSCCs, for the full induction of C-$X mRNA, since NMDA receptor whereas NMDA receptor activation results in Ca*+ influx di- antagonistsinhibit seizure induction of c-fos mRNA by 50-70% rectly through the NMDA channel itself. Since non-NMDA and (Labiner et al., 1990). NMDA receptor antagonistsalso inhibit NMDA receptors are colocalized at some and can induction of c-fos in other paradigms (Herrera and Robertson, be activated simultaneously by synaptically released glu- 1990). Activation of the NMDA receptor subtype is sufficient tamate, the induction of c-fos mRNA by each receptor may to induce expression of c-fos in vitro, but the mechanismby reflect the activation of identical and therefore redundant which NMDA produces this effect is unknown (Szekely et al., programs of gene expression. Alternatively, the varying 1989). The Ca2+dependence ofdepolarization-induced increas- routes of Ca2+ entry following stimulation of EAA receptor es of C-$X mRNA in PC12 cells (Morgan and Curran, 1986) subtypes may activate distinct signaling pathways that cul- together with the permeability of the NMDA receptor to Ca2+ (Macdermott et al., 1986; Ascher and Nowak, 1988; Iino et al., 1990) suggestedthat Ca2+might serve as the secondmessenger coupling this receptor to C-$X mRNA induction. Received July 30, 1991; revised Feb. 5, 1992; accepted Mar. 10, 1992. We examined the effects of selective activation of the iono- We are grateful to Dr. Ken McCarthy for the generous use of the image analysis tropic subtypesofglutamate receptorson C-$X mRNA. We used system for all of the calcium studies. We thank Drs. Doug Bonhaus and Jim Burke for helpful advice and discussions. This study was supported by NIH Grants GM an in vitro neuronal system, derived from postnatal rat 12049 and NS 1777 1 and by a grant from the Veterans Administration. dentate gyrus, to addressthe following questions: (1) is acti- Correspondence should be addressed to Dr. Leslie S. Lerea, Department of Medicine, Box 3676, 401 Bryan Research Building, Duke University, Durham, vation of either NMDA or non-NMDA receptors sufficient to NC 27710. induce c&s mRNA expression;(2) is Ca*+the secondmessenger Copyright 0 1992 Society for Neuroscience 0270-6474/92/122973-09%05.00/O that links activation of thesereceptors to c-fos mRNA expres- 2974 Lerea et al. * EAA Receptor-mediated Induction of c-fos mRNA

Figure 1. NMDA-stimulated changes in fura- fluorescence in cultured den- tate gyms cells. Pseudocolor-enhanced pictures showing Ca2+ responses in dentate gyrus neurons following appli- cation of NMDA. Dentate gyrus cells loaded with firm-2 were stimulated with 50 ELMNMDA, and changes in the fura- fluorescence ratio were monitored us- ing a dual-wavelength imaging system with excitation at 350 and 380 nm. The color bar ranges from purple to white and reflects Ca*+ levels from approxi- mately 50 to 500 nM. A, Phase-contrast photomicrograph ofthe field ofcells be- ing studied. Cells are maintained in buffer containing 2 mM Ca2+ unless oth- erwise specified. B, Pseudocolor rep- resentation of ratioed images prior to NMDA. Basal Ca:+ levels within den- tate cells ranged from 50 to 70 nM. C and D, Pseudocolor representations of ratioed images of fura- fluorescence, 7 and I3 set following addition of NMDA. Ca;+ increased to 300-400 nM in all neurons in response to NMDA. E, Pseudocolor representation of ra- tioed images 71 set after NMDA was washed out of the bath. F, Pseudocolor representation of ratioed images 8 set after NMDA addition in Ca*+-free buffer. Note the lack of an increase in Ca:+ when cells are stimulated in the absence of Ca2+.c

sion; and (3) if so, doesCa2+ originate from a single source and Immunocytochemistry. Cells were processed for immunocytochem- gain access to the via similar routes? istry according to the protocol of Lerea and McCarthy (I 990). Mouse monoclonal antibody against neurofilament was generously supplied by John Woods (Sandoz Laboratories, London, England). Sheep anti-glu- Materials and Methods tamic acid decarboxylase (anti-GAD) antiserum was generously sup- Cell preparutmn. Dentate gyrus cells were obtained from 4-d-old rat plied by Don Schmechel (Duke University, Durham, NC). Rabbit anti- pups according to the method of Mattson and Kater (I 989) and main- glial fibrillary acidic (anti-GFAP) antiserum was purchased from tained for 7-9 d rn vitro prior to use. Briefly, each hippocampus was Accurate Laboratories (Westbury, NY). Fluorescein isothiocyanate dissected and sectioned into 600-800-pm-thick transverse slices using (FITC)-conjugated rabbit anti-sheep IgG, FITC-conjugated goat anti- a Mclllwain tissue chopper. The dentate gyrus was separated from the rabbit IgG, and rhodamine-conjugated goat anti-mouse IgG were pur- hippocampal gyrus by microdissection using a dissecting microscope. chased from Cappel Laboratories (West Chester, PA). Efforts were made to dissect as close to the hilar border of the granule Calcium measurements. The Ca2+-sensitive indicator dye fura- (Mo- cell layer as possible, thereby excluding the hilus. All tissue was enzy- lecular Probes) was used to monitor changes in intracellular calcium matically dissociated with 0.25% trypsin for 30 min at 37°C. Following (Cat+) in individual cells as described by McCarthy and Salm (I 99 1). trypsinization, tissue was rinsed and dispersed into a single cell sus- Dentate gyrus cells grown on glass coverslips were loaded with 5 PM pension by gentle passage through a fire-polished Pasteur pipette. The fura- acetoxymethyl ester in growth media for 30-45 min at 37°C. The cell suspension was centrifuged for I O-l 2 min at 100 x g and the pellet coverslip was mounted into a viewing chamber and cells were rinsed resuspended in Modified Eagle’s Medium supplemented with 33 mM with HBSS+ (CaZ+/Mg2+/-free Hanks’ Balanced Salt Solution glucose, I mM pyruvic acid, 2 mM CaCl,, I5 mM KCI, and 10% fetal supplemented with 2 mM CaCI, and 5 PM ). Fura- loading under calf serum (MEM-C). An estimate of viable cells was obtained using these conditions yielded a uniform fluorescence within the cells. The the trypan blue dye exclusion test. Cells were plated in a small volume cells were visualized on a color monitor with a Zeiss ICM 405 micro- at a density of 4-6 x 10’ cells/mm2 onto poly-D-lysine-coated glass scope interlaced with an imaging system via an ISIT video camera. chamberslides or 22 mm glass coverslips. Cells were allowed to settle Background fluorescence was measured at both 350 nm and 380 nm and adhere for several hours prior to adding additional MEM-C. Cells excitation wavelengths and subtracted from cellular fluorescence during were maintained at 37°C in a humidified incubator with 5% CO,, 95% the course of data collection. Ratioed images (3501380 nm) were dis- 02. played as 256 pseudocolor-enhanced gray levels. Sixteen images were The Journal of Neuroscience, August 1992, 12(E) 2975

250 A K = 30 pM KAINIC ACID N = 50 @I NMDA E = 2.3 mM EGTA E - 1 mM EGTA _o 2200 a 0 &q 60- 2 g 60-

N I 40- 2 2 20- N- E- N- N- o- I I I I 1 0 500 1000 1500 2000 ‘- E- K- K- SECONDS I Ii &O l6bO 2400 32bO 40bo 200 - N = 50 JIM NMDA B SECONDS o 160 0 Ca = CALCIUM FREE BUFFER Figure 3. Effects of EGTA on the KA-stimulated calcium response. 2 160 Fura-2-loaded dentate gyrus cells were stimulated with KA (30 PM) in a140 n the presence ofCa:+ (2 mM), rinsed, and allowed to recover for I5 min. The same cells were restimulated with KA in the presence of 2.3 mM El20 EGTA. Cells were rinsed, returned to Ca’+ -containing buffer, allowed 2100 to recover, and stimulated a third time with KA. Fura- 350:380 nm ratios were collected at the times indicated. The bars indicate addition of or to the buffer. The fira- 3501380 ratio refers to the N mean, background corrected, 350:380 nm ratio. Values were obtained I 60 from cursor boxes placed over ratioed images of single cells. 2 40 3 N - 0 Ca- Solution (HBSS supplemented with 26 mM NaHCO,, 2.3 mM CaCI,, LL20 0 j, , N- I I I 10 mM HEPES, and 5 ELM glycine). Cells were returned to the 37°C 0 400 800 1200 incubator for 3-4 hr prior to stimulation. Cells were incubated with the designated treatment for 30 min at 37°C unless otherwise stated. Fol- SECONDS lowing treatment, cells were fixed with 4% paraformaldehyde at 4°C for 7-10 min, rinsed with HBSS supplemented with IO mM HEPES, and Figure 2. Effects of EGTA and 0 extracellular calcium on the NMDA- dehydrated through a series of ethanols. Cells were stored at -70°C stimulated calcium response. A, Fura-2-loaded dentate gyrus cells were until used for in situ hybridization. stimulated with NMDA (50 PM) in the presence of CaZ+ (2 mM), rinsed, Oligonucleotide probe. A 50-base pair oligonucleotide sequence of and allowed to recover for 15 min. The same cells were restimulated c:fos that exhibited minimal homology with known portions of the with NMDA in the presence of I mtvr Ca*+ plus 1 mM EGTA. Cells rodent genome was chosen. The sequence was complementary to nu- were rinsed, returned to Ca*+-containing buffer, allowed to recover, and cleotides 270-3 I9 of rat c-j& (Curran et al., 1987). Radiolabeled oli- restimulated a third time with NMDA. Fura- 350:380 nm ratios were gonucleotide was prepared with a 3’-terminal transferase reaction and collected at the times indicated. Values shown are from a single rep- ‘*P-a-dATP (6000 Ci/mmol; Du Pont-New England Nuclear) to yield resentative cell obtained from a cursor box placed over the ratioed image a specific activity of 2-4 x 10y cpm/pg. Evidence that this oligonucle- of the cell. This response is representative of data collected from 20 otide hybridizes selectively to c-fos mRNA has been previously dem- individual neurons. The bars indicate addition of agonist or drug to the onstrated and published by this laboratory (Simonato et al., 199 1). buffer. Thefuru-2 350/380 ratio refers to the mean, background cor- c-fos riboprobe. Antisense and sense c-j& riboprobes were prepared rected, 350:380 nm ratio. B, Fura-2-loaded dentate gyrus cells were from a full-length C&U cDNA insert (generously supplied by J. Morgan stimulated with NMDA in the presence of CaS+ (2 mM), rinsed, and and T. Cut-ran, Roche Institute of Molecular , Nutley, NJ). The allowed to recover for I5 min. The same cells were then placed in Ca2+- full-length c-fis cDNA insert is cloned in a pSP65 plasmid containing free buffer and restimulated with NMDA. Fura- 350:380 nm ratios the SP6 promoter and contains a single Xhol restriction site at base were collected at the times indicated. The bars indicate addition of 1353 from the 5’ end. The plasmid was linearized with Xhol and agonist or drug to the buffer. Values shown are from a single represen- riboprobes generated using a SP6 transcription assay in the presence of tative cell (n = 6). YS-rLJTP. Antisenseand sense riboprobes were hydrolyzed to fragments of approximately 200 base pairs with sodium carbonate at 60°C. In situ hybridizationfor lZP-oli~onucleotides. Slides were thawed grad- ually to room temperature and incubated successively in 95% EtOH, collected at each wavelength for each field of cells. The images were 70% EtOH, and 50% EtOH to rehydrate the cells. Cells were treated averaged and background fluorescence at each wavelength subtracted with 0.25% acetic anhydride diluted in 0.1 M triethanolamine, 0.9% prior to collecting ratios. All experiments were done in Mg2+ -free HBSS’ saline for IO min prior to hybridization. Cells were rinsed briefly in buffer. Vehicle or drug was added directly to the chamber bath as 10 x diethylpyrocarbonate-treated water, dehydrated in increasing concen- stock concentrations and changes in the 350/380 nm fura- ratio mon- trations of EtOH, and incubated for 3-4 hr at 37°C in prehybridization itored. All fura- experiments were done at room temperature. Follow- buffer [50% formamide; 0.6 M NaCI, 2 mM EDTA, 20 mM Tris (2x ing stimulation, cells were rinsed three times consecutively in HBSS’ NTE); 5 x Denhardt’s solution (Sigma); 500 &ml yeast tRNA (Sigma); and allowed at least a 15 min recovery period prior to restimulation. 500 &ml salmon sperm DNA (Sigma); and 0.05% sodium pyrophos- Calcium calibrations were performed using in vitro standards (Gryn- phate]. Cells were hybridized overnight (approximately 16 hr) at 37°C kiewcz et al., 1985). in the above buffer containing 100 &ml tRNA. 100 &ml DNA, and Cell treatment .for c-fos induction, MEM-C growth medium was re- 0.025 ng/Fl 32P-labeled c-fos oligonucleotide. Nonspecific hybridization moved from each culture well and replaced with Hanks’ Balanced Salt of radiolabeled probe to the cells was determined by including unlabeled 2976 Lerea et al. * EAA Receptor-mediated induction of c-fos mRNA

K+ - 60 mM POTASSIUM A ” - 10 (rM NIFEDIPINE

K+ - “- K+-n- ,.,- K+ - N- I OJ, 1 0 2000 4000 0000 a000 0 600 1200 1800 2400 3000 SECONDS SECONDS

120 C

EL

i 3 20 A- n- A- A- 0 ir , , , , , 0 so0 1600 2400 3200 4000 SECONDS Figure 4. Effects of nifedipine on calcium responses to various stimulations. A, Fura-2-loaded cells were stimulated with K+ (60 mM), rinsed, and restimulated with K+ in the presence of 10 NM nifedipine. Nifedipine was washed out and the same field of cells was stimulated for a third time with 60 rnM K+. Nifedipine markedly decreased the K+-induced Ca2+ response in a reversible manner. The cells were then stimulated with NMDA (50 FM) in the presence or absence of 10 PM nifedipine. Nifedipine had no effect on the NMDA-mediated Ca2+ response. The bars indicate drug additions to the extracellular buffer. B, Fura-2-loaded cells were pretreated with APV (100 PM) and stimulated with KA (30 KM). Cells were rinsed, allowed to recover, and restimulated with KA in the presence of 10 MM nifedipine. Cells were rinsed and stimulated a third time with KA in the absence of nifedipine. Nifedipine blocked the KA-mediated increase in Ca I+ in a reversible manner. C, Fura-2-loaded cells were pretreated with APV and stimulated with AMPA (100 NM). Cells were rinsed, allowed to recover, and restimulated with AMPA in the presence of 10 PM nifedipine. Cells were rinsed and stimulated a third time with AMPA in the absence of nifedipine. Nifedipine blocked the AMPA-mediated increase in Ca2f.

oligonucleotide probe in the prehybridization and hybridization buffer Results (10 x and 100 x concentrations, respectively) in adjacent wells. Follow- ing hybridization, cells were washed [20 min at room temperature in Characterization of dentate gyms cells in vitro 1 x NTE, 0.05% sodium pyrophosphate; 3 hr at 37°C in 10 mM Tris, An in vitro preparation of CNS neurons wasused to study changes 1 mM EDTA, 0.05% sodium pyrophosphate, 75 mM NaCl], dehydrated, in Ca:+ and c-fos mRNA following EAA receptor stimulation. and air dried. All slides were dipped in NTB-3 liquid emulsion (Kodak) and stored at 4°C for 17-21 d. of postnatal dentate gyrus yielded an enrichedneu- Emulsion-coated slides were developed in Kodak D- 19 (4 min), rinsed ronal population. Neurons were identified morphologically as in water (30 set), and fixed in Kodak fixer (6 min). Cells were stained having small cell somata with several well-defined elongated for Nissl substance, and silver grains were visualized and counted using neurites. This was confirmed by demonstrating that all mor- either bright-field or dark-field optics on a Leitz Laborlux-12 micro- phologically identified neurons were immunoreactive with an scope. Data are presented as silver grains per individual cell. In situ hybridization for -‘SS-riboprobe. Slides were treated as above. antibody directed against the 200 kDa neurofilament protein Cells were incubated for 3-t hr at 55°C with prehybridization buffer (data not shown). Since dentate gyrus contains both excitatory [50% formamide, 10% dextran sulfate, 3 x SSC (0.45 M NaCl, 0.045 M as well as inhibitory neurons,an antiserumto the GAD citric acid), 5 x Denhardt’s solution, 500 &ml yeast tRNA (Sigma), wasused to identify the percentageof inhibitory neuronspresent 500 &ml salmon sperm DNA (Sigma), and 10 mM dithiothreitol]. in the preparation. Lessthan 5% of the neurofilament-positive Cells were hybridized overnight (approximately 16 hr) at 55°C in the above buffer containing 60 &ml YS-labeled antisense riboprobe. Non- neurons stained with antisera directed againstGAD (data not specific hybridization was determined by using an j5S-labeled c-fos sense shown). Cultures prepared from whole hippocampuswere used riboprobe in adjacent wells. Following hybridization cells were rinsed as positive controls for GAD immunoreactivity. Sincethe ma- with 4 x SSC (three times, 15 min each) and RNase treated at 37°C for jority of cells are GAD negative (95%) and the dissection in- 30 min. Cells were then rinsed with 2 X, 1 x, and 0.5~ SSC (15 min each) and 0.1 x SSC at 55°C for 30 min. All slides were dipped in NTB-3 cludesonly the molecular and granule cell layers of the dentate liquid emulsion and stored at 4°C for 3-5 d. Slides were developed, gyrus, our cellular population most likely consistsof excitatory stained, and quantitated as described above. granule neurons. We used isolated dentate gyrus neurons for The Journal of Neuroscience, August 1992, Q(8) 2977

100 K = 30 /.IM KAINIC ACID 1 0 Na+= Na+ FREE BUFFER iii: J . Figure 5. Effect of removing Nat on 0 N = 50 jN NMDA KA- and NMDA-stimulated calcium 7 responses.Fura-2-loaded cells were F EO- 2 20- 0 Na+ pretreated with APV (100 FM) and K-N- 0 ? stimulated with KA (30 FM) in HBSS. 2 Cells were rinsed and allowed to recov- O- er before being placedin Na+-freebuffer $i 60- 0 600 800 ,000 ,200 ,400 1600 1800 2000 (sodium replaced with an equimolar CT) concentration of NMDG). Cells were \ SECONDS first stimulated with KA in the Na+-free z buffer. No increase in Cat+ occurred in CQ 40- response to KA in the Na+-free envi- ronment. Cells were then stimulated N with 50 PM NMDA while still in the I Na+-free buffer; NMDA stimulation K- 0 Na+ K- 4 K-N- yielded a dramatic rise in Ca;+ under a 20 these same conditions. Cells were rinsed, allowed to recover in buffer containing Na+, and restimulated with KA, yield- ing a normal Ca2+ response. Inset, An expanded time course of the Ca*+ re- sponses following KA and NMDA in 0 800 1200 1800 2400 3000 Na+-free buffer. The barsindicate the additions of agonists to, or changes in, SECONDS the extracellular buffer.

thesestudies because these neurons exhibit robust c-fos mRNA in a variety of paradigms in vivo (Morgan et al., 1987; Sonnen- Eflects of nifedipine on calcium responsesto EAA receptor berg et al., 1989; Simonato et al., 1991). Non-neuronal cells agonists were polygonal in shape and were identified as astroglia by The NMDA receptor is permeable to Ca2+ as well as to Na+ immunocytochemical staining with antisera to GFAP. and K+ (Macdermott et al., 1986). NMDA receptor activation may therefore induce Caz+ influx directly through the receptor Intracellular calcium responsesto EAA receptor agonists channel itself as well as indirectly through voltage-sensitivecal- Changesin Cat+ following bath application of EAA receptor cium channels (VSCCs) activated in responseto receptor-me- agonistsor depolarizing concentrations of potassium (K+) were diated cell depolarization. Nifedipine, a dihydropyridine Ca2+ monitored in individual dentate gyrus cells using the CaZ+-sen- , was usedto distinguish betweenthese two sites sitive dye fura-2. Neurons (n = 75) stimulated with NMDA of Ca*+ influx. Blocking dihydropyridine-sensitive voltage-de- exhibited a rapid increasein the fura- 350:380 nm ratio, in- pendent Ca2+channels with nifedipine (10 PM) did not inhibit dicating an increasein Cat+ (Fig. 1). The basalCay+ concentra- NMDA-induced increasesin Caf+. In contrast, nifedipine mark- tion ranged from 50 to 70 nM and increasedto a maximum of edly inhibited the Ca:+ responsesto KA, AMPA, and depolar- 300-400 nM after NMDA application. Intracellular CaZ+ re- izing concentrations of K+ (Fig. 4A-C). The remaining small mained elevated as long as the agonistwas present (2 min - Ca*+ responseobserved in the presenceof nifedipine may be imum time tested) and returned to basal levels upon NMDA due to Ca2+ influx through nifedipine-insensitive VSCCs or in- washout. Non-neuronal cells were never observed to respond flux through non-NMDA channelsdirectly. The role of VSCCs to NMDA. was further tested by replacing Na: with the nonpermeable The NMDA-induced increasein Cat+ was dependent on the cation N-methyl-D-glucamine (NMDG) to decreaseagonist-in- presenceof extracellular calcium (Ca%+).NMDA-mediated in- duced depolarization. The NMDA-induced Ca*+ responsewas creasesin Caz+ were markedly inhibited by either chelation of not affected by the removal of Na:, whereasthe Ca2+response Caj+ with EGTA (1 mM; Fig. 2A) or omission of Ca2+from the following KA stimulation was inhibited (Fig. 5). Together with extracellular buffer (Fig. 2B). This inhibition was reversedupon the resultsfound with nifedipine, these resultssuggest that the returning the cells to Ca2+-containing buffer. The non-NMDA principal route of CaZ+entry following NMDA application is receptor selective agonistskainic acid (KA; 30 PM) and AMPA through the NMDA receptor itself and does not require acti- (cu-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; 100 vation of VSCCs. In contrast, in this preparation the non-NMDA PM) also rapidly increasedCat+ (seebelow); this Ca*+ response receptor agonistsprimarily rely on the opening of VSCCs for wasalso blocked by addition of EGTA to the extracellular buffer the influx of Cat+. (Fig. 3). NMDA was applied in the presenceof 10 PM CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) to block non-NMDA Induction of c-fos mRNA by EAA receptoragonists receptors; non-NMDA receptor agonists were applied in the Dentate gyrus cells maintained in vitro were usedto study EAA presenceof APV [D(-)-2-amino-5phosphonopentanoic acid] induction of c-fos mRNA. NMDA induced c-fos mRNA lO- to block NMDA receptor activation. 15-fold over basallevels in individual neurons(n = 765). NMDA 2979 Lerea et al. * EAA Receptor-mediated Induction of c-fos mRNA

i14- B

g 12-

% Y 10 -

8 6 2 G 4- 0 ii _ u 2 ii! v) o-’ vohlcle NMDA NMDA NMDA UK601 MK601 IlUMl IlOuMl

l2 D i ti lo g e- . - NONSPECIFICTOTAL . - TOTAL . - NONSPECIFIC g 6- z

g 4- 5 3 2- iii o’, , , , , , , , 8 v I o-, I 0 10 20 30 40 50 60 70 60 90 100 0 5 10 16 20 25 30 1NMDAl uM MINUTES Figure 6. NMDA-stimulated changes in c-j& mRNA in cultured dentate gyrus cells. All cells were stimulated for 30 min and processed for in sifu hybridization as described in Materials and Methods. Silver grains were counted over individual cells. A and B, C-&X mRNA in dentate cells stimulated with NMDA (50 FM) in the absence or presence of APV (A) or Mk-801 (B). Each data point comes from at least 50 individual neurons. C, Dose-response curve of NMDA-induced c-j& mRNA. Cells were incubated with varying concentrations of NMDA for 30 min. Data are from a single representative experiment, repeated three times. The response at each dose is determined from counting silver grains over at least 20 individual neurons. D, Time course of NMDA-induced c-j& mRNA expression. Cells were treated with 50 PM NMDA for varying times. APV was added at the appropriate times to block NMDA receptor activation. Cells were fixed 30 min after the initial addition of NMDA. Data are from a single representative experiment, repeated at least three times. Data for each time point are from at least 50 individual neurons. Error bnrs refer to *SEM. induction oft-fis mRNA wasblocked by the competitive NMDA Effects nifedipine on c-fos induction APV (100 WM) aswell asthe noncompetitive qf antagonist Mk-80 1 (1 PM, 10 FM; Fig. 6A.B). Induction of c-fos Nifedipine blocked the induction of c-fos mRNA by K+ and mRNA occurred in a dose-dependent manner with near-max- KA, whereas it had little effect on NMDA-mediated c:fos in- imal expression being reached with 10 PM NMDA (Fig. 6c). duction (Fig. 9A-C). Nifedipine also blocked AMPA-induced c-j& mRNA induction by NMDA was reduced to basallevels C-$X mRNA (data not shown).These resultscorrelate well with by chelating CaZ+with EGTA, suggestingthat Ca2+influx from the Ca*+ data presented above where NMDA stimulation in- the extracellular environment is required for the response(Fig. creasedCa;+ even in the presenceof nifedipine. These findings 70). suggestthat NMDA-mediated induction of c-fis mRNA re- The NMDA induction of c-fos mRNA was time dependent quires Ca2+ influx directly through the NMDA channel. with near-maximal increasesobserved after a 10 min exposure (Fig. 60). A 1-3 min exposure to 50 FM NMDA was sufficient Discussion to induce an increase in c-fos mRNA detected 30 min after Three principal findings emergefrom this study. (1) Activation agonistexposure (Fig. 8). Shorter exposure times to NMDA (30 of either NMDA or non-NMDA subtypesof EAA receptorsis and 45 set) did not result in an increase in c-fos mRNA, even sufficient to induce the rapid and dramatic increase of c&s though this did result in clear increasesof Cat+ (Fig. 8, inset). mRNA in isolated dentate gyrus neurons. (2) c-fos mRNA in- Non-NMDA receptor agonistsalso induce c-fos mRNA in duction by either receptor subtype requires an increase of individual dentate neurons. AMPA (100 PM; n = 50) and KA Cat+. Receptor-mediatedincreases of both Cat+ and C-$XmRNA (30 PM; n = 141) induce C-$X mRNA 10-l 5-fold over basal are blocked by chelation of CaS+with EGTA. (3) The increases (Fig. 7A,B). Depolarizing the cells directly with K+, in the pres- of Ca:+ and c-fos mRNA triggered by KA or AMPA, but not ence of receptor antagonists,also resultsin an increasein c-fos NMDA, are prevented by nifedipine, a VSCC blocker. mRNA (60 mM; n = 161). C-$X mRNA induction following KA Our results are consistent with the report by Szekely et al. was also Ca*+ dependent, the responsebeing reduced to basal (1989) in which NMDA receptoractivation inducedc-fis mRNA levels by chelating Caf+ with EGTA (Fig. 7C). in cultured cerebellar granule cells. We extend those observa- The Journal of Neuroscience, August 1992, 12(E) 2979

Figure 7. KA and AMPA stimulated changes in c&s mRNA in dentate gy- rus cells. Cells were pretreated with 100 FM APV to block NMDA receptors pri- or to adding KA or AMPA. A and B, c-fos mRNA in response to 100 PM AMPA (A) or 30 /IM KA (B). Water is the vehicle used for KA and AMPA; TREATMENT HBSS is the vehicle used for NMDA. Each experiment was repeated at least three times. Each data point comes from counting silver grains over at least 50 C individual neurons. Data from NMDA (50 PM)-stimulated cells maintained in adjacent wells on the same slides are included in each panel as a positive control for the induction of c-fos mRNA. C and D, c-fos mRNA in re- sponse to 30 PM KA (C) or 50 fih4 NMDA (D) in the absence and presence of 2.3 mM EGTA. Water is the vehicle used for KA; HBSS is the vehicle used for NMDA. Chelation of Caz+ with EGTA blocked the induction of c-fos mRNA by both KA and NMDA. Each Kainic Acid EGTA NMDA EGTA Ktinic Acid NMDA experiment was repeated three times. TREATMENT TREATMENT Error bars refer to +-SEM. tions by demonstrating the absolute requirement for Ca*+ in the granule cells. Whether a reduced expression of KA/AMPA re- NMDA-mediated response. Our results differ from Szekely et ceptor subtypes, a difference in VSCCs or some other factor al. (1989) in that KA is also sufficient to induce c-j&s mRNA unique to the cerebellar granule cell preparation accounts for in dentate gyrus neurons whereas it was ineffective in cerebellar this difference is unknown. Figure8. Time course ofNMDA-me- diated increase in c-fos mRNA and in- tracellular calcium. Dentate gyrus neu- rons were treated with NMDA and processed for in situ hybridization to 12 T detect c-J&s mRNA or fura- imaging to detect changes in Ca:+ as described in Materials and Methods. NMDA treatment caused an increase in c-fos 10 mRNA in a time-dependent manner. Exposure to NMDA for less than 1 min did not result in an increase in c-fis I150 mRNA (same data shown in Fig. 60). a I Inset, Fura- Ca2+ responses and c-fos ;125 mRNA responses plotted on the same 2 graph following up to 2 min of NMDA ,100 6 1 treatment. N indicates the addition of 2 75 50 FM NMDA. For Ca*+ responses, ? changes in the fura- ratio were mon- p 50 itored immediately following addition 4 m of NMDA. For c-fos mRNA detection, i 25 m cells were exposed to NMDA for var- ious times from 0 to 120 set followed 0 0 60 120 by addition of APV to block continued receptor activation. Cells were incu- 2 SECONDS I bated for 30 min prior to being fixed ti and processed for in situ hybridization. The increase in Cay+ following NMDA 0 addition was immediate and sustained 0 5 10 15 20 25 30 for as long as agonist was present. In contrast, exposure to NMDA for less than 1 min did not induce c-fos mRNA. MINUTES Error bars refer to ?SEM. 2980 Lerea et al. * EAA Receptor-mediated Induction of c-fos mRNA

B 18- 4 = Total 16- 1 = Total I= Norqecie I= Nonspecif~ 14- 12 10 .a 6 4 2 0 HESS HOA )lifedphe DKO 6OinMK Mm WA &K

Figure 9. Effect of nifedipine on stimulated changes in c-fos mRNA in dentate cells. Cells were processed for in situ hybridization to detect c-&x mRNA as described in Materials and Methods. Non-NMDA receptor agonists and K+ were added in the presence of APV to block NMDA receptor activation. A, Cells were stimulated with NMDA (50 PM) in the absence or presence of nifedipine (10 PM). Nifedipine was dissolved in dimethyl sulfoxideand dilutedinto HBSS. Each data point comes from counting silver grains over at least 70 individual neurons. Nifedipine did not decrease the NMDA induction of c-Jx mRNA. B, Cells were stimulated with 60 mM K+ in the absence or presence of nifedipine (I 0 PM). Each data point comes from countingsilver grainsover at least70 individual neurons.Nifedipine blocked the K+-inducedincrease of c-fosmRNA. C, Cellswere stimulatedwith 30 PM KA in the absenceor presenceof nifedipine(10 PM). Eachdata point comesfrom countingsilver grains over at least50 individual neurons.Nifedipine blocked the KA inductionof c-fosmRNA. Error bars refer to +SEM.

Interestingly, NMDA-induced rises in Cat+ of ~1 min did nifedipine or by substituting Na: with NMDG, suggestingthat not induce c$s mRNA. A sustainedNMDA-induced increase Ca*+ enters the neurons principally through VSCCs activated in Cat+ was required for the maximal induction of c-fos mRNA, by neuronaldepolarization. Although Ca*+influx directly through 10 min of NMDA stimulation resulting in a more than twofold non-NMDA receptorshas recently been reported (Gilbertson et greater induction of c-fos mRNA compared to a 2 min stimu- al., 1991; Hollmann et al., 199l), we do not detect significant lation. Our data demonstrate that a transient Ca2+rise alone is Ca*+ influx directly through KA/AMPA receptorsin theseneu- not sufficient for the direct induction of c-fis mRNA, but more rons as measuredwith fura- imaging. In contrast, NMDA re- likely a sustainedincrease in Ca2+ is necessary.Sustained in- ceptor activation appearsto evoke CaZ+entry primarily through creasesin CaZ+lasting several minutes occur in hippocampal the NMDA receptor directly. Sodium ions account for approx- CA3 cells following a train of electrical stimuli (Muller and imately 88% of the depolarizing current induced by NMDA in Connor, 1991), as well as in hippocampal CA1 cells following neurons(Mayer and Westbrook, 1987). Removal of Na: should focal NMDA (Connor et al., 1988). Our findings may help ex- therefore greatly attenuate the depolarizing current following plain the low constitutive expressionof c-fos mRNA in normal NMDA stimulation. The persistent NMDA-induced increase rat where sustainedincreases in Ca2+presumably are not of Ca:+ in the presenceof the nonpermeablecation NMDG, as occurring under physiologic conditions. The nature of the sig- well as in the presenceof nifedipine, suggeststhat the majority naling mechanisms activated by the sustained Ca2+ rises is pres- of CaZ+enters directly through the NMDA receptor, not indi- ently unclear. rectly through VSCCs. Previous work from this laboratory indicated that NMDA Although the routes of Ca2+entry differ following activation receptor antagonistsinhibited kindled seizure induction of c-fos of non-NMDA and NMDA receptors, the rise of Cat+ triggered mRNA in dentate granule cells by 50-70% (Labiner et al., 1990). by each receptor induces a dramatic increasein c-fo.smRNA. Such resultssuggest that NMDA receptor activation is necessary The colocalization of non-NMDA and NMDA receptorsat in- for the full expression of seizure-induced c-fos mRNA. The dividual synapses(Bekkers and Stevens, 1989;Jones and Baugh- present findings demonstrate that NMDA receptor activation man, 1991) together with these findings suggeststhat synapti- is suficient to induce c-fos mRNA and that Ca2+is an essential cally releasedglutamate could trigger induction of c-fos mRNA second messengerin this signalingcascade. Our results suggest by Ca2+entering a cell via two distinct routes. The preponder- that synaptically releasedglutamate during a seizurecould evoke ance of VSCCs in the vicinity of the cell and proximal an increaseof c-fos mRNA by activating either NMDA or non- dendrites (Lipscombe et al., 1988; Westenbroek et al., 1990), NMDA receptor subtypes. This could account for the residual and of NMDA receptorsin the dendrites (Monaghan and Cot- 30-50% of c-fos mRNA found in the presenceof NMDA an- man, 1985; Jonesand Baughman, 199 l), suggeststhat increases tagonists (Labiner et al., 1990). in Ca:+ may be spatially separatewithin the cell (Tank et al., Our data indicate that the route of Ca2+ entry into dentate 1988; Regehr and Tank, 1990). Muller and Connor (1991) have gyrus neuronsdiffers following KA/AMPA and NMDA receptor recently demonstrated regional changesin Ca:+ levels within activation. KA-induced increasesin Cat+ can be eliminated by individual hippocampal cells following electrical stimulation. The Journal of Neuroscience, August 1992, I.?(8) 2981

Despite potentially distinct pools ofelevated Ca:+ resulting from types of glutamate receptors are concentrated at synapses on cerebral activation of each receptor, the increases of C-$X mRNA may cortical neurons in culture. Neuron 7:593-603. Labiner DM, Hosford DA, Shin C, Cao Z, Butler LS, McNamara JO reflect the independent activation of identical genetic programs, (1990) Dissociation of neuronal activity and seizure-induced C-$X providing redundancy in the genetic consequences. Altema- mRNA expression by NMDA receptor antagonist. Sot Neurosci Abstr tively, it has previously been demonstrated that increases in 16:848. CaZ+ arising from distinct sources can trigger different biochem- Lerea LS, McCarthy KD (1990) Neuron-associated astroglial cells ical responses (Brooks et al., 1989). Such differences could lead express @- and o, - receptors in vitro. Brain Res 52 1:7-l 4. Lipscombe D, Madison DV, Poenie M, Reuter H, Tsien RY, Tsien RW to the induction of distinct arrays of IEGs, c-fos mRNA being (1988) Spatial distribution of calcium channels and cytosolic calcium common to each. If so, distinct arrays of IEGs could regulate transients in growth cones and cell bodies of sympathetic neurons. expression of distinct arrays of target genes, thereby effecting Proc Nat1 Acad Sci USA 85:2398-2402. unique and receptor subtype-specific long-term phenotypic con- Macdermott AB, Mayer ML, Westbrook CL, Smith SJ, Barker JL (1986) NMDA-receptor activation increases cytoplasmic calcium concen- sequences. tration in cultured spinal cord neurons. Nature 32 I :5 19-522. Mattson MP, Kater SB (I 989) Development and selective neurode- References generation in cell cultures from different hippocampal regions. Brain Aniksztejn L, Ben-Ari Y (199 1) Novel form of long-term potentiation Res 490: I 10-l 25. produced by a K+ channel blocker in the hippocampus. Nature 349: Mayer ML, Westbrook CL (1987) Permeation and block of N-methyl- 67-69. o-aspattic acid receptor channels by divalent cations in mouse cul- Ascher P, Nowak L (I 988) The role ofdivalent cations in the N-meth- tured central neurones. J Physiol (Lond) 394:501-527. yl-o-aspartate responses ofmouse central neurons in culture. J Physiol McCarthy KD, Salm AK (199 1) Pharmacologically-distinct subsets (Lond) 399:247-26X of astroglia can be identified by their calcium response to neuroli- Bekkers JM, Stevens CF (1989) NMDA and non-NMDA receptors gands. Neuroscience 4 I :325-333. are co-localized at individual excitatory synapses in cultured rat hip- McNamara JO, Russell RD, Rigsbee L, Bonhaus DW (1988) Anti- pocampus. Nature 341:23C-233. and antiepileptogenic actions of Mk801 in the kindling Brooks RC, McCarthy KD, Lapetina EC, Morel1 P (1989) Receptor- and electroshock models. 27:563-568. stimulated phospholipase A, activation is coupled to influx ofextemal Monaghan DT, Cotman CW (1985) Distribution of N-methyl-o-as- calcium and not to mobilization of intracellular calcium in C62B partate-sensitive r$H]glutamate binding sites in rat brain. J Neurosci glioma cells. J Biol Chem 264:20 147-20 153. 5:2909-2919. CoIlingridge CL, Kehl SJ, McLennan H (1983) Excitatory amino acids Morgan JI, Curran T (1986) Role of ion flux in the control of c-fos in synaptic transmission in the Schaffer collateral-commissural path- expression. Nature 322:552-555. way of the rat hippocampus. J Physiol (Lond) 334:3346. Morgan JI, Curran T (1989) Stimulus-transcription coupling in neu- Connor JA, Wadman WJ, Hockberger PE, Wong RK (1988) Sustained rons: role ofcellular immediate-early genes. Trends Neurosci 12:459- dendritic gradients ofCa+ + induced by excitatory amino acids in CA1 462. hippocampal neurons. Science 240:649-653. Morgan JI, Cohen DR, Hempstead JL, Curran T (1987) Mapping Curran T, Morgan JI (1985) Superinduction of C-$X by nerve growth patterns of c-fos expression in the central nervous system after seizure. factor in the presence of peripherally active . Science Science 237: 192-197. 229:1265-1268. Muller W, Connor JA (I 991) Dendritic spines as individual neuronal Curran T, Gordon MB, Rubino KL, Sambucetti LC (1987) Isolation compartments for synaptic Ca2+ responses. Nature 354:73-76. and characterization of the c-fos (rat) cDNA and analysis of post- Regehr WC, Tank DW (1990) Postsynaptic NMDA receptor-medi- translational modification in vitro. 2:79-84. ated calcium accumulation in hippocampal CA 1 pyramidal cell den- Dragunow M, Robertson HA (1987) Kindling stimulation induces drites. Nature 345:807-8 IO. C$X protein(s) in granule cells of the rat dentate gyrus. Nature 329: Shin C, McNamara JO, Morgan JI, Curran T, Cohen DR (1990) In- 44 1-442. duction of c&s mRNA expression by afterdischarge in the hippo- Gilbertson TA, Scobey R, Wilson M (1991) Permeation of calcium campus of naive and kindled rats. J Neurochem 55:3050-3055. ions through non-NMDA glutamate channels in retinal bipolar cells. Simonato M, Hosford D, Labiner D, Shin C, Mansbach HH, McNamara Science 251:1613-1615. JO (199 1) Differential expression of immediate early genes in the Goelet P, Castellucci VF, Schacher S, Kandel ER (I 986) The long and hippocampus in the kindling model of epilepsy. Mol Brain Res 1 I: the short of long-term memory-a molecular framework. Nature 322: 115-124. 4 19-422. Sonnenberg JL, Mitchelmore C, Macgregor-Leon PF, Hempstead J, Grynkiewcz G, Poenie M, Tsien RY (1985) A new generation of Morgan JI, Curran T (I 989) Glutamate receptor agonists increase calcium indicators with greatly improved fluorescence properties. J the expression of Fos, Fra, and AP-I DNA binding activity in the Biol Chem 260:3440-3445. mammalian brain. J Neurosci Res 24:72-80. Harris EW, Ganong A, Cotman CW (1984) Long-term potentiation Stasheff SF, Anderson WW, Clark S, Wilson WA (1989) NMDA an- in the hippocampus involves activation of N-methyl-D-aspartate re- tagonists differentiate from seizure expression in an ceptors. Brain Res 323: 132-I 37. in vitro model. Science 245:648-65 1. Herrera DC, Robertson HA (1990) N-methyl-D-aspartate receptors Szekely AM, Barbaccia ML, Alho H, Costa E (1989) In primary cul- mediate activation of the c-fos proto-oncogene in a model of brain tures of cerebellar granule cells the activation of &‘-methyl-D-aspar- injury. Neuroscience 35:273-281. - tate-sensitive glutamate receptors induces c-fos mRNA expression. Hollmann M. Hartlev M. Heinemann S (1991) Cat+ uermeabilitv of Mol Pharmacol 35:401408. KA-AMPA-gated glutamate receptor channels depends on subunit Tank DW, Sugimori M, Connor JA, Llinas RR (1988) Spatially re- composition. Science 252:851-853. solved calcium dynamics of mammalian purkinje cells in cerebellar Iino M, Ozawa S, Tsuzuki K (I 990) Permeation of calcium through slice. Science 2421773-777. excitatory amino acid receptor channels in cultured rat hippocampal Westenbroek RE, Ahlijaman MK, Catterall WA (1990) Clustering of neurones. J Physiol (Lond) 424: I5 1-l 65. L-type Ca2+ channels at the base of major dendrites in hippocampal Jones KA, Baughman RW (1991) Both NMDA and non-NMDA sub- pyramidal neurons. Nature 347:28 l-284.