Previous Article | Next Article 
The Journal of Neuroscience, July 15, 1999, 19(14):5861-5874
Ca2+-Permeable AMPA Receptors Induce Phosphorylation
of cAMP Response Element-Binding Protein through a
Phosphatidylinositol 3-Kinase-Dependent Stimulation of the
Mitogen-Activated Protein Kinase Signaling Cascade in Neurons
Michael S.
Perkinton1,
Talvinder S.
Sihra2, and
Robert J.
Williams1
1 Biochemical Neuropharmacology Group, Neuroscience
Research Centre, Guy's, King's, and St. Thomas' Schools of
Biomedical Sciences, Guy's Campus, London, SE1 9RT, United Kingdom,
and 2 Department of Pharmacology, Medawar Building,
University College London, London, WC1E 6BT, United Kingdom
 |
ABSTRACT |
Ca2+-permeable AMPA receptors may play a key
role during developmental neuroplasticity, learning and memory, and
neuronal loss in a number of neuropathologies. However, the
intracellular signaling pathways used by AMPA receptors during such
processes are not fully understood. The mitogen-activated protein
kinase (MAPK) cascade is an attractive target because it has been shown
to be involved in gene expression, synaptic plasticity, and neuronal stress. Using primary cultures of mouse striatal neurons and a phosphospecific MAPK antibody we addressed whether AMPA receptors can
activate the MAPK cascade. We found that in the presence of cyclothiazide, AMPA caused a robust and direct (no involvement of NMDA
receptors or L-type voltage-sensitive Ca2+ channels)
Ca2+-dependent activation of MAPK through MAPK
kinase (MEK). This activation was blocked by GYKI 53655, a
noncompetitive selective antagonist of AMPA receptors. Probing the
mechanism of this activation revealed an essential role for
phosphatidylinositol 3-kinase (PI 3-kinase) and the involvement of a
pertussis toxin (PTX)-sensitive G-protein, a Src family protein
tyrosine kinase, and Ca2+/calmodulin-dependent
kinase II. Similarly, kainate activated MAPK in a PI 3-kinase-dependent
manner. AMPA receptor-evoked neuronal death and arachidonic acid
mobilization did not appear to involve signaling through the MAPK
pathway. However, AMPA receptor stimulation led to a
Ca2+-dependent phosphorylation of the nuclear
transcription factor CREB, which could be prevented by inhibitors of
MEK or PI 3-kinase. Our results indicate that
Ca2+-permeable AMPA receptors transduce signals from
the cell surface to the nucleus of neurons through a PI
3-kinase-dependent activation of MAPK. This novel pathway may play a
pivotal role in regulating synaptic plasticity in the striatum.
Key words:
AMPA; mitogen-activated protein kinase; phosphatidylinositol 3-kinase; CREB; wortmannin; LY 294002; pertussis
toxin; G-protein; PP2; cyclothiazide; glutamate; kainate; calcium/calmodulin-dependent kinase II; tyrosine kinase; AMPA toxicity; arachidonic acid; striatum; striatal neurons
| |
INTRODUCTION |
Glutamate is the major excitatory
neurotransmitter in the mammalian CNS and acts through two
classes of receptor: NMDA and non-NMDA [AMPA and kainate
(KA)] ionotropic receptors, and G-protein-coupled metabotropic
receptors (Hollmann and Heinemann, 1994
; Ozawa et al., 1998).
Ionotropic glutamate receptors mediate a wide range of physiological
processes such as fast excitatory synaptic transmission and neuronal
plasticity (Mayer and Westbrook, 1987; Monaghan et al., 1989; Lerma et
al., 1997; Ozawa et al., 1998), and their overactivation has been
implicated in excitotoxic neuronal loss that occurs during a number of
neuropathological conditions (Choi, 1992; Pellegrini-Giampietro et al.,
1997; Michaelis, 1998). It appears that extracellular
Ca2+ influx through glutamate receptor ion channels
plays a central role in gene expression associated with synaptic
plasticity (Deisseroth et al., 1996; Finkbeiner and Greenberg, 1996;
Xia et al., 1996), neuronal development, and survival (Bading et al.,
1993, 1995; Ikonomidou et al., 1999) and activity-dependent
synaptogenesis (Constantine-Paton and Cline, 1998; Lin and
Constantine-Paton, 1998). There is accumulating evidence suggesting
that these processes may require the transduction of signals through
the mitogen-activated protein kinase [(MAPK) ERK1 and ERK2] cascade,
which classically involves Shc tyrosine phosphorylation, recruitment of
the Grb2-Sos complex, and the subsequent sequential activation of Ras,
Raf kinase, and MAPK kinase (MEK) (Rosen et al., 1994; Fukunaga and Miyamoto, 1998). NMDA receptor-induced immediate early gene
transcription and induction of hippocampal long-term potentiation (LTP)
are both thought to involve a Ca2+-dependent
activation of the Ras-MAPK pathway in neurons (English and Sweatt,
1996; Finkbeiner and Greenberg, 1996; Xia et al., 1996; English and
Sweatt, 1997), however, studies regarding AMPA receptor activation of
this cascade are somewhat contradictory. Although AMPA and kainate do
not appear to activate MAPK in hippocampal neurons (Kurino et al.,
1995), it has been reported that AMPA receptors can stimulate MAPK in
cortical neurons through a mechanism that surprisingly requires
G-protein 
subunits (Wang and Durkin, 1995), which is classically
associated with receptors that couple to heterotrimeric G-proteins
(Crespo et al., 1994; Luttrell et al., 1997; Gutkind, 1998). Recent
studies have shown that phosphatidylinositol 3-kinase (PI 3-kinase)
plays a crucial role in signaling to MAPK at a target upstream
of Sos and Ras (Hawes et al., 1996; Lopez-Ilasaca et al., 1997),
although it is not known if this enzyme is required for AMPA receptor
signaling to MAPK.
Ca2+-permeable AMPA receptors can mediate NMDA
receptor-independent long-term potentiation in the amygdala (Mahanty
and Sah, 1998) and hippocampus (Feldmeyer et al., 1999) and probably
play an important role in striatal synaptic transmission and plasticity (Lovinger and Tyler, 1996; Williams and Glowinski, 1996). Furthermore, excessive influx of Ca2+ through AMPA receptors may
contribute to glutamate-induced neuronal cell death associated with
cerebral ischemia, seizure activity, and a number of chronic
neurodegenerative disorders such as Huntington's disease and
amyotrophic lateral sclerosis (Choi, 1995; Pellegrini-Giampietro et
al., 1997; Feldmeyer et al., 1999). However, a better understanding of
the specific synaptic and intracellular signals that occur during the
aforementioned AMPA receptor-mediated physiological and pathological
processes is required. A possible intracellular target is MAPK, thus,
the aim of this study was to investigate whether striatal AMPA
receptors signal to this cascade, and if so, to examine the mechanism
of this activation and to address possible functional consequences.
| |
MATERIALS AND METHODS |
Chemicals and reagents.
(1S,3R)-ACPD, (RS)-AMPA,
cyclothiazide, kainic acid, (+)-MK 801, nimodipine, and NMDA were all
obtained from Tocris Cookson (Bristol, UK). Anti-CREB
(phospho-Ser133-specific), Ionomycin, KN-62, LY
294002, PP2, PP3, Ro-31-8220, and wortmannin were from Calbiochem (La
Jolla, CA). Genistein and PD 98059 were from Alexis (San Diego, CA).
Fura-2 AM was from Molecular Probes (Eugene, OR). Enhanced
chemiluminescence reagent (ECL) and
[3H]arachidonic acid (200 Ci/mmol) were obtained
from Amersham (Little Chalfont, UK). Anti-ACTIVE MAPK polyclonal
antibody (pAb) was from Promega (Madison, WI), and ERK1/ERK2 pAb was
from Santa Cruz Biotechnology (Santa Cruz, CA). Vectastain Elite ABC
Kit and biotinylated goat anti-rabbit IgG were obtained from Vector
Laboratories (Burlingame, CA). Bio-Rad protein assay reagent
(concentrate) was obtained from Bio-Rad (Hemel Hempstead, UK).
Paraformaldehyde (methanol-free) was bought from TAAB Laboratories
(Aldermaston, UK). GYKI 53655 was a gift from Eli Lilly (Indianapolis,
IN). Bovine serum albumin (fatty acid-free),
poly-L-ornithine, and pertussis toxin were from Sigma
(Poole, UK). All other reagents were obtained from Sigma or Merck
(Poole, UK).
Neuronal cell culture. Primary cultures of mouse striatal
neurons were prepared essentially as described previously (El-Etr et
al., 1989). Briefly, striata were dissected from 15- to 16-d-old Swiss
mouse embryos (NIH, Harlan, UK) and mechanically dissociated using a fire-polished glass Pasteur pipette in PBS
(Ca2+- and Mg2+-free)
supplemented with glucose (33 mM). Cells were plated into either six- or 24-well Nunc (Roskilde, Denmark) multiwell plates that
had been coated previously overnight with 15 µg/ml
poly-L-ornithine and then with culture medium supplemented
with 10% fetal bovine serum (Life Technologies, Gaithersburg,
MD) for 2 hr. After removal of the final coating solution, cells
were seeded (106/ml) in a serum-free medium composed
of a mixture of DMEM and F-12 nutrient (1:1 v/v) (Life
Technologies) supplemented with (in mM) 33 glucose, 2 glutamine, 6.5 sodium bicarbonate, and 5 HEPES buffer, pH 7.4, 100 µg/ml streptomycin, and 60 µg/ml penicillin. A mixture of hormones
and salts composed of insulin (25 µg/ml), transferrin (100 µg/ml),
putrescine (60 µM), progesterone (20 nM), and
sodium selenate (30 nM) (all from Sigma) was also added to
the culture medium. Cells were cultured at 37°C in a humidified atmosphere of 95% air and 5% CO2 and were used after 6-7
days in vitro (DIV) when the majority of cells were
neuronal and there were no detectable glial elements.
Immunoblotting. Immunoblot analysis was performed
essentially as described previously (Samanta et al., 1998) with minor
modifications. Neurons were cultured in six-well plates, placed into a
thermostatted water bath at 37°C, and left for 5 min to equilibrate.
After this period, the culture medium was removed and replaced with
HEPES-buffered incubation medium (HBM) (in mM): NaCl 140, KCl 5, NaHCO3 5, MgCl2 · 6H20
1.1, Na2HP04 1.2, CaCl2 1.2, glucose 5.5, and HEPES 20, pH 7.4. For NMDA agonist treatments and
Ca2+ dependency studies, MgCl2 and
CaCl2 were omitted from the HBM, respectively. Other
additions before agonist/depolarization treatments were made as
detailed in the legends to the Figures. After 10 min of incubation,
fresh HBM containing AMPA (50 µM) (in the presence or
absence of 50 µM cyclothiazide), glutamate (100 µM), kainate (100 µM), NMDA (100 µM), (1S,3R)-ACPD (200 µM), KCl (50 mM), or ionomycin (5 µM) was added to the neurons. After 5 min, the HBM was
removed, and the plates were quickly washed with ice-cold PBS, pH 7.4, (Ca2+-free) and placed immediately on ice. The cell
monolayer was rapidly scraped in ice-cold lysis buffer (200 µl/well)
[50 mM Tris, pH 7.5, 150 mM NaCl, 1% Tergitol
(type NP-40), 2 mM EDTA, 2 mM EGTA, 0.5 mM PMSF, 10 µg/ml leupeptin, 10 µg/ml antipain, 1 µg/ml pepstatin A, 1 mM
Na3V04, and 50 mM NaF],
left on ice for 10 min, then homogenized with 20 full strokes in a
glass-on-glass hand-held homogenizer and centrifuged at 1000 × g for 4 min at 4°C to remove cell debris. The supernatant
(crude homogenate) was kept, and protein concentration was determined
by the assay of Bradford (1976). Samples were then boiled for 5 min in
boiling buffer, giving 62.5 mM Tris, pH 6.8, 2%
SDS, 5% 2-mercaptoethanol, 10% glycerol, and 0.0025%
bromophenol blue in the final sample. Boiled samples (15 µg/lane)
were run on 7.5% SDS-polyacrylamide gels, and proteins were
transferred to nitrocellulose membrane (Hybond-C pure; Amersham) by
electroblotting. The equal loading and efficient transfer of proteins
was confirmed by staining the nitrocellulose with Ponceau Red (Sigma).
The nitrocellulose membrane was then incubated in blocking buffer (20 mM Tris, pH 7.5, 150 mM NaCl; TBS) containing
5% skimmed milk powder, for 30 min at room temperature on a
three-dimensional rocker, followed by two 10 min washes in TBS
supplemented with 0.1% (v/v) Tween 20 (TTBS). Blots were then
incubated with anti-ACTIVE MAPK pAb (1:12,500 dilution of stock),
anti-ERK1/ERK2 pAb (1:750 dilution of stock), or
anti-phospho-Ser133 CREB pAb (1:1500 dilution of
stock) in TTBS containing 1% skimmed milk powder (antibody buffer),
overnight at room temperature on a three-dimensional rocker. The blots
were washed (2 × 10 min) in TTBS and then incubated with goat
anti-rabbit IgG conjugated to horseradish peroxidase (1:1000 dilution
of stock) in antibody buffer for 1 hr. Finally, the blots were washed
in TTBS (2 × 10 min), rinsed in TBS, and exposed to ECL reagent
for 1 min, as described in the Amersham protocol. Excess reagent was
removed, and the blots were then exposed to Hyperfilm ECL (Amersham)
for 2 min in an autoradiographic cassette and developed. Bands were analyzed using BioImage Intelligent Quantifier software (Ann Arbor, MI)
Molecular weights of ERK1/ERK2 and CREB were calculated from comparison
with prestained molecular weight markers (MW 27,000-180,000) that were
run in parallel with samples.
Immunocytochemistry. For immunocytochemical studies, neurons
were cultured in 24-well plates and incubated in HBM, pH 7.4, at
37°C, as described for immunoblot analysis. In experiments involving
AMPA receptor stimulation, (+)-MK 801 (2 µM) was present in the HBM to block NMDA-type glutamate receptors. Other additions before agonist treatment were made as detailed in the legends to
Figures 9 and 10B. After 10 min of incubation, fresh
HBM containing AMPA (50 µM) (in the presence of 50 µM cyclothiazide) was added to the neurons. After a
further 5 min, the HBM was removed, and the plates were quickly rinsed
with ice-cold PBS, pH 7.4, (Ca2+-free), and neurons
were fixed on ice in PBS containing 4% paraformaldehyde for 1 hr. All
other procedures were performed at room temperature. Neurons were
permeabilized with 0.2% Triton X-100 made up in PBS containing 1%
goat serum (blocking solution) for 1 hr and then incubated with
anti-ACTIVE MAPK pAb (1:500 dilution of stock), ERK1/ERK2 pAb (1:200
dilution of stock), or anti-phospho-Ser133 CREB pAb
(1:500 dilution of stock) overnight with constant agitation. After
3 × 10 min washes in PBS, neurons were incubated with
biotinylated goat antibody to rabbit IgG (1:200 dilution of stock) for
1 hr, washed 3 × 10 min with PBS, and neuronal staining was
localized using the avidin-biotin-peroxidase detection system (Vector
Laboratories). Neurons were visualized using a Leica Fluovert microscope.
Intracellular Ca2+ measurements. Striatal
neurons prepared as described above, were plated into 24-well Nunc
multiwell plates containing 13 mm diameter glass coverslips that had
been previously coated overnight with 15 µg/ml
poly-L-ornithine and then with 2 µg/ml laminin (Life
Technologies) for 4 hr. After 6-7 DIV, neurons were loaded with the
cell-permeant fluorescent Ca2+-probe fura-2 AM (5 µM) in HBM, pH 7.4, for 45 min at 37°C. Subsequent procedures were performed at room temperature (20-25°C). Coverslips with fura-2-loaded neurons were clamped to a purpose-made recording chamber attached to a Nikon Diaphot 200 inverted microscope stage and
preincubated in HBM, pH 7.4, with or without inhibitors, as detailed in
the legend to Table 1. Neurons were
stimulated with AMPA (50 µM) or AMPA/cyclothiazide (50 µM/50 µM), and to ensure that AMPA
receptor-mediated changes in intracellular Ca2+ were
being observed in the absence of possible NMDA receptor responses, 1 mM MgCl2 and 2 µM (+)-MK 801 were
included in the HBM. Neurons were alternately excited at 340 and 380 nm
using a Cairn filter changer. Fluorescence at each excitation
wavelength was filtered using a 420 nm barrier filter/400 nm dichroic
mirror set and collected using a C4880-81Hamamatsu multiformat CCD
camera (Hamamatsu Photonics K.K., Hamamatsu City, Japan). Data
acquisition and analysis was performed using Acquisition Manager
(version 3.1) and Lucida Analyze (version 3.51) software from Kinetic
Imaging (Liverpool, UK). To improve the signal-to-noise ratio, the
collection period during 340 nm excitation (600 msec) was set at three
times that used at 380 nm (200 msec). Ratio values in the data
presented have not been corrected for this difference. Fluorescence
responses from at least six cells in a single field were analyzed on
each coverslip. For each experimental condition, cells were analyzed on
a minimum of three coverslips from at least two separate striatal cultures. Means ± SEM of the 340 nm/380 nm ratio traces are
reported in Figure 3, and means ± SEM of the agonist-induced
increases in the ratio values evaluated in Table 1.
View this table:
[in this window]
[in a new window]
|
Table 1.
AMPA receptor-evoked increases in intracellular
Ca2+ are insensitive to the protein kinase inhibitors
KN-62, PD 98059, and genistein, and the lipid kinase inhibitor
wortmannin
|
|
Arachidonic acid release. Arachidonic acid release was
performed essentially as described previously (Samanta et al., 1998). Neurons, cultured in 24-well plates were incubated overnight with 0.5 µCi/ml [3H]arachidonic acid. Unincorporated
label was removed by three successive washes in HBM, pH 7.4, containing
fatty acid-free BSA (2 mg/ml; HBM/BSA), and neurons were then
preincubated for 10 min in HBM/BSA at 37°C in the presence or absence
of enzyme inhibitors, with or without cyclothiazide (50 µM). After removal of the preincubation buffer, neurons
were incubated for 15 min in HBM/BSA containing either vehicle or
AMPA/cyclothiazide (50 µM/50 µM), at 37°C
in the continued presence or absence of enzyme inhibitors, as detailed in the legend to Figure 7. The extracellular medium was removed and
centrifuged at 400 × g for 5 min to remove dislodged
cells and the radioactivity in the supernatant was measured by
scintillation counting. Because arachidonic acid is poorly metabolized
in primary cultured mouse neurons (Oomagari et al., 1991), the
3H released was assumed to be essentially
[3H]arachidonic acid.
Neuronal viability. Experiments were performed using neurons
cultured in 24-well plates. The effect of AMPA on neuronal viability was assessed by a colorimetric assay based on the cleavage of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) into
a blue-colored formazan product by mitochondrial succinate dehydrogenase. Additions and agonist treatments were made directly to
the neuronal culture medium for 4 hr, as detailed in the legend to
Figure 8. Neurons were then washed twice with HBM, pH 7.4, and
incubated for 45 min at 37°C in HBM containing MTT (0.5 mg/ml). After
this period, the HBM was carefully removed, and the blue formazan
product was solubilized in 300 µl of 100% dimethyl sulfoxide. The absorbance of each well was read at 570 nm. Morphological assessment of the effect of AMPA on neuronal viability was made by
phase contrast microscopy as detailed in the legend to Figure 8.
Data handling. Immunoblot band intensity,
[3H]arachidonic acid release, neurotoxicity, and
intracellular Ca2+ data were analyzed by unpaired
two-tailed Student's t tests. Differences were considered
to be of statistical significance when p < 0.05.
| |
RESULTS |
The non-NMDA glutamate receptor desensitization inhibitor
cyclothiazide unmasks an AMPA-induced activation of MAPK in cultured
striatal neurons
The glutamate receptor agonist AMPA can evoke a range of
physiological and neurotoxic processes in striatal neurons (Lovinger and Tyler, 1996; Williams and Glowinski, 1996; Calabresi et al., 1998)
by binding to a heterogenous population of postsynaptic AMPA-type
glutamate receptors assembled from different combinations of
GluR1-GluR4 subunits (Bernard et al., 1997; Kwok et al., 1997). The
precise nature of the intracellular mechanisms activated subsequent to
AMPA receptor stimulation are not fully understood. Recently, much
attention has been focused on the possible physiological and
pathological roles that the MAPK cascade may play at glutamatergic synapses (Fukunaga and Miyamoto, 1998; Murray et al., 1998; Sgambato et
al., 1998). Indeed, it has been demonstrated in many neuronal types,
including striatal neurons, that stimulation of NMDA-type glutamate
receptors causes a Ca2+-dependent activation of MAPK
(Xia et al., 1996; Vincent et al., 1998). Although postsynaptic AMPA
receptors are classically associated with low Ca2+
permeability (Hollmann and Heinemann, 1994)
Ca2+-permeable AMPA receptors are present in the
striatum (Williams and Glowinski, 1996) and, thus, there may be a
population of striatal AMPA receptors that can couple to MAPK in a
Ca2+-dependent manner. However, a direct coupling of
Ca2+-permeable AMPA receptors to MAPK has not been
shown in striatal neurons.
To test whether stimulation of striatal AMPA receptors leads to
activation of the ERK1/ERK2 forms of MAPK, immunoblotting of neuronal
homogenates with an anti-ACTIVE MAPK pAb, which recognizes the dually
phosphorylated Thr/Glu/Tyr region within the catalytic core of the
active form of ERK1 (p44) and ERK2 (p42) was performed. Treatment of
striatal neurons with AMPA (50 µM) for 5 min failed to
activate MAPK (Fig. 1A,
lane 3) because there was no significant increase in
immunodetectable activated ERK1 and ERK2 (seen as two bands of ~44
and 42 kDa, respectively) above basal levels (Fig.
1A, lane 1). However, we have
previously shown that in cultured striatal neurons, AMPA
receptor-mediated responses are rapidly desensitizing (Williams and
Glowinski, 1996). To establish whether an apparent lack of activation
of MAPK by AMPA in these neurons is caused by rapid AMPA receptor
desensitization, striatal neurons were preincubated with the AMPA
receptor-specific desensitization inhibitor cyclothiazide (50 µM). In the presence of cyclothiazide, AMPA evoked a
large increase in the intensity of the two immunodetectable bands (44 and 42 kDa) compared with basal levels, corresponding to the activated
forms of MAPK, ERK1 and ERK2 (Fig. 1A, lane
4, B). Cyclothiazide alone did not activate ERK1
or ERK2 (Fig. 1A, lane 2),
indicating that it was acting specifically at AMPA receptors and was
not directly activating MAPK. A set of parallel immunoblots with an
antibody that detects total MAPK (active and inactive ERK1/ERK2) were
performed to show that the total levels of MAPK were not altered by any
of the drug treatments (Fig. 1A, lanes 5-8). Complementing our results with the
anti-ACTIVE MAPK pAb, AMPA/cyclothiazide treatment revealed a doublet
of 42 kDa immunoreactivity (seen as a faint slower-migrating band just
above the main 42 kDa band) (Fig. 1A, lane
8), indicative of an AMPA-mediated phosphorylation of p42
MAPK.
View larger version (39K):
[in this window]
[in a new window]
|
Figure 1.
Cyclothiazide unmasks an AMPA-mediated activation
of MAPK in striatal neurons. A, Crude homogenates (15 µg of each), prepared from striatal neurons preincubated for 5 min in
the absence (lanes 1,
3, 5, 7) or
presence (lanes 2, 4, 6,
8) of 50 µM cyclothiazide
(Cyz), and then exposed for 5 min to vehicle
(lanes 1, 5), 50 µM Cyz
(lanes 2, 6), 50 µM
AMPA (lanes 3, 7), or 50 µM AMPA and 50 µM Cyz (lanes
4, 8), were immunoblotted with an antibody that
specifically recognizes the dually phosphorylated Thr/Glu/Tyr region
within the catalytic core of the active form of ERK1 and ERK2
(anti-ACTIVE MAPK pAb) (lanes
1-4), or with an antibody that recognizes
total levels of ERK1 and ERK2 (anti-ERK1/ERK2 pAb) (lanes
5-8). B, Data obtained from
immunoblot experiments represented in lanes
1-4 was analyzed using BioImage Intelligent
Quantifier software. Each column is the mean ± SEM value of 6-12
independent cultures (n = 6-12).
**AMPA/Cyz was significantly different
from control (Con) (p < 0.001, unpaired two-tailed Student's t test).
|
|
Activation of MAPK by AMPA/cyclothiazide could also be demonstrated at
the single cell level. We first showed, by immunocytochemistry, the
presence of the ERK1 and ERK2 forms of MAPK in striatal neurons (see
Fig. 9Ai). The majority of the neurons were immunopositive for MAPK, with the highest levels of staining found at the neuronal soma, and some weaker staining in the neurites. After closer
examination of the cell body, a narrow outer ring of intense staining
was visible in the majority of neurons (see Fig. 9Ai,
white arrows), suggesting a cytosolic location for
MAPK. Some nuclear staining was also detectable, but to a lesser
extent. In a set of control experiments, omission of the primary
antibody completely abolished staining for MAPK (see Fig.
9Aii). Using the anti-ACTIVE MAPK pAb we showed that, under
control conditions, basal levels of activated MAPK are very low (see
Fig. 9Bi), which may indicate that MAPK activation is
tightly regulated in striatal neurons. However, single neurons
containing activated MAPK were visible after treatment of cultures with
AMPA in the presence of cyclothiazide (see Fig. 9Bii,
black arrow). When neurons were preincubated with the
2,3-benzodiazepine GYKI 53655 (100 µM), a noncompetitive
and selective antagonist of AMPA receptors, before application of AMPA/cyclothiazide, no staining for activated MAPK could be detected (see Fig. 9Biii). Thus, we have demonstrated, for the first
time, an AMPA receptor-mediated activation of MAPK (ERK1/ERK2) in
single isolated neurons.
Ca2+ influx through AMPA receptors triggers
activation of MAPK in striatal neurons
Activation of the MAPK cascade can be a
Ca2+-dependent process and in neurons, the
activation of ERK1/ERK2 by NMDA is completely dependent on the influx
of Ca2+ through the receptor (Kurino et al., 1995;
Xia et al., 1996), thus, the possibility that our observed stimulation
of MAPK by AMPA receptors also required Ca2+ was
tested. In Ca2+-containing incubation buffer,
glutamate (100 µM) (Fig.
2A, lane 2), AMPA/cyclothiazide (50 µM/50
µM) (Fig. 2A, lane 3),
and NMDA (100 µM) (data not shown) strongly activated
MAPK, as indicated by a large increase in the intensity of the 44 kDa
(ERK1) and 42 kDa (ERK2) immunodetectable bands, compared with basal
levels (Fig. 2A, lane 1). However,
in the absence of extracellular Ca2+, but the
presence of EGTA (200 µM) to chelate any residual
Ca2+, all three agonists [glutamate (Fig.
2A, lane 5), AMPA/cyclothiazide (Fig. 2A, lane 6), and NMDA
(data not shown)] failed to elicit an increase in the 44 kDa (ERK1)
and 42 kDa (ERK2) band intensities above basal levels (in the absence
of Ca2+) (Fig. 2A, lane
4). Further evidence that extracellular
Ca2+ influx is a trigger for MAPK activation in
striatal neurons came from the observation that the
Ca2+ ionophore ionomycin (5 µM), which
causes delocalized Ca2+ entry after randomly
intercalating at sites all over the neuronal plasma membrane, also
activated MAPK [(Fig. 2B, lane 3 (ionomycin); lane 1 (basal)]. It is noticeable
that ionomycin treatment did not appear to activate ERK1/ERK2 as
efficiently as glutamate (Fig. 2B, lane
2).
View larger version (26K):
[in this window]
[in a new window]
|
Figure 2.
Ca2+ dependence of AMPA
receptor-evoked MAPK activation in striatal neurons. A,
Crude homogenates (15 µg of each), prepared from striatal neurons
preincubated for 5 min in the absence (lanes 1,
2, 4, 5) or presence
(lanes 3, 6) of 50 µM cyclothiazide (Cyz), and then exposed for 5 min to vehicle (lanes 1, 4), 100 µM glutamate (lanes 2, 5),
or 50 µM AMPA and 50 µM Cyz (lanes
3, 6) in the presence (lanes
1-3) or absence (lanes
4-6) of CaCl2 (1 mM), were immunoblotted with an antibody that specifically
recognizes the dually phosphorylated Thr/Glu/Tyr region within the
catalytic core of the active form of ERK1 and ERK2 (anti-ACTIVE MAPK
pAb). In experiments in which CaCl2 was omitted from the
incubation medium, EGTA (200 µM) was added 1 min before
stimulation. B, Crude homogenates (15 µg of each),
prepared from striatal neurons exposed to vehicle (lane
1), 100 µM glutamate (lane 2), or
5 µM ionomycin (lane 3) for 5 min, were
immunoblotted with anti-ACTIVE MAPK pAb. C, Data
obtained from immunoblot experiments with AMPA/Cyz was analyzed using
BioImage Intelligent Quantifier software. Each column is the mean ± SEM value of six independent cultures (n = 6).
*AMPA/Cyz in the absence of CaCl2
(AMPA/Cyz-Ca2+)
was significantly different from AMPA/Cyz in the presence of
CaCl2 (AMPA/Cyz)
(p < 0.05, unpaired two-tailed Student's
t test).
|
|
To confirm that AMPA/cyclothiazide-evoked MAPK activation resulted from
the influx of Ca2+ through AMPA receptors,
intracellular Ca2+ measurements were performed using
fura-2-loaded neurons. AMPA caused a small rise in intracellular
Ca2+, which was strongly potentiated (approximately
threefold) in the presence of cyclothiazide (Fig.
3). This response was completely blocked
by preincubation with the selective AMPA receptor antagonist GYKI 53655 (100 µM) (Fig. 3, Table 1).
View larger version (19K):
[in this window]
[in a new window]
|
Figure 3.
AMPA receptor-evoked increase in intracellular
Ca2+ is enhanced by cyclothiazide and blocked by
GYKI 53655. Fura-2-loaded striatal neurons were exposed to 50 µM AMPA, 50 µM AMPA in the presence of 50 µM cyclothiazide
(AMPA/Cyz), or AMPA/Cyz in the presence
of 100 µM GYKI 53655 added 5 min before stimulation.
Fura-2 fluorescence (340 nm/380 nm ratio) was monitored as described in
Materials and Methods. Results are means ± SEM of responses from
the number of neurons (n) shown in
parentheses.
|
|
AMPA receptor activation of MAPK occurs through MEK and is direct,
with no involvement of NMDA receptors and/or L-type
voltage-sensitive Ca2+ channels
We have shown at the single cell level that GYKI 53655, a
selective antagonist at AMPA receptors can block
AMPA/cyclothiazide-induced MAPK activation (see Fig. 9B).
This finding was mirrored in large populations of cultured striatal
neurons because the robust activation of ERK1/ERK2 by
AMPA/cyclothiazide (Fig.
4A, lane
3) was completely blocked by preincubation with GYKI 53655 (100 µM) before AMPA/cyclothiazide treatment (Fig.
4A, lane 4, C). The
direct upstream activator of MAPK is MAPK kinase (MEK) (Seger and
Krebs, 1995). To establish whether or not striatal AMPA receptors
activate MAPK via the conventional pathway, i.e., phosphorylation of
MAPK by MEK, we performed experiments with PD 98059, a specific
inhibitor of the activation of MEK (Alessi et al., 1995). Preincubation
of neurons with PD 98059 (50 µM) completely blocked
AMPA/cyclothiazide-evoked activation of MAPK (ERK1/ERK2) (Fig.
4A, lane 5, C).
View larger version (26K):
[in this window]
[in a new window]
|
Figure 4.
AMPA/cyclothiazide activation of MAPK through MEK
is direct and independent of NMDA receptor and/or L-type
voltage-sensitive Ca2+ channel activation.
A, Crude homogenates (15 µg of each), prepared from
striatal neurons preincubated for 5 min in the absence (lanes
1, 2) or presence (lanes
3-5) of 50 µM cyclothiazide
(Cyz), and then exposed for 5 min to vehicle (lane
1), 50 µM AMPA (lane 2), 50 µM AMPA and 50 µM Cyz (lane
3), 50 µM AMPA and 50 µM Cyz in the
presence of 100 µM GYKI 53655 added 5 min before AMPA/Cyz
(lane 4), or 50 µM AMPA and 50 µM Cyz in the presence of 50 µM PD 98059 added 5 min before AMPA/Cyz (lane 5), were immunoblotted
with an antibody that specifically recognizes the dually phosphorylated
Thr/Glu/Tyr region within the catalytic core of the active form of ERK1
and ERK2 (anti-ACTIVE MAPK pAb). B, Crude homogenates
(15 µg of each), prepared from striatal neurons preincubated for 5 min in the absence (1) or presence (lanes
2-5) of 50 µM Cyz and then exposed
for 5 min to vehicle (lane 1), 50 µM AMPA
and 50 µM Cyz (lane 2), 50 µM AMPA and 50 µM Cyz in the presence of
100 µM GYKI 53655 added 5 min before AMPA/Cyz
(lane 3), 50 µM AMPA and 50 µM Cyz in the presence of 2 µM (+)-MK 801 added 5 min before AMPA/Cyz (lane 4), or 50 µM AMPA and 50 µM Cyz in the presence of 5 µM nimodipine added 5 min before AMPA/Cyz (lane
5), were immunoblotted with anti-ACTIVE MAPK pAb.
C, Data obtained from immunoblot experiments represented
in A (lanes 1-5) was
analyzed using BioImage Intelligent Quantifier software. Each column is
the mean ± SEM value of five to eight independent cultures
(n = 5-8). *AMPA/Cyz in the presence of GYKI 53655 or PD 98059 was significantly different from AMPA/Cyz
(p < 0.05, unpaired two-tailed Student's
t test).
|
|
Further pharmacological characterization of the AMPA receptor effect
was performed to establish whether AMPA receptors are directly coupled
to the MAPK cascade. Ca2+ entry through NMDA
receptors and/or L-type voltage-sensitive Ca2+
channels (VSCC) can activate the MAPK pathway (Rosen et al., 1994; Xia
et al., 1996), thus, the ability of AMPA/cyclothiazide to activate MAPK
in the presence of the NMDA receptor antagonist (+)-MK 801 (2 µM) or the L-type VSCC blocker nimodipine (5 µM) was examined. The activation of MAPK by
AMPA/cyclothiazide (Fig. 4B, lane 2)
was unaltered in the presence of either (+)-MK 801 (Fig.
4B, lane 4) or nimodipine
(Fig. 4B, lane 5), indicating that
AMPA receptor stimulation does not evoke a secondary activation of MAPK
from Ca2+ entry through NMDA receptors or L-type
VSCC. Elevation of the external [K+] in the
incubation buffer to 50 mM, which causes neuronal
depolarization and influx of Ca2+ through VSCC, did
not stimulate MAPK activity above basal levels (data not shown), thus,
providing further evidence that L-type VSCC cannot be involved in AMPA
receptor activation of MAPK in cultured striatal neurons. Activation of
neuronal MAPK may arise from the mobilization of intracellular
Ca2+ stores after stimulation of group I
metabotropic glutamate receptors (mGluRs). However, the possibility
that AMPA was acting as a ligand at group I mGluRs or that AMPA
receptor stimulation was indirectly activating mGluRs was ruled out
because, consistent with previous reports (Wang and Durkin, 1995;
Vincent et al., 1998), the group I mGluR agonist
(1S,3R)-ACPD (200 µM) failed to
activate MAPK (data not shown).
Regulation of AMPA receptor activation of MAPK by a Src family
protein tyrosine kinase and
Ca2+/calmodulin-dependent kinase II, but not protein
kinase C
It is known that MEK needs to be serine phosphorylated for its
activation, and a likely kinase for performing this role is Raf
(B-Raf or c-Raf-1) (Seger and Krebs, 1995). Furthermore, in non-neuronal cells at least, it is known that Raf kinase is itself phosphorylated and consequently activated by one or more upstream kinases, with PKC and the protein tyrosine kinase (PTK) family being
the most likely candidates (Seger and Krebs, 1995). The involvement of
PKC and/or protein tyrosine phosphorylation in AMPA receptor-evoked
activation of MAPK in striatal neurons was tested using kinase
inhibitors. Although NMDA receptor-induced activation of neuronal MAPK
has been shown to be partially prevented by PKC inhibitors (Kurino et
al., 1995; Vincent et al., 1998), preincubation of striatal neurons
with the selective PKC inhibitor Ro-31-8220 (5 µM)
before AMPA/cyclothiazide treatment did not alter the level of MAPK
activation [Fig. 5A,
lane 3 vs lane 2 (AMPA/Cyz),
C], arguing against an involvement of PKC in AMPA receptor
activation of MAPK.
View larger version (26K):
[in this window]
[in a new window]
|
Figure 5.
Involvement of a Src family protein tyrosine
kinase and CaM-KII, but not protein kinase C, in AMPA receptor-evoked
activation of MAPK. A, Crude homogenates (15 µg of
each), prepared from striatal neurons preincubated for 5 min in the
absence (lane 1) or presence (lanes
2-5) of 50 µM cyclothiazide
(Cyz) and then exposed for 5 min to vehicle (lane
1), 50 µM AMPA and 50 µM Cyz
(lane 2), 50 µM AMPA and 50 µM Cyz in the presence of 5 µM Ro-31-8220
added 5 min before AMPA/Cyz (lane 3), 50 µM AMPA and 50 µM Cyz in the presence of 5 µM KN-62 added 5 min before AMPA/Cyz (lane
4), or 50 µM AMPA and 50 µM
Cyz in the presence of 50 µM genistein added 5 min before
AMPA/Cyz (lane 5), were immunoblotted with an antibody
that specifically recognizes the dually phosphorylated Thr/Glu/Tyr
region within the catalytic core of the active form of ERK1 and ERK2
(anti-ACTIVE MAPK pAb). B, Crude homogenates (15 µg of
each), prepared from striatal neurons preincubated for 5 min in the
absence (lane 1) or presence (lanes
2-4) of 50 µM Cyz, and then
exposed for 5 min to vehicle (lane 1), 50 µM AMPA and 50 µM Cyz (lane
2), or 50 µM AMPA and 50 µM Cyz in
the presence of either 10 µM PP2 (lane 3)
or 10 µM PP3 (lane 4) both added 5 min before AMPA/Cyz, were immunoblotted with anti-ACTIVE MAPK pAb.
C, Data obtained from immunoblot experiments represented
in A (lanes 1-5) was
analyzed using BioImage Intelligent Quantifier software. Each column is
the mean ± SEM value of four to six independent cultures
(n = 4-6). *AMPA/Cyz in the presence of KN-62 or
genistein was significantly different from AMPA/Cyz
(p < 0.05, unpaired two-tailed Student's
t test).
|
|
Two families of nonreceptor PTKs, the Src family and the focal adhesion
kinase (FAK) family, contain members that are prominently expressed in
neurons (Boxall and Lancaster, 1998) and it is believed, although not
yet proven, that these Ca2+-dependent enzymes
subserve specific physiological roles in the CNS beyond mere
housekeeping functions. For example, members of the Src and FAK
families have been strongly implicated in controlling activation of the
Ras/MAPK signaling pathway (Lev et al., 1995; Finkbeiner and Greenberg,
1996; Hayashi et al., 1999), and it appears that protein tyrosine
phosphorylation may be important at several points within the MAPK
cascade. To test the involvement of tyrosine phosphorylation in AMPA
receptor-evoked activation of MAPK, neurons were pretreated with the
broad range tyrosine kinase inhibitor genistein (50 µM)
or the Src family-selective tyrosine kinase inhibitor PP2 (10 µM) (Hanke et al., 1996) before AMPA/cyclothiazide. No
activation of ERK1/ERK2 by AMPA/cyclothiazide was detected in the
presence of either genistein (Fig. 5A, lane 5, C) or PP2 [Fig. 5B, lane
3 vs lane 2 (AMPA/Cyz)], indicating that
stimulation of a Src family protein tyrosine kinase activity is
required for AMPA receptor coupling to MAPK. Preincubation of striatal
neurons with the pyrazolo[3,4-d]pyrimidine, PP3 (10 µM), a proposed negative control for PP2 (Traxler et al.,
1997), did cause some inhibition of AMPA receptor-evoked activation of ERK1/ERK2 (Fig. 5B, lane 4),
however, it was not as great as that caused by PP2.
Ca2+/calmodulin-dependent kinase II (CaM-KII), a
major protein present at the postsynaptic density region of neurons, is
activated following increases in intracellular Ca2+
(Hanson and Schulman, 1992), thus, we considered that CaM-KII may play
a role in our observed AMPA receptor activation of MAPK. Preincubation
of striatal neurons with the selective CaM-KII inhibitor KN-62 (5 µM) before AMPA/cyclothiazide caused a robust inhibition of AMPA receptor-evoked MAPK activation (Fig. 5A, lane
4, C), indicating that CaM-KII may play an
important regulatory role in the activation of MAPK by
Ca2+-permeable AMPA receptors.
An essential role for a pertussis toxin-sensitive G-protein and
phosphatidylinositol 3-kinase (PI 3-kinase) in AMPA receptor coupling
to MAPK
Application of AMPA to rat cerebral cortical neurons has been
shown to lead to a rapid increase in Ras activity and activation of
MAPK (Wang and Durkin, 1995). Ras-dependent activation of MAPK is
usually associated with seven transmembrane receptors that couple to
heterotrimeric G-proteins (Gutkind, 1998), however, it has been
demonstrated in rat cortical neurons that AMPA activates ERK2 (p42) by
causing a Ca2+-dependent association of G-protein
subunits, probably Gi, with a Ras, Raf
kinase, MEK complex (Wang and Durkin, 1995; Wang et al., 1997). This
novel involvement of a heterotrimeric G-protein in ionotropic AMPA
receptor signaling was examined. Striatal neurons were pretreated with
pertussis toxin (PTX) (1 µg/ml, made up in sterile PBS) or PBS
vehicle for 24 hr before experiments with AMPA/cyclothiazide. In
agreement with the results of Wang and Durkin (1995), PTX treatment
abolished AMPA receptor activation of MAPK [Fig.
6A, lane
3 vs lane 2 (AMPA/Cyz), C),
indicating a role for a Gi or Go-type G-protein
in the activation of MAPK by AMPA receptors in striatal neurons.
View larger version (23K):
[in this window]
[in a new window]
|
Figure 6.
Essential role for PI 3-kinase and a
Gi/Go-type G-protein in the activation
of MAPK by Ca2+-permeable AMPA receptors.
A, Crude homogenates (15 µg of each), prepared from
striatal neurons preincubated for 5 min in the absence (lane
1) or presence (lanes 2-4) of 50 µM cyclothiazide (Cyz) and then exposed for 5 min to vehicle (lane 1), 50 µM AMPA and 50 µM Cyz (lane 2), 50 µM AMPA
and 50 µM Cyz in the presence of 1 µg/ml pertussis
toxin (PTX) added 24 hr before AMPA/Cyz (lane 3), or 50 µM AMPA and 50 µM Cyz in the presence of
either 100 nM wortmannin (lane 4) or
50 µM LY 294002 (C) both added 5 min before AMPA/Cyz, were immunoblotted with an antibody that
specifically recognizes the dually phosphorylated Thr/Glu/Tyr region
within the catalytic core of the active form of ERK1 and ERK2
(anti-ACTIVE MAPK pAb). B, Crude homogenates (15 µg of
each), prepared from striatal neurons exposed for 5 min to vehicle
(lane 1), 100 µM KA (lane
2), 100 µM KA in the presence of 100 µM GYKI 53655 added 5 min before KA (lane
3), 100 µM KA in the presence of 1 µg/ml PTX
added 24 hr before KA (lane 4), or 100 µM KA in the presence of 100 nM wortmannin
added 5 min before KA (lane 5), were immunoblotted with
anti-ACTIVE MAPK pAb. C, Data obtained from immunoblot
experiments represented in A (lanes
1-4) was analyzed using BioImage
Intelligent Quantifier software. Each column is the mean ± SEM
value of five to six independent cultures (n = 5-6).
*AMPA/Cyz in the presence of PTX or wortmannin was significantly
different from AMPA/Cyz (p < 0.05, unpaired
two-tailed Student's t test) and **AMPA/Cyz in the
presence of LY 294002 was significantly different from AMPA/Cyz
(p < 0.001, unpaired two-tailed Student's
t test).
|
|
More recently, additional links between G subunits and the
Ras-MAPK cascade have been identified. Wortmannin, a selective (low
nanomolar range) inhibitor of PI 3-kinase has been reported to block
Gi-coupled receptor-induced MAPK activation (Hawes et al.,
1996), and experiments have indicated that PI 3-kinase is recruited by
free G subunits and coordinates the activation of Src-like
nonreceptor PTKs involved in activating the Shc-Grb2-Sos complex
(Lopez-Ilasaca et al., 1997; Gutkind, 1998). Furthermore, it has been
shown that PI 3-kinase activity can be regulated by Ca2+/calmodulin (Joyal et al., 1997) and therefore,
the possibility that PI 3-kinase is required for AMPA receptor-mediated
MAPK signaling was examined. Pretreatment of striatal neurons with
wortmannin (100 nM) (Fig. 6A, lane
4, C) or the highly selective PI 3-kinase inhibitor LY 294002 (50 µM) (Fig. 6C)
completely blocked the activation of ERK1/ERK2 MAPK by
AMPA/cyclothiazide. This is the first report that AMPA receptor
coupling to MAPK is dependent on PI 3-kinase.
Complementing the AMPA/cyclothiazide data, in a set of additional
experiments we showed that the nondesensitizing AMPA receptor agonist
KA (100 µM) also activated MAPK, exclusively through AMPA receptors (no involvement of kainate receptors was indicated by a
complete sensitivity of the KA effect to GYKI 53655), in a PTX- and
wortmannin-sensitive manner (Fig. 6B).
We considered it possible that the kinase inhibitors KN-62, PD 98059, genistein and wortmannin blocked AMPA/cyclothiazide-evoked MAPK
activation by inhibiting the entry of Ca2+ through
AMPA receptors. To address this, intracellular Ca2+
measurements were made using fura-2-loaded neurons.
AMPA/cyclothiazide-evoked increases in cytosolic
Ca2+ were not altered by any of the kinase
inhibitors tested (Table 1), and, thus, it seems likely that CaM-KII,
MEK, tyrosine kinase, and PI 3-kinase activities all have sites of
action at points downstream of Ca2+ influx.
Is cytosolic phospholipase A2
(cPLA2) a putative downstream target for ERK1/ERK2
MAP kinases in striatal neurons?
We have previously shown that stimulation of
Ca2+-permeable cyclothiazide-sensitive AMPA
receptors leads to a release of arachidonic acid (AA) from striatal
neurons (Williams and Glowinski, 1996). This release can be prevented
by removing extracellular Ca2+ or by preincubation
with the selective cPLA2 inhibitor AACOCF3 before AMPA receptor stimulation (M. S. Perkinton and R. J. Williams, unpublished observations), suggesting that striatal AMPA
receptors mobilize AA through the activation and
Ca2+-dependent translocation of cPLA2.
In non-neuronal cells, MAP kinases phosphorylate cPLA2 on
Ser505 and/or Ser727, and it has
been proposed that this may be crucial for agonist-induced activation
of cPLA2 (Lin et al., 1993; Borsch-Haubold et al., 1998).
Thus, it is possible that one of the functional consequences of AMPA
receptor-mediated activation of MAPK in striatal neurons is the
phosphorylation and activation of cPLA2 leading to AA
mobilization. To examine this possibility, AMPA receptor-evoked
[3H]AA release from striatal neurons was assayed
in the absence or presence of inhibitors of MAPK activity. In control
conditions, treatment of neurons with AMPA/cyclothiazide (50 µM/50 µM) produced a 175 ± 8.5%
release of [3H]AA above basal levels (Fig.
7). Preincubation of neurons with the MEK
inhibitor PD 98059 (50 µM) or the PI 3-kinase inhibitor wortmannin (100 nM), both of which completely inhibit AMPA
receptor-evoked activation of the ERK1 and ERK2 forms of MAPK (Figs.
4A, 6A, respectively), had no
significant effect on AMPA receptor-evoked [3H]AA
release (Fig. 7; AMPA/cyclothiazide + wortmannin, 195 ± 20% of
paired basal; AMPA/cyclothiazide + PD 98059, 196 ± 6% of paired basal). These data strongly suggest that ERK1/ERK2 do not regulate agonist-evoked AA mobilization in cultured striatal neurons, although it should be noted that PD 98059 increased basal
[3H]AA levels by 30-40% through an as yet
unknown mechanism (Fig. 7).
View larger version (26K):
[in this window]
[in a new window]
|
Figure 7.
AMPA receptor-evoked mobilization of arachidonic
acid in striatal neurons is independent of the MAPK cascade. Striatal
neurons were preincubated for 10 min in HEPES-buffered medium
containing fatty acid-free BSA (2 mg/ml; HBM/BSA) in the presence or
absence of enzyme inhibitors (100 nM wortmannin or 50 µM PD 98059), with or without 50 µM
cyclothiazide (Cyz) and then incubated for 15 min in HBM/BSA
containing either vehicle (Basal), or 50 µM AMPA and 50 µM Cyz
(AMPA/Cyz) in the continued presence or
absence of enzyme inhibitors as shown. Both AMPA and Cyz alone failed
to release [3H]AA (data not shown). Results are
expressed as the percent release of [3H]AA
relative to basal values, which was defined as 100%. Data are the
mean ± SEM values of three to five independent cultures
(n = 3-5), each performed in triplicate.
|
|
Inhibition of PI 3-kinase or MAPK activity does not modulate
chronic AMPA receptor-evoked neurotoxicity in striatal neurons
Activation of MAPK may be important during oxidative and
excitotoxic stimuli (Guyton et al., 1996; Murray et al., 1998; Samanta et al., 1998), thus, we considered it possible that our observed AMPA
receptor-evoked activation of MAPK through PI 3-kinase could represent
an early neurotoxic or survival signal that might be significant during
AMPA/cyclothiazide-induced excitotoxicity. Application of AMPA (50 µM) to striatal neurons for 4 hr produced no visible
morphological effects [(Fig.
8Aii vs Ai
(control)], and the capacity of striatal neurons to metabolize the
tetrazolium salt MTT into formazan using mitochondrial succinate
dehydrogenase, a marker for neuronal viability, was unaltered (Fig.
8B). However, in the presence of cyclothiazide (50 µM; which alone had no deleterious effects), AMPA induced
profound morphological changes to striatal neurons, including a loss of
neurites and cell body integrity (Fig. 8Aiii).
Furthermore, the ability of neurons to metabolize MTT into formazan was
significantly reduced (Fig. 8B). Both of these
effects were completely blocked by preincubation with the AMPA receptor
antagonist GYKI 53655 (100 µM) (Fig.
8Aiv; data not shown), demonstrating that toxicity
was mediated through cyclothiazide-sensitive AMPA receptors as
previously proposed (May and Robison, 1993; Cebers et al., 1997).
Chronic exposure (4 hr) of striatal neurons to either the MEK inhibitor
PD 98059 (50 µM) or the PI 3-kinase inhibitor wortmannin
(100 nM) had no effect on neuronal morphology (data not
shown) or MTT turnover (Fig. 8B) and neither
accelerated nor attenuated the neurotoxic actions of AMPA/cyclothiazide
(Fig. 8B). These
data suggest that these kinases do not play a role in neurotoxicity
evoked by chronic AMPA receptor stimulation in striatal neurons.
View larger version (61K):
[in this window]
[in a new window]
|
Figure 8.
Onset of Ca2+-permeable AMPA
receptor-induced neurotoxicity is neither reversed nor accelerated by
inhibition of MAPK/PI 3-kinase activity. A, Striatal
neurons were exposed to vehicle (i), 50 µM AMPA (ii), 50 µM AMPA in
the presence of 50 µM cyclothiazide (Cyz)
(iii), or 50 µM AMPA and 50 µM Cyz in the presence of 100 µM GYKI 53655 added 5 min before AMPA/Cyz (iv) for 4 hr and then
photographed under phase contrast microscopy (320× magnification).
B, Striatal neurons were exposed to vehicle or AMPA,
with or without Cyz (as described above), in the presence or absence of
50 µM PD 98059 or 100 nM wortmannin
(Wort) which were added 5 min before treatments. After 4 hr exposure, the culture medium was removed and MTT added as described
in Materials and Methods. Results are expressed as the percent of
formazan production relative to control, which was defined as 100%.
Each column is the mean ± SEM value of four independent cultures
(n = 4), each performed in triplicate. *AMPA/Cyz
was significantly different from AMPA (p < 0.05, unpaired two-tailed Student's t test).
|
|
View larger version (59K):
[in this window]
[in a new window]
|
Figure 9.
Localization of MAPK and activation by AMPA
receptor stimulation in cultured striatal neurons. A,
Striatal neurons were immunostained with an antibody to the MAPK
enzymes ERK1 and ERK2 (anti-ERK1/ERK2 pAb) (i),
or without primary antibody (ii), as described in
Materials and Methods. White arrows indicate intense
cytosolic staining of ERK1 and ERK2. B, Striatal neurons
were either unstimulated (i) or, after incubation
with 50 µM cyclothiazide for 5 min, stimulated with 50 µM AMPA (in the presence of 50 µM
cyclothiazide) for 5 min in the absence (ii) or presence
(iii) of 100 µM GYKI 53655 added 5 min
before stimulation. After stimulation, neurons were immunostained with
an antibody that specifically recognizes the dually phosphorylated
Thr/Glu/Tyr region within the catalytic core of the active form of ERK1
and ERK2 (anti-ACTIVE MAPK pAb) as described in Materials and Methods.
Black arrow indicates neurons that are immunopositive
for anti-ACTIVE MAPK pAb.
Figure 10.
Involvement of the MAPK cascade in AMPA
receptor-induced phosphorylation of the nuclear transcription factor
CREB in cultured striatal neurons. A, Crude homogenates
(15 µg of each), prepared from striatal neurons preincubated for 5 min in the absence (lane 1) or presence (lanes
2-5) of 50 µM cyclothiazide
(Cyz) and then exposed for 5 min to vehicle (lane
1), 50 µM AMPA and 50 µM Cyz
(lane 2), 50 µM AMPA and 50 µM Cyz in the absence of CaCl2 (1 mM) (lane 3), or 50 µM AMPA
and 50 µM Cyz in the presence of either 50 µM PD 98059 (lane 4) or 100 nM wortmannin (lane 5) both added 5 min
before stimulation, were immunoblotted with an antibody that detects
CREB when phosphorylated at Ser133
(anti-phospho-Ser133 CREB pAb). In experiments in
which CaCl2 was omitted from the incubation medium, EGTA
(200 µM) was added 1 min before stimulation.
B, Striatal neurons were either unstimulated
(i) or, after incubation with 50 µM
cyclothiazide for 5 min, treated with 50 µM AMPA (in the
presence of 50 µM cyclothiazide) for 5 min in the absence
(ii) or presence (iii) of 100 µM GYKI 53655 added 5 min before stimulation. Neurons
were immunostained with an anti-phospho-Ser133 CREB
pAb as described in Materials and Methods.
|
|
AMPA receptors induce phosphorylation of the nuclear transcription
factor CREB through a MAPK-dependent mechanism in striatal neurons
It has been reported that glutamate can induce phosphorylation and
activation of the nuclear transcription factor CREB on the regulatory
site Ser133 in different brain regions (Ginty et
al., 1993; Xia et al., 1996; Obrietan et al., 1998; Sgambato et al.,
1998; Vanhoutte et al., 1999), and it is thought that CREB activation
may be an important step in processes underlying long-term memory
(Frank and Greenberg, 1994; Deisseroth et al., 1996; Ginty, 1997).
Although NMDA-type glutamate receptors have been implicated in
Ca2+-dependent activation of CREB (Ginty et al.,
1993; Xia et al., 1996; Deisseroth et al., 1996), it is not known
whether CREB is also a downstream target for AMPA receptor-activated
signaling pathways. We tested whether activation of
Ca2+-permeable AMPA receptors leads to stimulation
of CREB in cultured striatal neurons, using an antibody that detects
CREB when it is phosphorylated on the regulatory site
Ser133 (anti-pCREB). Immunoblotting of neuronal
homogenates, containing the nuclear fraction, and single cell
immunocytochemistry with anti-pCREB showed that AMPA/cyclothiazide (50 µM/50 µM) induced a robust phosphorylation
of CREB on Ser133 in the nucleus of striatal neurons
(Fig. 10A, lane 2, Bii).
This effect was entirely dependent on the presence of extracellular Ca2+ (Fig. 10A, lane
3), was blocked by the AMPA receptor antagonist GYKI 53655 (100 µM) (Fig. 10Biii), and was
strongly inhibited by PD 98059 (50 µM) and wortmannin
(100 nM) (Fig. 10A, lanes
4 and 5, respectively), indicating that
stimulation of CREB by AMPA receptors is dependent on a PI
3-kinase-sensitive activation of the MAPK cascade.
| |
DISCUSSION |
There is some discrepancy in the literature with regard to the
ability of AMPA receptors to activate the MAPK cascade. It has been
reported that AMPA receptors can couple to the MAPK pathway in cortical
neurons (Wang and Durkin, 1995), whereas in hippocampal neurons AMPA
and kainate do not appear to activate MAPK (Kurino et al., 1995). In
the latter study, a lack of AMPA effects could be caused by either
rapid receptor desensitization or reduced Ca2+
permeability resulting from the presence of Q/R-edited GluR2 subunits
in the native receptor complexes: Ca2+ influx
through glutamate receptor channels is thought to be the trigger for
MAPK activation (Xia et al., 1996). Although it has been reported that
AMPA and NMDA induce comparable increases in MAPK activity in cultured
striatal neurons (Vincent et al., 1998), a direct
Ca2+-permeable AMPA receptor coupling to MAPK cannot
be concluded because of a lack of characterization of the AMPA response
along with no antagonist data. Furthermore, this increase in MAPK
activity with AMPA is somewhat surprising, as it has previously been
shown that AMPA evokes either very weak or no functional responses in striatal neurons (Tence et al., 1995; Williams and Glowinski, 1996)
unless rapid receptor desensitization is blocked with cyclothiazide (Petitet et al., 1995; Williams and Glowinski, 1996). Similarly, in
this study we show that AMPA alone does not significantly activate MAPK. However, a robust AMPA receptor-evoked activation of ERK1 and
ERK2 was unmasked with cyclothiazide. Intracellular
Ca2+ measurements revealed that AMPA produced only a
small rise in cytosolic Ca2+, which may explain our
observed lack of activation of MAPK with AMPA alone. By contrast, in
the presence of cyclothiazide, a large threefold increase in
AMPA-evoked rises in cytosolic Ca2+ was observed,
and this enhanced response probably underlies the unmasking of AMPA
receptor-evoked MAPK stimulation. Further pharmacological characterization of AMPA/cyclothiazide-evoked stimulation of ERK1/ERK2 demonstrated that activation resulted from the direct influx of Ca2+ through the AMPA receptor itself with no
secondary facilitatory involvement of either NMDA receptors or L-type
voltage-sensitive Ca2+ channels.
AMPA receptors appear to activate the MAPK cascade in cortical neurons
through a novel mechanism involving the subunits of
PTX-sensitive G-proteins (Wang and Durkin, 1995). Our finding that AMPA
receptor-evoked activation of MAPK in striatal neurons was sensitive to
PTX supports the concept that ionotropic receptors can transduce
signals through Gi/Go-type G-proteins
(Wang et al., 1997; Rodriguez-Moreno and Lerma, 1998). Recently,
seven-transmembrane receptors coupled to inhibitory G-proteins have
been shown to signal to MAPK in non-neuronal cells through a mechanism
requiring the recruitment of PI 3-kinase (Hawes et al., 1996;
Lopez-Ilasaca et al., 1997; Gutkind, 1998). Interestingly, distinct
from the well established phosphotyrosine-dependent activation of PI
3-kinase, rises in cytosolic Ca2+ can activate PI
3-kinase in chinese hamster ovary cells by inducing a direct
association of calmodulin with both SH2 domains of the regulatory
subunit (p85) of the kinase (Joyal et al., 1997). We, therefore,
considered that PI 3-kinase could be an important intermediate in our
observed Gi/Go- and
Ca2+-dependent activation of MAPK by AMPA receptors.
Indeed, AMPA receptor activation of MAPK was completely inhibited by
the PI 3-kinase inhibitors wortmannin and LY 294002, indicating an
essential role for PI 3-kinase in AMPA receptor-stimulated ERK1 and
ERK2 activation in striatal neurons. This is the first report of an involvement of PI 3-kinase in the regulation of postsynaptic AMPA receptor signaling in the CNS. MAPK activation was not modified by the
PKC inhibitor Ro-31-8220, which is consistent with previous reports in
non-neuronal cells that signaling to MAPK via G subunits and PI-3
kinase is independent of PKC (Luttrell et al., 1997). Thus, although
the specific protein-protein interactions that lead to activation of
the Ras-MAPK pathway by AMPA receptors are not currently known, it
seems reasonable to propose that AMPA receptor-evoked rises in
cytosolic Ca2+ may trigger activation of PI 3-kinase
and then recruitment of the lipid kinase to the MAPK cascade may, as is
the case with seven-transmembrane
Gi/Go-type G-protein linked receptors,
be orchestrated by free G subunits. The specific exchange factors regulating Ras activity after AMPA receptor stimulation also remain to
be determined. An involvement of the neuron-specific guanine nucleotide
exchange factor, Ras-GRF, seems plausible because it has recently been
demonstrated that Ras-GRF can be activated in response to increases in
intracellular Ca2+ (Farnsworth et al., 1995;
Finkbeiner and Greenberg, 1996) and/or free G-protein subunits
that induce phosphorylation of Ras-GRF by as yet unknown kinases
(Mattingly and Macara, 1996).
However, Ca2+/calmodulin-dependent activation of
Ras-GRF does not appear to involve PTKs (Farnsworth et al., 1995),
thus, our results indicating that tyrosine phosphorylation may be an
important step in AMPA receptor activation of MAP kinase suggests that
additional Ca2+-dependent routes to Ras may be
activated. It has been proposed that PI 3-kinase can coordinate the
activation of Src-like nonreceptor PTKs (Lopez-Ilasaca et al., 1997;
Gutkind, 1998). Furthermore, the Src and FAK families of PTKs have been
strongly implicated in controlling Ca2+-dependent
activation of the Ras/MAPK signaling pathway, however, it is not clear
whether members of these two families act independently or together to
coordinate association of the Grb2-Sos complex, which leads ultimately
to the sequential activation of Ras, Raf kinase, and MEK (Lev et al.,
1995; Finkbeiner and Greenberg, 1996). We found that AMPA
receptor-evoked activation of MAPK was completely blocked in the
presence of the tyrosine kinase inhibitors genistein and PP2. It has
recently been reported that PP2 selectively inhibits tyrosine kinases
of the Src family, namely Src, Fyn, and Lck (all of which are present
in the brain), while having negligible inhibitory activity against Jak2
or Zap PTKs (Hanke et al., 1996). Our results with PP2 agree with
previous work demonstrating that Src-like protein tyrosine kinases
regulate Ca2+-dependent activation of the MAPK
cascade (Rusanescu et al., 1995; Finkbeiner and Greenberg, 1996) and
synaptic plasticity involving AMPA receptor activation (Boxall and
Lancaster, 1998; Hayashi et al., 1999). Whether or not PYK2, a member
of the FAK family that has been shown to be important in
Ca2+-dependent activation of the MAPK cascade (Lev
et al., 1995), is involved in neuronal AMPA receptor signaling remains
to be determined.
CaM-KII is enriched at the postsynaptic density of glutamatergic
synapses where one of its major roles is thought to be in the
regulation of glutamate receptor responses (Fukunaga et al., 1992). It
has been shown that CaM-KII can phosphorylate AMPA receptor subunits
(Mammen et al., 1997), resulting in enhanced receptor currents
(McGlade-McCulloh et al., 1993; Tan et al., 1994), and this has been
implicated in the strengthening of postsynaptic responses associated
with synap
TOP
ABSTRACT
INTRODUCTION
