J .P. Smith et al. Brain Research 887 2000 98 –109
99
physiological and molecular techniques to better under- positive, protoplasmic astrocytes, as assessed using im-
stand their individual contributions and interactions. munofluorescent staining see below. All experiments
21
While [Ca ] responses have been identified in various
were performed on astrocytes that had been maintained in
i
types of glia in response to ionotropic glutamate receptor culture for 17–30 days. Data from cortical and hippocam-
agonists [13,18,30], the mechanisms responsible for these pal astrocytes were always analyzed separately, but were
responses remain to be clearly established in protoplasmic never found to be significantly different, and therefore
astrocytes. Our results demonstrate that AMPAR activa- were combined and presented simply as data from cerebral
tion, under non-desensitizing conditions, results in a astrocytes.
21
substantial elevation of [Ca ] , which is dependent on the
i
presence of extracellular calcium. We provide evidence 2.2. Immunocytochemistry
21
that the majority of this Ca influx is unlikely to be
carried directly by the AMPAR channel, given the rectifi- Astrocyte cultures were fixed with 4 paraformal-
cation properties of these AMPARs, the abundant expres- dehyde, blocked in 4 normal donkey serum Chemicon
sion of the GluR2 AMPAR subunit by these astrocytes, International Inc., Temecula, CA, 1 bovine serum
and the inhibition of the entire influx by agents that are not albumin BSA, and 0.4 Triton X-100 Sigma. Cells
known to interact with AMPARs. While it is possible that were incubated overnight at 48C in phosphate-buffered
AMPAR-mediated depolarization could activate VGCCs to saline PBS plus 0.4 Triton X, and 1.0 bovine serum
21
allow transmembrane Ca entry, we found no evidence
albumin BSA containing a 1:200 dilution of guinea pig for the expression of VGCCs in these cultured astrocytes,
anti-GFAP antibody 031223; Advanced Immunochemical, as assessed using both electrophysiological and molecular
Long Beach, CA or a 1:1500 dilution of rabbit anti-GluR2
21
methods. However, Ca entry following AMPAR stimula-
polyclonal antibody AB1768; Chemicon International. tion was found to be completely blocked by benzamil and
Cells were rinsed and incubated under subdued light for 1
1 21
U-73122, two pharmacological inhibitors of Na Ca h at room temperature in PBS containing a 1:150 dilution
1
exchanger activity, suggesting that Na influx through the
of fluorescein-conjugated, affinity-purified, donkey anti- AMPAR may drive reverse-mode operation of this ex-
guinea pig IgG Jackson Immuno Research Laboratories changer, thereby mediating the rapid and significant in-
Inc., West Grove, PA and 1.0 mg ml Hoescht nuclear
21
crease in [Ca ]
that we observed. We also provide
i
stain. Alternatively, biotinylated anti-GFAP secondary
21
evidence that AMPAR-dependent transmembrane Ca antibodies were visualized using the ABC-peroxidase
1 21
influx via reverse mode operation of the Na Ca method Vecor Elite kit, Vector Laboratories Inc., Burling-
21
exchanger, is associated with increased [Ca ] , through
i
ame, CA. Cells were mounted on glass slides and
21
the release of Ca from thapsigargin-sensitive intracellu-
photographed. Cells were .95 GFAP positive and lar stores. We conclude that the AMPAR and reverse mode
exhibited a flattened morphology typical of protoplasmic
1 21
Na Ca exchanger activity may be functionally linked
astrocytes.
21
in the co-regulation of intracellular Ca homeostasis in
astrocytes under certain conditions. 2.3. Reverse transcription–polymerase chain reaction
rtPCR
2. Materials and methods
Total RNA was isolated and purified from cultured 2.1. Cell culture
astrocytes by
a modification
of the
method of
Chomczynsky [8], and quantified by spectrophotometry Cerebral astrocytes were grown from the cerebral cortex
using a Beckman DU-64 spectrophotometer. All samples and hippocampus of newborn Sprague–Dawley rats using
produced absorbance ratios between 1.5 and 2.0 l5260 a modification of the method of McCarthy and de Vellis
280 nm. Samples were immediately reverse transcribed [29]. Briefly, on the day of birth postnatal day 0, rat pups
into cDNA using the Superscript preamplification system were decapitated, meninges were removed, and cerebral
according to the manufacturers recommended protocol cortices or hippocampi were dissected. Tissues were
Gibco BRL. RNA templates were digested with Es- mechanically dissociated and plated in Corning T75 or T25
cherichia coli RNase H 2 units ml, GibcoBRL. cDNAs
6 2
tissue culture flasks at a density of 0.75310 cells cm . were stored for less than 2 weeks at 48C before use. Two
Growth media consisted of Dulbecco’s Modified Eagle’s to four microliters of the cDNA reverse transcription
Medium with F-12 supplements Gibco BRL, Gaithers- product were used as polymerase chain reaction PCR
burg, MD plus 10 fetal calf serum Sigma, St. Louis, templates using primers specific for the GluRs, the a1
MO, lot 46H46003, containing 0.1 mg ml gentamycin. subunit of L-type calcium channels, or b-actin sequences,
Cells were grown to confluence 12–14 days and then as described below. All PCR reactions were performed
enriched for cerebral astrocytes by shaking at 250 rota- using the Perkin–Elmer Geneamp PCR kit, according to
tions min for 1–3 days. Cultures were .95 GFAP- the manufacturer’s protocol, and a Perkin–Elmer Genome
100 J
PCR Thermocycler System 9600 Perkin–Elmer, Branch- 40 mg protein per lane for electrophoresis. Gel running
burg, NJ. buffer consisted of 25 mM Tris, 192 mM glycine, and
0.1 SDS. Proteins were electrophoretically transferred to 2.3.1. GluR subunits
immobilon-P membranes Millipore Corp. at 48C. GluR subunit cDNA sequences were detected using the
Membranes were blocked for 1.5 h at room temperature following primers specific for the conserved sequences
in blocking buffer consisting of 5.0 nonfat dry milk, 10 shared by GluR1, 2, 3 and 4 cDNAs as described by
mM Tris pH 7.4, 150 mM NaCl, and 0.05 Tween-20. Lambolez [25]: sense 59-CCTTTGGCCTATGAGATCTG-
Membranes were probed with primary rabbit polyclonal GATGTG-39 and antisense 59-TCGTACCACCATTTGTT-
antibodies at room temperature for 1.5 h and then rinsed. TTTCA-39. cDNA was amplified by 35 cycles of 30 s at
Primary antibodies were used as follows all from 948C, 30 s at 498C and 30 s at 728C, which yielded an
Chemicon International:
2.5 mg ml
anti-GluR1 amplified product of expected 750 base pairs. To provide
AB1504, 2.0 mg ml anti-GluR2 AB1768, 1 mg ml enough material for restriction analysis, the GluR PCR
anti-GluR2 3 AB1506,
and 3
mg ml anti-GluR4
product was reamplified using a 1:7000 dilution of initial AB1508. Membranes were incubated in blocking buffer
PCR product for 35 cycles of 30 s at 968C, 30 s at 558C containing anti-rabbit Ig, horseradish peroxidase-linked
and 30 s at 728C. To distinguish between the different HRP secondary antibody NIF824; Amersham, Arlington
GluR subunit sequences in this reaction mixture, four Heights, IL for 90 min at room temperature and then
equal aliquots were each digested with one of the follow- rinsed. Secondary antibodies were diluted 1:4000. Western
ing restriction enzymes: BgII, Bsp1286I, EcoR47II, or blots were detected using the ECL Western Blotting
EcoRI, each of which recognizes a restriction site unique Analysis System Amersham, and developed with Hy-
to the GluR1, 2, 3, or 4 rtPCR products, respectively, perfilm-ECL Amersham. Controls for non-specific sec-
yielding two fragments of 300 and 449 bp for GluR1 ondary antibody binding were completed in parallel re-
BgII, 478 and 271 bp for GluR2 Bsp1286I, 359 and 396 actions with the omission of the primary antibody solu-
bp for GluR3 EcoR47II, and 411 and 338 bp for GluR4 tions.
EcoRI [25,33]. Digestions were carried out to completion at 378C for 2 h in the appropriate restriction buffers.
2.5. Whole-cell patch clamp 2.3.2. a subunit of the L-type calcium channel
Membrane currents were measured by the whole-cell
1
The following primers were used to amplify sequences patch clamp technique using an EPC-7 amplifier Medical
that are conserved among the a subunits of all L-type
Systems, Darmstadt, Germany and pCLAMP6 software
1
calcium channels: sense 59-CGAAGCTTCTTCATGATG- Axon Instruments Inc., Foster City, CA. Recordings were
AACATCTT-39, antisense 59-GCGGATCCATGTAGAA- made in extracellular recording solution containing in
GCTGATGAA-39. Amplification reactions were carried mM: 140 NaCl, 5 KCl, 2 CaCl , 1 MgCl , 10 HEPES, 25
2 2
21
out at 35 cycles of 30 s at 948C, 30 s at 508C, and 60 s at dextrose, pH 7.4, |320 mosmol. For detection of Ba
728C, which yielded an expected 974 bp product [36]. currents, cells were rapidly superfused with barium and
tetraethylammonium Ba-TEA solution containing in 2.3.3. b-Actin
mM: 10 TEA–Cl, 140 NaCl, 5 KCl, 2 or 12 BaCl , 1
2
The following primer pair was used to amplify b-actin MgCl , 10 HEPES, 25 dextrose, pH 7.4, |320 mosmol.
2
cDNA sequences
as a
positive control:
sense 59-
The pipette solution contained in mM: 140 KCl, 4 TCCTTCCTGGGTATGGAATC-39, antisense 39-ACTCA-
MgCl , 0.1 EGTA, 2 ATP, 10 HEPES, pH 7.4, |290
2
TCGTACTCCTGCTTG-59. Amplification reactions were mosmol. Cells were stepped from a holding potential of
carried out as described above for each reaction in which 260 mV to the indicated test potentials for 25 ms. Data
this primer pair was used. were low pass filtered at 30 kHz and corrected for leak
currents by graphic linear leak subtraction. Since we were 2.4. Western blot analysis
interested in the amplitude of slow AMPA and L-type calcium currents and we did not attempt to measure
Confluent cultures were harvested by scraping and then reversal potentials at large holding potentials, series resist-
collected in ice cold HEPES buffer 0.5 mM containing 1 ance was not an important consideration and was not
mg ml aprotinin and 10 mg ml leupeptin. After centrifu- compensated. Experiments were carried out at room tem-
gation at 14 0003g at 48C, samples were resuspended, perature 22–278C.
quantitated by the Lowry method, and stored at 48C for a A gravity-driven Y-tube [31] with a 100 mm diameter,
maximum of 2 days, or directly electrophoresed. Samples was placed within 10 cell diameters of an astrocyte to
were mixed 1:4 with loading buffer consisting of 62.5 mM permit the rapid exchange of the extracellular solution.
Tris–HCl, 10 glycerol, 2 sodium dodecyl sulfate This system was used to continuously deliver HPP-sper-
SDS, 200 mM b-mercaptoethanol, and 0.00125 bromo- mine, NBQX, or CTZ. A pneumatically driven picospritzer
phenol blue, and boiled for 3–5 min prior to loading into PV820 World Precision Instruments, Sarasota, FL with a
9.0 SDS polyacrylamide mini-gels with 4 stacking gels heat-polished glass capillary tube of 0.5 mm diameter was
J .P. Smith et al. Brain Research 887 2000 98 –109
101
positioned within 1.5 cell diameters of an astrocyte and were continuously perfused during bath exchange, and
used for rapid application of AMPA. Dose–response AMPA or TGN were added directly to the bath to
curves were generated from peak AMPA-induced inward transiently produce the desired bath concentration. Cells
current amplitudes using a least squares regression iterated were pretreated with benzamil and U73122 for no more
by the Marquardt–Levenburg method Prostat V 1.02 than 15 min during which time fluorescence ratios reached
software, Poly Software International, Salt Lake City, UT. a stable baseline, cells showed no changes in morphology,
Because A7r5 cells are well characterized for their con- and ionomycin responses remained robust indicating that
21
stitutive expression of L-type Ca channels [28], these
there were no toxic effects of these drugs at the con- cells were used as a positive control for VGCC-mediated
centrations used. currents in our patch clamp experiments.
2.6. Characterization of AMPARs
3. Results
