Developmental changes in metabotropic glutamate receptor-mediated calcium homeostasis

Developmental Changes in MetabotropicGlutamate Receptor–Mediated

Calcium Homeostasis

LANCE ZIRPEL,* MARY A. JANOWIAK, DWAN A. TAYLOR, AND

THOMAS N. PARKS

Department of Neurobiology and Anatomy, University of Utah School of Medicine,Salt Lake City, Utah 84132

ABSTRACTNeurons of the chick cochlear nucleus, nucleus magnocellularis (NM), require eighth

nerve activation of metabotropic glutamate receptors (mGluRs) for maintenance of intracel-lular calcium homeostasis. Interrupting this activation results in an increase in intracellularcalcium concentration ([Ca21]i) followed by cell atrophy, degeneration, and death of manyneurons. Although these phenomena are well characterized in late embryonic and posthatchchicks, little is known about the role of mGluRs and calcium homeostasis during the devel-opment of synaptic activity in NM. Using Fura-2 imaging, fluorescent immunohistochemis-try, and Western immunoblotting, we investigated (1) the expression and function of group ImGluRs and their role in calcium regulation during development of NM, and (2) the expres-sion of two other key molecules involved in regulating neuronal [Ca21]i : inositol trisphos-phate receptors (IP3Rs) and sarcoplasmic/endoplasmic reticulum calcium ATPases (SER-CAs). Confocal imaging of Fluo-3-labeled NM was used to investigate the kinetics of globalNM neuron calcium signals. Measurements were made at four ages that extend from beforesynaptic function begins in NM, through functional onset, to mature patterns of spontaneousactivity, namely, embryonic days (E) 10, 13, 15, and 18. mGluR5, mGluR1, and SERCAexpression peaked at E13 and then decreased with age. IP3R expression increased to peak atE18. [Ca21]i response to mGluR activation increased with age. The rise time of [Ca21]i signalsin NM neurons did not change with development, but E13 neurons were slower to reestablishbaseline [Ca21]i. These results suggest that the mGluR-mediated calcium homeostasis of NMneurons develops in parallel with synaptic activity and appears to be refined with increasingsynaptic activity. J. Comp. Neurol. 421:95–106, 2000. © 2000 Wiley-Liss, Inc.

Indexing terms: nucleus magnocellularis; activity dependent maintenance; deafferentation;

glutamate receptor; metabotropic glutamate receptors; mGluR1; mGluR5;

auditory neurons; inositol trisphosphate receptor; sarcoplasmic/endoplasmic

reticulum calcium ATPase

Developing neurons require greater Ca21 fluxes andhigher intracellular Ca21 concentration ([Ca21]i) than theirmature counterparts to facilitate survival, synapse forma-tion, dendrite growth, and other cellular functions (reviewedby Franklin and Johnson, 1992; Ghosh and Greenberg, 1995;Spitzer, 1995). Because the same Ca21 concentrations can bedamaging or fatal to mature neurons, these cells use power-ful regulatory mechanisms to maintain [Ca21]i at submicro-molar levels. The calcium-handling capacity of a neuron caninfluence neurite outgrowth (Mills and Kater, 1990), firingproperties (Turrigiano et al., 1994), gene transcription (Bin-dokas et al., 1998), and survival (Scharfman and Schwartz-kroin, 1989).

Conversely, activity can influence the Ca21 bufferingcapacity of neurons. In cultured nerve cord explants, ton-ically active motor axons display greater Ca21 bufferingcapacity than those firing phasically (Lnenicka et al.,1998). In chick dorsal root ganglion neurons, both action

Grant sponsor: NIH; Grant numbers: HD07491, DC00144.*Correspondence to: Lance Zirpel, Department of Neurobiology and

Anatomy, University of Utah School of Medicine, Salt Lake City, UT84132. E-mail: [email protected]

Received 22 June 1999; Revised 18 January 2000; Accepted 18 January2000

THE JOURNAL OF COMPARATIVE NEUROLOGY 421:95–106 (2000)

© 2000 WILEY-LISS, INC.

potential amplitude and frequency have significant effectson Ca21 fluxes (Park and Dunlap, 1998), and in rat hip-pocampal slices, synaptic activity alters the ability of mi-tochondria to buffer intracellular Ca21 (Bindokas et al.,1998).

Neurons of the chick cochlear nucleus, nucleus magno-cellularis (NM), depend on eighth nerve activation ofmetabotropic glutamate receptors (mGluRs) for mainte-nance of [Ca21]i (Zirpel and Rubel, 1996). Disruption ofthis mGluR-mediated calcium homeostasis, either bydeafferentation or pharmacologic block of mGluRs, resultsin an increase in [Ca21] that leads to cell degeneration anddeath of many NM neurons (Zirpel et al., 1995a; Zirpel etal., 1998a; reviewed in Zirpel et al., 1997). Although thesephenomena are well characterized in late embryonic andearly posthatch chicks that have fully developed, func-tional auditory systems, little is known about the devel-opment of this mGluR-mediated [Ca21]i homeostasis dur-ing the period when synaptic function matures betweenembryonic days (E) 10 and 18 (Jackson et al., 1982; Lippe,1994).

Because NM neurons receive such high levels of gluta-matergic, afferent input, it is plausible to hypothesize thatcalcium regulatory mechanisms are implemented in NMneurons before or during the time that these activity lev-els result in an unhealthy, hypercalcemic environment.Because mGluRs mediate a large component of NM neu-ron global Ca21 homeostasis (Zirpel and Rubel, 1996), thegoal of this study was to test the hypothesis that imple-mentation of mGluR-mediated mechanisms of calciumregulation by NM neurons develops in parallel with syn-aptic activity. During the development of synaptic activityin NM, we examined the protein expression and functionof group I mGluRs, protein expression of inositol trisphos-phate receptors (IP3Rs) and sarcoplasmic/endoplasmic re-ticulum calcium ATPases (SERCAs) (two major compo-nents of mGluR-mediated calcium homeostasis in NMneurons), and the temporal kinetics of global calcium buff-ering. Portions of this work have been reported in prelim-inary form (Zirpel et al., 1998b).

MATERIALS AND METHODS

Fura-2 imaging

Tissue preparation. Live in vitro tissue slices (300mm, containing NM) from each of the four ages wereacquired as previously described (Zirpel et al., 1995b; Zir-pel and Rubel 1996; Zirpel et al., 1998a) and incubated inoxygenated artificial cerebrospinal fluid (ACSF) contain-ing 10 mM (E10), 8 mM (E13), or 6 mM (E15 and E18)Fura-2 (Molecular Probes, Eugene, OR), 1.7% anhydrousdimethylsulfoxide (DMSO), and 0.03% Pluronic (Molecu-lar Probes) for 30 minutes at room temperature. Sliceswere placed in the imaging chamber and bathed in nor-mal, oxygenated ACSF for approximately 5 min before theinitiation of data acquisition.

Imaging. Ratiometric fluorescence imaging tech-niques used in this study were similar to those describedpreviously (Zirpel and Rubel 1996; Zirpel et al., 1995a,b,1998a). Fura-2 loaded NM neurons were alternately ex-cited by a fiberoptic illumination system with 350-nm and380-nm wavelengths of light from a xenon source using acomputer-controlled shutter and filter wheel. Exposuretimes for each wavelength were between 100 and 500 ms.

Fluorescence emission was acquired through a 203 Fluorobjective (Nikon), filtered by a 495-nm long-pass filter,and collected with a cooled CCD (Photometrics). Paired350/380 excitation images were acquired approximatelyevery 3–5 seconds or at 1-minute intervals and were ra-tioed on a pixel-by-pixel basis. Intracellular calcium con-centration was estimated according to Grynkiewicz et al.(1985) using NIH Image (public domain software devel-oped at the U. S. National Institutes of Health and avail-able on the Internet at http://rsb.info.nih.gov/nih-image).Five-point external calibrations were performed on theimaging system routinely and the Kd for Fura was calcu-lated with each calibration. The range of Kd values was75–172 nM. Up to 10 neurons were analyzed for any givenexperiment. However, each slice was considered n 5 1,and the data from all neurons in an experiment wereaveraged. Cells were not included in the analysis if base-line [Ca21]i was greater than 250 nM because this concen-tration is indicative of a dying cell (Zirpel et al., 1998a).Cells were neither added to nor subtracted from the anal-ysis of an experiment except when a complete loss offluorescence was observed, indicating cell death (Johnsonet al., 1994).

Drug application. Brainstem slices containing NMwere placed in a custom chamber fitted onto the stage ofthe microscope. The floor of the chamber was a 1-mm-thick coverslip. The slice was stabilized with a weighted,stainless steel net and continuously bathed with oxygen-ated ACSF at a rate of approximately 5–8 ml/min. Thegravity-fed superfusion port was placed approximately 1mm above the tissue slice, and the vacuum-removal portwas located on the floor of the chamber at the greatestpossible distance from the slice. This configuration al-lowed for rapid and efficient drug delivery to and removalfrom the tissue. Kainate (KA) was applied for 20–30 sec-onds followed by a 1–6-minute washout with ACSF. Cellsthat failed to respond to KA application with $100% in-crease in [Ca21]i followed by recovery to baseline wereconsidered unhealthy and were excluded from analysis(Zirpel and Rubel, 1996; Zirpel et al., 1998a).

Data analysis. Intracellular calcium concentrations([Ca21]i) were plotted as a function of time using EXCEL(Microsoft, Redmond, WA) and Cricket Graph (CricketSoftware, Malvern, PA). In all experiments, cells fromeach slice were averaged and treated as a single observa-tion (n 5 1). Data are presented as means 6 1 SEM unlessotherwise indicated. For acute [Ca21]i responses, baselinevalues were obtained by averaging the [Ca21]i for 10 sec-onds before drug application. Peak response was mea-sured within the following 30–60 seconds. Age-dependentchanges in protein expression and [Ca21]i response totrans-aminocyclopentane dicarboxylate (ACPD) were an-alyzed by examining the interaction term in the one-factoranalysis of variance. When the interaction term was sig-nificant, post hoc comparisons were made to determine theages at which the level of protein or [Ca21]i was signifi-cantly changed. Two-tailed t-tests and analyses of vari-ance were performed using Statview (SAS Institute, Cary,NC).

Electrophysiology

Methods for stimulating and recording from in vitrochick brainstem slices have been previously described(Jackson et al., 1982; Hyson and Rubel, 1989; Zirpel andRubel, 1996; Zirpel et al., 1998a). A concentric bipolar

96 L. ZIRPEL ET AL.

stimulating electrode (FHC, Bowdoinham, ME) wasplaced on the eighth nerve dorsal and lateral to NM.Stimulation pulses (50–100-ms duration, 60–80V) weredelivered at a rate of 0.1–1.0 Hz. Responses were recordedwith an ACSF-filled microelectrode ('1 MV) attached toan intracellular amplifier. Our results paralleled thosereported by Jackson et al. (1982), with the exception thatwe were unable to record an afferent volley component ofthe field potential at E10, regardless of electrode place-ment.

Immunohistochemistry

Tissue preparation. Embryos from each of the fourages were perfused transcardially with 0.9% saline con-taining 1,000 IU heparin and 2 mg/ml NaNO2 followed by4% paraformaldehyde in 0.1 M phosphate-buffered saline(PBS, pH 7.4). The brainstem was dissected free andplaced in 4% paraformaldehyde for 3 hours and then wasstored in 30% sucrose in 0.1 M PBS at 4°C overnight.Tissue was then embedded and frozen in OCT (VWR, SaltLake City, UT). Using a cryostat, 18-mm-thick coronalsections were cut through the brainstem at the level of theauditory nuclei and were placed on microscope slides.Tissue sections were blocked for 20 minutes with 1% nor-mal goat serum and 0.4% Triton X-100 in PBS. Sectionswere then incubated overnight at 4°C with primary anti-body diluted in blocking buffer (pH 7.3–7.4). Primary an-tibodies were as follows: 1 mg/ml polyclonal anti-mGluR5(Upstate Biotechnology, Lake Placid, NY); 1 mg/ml poly-clonal anti-mGluR1 (Upstate Biotechnology); 1 mg/mlmonoclonal anti-IP3R (Chemicon International, Te-mecula, CA); and 1:1,000 monoclonal anti-SERCA2 [CaS-3H2; (Kaprielian and Fambrough, 1987); generously pro-vided by Dr. Douglas Fambrough]. Primary antibodyincubation was continued for an additional 1–3 hours atroom temperature followed by three 10-minute washes inPBS. Alexa 594 secondary antibody (1:800 in 0.4% normalgoat serum (NGS)/PBS; Molecular Probes, Eugene, OR) orTRITC secondary antibody (1:120 in 0.4% NGS/PBS; Jack-son Immunoresearch, West Grove, PA) was applied for 90minutes at room temperature followed by three 10-minutewashes with PBS. Slides were coverslipped with Fluor-save (Calbiochem, La Jolla, CA) and stored in the dark at4°C until analyzed. Controls for each age were processedidentically except that instead of a primary antibody in-cubation, they were incubated in blocking solution.

Analysis. For each experiment, tissue sections fromall four ages were processed simultaneously under identi-cal conditions. The tissue section most closely approximat-ing the 50th percentile of the anterior–posterior extent ofNM was chosen for analysis for each embryo. Using con-focal laser scanning microscopy (CLSM; BioRad, Hercules,CA), this tissue section was “optically” sectioned throughthe extent of fluorescent labeling. The optical section half-way through this thickness was chosen for analysis. NMlabeling was imaged through a Fluor 403 oil immersionobjective (1.3 n.a., Nikon). The resolution of the image wasset at 1,024 3 1,024 pixels and the settings (magnification,gain, black level, laser intensity, and iris diameter) wereheld constant for analysis of tissue from each age in agiven experiment. Typical settings using the Alexa 594secondary were 3–10% laser intensity, 3.0–3.2 iris, 0 blacklevel, and 1,057 gain. All images were Kalman filtered bya factor of 4.

Images were converted to TIFF files and analyzed using

NIH Image. No-primary control tissue using the Alexa 594secondary showed virtually no background fluorescence,and images were not corrected. The TRITC secondaryantibody control showed considerable, uniform back-ground fluorescence that was averaged over 10 NM neu-rons in a control section. This average value was sub-tracted from the analyzed images. A typical TRITCbackground value was 15–18 on a 0–255 scale of fluores-cence. Fluorescence intensity was measured over the en-tire area of the easily identifiable NM neurons, excludingthe nucleus, which never exhibited fluorescence. In thecase of E18 mGluR labeling, fluorescence was also mea-sured over the fluorescently labeled area only. Fluores-cence was then expressed per unit area. A minimum of 10measurements per image were made. These 10 measure-ments were averaged, and the same analysis was per-formed on the images from all four ages in any givenexperiment. The age showing the greatest value of fluo-rescence per unit area was designated 100%, and the otherthree ages were expressed as a percentage of that value.Thus, an experiment (n 5 1) consisted of fluorescencevalues normalized to a maximum from each of the fourages examined in a single run of immunocytochemistry.These percentage values were averaged across experi-ments and are presented as means 6 1 SEM. Images usedfor presentation were processed in Adobe Photoshop (v.5.0.2, Adobe Systems, San Jose, CA).

Western immunoblots. NM from E10, E13, E15, andE18 chicks were dissected from brainstems, placed intoharvesting buffer, homogenized, and sonicated. Connec-tive tissue was removed from preparations by centrifuga-tion (14,000 rpm). Protein concentration of samples wasdetermined using a bicinchoninic acid assay (BioRad). Astandard curve was generated with 0.5, 1, 2, 4, 6, 8, and10-mg bovine serum albumin each time the protein assaywas run. The r2 value for the curve was always $0.99. NMlysate protein concentrations (micrograms per milliliter)were determined in triplicate and averaged. Protein con-centrations of samples run on the same gel (i.e., samplesfor the four different ages) were determined in the samespecific protein assay. Equal amounts of NM protein wereloaded into the wells of the gel, thus allowing normaliza-tion of the protein of interest to total protein concentra-tion. Five to 50 mg of protein were resolved via 7.5%sodium dodecyl sulfate polyacrylamide gel electrophoresisat 100V for 2 hours at room temperature before transferonto Immobilon-P membrane (Fisher) at 100V for 2 hoursat 4°C. Blots were probed with the appropriate antibody(see Results section and figure legends). Bound primaryantibodies were detected using Amersham (ArlingtonHeights, IL) Enhanced Chemiluminescence protocol. Gelswere scanned into a computer and the relative levels ofmGluR1, mGluR5, SERCA-2, or IP3R were determined bydensitometry using NIH Image. To ensure consistent mea-surements, and to include the presumed glycosylationbands (see Results section), a box of defined size was usedto measure the optical density of each band in a given laneat the same position in the y dimension on each gel. Aswith the immunohistochemistry described above, thesevalues were converted to percent maximum of the highestvalue for each experiment to allow for comparison be-tween experiments.

Thapsigargin. Tissue slices (300 mm, containing NM)from each of the four ages were acquired as describedabove for Fura-2 imaging. Slices were incubated in oxy-

97ONTOGENY OF AUDITORY NEURON Ca21 REGULATION

genated ACSF containing 20 mM BODIPY FL thapsigar-gin (Molecular Probes). Thapsigargin irreversibly binds toSERCAs (Lytton et al., 1991). CLSM images were ac-quired while the tissue was superfused with oxygenatedACSF. Images were processed and analyzed as describedabove. Controls were performed by first incubating slicesfor 20 minutes (25°C) with 200 mM non-fluorescent thap-sigargin (Calbiochem) before incubation with 20 mMBODIPY FL-conjugated thapsigargin. Control slicesshowed neither significant nor specific labeling abovebackground levels.

Temporal kinetics

The rise and fall times of a transient [Ca21]i response inNM neurons were examined using in vitro slices that hadbeen incubated in oxygenated ACSF containing 5 mMFluo-3, 1.7% dimethyl sulfoxide, and 0.03% Pluronic. Ini-tially, the XT linescan mode (6 msec/line) of the CLSMwas used to monitor 200 (650)% fluorescence increaseselicited by varying concentrations of KA at the differentages. However, the transients were slow and could beresolved using the FAST XY scan mode of the CLSM.(Images were #512 3 512 pixels and acquisitionwas #750 msec/image.) Therefore, the remainder of theexperiments were conducted using this mode of imageacquisition. The fluorescence values were averaged for allcells in a given slice and treated as a single observation.To assess the temporal aspects of the increase of the global[Ca21]i transient, the time for the fluorescence values toincrease from 10% to 90% of the maximum response to KAwas analyzed (10–90 rise time). Similarly, to assess thetemporal aspects of the global buffering of the [Ca21]itransient, the time for the fluorescence to recover from90% to 10% of the maximum was analyzed. These valueswere compared across ages using one-factor analysis ofvariance.

Drugs, chemicals, media

Harvesting buffer (pH 7.4) for the Westerns consisted of50 mM Trizma (Sigma Chemical Co.), 1 mM EDTA, 7mg/ml CLAP (chymostatin, leupeptin, antipain, and pep-statin), 350 mg/ml phenylmethylsulfonyl fluoride, 5 mg/mlaprotinin, and 1% Triton X-100. ACSF consisted of (inmM): 130 NaCl; 3 KCl; 2 CaCl2; 2 MgCl2; 26 NaHCO3;1.25 NaH2PO4; and 10 D-glucose. Ca21-free ACSF hadCaCl2 replaced with MgCl2 and addition of 1 mM EGTA.KA and ACPD were from Tocris (St. Louis, MO). Fura-2/AM and Fluo-3/AM were from Molecular Probes. Allother reagents were of tissue culture grade. Solutionswere prepared within 48 hours of use.

RESULTS

Activity-dependent [Ca21

]i homeostasis

Previous studies have shown that late embryonic NMneurons require eighth nerve activation of mGluRs tomaintain physiologic levels of intracellular Ca21 (Zirpeland Rubel, 1996). If NM neurons at earlier ages alsodepend on eighth nerve activity for [Ca21]i homeostasis,one would expect activity deprivation to disrupt that ho-meostasis. Figure 1 shows the effect of the absence ofeighth nerve stimulation on the [Ca21]i of NM neurons atfour different ages: E18 (h), E15 (Œ), E13 ({), and E10 (F).E18 and E15 NM neurons showed an increase in [Ca21]i

with time in the absence of eighth nerve activity. During2 hours of activity deprivation, E18 neurons showed anincrease from 100 6 13 nM to 271 6 22 nM, whereas E15NM neurons increased from 89 6 12 nM to 185 6 20 nM.Both of these changes were significant (P , 0.05, twotailed t-test). Stimulation of the eighth nerve at 0.1 to 1.0Hz prevented the increase in [Ca21]i in E18 (data notshown; see also Zirpel and Rubel, 1996) and E15 (data notshown) NM neurons. Neither E10 nor E13 NM neuronsshowed changes in [Ca21]i during 2 hours of activity de-privation (66 6 14 nM vs. 52 6 9 nM, and 85 6 38 vs. 97 640 nM, respectively). Eighth nerve stimulation had noeffect on the [Ca21]i of E13 or E10 neurons during 2 hours(data not shown). The initial baseline [Ca21]i of NM neu-rons did not differ significantly across the four ages(F[3,12] 5 0.511, P 5 0.682).

It is interesting to note that at E13, when the variabilityof [Ca21]i is twice that of any other age, NM neurons areundergoing morphological and functional transformationsalong a spatiotemporal gradient that results in relativelymature neurons rostromedially and relatively immaturecells caudolaterally within the nucleus (reviewed by Rubeland Parks, 1988). Nucleus magnocellularis neurons atE13 also occasionally showed spontaneous (two of eightslices) and ACPD-induced (one of eight) oscillations in[Ca21]i (inset of Figure 1). These three slices were notincluded in the analysis of the data presented in Figure 1.[Ca21]i oscillations were not observed in NM neurons ofslices at any other age. It is interesting to note that thetemporal kinetics of these oscillations are quite differentthan those observed in response to KA application (firstpeak in inset of Fig 1). The mechanism of these oscilla-tions is unknown at this time.

Fig. 1. Development of deafferentation-induced increase in[Ca21]i. Graph shows average nucleus magnocellularis (NM) [Ca21]iin the absence of stimulation at the four developmental stages used inthis study. t 5 0 is the initiation of the imaging experiment and doesnot reflect the 40-minute tissue preparation time. Embryonic day 18(E18), n 5 4 (39 cells); E15, n 5 4 (38 cells); E13, n 5 5 (43 cells); andE10, n 5 3 (28 cells). Error bars represent the standard error of themean and are #10 nM where not seen. Inset graph shows theaverage [Ca21]i of 10 NM neurons from an E13 slice. First horizontalbar denotes application of 30 mM Kainate, second (long) horizontal bardenotes Ca21-free medium, and third horizontal bar denotes applica-tion of 500 mM trans-aminocyclopentane dicarboxylate (ACPD). Noteinitiation of oscillations of [Ca21]i on application of Ca21-free mediumand that these oscillations increase in frequency and become moreregular on stimulation of metabotropic glutamate receptors withACPD.

98 L. ZIRPEL ET AL.

Expression levels of group I mGluRs

Because mGluRs are critical for the activity-dependentmaintenance of NM neuron [Ca21]i, we sought to examinethe protein levels of group I mGluRs during the develop-ment of synaptic activity. Staining of NM neurons formGluRs 1 or 5 revealed an interesting change in distribu-tion of label over development. Figure 2A shows mGluR5staining in an E10 NM. At this age, both mGluR1 and 5showed diffuse staining throughout the cytoplasm of thecell but not the nucleus. mGluR1 and mGluR5 stainingshowed identical patterns at all ages. The cytoplasmicstaining was not uniform because areas of intense fluores-cence were frequently seen surrounding the nucleus. Theintensity of staining decreased toward the periphery and

in the dendritic areas. Labeling could be seen extendinginto what appeared to be the axons. Bright, punctatelabeling was observed in the dendritic areas and possiblyon the surrounding glia. At E13 (Figure 2B), mGluR la-beling was more uniform throughout the cytoplasm of theNM neurons and appeared to extend into the axon initialsegment, but not beyond. The dendritic areas could still beseen but were much smaller and less intensely stainedthan at E10. Punctate labeling was not observed at thisage. By E15, dendritic labeling was absent because theNM neurons no longer possess dendrites (Rubel andParks, 1988). Staining was diffuse throughout the cyto-plasm of the somata in a manner similar to E10 staining,but without extending into the axonal processes. E18 la-

Fig. 2. Metabotropic glutamate receptor 5 immunofluorescent labeling of nucleus magnocellularisneurons at embryonic day 10 (E10) (A), E13 (B), E18 (C), and E18 at high power (D). Scale bars 5 30 mm.


beling (Figure 2C,D) was restricted to the peripheral areasof the NM neurons, forming an annular type pattern sur-rounding the cells. The cytoplasm was mostly devoid ofstaining, with the rare exception of what appeared to belight staining of the initial part of axonal processes. Areasof intense staining were frequently observed on the edge ofcells; these may represent synaptic zones. Intense stain-ing in the inter- and pericellular areas may representpresynaptic or glial labeling.

Quantification of the fluorescent labeling yielded rela-tive values almost identical to those obtained from immu-noblots on NM isolated from brainstems. Therefore, webelieve that this verifies the validity of each technique(although they function through the same mechanism ofprimary antibody binding). Immunoblots were normalizedto total protein and immunofluorescence was normalizedto cell area. Both methods produced parallel quantifica-tion curves and thus provide converging lines of evidence

for developmental changes in mGluR expression. There-fore, these data were expressed as percent maximum andpooled for statistical analysis. It should be noted thatalthough statistics are presented for pooled data, analysisof variance performed on the data from both experimentalmethods for each protein quantified yielded a significantinteraction term (P # 0.05). Figure 3 shows a typicalimmunoblot, which includes lanes for NM from each of thefour ages studied and rat forebrain as a positive control,obtained using a polyclonal antibody against mGluR5.The molecular weight of mGluR5 is approximately 128kD, and the band above the main mGluR5 band (which isgenerally regarded as representing glycosylated receptors,Abe et al., 1992; Shigemoto et al., 1993; Standley et al.,1998) was included in the data analysis. Twenty micro-grams of protein from rat liver was also run as a negativecontrol and showed no labeling in the vicinity of themGluR5 band (data not shown). The top graph in Figure 3shows the data from immunoblots and fluorescent immu-nolabeling plotted on respective ordinates. The bottomgraph in Figure 3 shows the pooled data expressed aspercent maximum. Peak mGluR5 expression occurred atE13. This is a significant increase from E10 (62 6 6% ofmaximum). mGluR5 expression decreased between E13and E15 to 40 6 6% of maximum. Expression continued todecrease to E18 when levels were 10 6 2% of E13 levels.Analysis of variance showed a significant interaction term(F[3,20] 5 72.4, P , 0.0001) for mGluR5 levels over age.Post hoc analysis revealed that the level at each age wassignificantly different from levels at every other age (Fish-er’s protected least significant difference (PLSD), P #0.0001 except between E10 and 15, where P 5 0.0019).

Figure 4 shows a typical mGluR1 immunoblot, whichincludes lanes for NM and rat cerebellum as a positivecontrol. Twenty micrograms of protein from rat liver wasalso run as a negative control and showed no labeling nearthe mGluR1 band (data not shown). The molecular weightof mGluR1 is 133 to 155 kD (Houamed et al., 1991; Masuet al., 1991). Figure 4 shows that peak expression ofmGluR1 also occurred at E13. This represents a signifi-cant increase from E10 (38 6 9% maximum). E15 expres-sion levels were 78 6 12% and E18 levels were 15 6 8% ofE13 levels. Analysis of variance showed a significant in-teraction between age and expression level of mGluR1(F[3,16] 5 20.8, P , 0.0001). Post hoc analysis (Fisher’sPLSD) revealed that the increase in mGluR1 expressionfrom E10 to E13 was significant (P , 0.0001), as was thedecrease in expression from E15 to E18 (P , 0.0001).

E18 NM neurons can release Ca21 from intracellularstores in response to mGluR activation (Zirpel et al.,1995b). This release is critical for the overall maintenanceof stable [Ca21]i homeostasis (Zirpel and Rubel, 1996).Given the large changes in mGluR expression at differentages described in the previous section, we sought to deter-mine whether there was a parallel change in ability ofmGluRs to release Ca21 from intracellular stores. [Ca21]iwas monitored in slices loaded with Fura-2, whereas 1, 10,100, 1,000, or 10,000 mM ACPD, an mGluR agonist, wasapplied in Ca21-free medium. Low Ca21 (0.5 mM) andCa21-free media prevent synaptic transmission from theeighth nerve to NM neurons (Jackson et al., 1982; Jacksonand Parks, 1988; Hyson and Rubel, 1989; L. Zirpel, un-published observations). The average responses of NMneurons at the different ages are plotted in Figure 5A.Between E10 and 13, there is a large increase in the

Fig. 3. Relative levels of metabotropic glutamate receptor 5(mGluR5) in nucleus magnocellularis (NM) during development. Im-munoblot analysis using 20-mg protein samples from isolated NM ateach age. Twenty micrograms of rat forebrain (f. b.) protein was usedas positive control. Top graph shows the average values for mGluR5 ateach age as determined by immunoblot (n 5 3, {) and immunofluo-rescence (n 5 3; Œ). Bottom graph shows these data pooled andexpressed as percent of maximum expression. Error bars representthe standard error of the mean.

100 L. ZIRPEL ET AL.

amplitude of the Ca21 signal released from stores byACPD activation of mGluRs. Figure 5B more clearly illus-trates this point by showing the amplitude of the Ca21

signal elicited by application of 1,000 mM ACPD at thedifferent ages. Analysis of variance showed a significantinteraction between age and response amplitude (F[3,15]5 8.60, P 5 0.0015). Post hoc analysis (Fisher’s PLSD)revealed that the increase in amplitude of [Ca21]i re-sponse to ACPD between E10 and 13 was significant (P 50.0029), but that the response amplitudes at E13, E15,and E18 were not significantly different from one another(P $ 0.10).

SERCA and IP3 receptor expression levels

Because our data indicate an increase in the ability ofmGluR activation to release Ca21 from stores despite adecrease in mGluR expression, we sought to determinewhether other components of the Ca21 regulatory machin-ery of NM neurons underwent changes in expression thatcould account for these surprising observations.

Sarcoplasmic/endoplasmic reticulum Ca21 ATPases(SERCAs) pump Ca21 from the cytoplasm into the ER,thereby functioning as a dynamic Ca21 buffering mecha-nism that permits a cell to reduce [Ca21]i (Benham et al.,1992; Pozzan et al., 1994). Campbell et al. (1993) reportedthat SERCAs are strongly expressed in chick brainstemauditory nuclei. We examined the protein levels of SER-CAs in NM neurons at the four ages. This was accom-plished using immunoblots and fluorescent quantificationof BODIPY-conjugated thapsigargin-stained live tissue.Thapsigargin is a potent and irreversible inhibitor of SER-CAs (Lytton et al., 1991). Figure 6A shows typical labelingof NM neurons with BODIPY thapsigargin. Labeling waslocated in the cytoplasm of the neurons (and some glia)and was absent from the nucleus. Regions of the cyto-plasm often showed more intense labeling (Fig. 6A, inset)consistent with localization to the endoplasmic reticulum

Fig. 4. Relative levels of metabotropic glutamate receptor 1(mGluR1) in nucleus magnocellularis (NM) during development. Im-munoblot analysis using 20 mg protein samples from isolated NM ateach age. Five micrograms of rat cerebellum was used as positivecontrol. Top graph shows average relative levels of mGluR1 expres-sion at each age as determined by immunoblot (n 5 2; 1) and immu-nofluorescent labeling analyses (n 5 3; E). Bottom graph shows thesedata pooled and expressed as percent of maximum expression. Errorbars represent the standard error of the mean.

Fig. 5. Increasing function of Group I metabotropic glutamatereceptors in nucleus magnocellularis (NM) neurons during develop-ment. A: Dose–response relationship between release of Ca21 fromintracellular stores and concentration of trans-aminocyclopentane di-carboxylate (ACPD). [Ca21]i was monitored while indicated concen-trations of ACPD were applied for 30 seconds to slices of differing agesthat were bathed in Ca21-free medium containing 1 mM EGTA for 2minutes before ACPD application. Data points represent the meanvalue of five slices at each age. B: Bar graph shows increasing sensi-tivity of NM neuron Ca21 release to 1,000 mM ACPD (right axis).Thirty micromolar kainate (KA) was applied in each experiment as acontrol for tissue viability (left axis). Error bars represent 1 SD.Embryonic day 10 (E10) n 5 8 for KA, 5 for ACPD; E13 n 5 5 for KA,4 for ACPD; E15 n 5 8 for KA, 7 for ACPD; E18 n 5 5 for KA, 3 forACPD.


(ER). Labeling could also be seen extending down theaxonal processes. Immunoblot (top of Fig. 6B) data andfluorescent quantification of SERCAs showed identicaltrends (Fig. 6B, top graph); therefore, data were pooledand are shown in the graph of Figure 6C. SERCA expres-sion increases from E10 (65 6 10% maximum) to peak atE13. SERCA expression then decreases from E13 to E15(50 6 6% maximum). Analysis of variance showed a sig-nificant interaction between age and SERCA expression(F[3,16] 5 17.0, P , 0.0001). Post hoc analysis (Fisher’sPLSD) revealed that the increase from E10 to E13 wassignificant (P 5 0.0021) and the decrease from E13 to E15also was significant (P , 0.0001). The change in SERCAexpression from E15 to E18 was not statistically signifi-cant (P 5 0.126).

mGluRs release Ca21 from intracellular stores in NMneurons by activating phospholipase C, which generatesIP3 (Zirpel et al., 1994; Zirpel et al., 1998a), which binds toreceptors on the ER, thus allowing Ca21 influx into thecytoplasm. Using immunoblot analysis and fluorescentimmunohistochemistry, we examined the expression lev-els of IP3R in NM neurons at the four different ages.Immunohistochemical staining of NM neurons for IP3Rrevealed a labeling pattern identical to that observed withBODIPY thapsigargin, consistent with localization to theER. Fluorescence quantification and immunoblot data(Fig. 6B, bottom graph) were pooled and are shown in thegraph in Figure 6C. Similar to SERCA expression, IP3Rexpression increased significantly from E10 (42 6 8%maximum) to E13 (80 6 9% maximum). However, unlikeSERCA expression, IP3R expression continued to increaseto peak at E18. Analysis of variance showed a significantinteraction between age and expression level (F[3,20] 510.2, P 5 0.0003). Post hoc analysis revealed the changefrom E10 to E13 to be significant (P 5 0.002).

Temporal kinetics of NM neuron globalcalcium signaling

Because multiple components of the NM neuron calciumsignaling machinery are changing during development,interpretation of the increasing Ca21 release in responseto ACPD is difficult at best. To clarify this point, weexamined the kinetics of the overall Ca21 signal in NMneurons at the different ages. If overall buffering changesdramatically during development, then the changes inACPD-induced Ca21 release we observed would be diffi-cult to interpret since it would not be clear what changeswithin the Ca21 signaling machinery. However, if theoverall buffering remains the same over development, de-spite various components changing, it is likely that theobserved increase in ACPD-induced Ca21 release is real,even if the exact mechanism responsible for it is unknown.KA-elicited [Ca21]i signals from Fluo-3 loaded NM neu-rons were initially monitored using the XT mode (6 ms/line) of a CLSM. The Ca21 signals in NM neurons showedtime courses on the order of minutes, thus obviating theneed for the degree of temporal resolution of the XT scan.Conventional CLSM Fluo-3 imaging was performed forthe remainder of the experiments. At each age, the con-centration of KA required to elicit a 200 (650)% increasein Fluo-3 fluorescence was determined and used for theremainder of experiments. Figure 7 shows the averagerise and fall times of these KA-elicited Ca21 signals at thedifferent ages. There were no consistent age-related dif-ferences in the shape of the KA-elicited Ca21 signals.

Fig. 6. Inositol trisphosphate receptors (IP3R) and sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA) expression changewith development. A: Nucleus magnocellularis (NM) neurons from alive slice stained with BODIPY conjugated thapsigargin. Note cyto-plasmic staining in both neurons and glia. Inset: Cell image showsthat staining extends well into the axonal process. B: Top. Relativelevels of SERCA-2 expression at each age using 25 mg protein fromisolated NM for immunoblots (n 5 2; F) and BODIPY-conjugatedthapsigargin in live slices (n 5 3; h). B: Bottom. Graph representsrelative levels of IP3R expression as determined by immunofluores-cent labeling (n 5 5; i) and immunoblot (n 5 1; ¿). Error barsrepresent the SEM, except for IP3R immunoblot, where they repre-sent 1 SD. C: Graph shows relative levels of SERCA-2 (}) and IP3R(E) expression at the different ages as percent maximum expression.Error bars represent the S.E.M. Scale bar 5 30 mm in A.


Analysis of variance showed that there is not a significantinteraction between age and rise time of Ca21 signal(F[3,12] 5 0.847, P 5 0.494). However, there was a signif-icant interaction between age and fall time of Ca21 signal(F[3,12] 5 4.99, P 5 0.0179). Post hoc analysis revealedthat the E13 fall time is significantly (P , 0.05) differentfrom the fall times at the other ages. This indicates thatE13 NM neurons require more time to return [Ca21]i tobaseline levels after a Ca21 signal.

In summary, we have shown that NM neurons becomedependent on afferent activity for maintenance of [Ca21]ihomeostasis at an age (E15) after anatomical synapticconnections are made and when spontaneous synapticactivity becomes prominent. Expression of mGluR1,mGluR5, and SERCAs peaks earlier, at E13, and theability of mGluRs to release Ca21 from intracellular storesincreases significantly from E10 to E13. From E13 to E18,a time of increasing synaptic activity, levels of mGluR1,mGluR5, and SERCAs decrease, but the mGluR-mediatedrelease of Ca21 does not. Also during this time, IP3Rexpression continues to increase after a large increasefrom E10 to E13. Surprisingly, the rise time of globalcalcium signals did not change with age, but E13 NMneurons were slower to buffer [Ca21]i back to baselinelevels. These results show (1) that NM activity-dependent,

mGluR-mediated [Ca21]i homeostasis develops in parallelwith synaptic activity and becomes more efficient withincreasing activity; and (2) that E13 NM neurons imple-ment mechanisms that serve to increase the magnitudeand time course of [Ca21]i signals at a time when synapticactivity is increasing.

DISCUSSION

Our results are consistent with the hypothesis that thedevelopment of synaptic activity results in the implemen-tation by NM neurons of mGluR-mediated mechanisms tocope with increasing activity and Ca21 fluxes. It is nowwell established that certain aspects of development insome neural systems are activity dependent (reviewed byCrair, 1999 and Buonanno and Fields, 1999). Sensorysystems in particular exhibit spontaneous activity pat-terns early in development, presumably to consolidateconnections and provide a signal of impending sensoryinput (Lippe, 1994; Katz and Shatz, 1996; Wong et al.,1998). In the chick auditory system, rhythmic bursts ofactivity generated in the inner ear are propagated viacochlear nerve synapses to NM neurons as early as E14(Lippe, 1994). Lippe also showed that the duration of therhythmic bursts (and thus the total amount of electricalactivity) increases steadily from E14 until the high con-tinuous level of spontaneous activity that characterizesmature auditory systems is reached at E18. In addition toavian cochlear nucleus neurons (Zirpel et al., 1995a), im-mature rat retinal cells and cortical neurons depend onactivation of mGluRs for regulation of [Ca21]i (Price et al.,1995). Neurons of the rat striatum and hippocampus alsoshow distinct changes in inositol phospholipid signalingpathways after deafferentation (Nicoletti et al., 1987). NMneurons show an increasing dependence on eighth nerveactivity for maintenance of [Ca21]i homeostasis that ap-pears at E15. At this point in development, NM neuronshave become adendritic spheres and receive calyceal glu-tamatergic endings from the cochlear nerve (reviewed byRubel and Parks, 1988). Assuming that NM neurons atE15 must maintain relatively low levels of [Ca21]i forhealth and survival (Zirpel et al., 1995a; Zirpel and Rubel,1996; Zirpel et al., 1998a) despite a high level of glutamatereceptor–mediated depolarization, it seems appropriatethat it is at this stage that the mGluR-mediated regula-tion of [Ca21]i becomes prominent.

Many studies have investigated the developmental ex-pression and function of mGluRs in various tissues. In ratretinal bipolar cells, mGluR6 staining shows a redistribu-tion from somatic and dendritic labeling to restricted lo-calization at the postsynaptic site (Nomura et al., 1994).Furthermore, group I mGluRs show a similar redistribu-tion with increasing synaptic activity in the mouse ventralposterior thalamic nucleus (Liu et al., 1998). This is sim-ilar to the change in staining pattern of group I mGluRs inNM reported here. Although we cannot rule out that theannular mGluR staining at E18 may be presynaptic orglial, there are several lines of evidence against this. First,the mGluR staining at earlier ages is localized to NMneurons, and it seems unlikely that expression wouldcease at a time when function is greatest. Second, thepharmacology experiments characterizing mGluR func-tion show that [Ca21]i signals are clearly localized to thepostsynaptic NM neurons (Zirpel et al., 1995b). The factthat NM neurons respond to ACPD in a Ca21-free medium

Fig. 7. Temporal aspects of global [Ca21]i homeostasis in nucleusmagnocellularis neurons. Graphs show the rise (top) and fall (bottom)times of a 200 (650)% Fluo-3 fluorescence increase elicited by kainate(KA) at the four different ages. KA concentrations used for 20-secondapplications: embryonic day 10 (E10), 25 mM (n 5 4); E13, 15 mM (n 55); E15, 15 mM (n 5 4), and E18, 10 mM (n 5 3). Error bars representthe S.E.M.


(Fig. 5), which prevents synaptic transmission, indicatesthat at least some mGluRs are located on postsynapticNM neurons. Third, acutely dissociated NM neurons de-void of presynaptic terminals show [Ca21]i release in re-sponse to ACPD in Ca21-free medium (L. Zirpel, unpub-lished observations). Finally, 24 hours and 1 week aftercochlea removal, which results in the death of the gan-glion cells and their axons, NM neurons show the sameannular staining pattern for mGluR5 (P. Monsivais and E.W Rubel, personal communication). It therefore is likelythat the redistribution of labeling of group I mGluRs inNM is an activity-dependent relocalization to postsynapticsites under the eighth nerve calyces. Immunohistochemi-cal electron microscopy and multilabel confocal studiesare currently under way to definitively clarify this issue.

Expression of mGluR5 decreases dramatically from P7to adult in the rat brain (Romano et al., 1996); group ImGluR expression decreases over development in cat vi-sual cortex (Reid et al., 1997), and several studies haveshown a decrease in group I mGluR function with increas-ing age (Nicoletti et al., 1986; Dudek and Bear, 1989;Dudek et al., 1989). Dudek and Bear (1989) report a peakin mGluR function at the peak of kitten visual cortexplasticity, whereas Reid et al. (1997) report that the ex-pression of mGluR1 and mGluR5 peaks much sooner. Ourresults are consistent with these findings in that group ImGluR expression peaks at E13, before the peak in func-tion at E18. Although mGluRs have been shown not to beinvolved in ocular dominance plasticity in visual cortex(Hensch and Stryker, 1996; Reid et al., 1997; Huber et al.,1998), our data are consistent with the hypothesis thatgroup I mGluRs are critical components of activity-dependent regulation of chick cochlear nucleus neuronsduring the developmental period of susceptibility to deaf-ferentation.

Immature and developing neurons require influx ofCa21 for survival, differentiation, and consolidation ofconnections (reviewed by Franklin and Johnson, 1992;Ghosh and Greenberg, 1995; Spitzer, 1995). Fully differ-entiated, mature neurons, on the other hand, maintainrelatively low [Ca21]i, and perturbations to the Ca21 ho-meostatic machinery that cause increases in [Ca21]i areoften pathological (reviewed in Choi, 1988; 1992). Thus, asneurons mature, they progress from requiring high Ca21

fluxes to expending considerable energy to maintain low[Ca21]i. For example, in cultured rat superior cervicalganglion neurons, immature cells show dependence onnerve growth factor (NGF)-induced increases in [Ca21]iand sustained [Ca21]i oscillations to caffeine application,whereas mature cells are not dependent on NGF-induced[Ca21]i increases and minimize the transient [Ca21]i re-sponse to caffeine application (Itoh et al., 1998). E13 NMneurons are completing their morphological changes andbeginning to receive and process auditory input (Rubeland Parks, 1988). We hypothesize that E13 NM neuronsrequire high levels of Ca21 fluxes to complete these tasks.Consistent with this hypothesis, E13 NM neurons showedthe highest levels of group I mGluR expression and thegreatest increase in mGluR-mediated release of Ca21 fromintracellular stores. That E13 NM neurons also showedsignificantly slower buffering kinetics supports the notionof a larger Ca21 requirement. Returning [Ca21]i to base-line levels at a slower rate functionally increases a givenCa21 signal by increasing the integrated signal, which isusually of more biological importance than amplitude. For

example, neurons exhibit large calcium spikes in responseto depolarization with few, if any, adverse effects, but asustained Ca21 signal of lesser amplitude results in de-generation or death (Carafoli, 1987; Blaustein, 1988;Ghosh and Greenberg, 1995; Zirpel et al., 1998a). E13 NMneurons also exhibited a behavior not observed at anyother age: spontaneous and induced [Ca21]i oscillations.Spontaneous activity in developing neurons has been ob-served in a number of neural circuits and is proposed toassist in the consolidation of precise connections (Lippe,1994; Katz and Shatz, 1996; Wong et al., 1998). Activationof mGluRs induces electrical oscillations in hippocampaland cortical neurons (Taylor et al., 1995; Whittington etal., 1995) and [Ca21]i oscillations in transfected cells(Kawabata et al., 1996). The [Ca21]i oscillations observedin E13 NM neurons may be mediated by mGluRs andcould increase the overall [Ca21]i to facilitate the consol-idation of the cochlear nerve afferent synapses from im-mature, multiple axodendritic contacts to the one to twovery large, axosomatic calyces seen at maturity.

The increase in mGluR-mediated function observed inNM neurons during a period of decreasing mGluR proteinexpression might be explained by receptor insertion intothe plasma membrane. Our studies did not differentiatebetween cytosolic and membrane-bound receptor proteinand thus, from E10 to E18, there may be less total recep-tor protein but more functional receptors inserted into theplasma membrane. Alternatively, the increasing efficacyof ACPD to elicit Ca21 responses despite decreasing re-ceptor protein levels may be explained by the parallelincrease in IP3R expression. Thus, the IP3 generated bythe activation of any given mGluR has a greater probabil-ity of binding to and activating an IP3 receptor and releas-ing Ca21 from the ER, thus increasing the efficiency ofmGluR coupling to Ca21 release. It may initially seemparadoxical that increased efficiency of releasing Ca21

from stores ultimately contributes to strict [Ca21]i ho-meostasis of NM neurons. [Ca21]i homeostasis is a bal-ance of mechanisms that increase [Ca21]i (such as influxthrough voltage- and ligand-gated Ca21 channels and re-lease from stores) and mechanisms that decrease [Ca21]i(such as extrusion, pumping, and binding to proteins).Release of Ca21 is one of several mechanisms that regu-late [Ca21]i during any physiological state, in this casecochlear nerve activation of glutamate receptors. Secondmessenger systems are known to interact with one an-other (Hille, 1992), and mGluR-mediated signaling path-ways in NM neurons clearly do so to result in dynamicregulation of [Ca21]i during synaptic activation (Zirpel etal., 1998a).

It is surprising that SERCA expression decreases in NMneurons after E13 whereas overall [Ca21]i homeostasisbecomes more tightly regulated. Because SERCAs arehigh-affinity, low-capacity Ca21 pumps, they may be re-placed during development by more efficient Ca21 bufferssuch as Na1/Ca21 exchangers (Mattson et al., 1989). Con-sistent with this idea, Werth et al. (1996) showed thatinhibition of SERCAs with thapsigargin had no effect onthe ability of rat dorsal root ganglion neurons to buffer adepolarization-induced [Ca21]i transient, but inhibition ofplasma membrane calcium ATPases (PMCAs) with a pep-tide inhibitor significantly slowed recovery to baseline[Ca21]i. Also, the ability of E18 NM neurons to recoverfrom a depolarization-induced [Ca21]i transient is unaf-fected by SERCA inhibition (Mostafapour et al., 1997).


The roles of Na1/Ca21 exchange and PMCAs in NM neu-ron [Ca21]i homeostasis need to be investigated.

Our results are consistent with the hypothesis that in-creasing synaptic activity during the development of thechick cochlear nucleus results in mGluR-mediated [Ca21]ihomeostasis. This is implemented first by an increase inexpression of group I mGluRs and then by more efficientcoupling of mGluRs to release of Ca21 from intracellularstores. This release of Ca21 ultimately results in moreefficient [Ca21]i homeostasis by Ca21-dependent kinasesand phosphatases. We hypothesize that E13 NM neuronstransiently express mechanisms that ensure a largeenough intracellular Ca21 signal to facilitate the expres-sion and function of the appropriate mechanisms to ac-commodate the increasing glutamatergic, afferent inputand large Ca21 fluxes. After the implementation of thesemechanisms, the NM neurons, in the face of increasingactivity levels, then implement new or additional mecha-nisms that allow the activity-dependent maintenance oflow [Ca21]i. These hypotheses can be tested experimen-tally in a model where all spontaneous activity in thebrainstem auditory pathway is eliminated by early bilat-eral removal of the otocysts (e.g., Parks et al., 1997).

ACKNOWLEDGMENTS

The authors thank Dr. S. B. Kater for allowing us to usehis imaging facilities and Dr. Douglas Fambrough forgenerously providing the CaS-3H2 monoclonal antibody.We also thank Alan C. Peterson for technical assistance.

LITERATURE CITED

Abe T, Sugihara H, Nawa H, Shigemoto R, Mizuno N, Nakanishi S. 1992.Molecular characterization of a novel metabotropic glutamate receptormGluR5 coupled to inositol phosphate/Ca21 signal transduction. J BiolChem 267:13361–13368.

Benham CD, Evans ML, McBain CJ. 1992. Ca21 efflux mechanisms fol-lowing depolarization evoked calcium transients in cultured rat sen-sory neurones. J Physiol (Lond) 455:567–583.

Bindokas VP, Lee CC, Colmers WF, Miller RJ. 1998. Changes in mitochon-drial function resulting from synaptic activity in the rat hippocampalslice. J Neurosci 18:4570–4587.

Blaustein MP. 1988. Calcium transport and buffering in neurons. TrendsNeurosci 11:438–443.

Buonanno A, Fields RD. 1999. Gene regulation by patterned electricalactivity during neural and skeletal muscle development. Curr OpinNeurobiol 9:110–120.

Campbell AM, Wuytack F, Fambrough DM. 1993. Differential distributionof the alternative forms of the sarcoplasmic/endoplasmic reticulumCa21-ATPase, SERCA2b and SERCA2a, in the avian brain. Brain Res605:67–76.

Carafoli E. 1987. Intracellular calcium homeostasis. Annu Rev Biochem56:395–433.

Choi DW. 1988. Glutamate neurotoxicity and diseases of the nervoussystem. Neuron 1:623–634.

Choi DW. 1992. Excitotoxic cell death. J Neurobiol 23:1261–1276.Crair MC. 1999. Neuronal activity during development: permissive or

instructive? Curr Opin Neurobiol 9:88–93.Dudek SM, Bear MF. 1989. A biochemical correlate of the critical period for

synaptic modification in the visual cortex. Science 246:673–675.Dudek SM, Bowen WD, Bear MF. 1989. Postnatal changes in glutamate

stimulated phosphoinositide turnover in rat neocortical synaptoneuro-somes. Brain Res Dev Brain Res 47:123–128.

Franklin JL, Johnson EM Jr. 1992. Suppression of programmed neuronaldeath by sustained elevation of cytoplasmic calcium. Trends Neurosci15:501–508.

Ghosh A, Greenberg ME. 1995. Calcium signalling in neurons: molecularmechanisms and cellular consequences. Science 268:239–247.

Grynkiewicz G, Poenie M, Tsien RY. 1985. A new generation of Ca21

indicators with greatly improved fluorescence properties. J Biol Chem260:3440–3450.

Hensch TK, Stryker MP. 1996. Ocular dominance plasticity under metabo-tropic glutamate receptor blockade. Science 272:554–557.

Hille B. 1992. G protein-coupled mechanisms and nervous signaling. Neu-ron 9:187–195.

Houamed KM, Kuijper JK, Gilbert TL, Haldeman BA, O’Hara PJ, Mulvi-hill ER, Almers W, Hagen FS. 1991. Cloning, expression and genestructure of a G protein-coupled glutamate receptor from rat brain.Science 252:1318–1321.

Huber KM, Sawtell NB, Bear MF. 1998. Effects of the metabotropic glu-tamate receptor antagonist MCPG on phosphoinositide turnover andsynaptic plasticity in visual cortex. J Neurosci 18:1–9.

Hyson RL, Rubel EW. 1989. Transneuronal regulation of protein synthesisin the brain stem auditory system of the chick requires synaptic acti-vation. J Neurosci 9:2835–2845.

Itoh T, Niwa H, Nagamatsu M, Mitsuma T, Miyakawa A, Pleasure D,Sobue G. 1998. Nerve growth factor maintains regulation of intracel-lular calcium in neonatal sympathetic neurons but not in mature oraged neurons. Neuroscience 82:641–651.

Jackson H, Parks TN. 1988. Induction of aberrant functional afferents tothe chick cochlear nucleus. J Comp Neurol 271:106–114.

Jackson H, Hackett JT, Rubel EW. 1982. Organization and development ofbrain stem auditory nuclei in the chick: ontogeny of postsynaptic re-sponses. J Comp Neurol 210:80–86.

Johnson ME, Gores GJ, Uhl CB, Sill JC. 1994. Cytosolic free calcium andcell death during metabolic inhibition in a neuronal cell line. J Neuro-sci 14:4040–4049.

Kaprielian Z, Fambrough DM. 1987. Expression of fast and slow isoformsof the Ca21-ATPase in developing chick skeletal muscle. Dev Biol124:490–503.

Katz LC, Shatz CJ. 1996. Synaptic activity and the construction of corticalcircuits. Science 274:1133–1138.

Kawabata S, Tsutsumi R, Kohara A, Yamaguchi T, Nakanishi S, Okada M.1996. Control of calcium oscillations by phosphorylation of metabo-tropic receptors. Nature 383:89–92.

Lippe WR. 1994. Rhythmic spontaneous activity in the developing avianauditory system. J Neurosci 14:1486–1495.

Liu XB, Munoz A, Jones EG. 1998. Changes in subcellular localization ofmetabotropic glutamate receptor subtypes during postnatal develop-ment of mouse thalamus. J Comp Neurol 395:450–465.

Lnenicka GA, Arcaro KF, Calabro JM. 1998. Activity-dependent develop-ment of calcium regulation in growing motor axons. J Neurosci 18:4966–4972.

Lytton J, Westlin M, Hanley MR. 1991. Thapsigargin inhibits the sarco-plasmic or endoplasmic reticulum Ca-ATPase family of calcium pumps.J Biol Chem 266:17067–17071.

Masu M, Tanabe Y, Tsuchida K, Shigemoto R, Nakanishi S. 1991. Se-quence and expression of a metabotropic glutamate receptor. Nature349:760–765.

Mattson MP, Guthrie PB, Kater SB. 1989. A role for Na1-dependent Ca21

extrusion in protection against neuronal excitotoxicity. FASEB J3:2519–2526.

Mills LR, Kater SB. 1990. Neuron-specific and state-specific differences incalcium homeostasis regulate the generation and degeneration of neu-ronal architecture. Neuron 2:149–163.

Mostafapour SP, Lachica EA, Rubel EW. 1997. Mitochondrial regulation ofcalcium in the avian cochlear nucleus. J Neurophysiol 78:1928–1934.

Nicoletti F, Iadarola MJ, Wroblewski JT, Costa E. 1986. Excitatory aminoacid recognition site coupled with inositol phospholipid metabolism:developmental changes and interaction with a1-adrenoreceptors. ProcNatl Acad Sci USA 83:1931–1935.

Nicoletti F, Wroblewski JT, Alho H, Eva C, Fadda E, Costa E. 1987.Lesions of the putative glutamatergic pathways potentiate the increaseof inositol phospholipid hydrolysis elicited by excitatory amino acids.Brain Res 436:103–112.

Nomura A, Shigemoto R, Nakamura Y, Okamoto N, Mizuno N, NakanishiS. 1994. Developmentally regulated postsynaptic localization of ametabotropic glutamate receptor in rat rod bipolar cells. Cell 77:361–369.

Park D, Dunlap K. 1998. Dynamic regulation of calcium influx byG-proteins, action potential waveform, and neuronal firing frequency.J Neurosci 18:6757–6766.


Parks TN, Code RA, Taylor DA, Solum D, Strauss KI, Jacobowitz D,Winsky L. 1997. Calretinin expression in the chick brainstem auditorynuclei develops and is maintained independently of cochlear nerveinput. J Comp Neurol 383:112–121.

Pozzan T, Rizzuto R, Volpe P, Meldolesi J. 1994. Molecular and cellularphysiology of intracellular calcium stores. Physiol Rev 74:595–636.

Price MT, Romano C, Fix AS, Tizzano JP, Olney JW. 1995. Blockade of thesecond messenger functions of the glutamate metabotropic receptor isassociated with degenerative changes in the retina and brain of imma-ture rodents. Neuropharmacology 34:1069–1079.

Reid SN, Romano C, Hughes T, Daw NW. 1997. Developmental andsensory-dependent changes of phosphoinositide-linked metabotropicglutamate receptors. J Comp Neurol 389:577–583.

Romano C, van den Pol AN, O’Malley KL. 1996. Enhanced early develop-mental expression of the metabotropic glutamate receptor mGluR5 inrat brain: protein, mRNA splice variants, and regional distribution.J Comp Neurol 367:403–421.

Rubel EW, Parks TN. 1988. Organization and development of the avianbrain-stem auditory system. In: Edelman GM, Gall WE, Cowan WM,editors. Auditory function: neurobiological bases of hearing. New York:John Wiley & Sons. p 3–92.

Scharfman HE, Schwartzkroin PA. 1989. Protection of dentate hilar cellsfrom prolonged stimulation by intracellular calcium chelation. Science246:257–260.

Shigemoto R, Nomura S, Ohishi H, Sugihara H, Nakanishi S, Mizuno N.1993. Immunohistochemical localization of a metabotropic glutamatereceptor, mGluR5, in the rat brain. Neurosci Lett 163:53–57.

Spitzer NC. 1995. Spontaneous activity: function of calcium transients inneuronal differentiation. Perspect Dev Neurobiol 2:379–386.

Standley S, Tocco G, Wagle N, Baudry M. 1998. High- and low-affinity alpha-[3H]amino-3-hydroxy-5-methylisoxazole-4-propionic acid ([3H]AMPA)binding sites represent immature and mature forms of AMPA receptorsand are composed of differentially glycosylated subunits. J Neurochem70:2434–2445.

Taylor GW, Merlin LR, Wong RKS. 1995. Synchronized oscillations inhippocampal CA3 neurons induced by metabotropic glutamate receptoractivation. J Neurosci 15:8039–8052.

Turrigiano G, Abbott LF, Marder E. 1994. Activity-dependent changes inthe intrinsic properties of cultured neurons. Science 264:974–977.

Werth JL, Usachev YM, Thayer SA. 1996. Modulation of calcium effluxfrom cultured rat dorsal root ganglion neurons. J Neurosci 16:1008–1015.

Whittington MA, Traub RD, Jefferys JGR. 1995. Synchronous oscillationsin interneuron networks driven by metabotropic glutamate receptoractivation. Nature 373:612–615.

Wong WT, Sanes JR, Wong ROL. 1998. Developmentally regulated spon-taneous activity in the embryonic chick retina. J Neurosci 18:8839–8852.

Zirpel L, Rubel EW. 1996. Eighth nerve activity regulates the intracellularcalcium concentration of avian cochlear nucleus neurons via a metabo-tropic glutamate receptor. J Neurophysiol 76:4127–4139.

Zirpel L, Lachica EA, Lippe WR. 1995a. Deafferentation increases theintracellular calcium of cochlear nucleus neurons in the embryonicchick. J Neurophysiol 74:1355–1357.

Zirpel L, Lachica EA, Rubel EW. 1995b. Activation of a metabotropicglutamate receptor increases intracellular calcium concentrations inneurons of the avian cochlear nucleus. J Neurosci 15:214–222.

Zirpel L, Lachica EA, Rubel EW. 1997. Afferent regulation of neurons inthe avian cochlear nucleus: intracellular mechanisms and signal trans-duction pathways. In: Berlin C, editor. Neurotransmission and hearingloss: basic science, diagnosis and management. San Diego: SingularPublishing Group. p 47–76.

Zirpel L, Lippe WR, Rubel EW. 1998a. Activity-dependent regulation ofintracellular calcium in avian cochlear nucleus neurons: roles of pro-tein kinases A and C and correlation with cell death. J Neurophysiol79:2288–2302.

Zirpel L, Nathanson NM, Rubel EW, Hyson RL. 1994. Glutamate stimu-lated phosphatidylinositol metabolism in the avian cochlear nucleus.Neurosci Lett 168:163–168.

Zirpel L, Peterson AC, Taylor DA, Parks TN. 1998b. Developmentalchanges in mGluR-mediated calcium homeostasis. Soc Neurosci Abstr24:796.


Documents

Developmental changes in metabotropic glutamate receptor-mediated calcium homeostasis