Journal of Neurochemistry, 2001, 77, 1001±1009
Extracellular ATP stimulates an inhibitory pathway towards
growth factor-induced cRaf-1 and MEKK activation in astrocyte
cultures
Guido Lenz,* Daniela GoncËalves,² Zhijun Luo,³ Joseph Avruch,³ Richard Rodnight² and
Joseph T. Neary§
Departments of *BiofõÂsica and ²BioquõÂmica, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil
³Diabetes Unit and Department of Molecular Biology, Massachusetts General Hospital, and Department of. Medicine, Harvard
Medical School, Boston, Massachusetts, USA
§Research Service, VA Medical Center and Departments of Pathology and Biochemistry & Molecular Biology, University of Miami
School of Medicine, Miami, Florida, USA
Abstract
ATP, acting via P2Y, G protein-coupled receptors (GPCRs), is
a mitogenic signal and also synergistically enhances ®broblast
growth factor-2 (FGF-2)-induced proliferation in astrocytes.
Here, we have examined the effects of ATP and FGF-2
cotreatment on the main components of the extracellularsignal regulated protein kinase (ERK) cascade, cRaf-1,
MAPK/ERK kinase (MEK) and ERK, key regulators of cellular
proliferation. Surprisingly, ATP inhibited activation of cRaf-1
by FGF-2 in primary cultures of rat cortical astrocytes. The
inhibitory effect did not diminish MEK and ERK activation;
indeed, cotreatment resulted in a greater initial activation of
ERK. ATP inhibition of cRaf-1 activation was not mediated by
an increase in cyclic AMP levels or by protein kinase C
activation. ATP also inhibited the activation of cRaf-1 by other
growth factors, epidermal growth factor and platelet-derived
growth factor, as well as other MEK1 activators stimulated by
FGF-2, MEK kinase 1 (MEKK1) and MEKK2. Serotonin, an
agonist of another GPCR coupled to ERK, did not inhibit
FGF-2-induced cRaf-1 activation, thereby indicating speci®city
in the ATP-induced inhibitory cross-talk. These ®ndings
suggest that ATP stimulates an inhibitory activity that lays
upstream of MEK activators and inhibits growth factor-induced
activation of cRaf-1 and MEKKs. Such a mechanism might
serve to integrate the actions of receptor tyrosine kinases and
P2Y-GPCRs.
Keywords: ATP, cross-talk, ERK, FGF, MAP kinase, Raf.
J. Neurochem. (2001) 77, 1001±1009.
Signal transduction pathways have been generally treated as
a linear sequence of events, from a receptor, through a
second messenger, to an effector, resulting in a speci®c set
of effects on the cell. With the discovery of new pathways
and regulatory mechanisms, it became clear that these linear
pathways actually were part of large and extremely complex
systems, with signaling capabilities much beyond the simple
linear transference of information from the plasma membrane to cellular components (Bhalla and Iyengar 1999).
One of the most important characteristics of these complex
signaling systems is the numerous points of cross-talk and
feedback, providing them with multiple and dynamic
features for signal integration.
Mitogen-activated protein kinase (MAPK) cascades are
examples of complex signaling systems. These cascades
are comprised of at least three protein kinases. In the case of
the extracellular signal regulated protein kinase (ERK)
cascade, Raf phosphorylates MEK1/2 which in turn
phosphorylates ERK1/2; these key enzymes are involved
Received October 26, 2000; revised manuscript received February 1,
2001; accepted February 6, 2001.
Address correspondence and reprint requests to J. T. Neary, Research
Service 151, VA Medical Center, 1201 NW 16th Street, Miami, FL
33125, USA. E-mail: [email protected]
Abbreviations used: AEBSF, 4-(2-aminoethyl)benzenesulfonyl¯uride; EGF, epidermal growth factor; ERK, extracellular signal regulated
protein kinase; FGF: ®broblast growth factor; GPCR, G protein-coupled
receptor; GST, glutathione S transferase; MAPK, mitogen-activated
protein kinase; MBP, myelin basic protein; MEK, MAPK/ERK kinase;
MEKK, MEK kinase; PDGF, platelet-derived growth factor; PKA,
cAMP-dependent protein kinase; PKC, protein kinase C; PKG,
cGMP-dependent protein kinase; RTK, receptor tyrosine kinase.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
1001
1002 G. Lenz et al.
in cellular proliferation and differentiation (reviewed in
Seger and Krebs 1995). MEK can also be activated, among
others, by the MEK kinases (MEKKs), subtypes MEKK 1, 2
and 3 (Schlessinger et al. 1998). Besides the classical
activation mechanism of the growth factor receptors, the
ERK cascade can be activated by G protein-coupled
receptors (GPCRs) and can be modulated by representative
components of several signal transduction pathways such as
PKC, PKA, PKG, Akt, G protein subunits and Ca21 (Wu
et al. 1993; Farnsworth et al. 1995; Mattingly and Macara
1996; SchoÈnwasser et al. 1998; Suhasini et al. 1998;
Zimmermann and Moelling 1999).
Extracellular ATP is an important neurotransmitter that
interacts with both neurons and astrocytes through a variety
of purinoceptors (Burnstock 1997). ATP is released upon
electrical stimulation of hippocampal slices (Wieraszko et al.
1989) and is co-released from nerve endings with norepinephrine or acetylcholine (reviewed in Burnstock 1990).
Considering the high intracellular concentration of ATP, this
molecule is also an important signal resulting from cell
damage in several types of brain injury, potentially stimulating responses in astrocytes, such as reactive gliosis
(reviewed in Neary et al. 1996). P2Y purinergic receptors
are expressed in glial cells in vivo (Deng et al. 1998; Webb
et al. 1998) and activation of P2Y receptors via microinfusion of an ATP analog, 2-methylthioATP, into rat brain
led to increased glial ®brillary acidic protein immunoreactivity, astrocyte hypertrophy and proliferation, key
features of gliosis (Franke et al. 1999). In astrocytes, the
ERK cascade is fundamental for the proliferation induced by
ATP (Neary et al. 1999a). ATP, acting on P2Y receptors
that are members of the GPCR superfamily, activates ERK
by a pathway that is dependent on PLD and PKCd but
independent of cRaf-1, using an as yet unknown MEK
activator (Neary et al. 1999a; Lenz et al. 2000).
Signaling mediated by ®broblast growth factor (FGF) is
involved in several physiological and pathological processes. Various observations indicate that FGF plays an
important role in lesions of the central nervous system. For
example, FGF-2 is increased after brain injury (Finklestein
et al. 1988) and also potentiates the induction of reactive
gliosis (Eclancher et al. 1996). Since ATP and FGF-2 are
involved in reactive gliosis, the presence of one of them
alone, or of both together, could indicate different
physiopathological conditions. Indeed, synergistic interactions between ATP and FGF have been reported. For
example, ATP and FGF-2 induced incorporation of
[3H]thymidine into astrocytes twofold and 14-fold, respectively, but when ATP and FGF-2 were added concurrently, a
52-fold increase was observed (Neary et al. 1994). This
indicated the existence of interactions between the signal
transduction pathways activated by ATP and FGF-2 that
couple the receptors to the cell cycle regulatory machinery.
In the present work we have focused on the interactions
between the ATP and FGF-2 signal transduction pathways
that are linked to the ERK cascade. This interaction was
more clearly seen at the level of cRaf-1, the activation of
which by FGF-2 was inhibited by treatment with ATP, an
effect probably mediated by a purinoceptor similar to that
involved in ERK activation, without involving cAMP and
PKC. The use of other growth factors, the assay of other
MEK1 activators and kinetic evidence suggested that the
inhibition occurred between the growth factor receptor and
cRaf-1. The possible involvement of this cross-talk in the
synergism between these two signaling molecules observed
in proliferation is discussed.
Materials and methods
Reagents
All reagents were reagent grade or better. Anti-MEKK1 (sc-252),
anti-MEKK2 which also recognizes MEKK3 (sc-1089), and protein
A/G-Agarose (sc-2003) were from Santa Cruz Biotechnology
(Santa Cruz, CA, USA). Anti-cRaf-1 was produced as described
(Kyriakis et al. 1992). MEK-1 and ERK-1 polypeptides were
bacterially expressed as GST fusion proteins; after puri®cation by
glutathione-sepharose, free inactive ERK-1 was obtained by
thrombin cleavage (Luo et al. 1996). A polyhistidine-tagged
ERK-2 (sc-4024; Santa Cruz Biotechnology) was used in some
experiments with similar results. ATP, myelin basic protein (MBP),
EGF, phorbol 12-myristate 13-acetate (PMA), aprotinin, leupeptin
and pepstatin were obtained from Sigma (St Louis, MO, USA).
[g-32P]ATP was purchased from DuPont±NEN (Boston, MA,
USA) and Amersham (Amersham Pharmacia Biotech, Piscataway,
NJ, USA). Antibodies that recognize dually phosphorylated
ERK1/2 and total ERK1/2 were obtained from Promega (Madison,
WI, USA) and Santa Cruz Biotechnology, respectively. Human
recombinant FGF-2 was purchased from R & D Systems
(Minneapolis, MN, USA). PDGF was from Amgen Inc. (Thousand
Oaks, CA, USA). 4-(2-Aminoethyl) benzenesulfonyl¯uoride
(AEBSF), GF109203X and Ro 31±8220 were obtained from
Calbiochem (San Diego, CA, USA).
Cell culture and treatment
Primary astrocytes were obtained from neonatal rat cerebral
cortices as previously described (Neary et al. 1994). At least 99%
of the cell population were astrocytes, as determined by staining
with cell-speci®c markers (Neary et al. 1994). Experiments were
conducted with 3±5-week-old cultures. Prior to treatment with
nucleotides or growth factors, cells that had been maintained in
Dulbecco's modi®ed Eagle's medium (DMEM) containing 10%
horse serum were shifted to the quiescent phase by incubation in
DMEM containing 0.5% horse serum for 48 h. Stock solutions of
nucleotides and polypeptide growth factors were divided into
single-use aliquots and stored at 2 808C. In the case of cotreatment
with ATP and FGF-2, ATP was added immediately (about 2 s)
after the addition of FGF-2.
MAPKKK, MEK and ERK assays
Activation of cRaf-1 and other MAKKKs was measured by means
of a coupled assay in which inactive GST-MEK1 and inactive
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
Purinergic inhibition of MEK activators 1003
ERK1 were added to immunoprecipitates of cRaf-1 or other
MAPKKKs, as previously described (Kyriakis et al. 1993; Luo et al.
1996) with minor modi®cations (Lenz et al. 2000). MEK and ERK
activities were based on the phosphorylation of ERK1 and MBP,
respectively, as described (Lenz et al. 2000). It is important to note
that, due to differences in assay sensitivities, the amplitude of
activation of ERK, MEK and cRaf-1 cannot be directly compared.
Phosphorylated and total ERK1/2 were measured by immunoblotting as described (Neary et al. 1999a).
cAMP determination
cAMP was assayed using the binding of [3H]cAMP to protein
kinase A as described (Lenz et al. 2000).
Statistical analyses
Data were analyzed by Student's t-tests for two groups or anova
followed by post hoc comparisons for multiple groups with an
Instat software package (GraphPad Software Inc., San Diego, CA,
USA). Dose±response curves were analyzed with GraphPad
PrismR.
Results
Extracellular ATP inhibits the FGF-2 induced activation
of cRaf-1
Both ATP and FGF-2 activate the ERK/MAPK pathway and
increase the incorporation of [3H]thymidine into primary
astrocyte cultures (Neary and Zhu 1994; Neary et al. 1994).
Interestingly, both molecules synergistically increase the
proliferation of these cells (Neary et al. 1994). Since the
ERK pathway is fundamental for the proliferation induced
by ATP (Neary et al. 1999a), we decided to investigate
whether these two mitogens interact in their signal
transduction mechanisms, mainly at the ERK cascade,
which could be involved in the above mentioned synergism.
As demonstrated previously (Lenz et al. 2000), FGF-2, but
not ATP, activates cRaf-1 in astrocytes. Surprisingly, 5 min
after the concurrent addition of ATP and FGF-2, the
activation of cRaf-1 induced by FGF-2 was reduced by
ATP (Fig. 1). Although Raf activation by FGF-2 was
inhibited by ATP, at this time point the activation of MEK
and ERK in cells cotreated with ATP and FGF-2 was similar
to that in cells treated with FGF-2 alone (Fig. 1); this may be
due at least partly to the ability of ATP to activate MEK and
ERK independently of FGF-2 signaling (Neary and Zhu
1994; Lenz et al. 2000). The relatively low stimulation of
ERK (about twofold, see Fig. 1) was probably due to the
low sensitivity of the substrate used in the ERK activity
assay, myelin basic protein, because in immunoblots probed
with antibodies for phosphorylated ERK1/2, the extent of
activation of both ERK1 and ERK2 by either ATP or FGF-2
alone was much greater (see Fig. 3 in Neary et al. 1999a;
and Fig. 1 in Lenz et al. 2000). Consistent with these
immunoblot results, we have found an 8±10-fold activation
of ERK by ATP with a peptide substrate that is highly
selective for ERK1/2 (J. T. Neary et al., unpublished
observations). Because the ERK enzyme activity assay used
does not distinguish between ERK1 and ERK2, we used
antibodies that recognize dually phosphorylated ERK1 and
ERK2 (Thr183 and Tyr185) to determine the effect of
cotreatment with ATP 1 FGF-2 on ERK1/2 phosphorylation. Cultures were treated with FGF-2 (25 ng/mL) and ATP
(100 mm), either alone or in combination, for 5 min; cells
were lysed and immunoblots were prepared and probed for
phosphorylated ERK1 and ERK2. We found that cotreatment with ATP and FGF-2 also stimulated phosphorylation
of both ERK1 and ERK2 (data not shown).
We then measured the EC50 for ATP-induced activation
of ERK and inhibition of cRaf-1. Figure 2 shows that the
concentration±response curves for the ATP-induced inhibition of cRaf-1 and activation of ERK were very similar, with
half-effective doses at the same levels (EC50 of 2.1 mm for
ERK activation and IC50 of 1.3 mm for cRaf-1 inhibition).
The similarity of these EC50/IC50 values suggests that the
effect of ATP was purinoceptor-mediated, and not due to a
non-speci®c interaction of ATP with the FGF-2 peptide or
with the FGF-2 receptor. Preliminary experiments
indicated that the general P2 receptor antagonist suramin
(Ralevic and Burnstock 1998) could not be used to block
Fig. 1 ATP inhibits the FGF-2 induced activation of cRaf-1. (a) Primary cultures of rat cortical astrocytes were treated 5 min with
100 mM ATP, 25 ng/mL of FGF-2, or ATP 1 FGF-2 (A 1 F), after
which the cells were lysed. Lysates containing equivalent amounts
of protein were immunoprecipitated with cRaf-1 polyclonal antibody,
and the immunoprecipitates were assayed in a coupled assay, using
GST-MEK, ERK, MBP, and [g-32P]-ATP as described in Materials
and methods. In order to exclude possible MEK activator contamination in the immunoprecipitate, an assay without the addition of GSTMEK was performed simultaneously, and the result from this assay
was subtracted from the one containing GST-MEK. MEK and
ERK assay were done as described in Materials and methods. (b)
Summary of results obtained from at least six experiments.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
1004 G. Lenz et al.
Table 1 Effect of NE, serotonin and ATP on cAMP production
cAMP (nmol/mg protein; SEM)
Basal
ATP
5-HT
NE
Fig. 2 Concentration±response for the activation of ERK and the
inhibition of cRaf-1 by ATP. (a) Primary cultures of rat astrocytes
were treated with FGF-2 and increasing concentrations of ATP for
5 min, and cRaf-1 activity was measured as described in Materials
and methods. (b) Primary cultures of rat astrocytes were treated for
2 min with the indicated concentrations of ATP, and ERK activity (W)
was assayed using MBP as substrate. For cRaf-1 (O), cells were
treated with FGF-2 (which was used as 1 to calculate the relative
stimulation) or FGF-2 plus the indicated concentrations of ATP for
5 min and cRaf-1 coupled assays were performed. Data (means
^ SEM) were obtained from a minimum of two experiments and
were ®tted in a sigmoid dose±response curve by PRISM v2.0
(GraphPad).
the effect of ATP because suramin also altered growth factor
responses.
A time course of the inhibitory effect of ATP on the
activation of cRaf-1 by FGF-2 showed that the inhibition
was not present at 1 min, started between 1 and 2 min and
reached maximum around 5 min (Fig. 3). The inhibition
was still present at 60 min. The fast kinetics of the inhibition
excluded mechanisms that are dependent on ERK activation
or gene expression, pointing to mechanisms that are
activated by the early signaling events stimulated by ATP.
We therefore searched for mechanisms that might inhibit
cRaf-1, or upstream components of the cascade and that
Fig. 3 cRaf-1 stimulated by FGF-2 is rapidly blocked by ATP. Cells
were treated with ATP (A), FGF-2 (X) or ATP 1 FGF-2 (K) for the
indicated times and cRaf-1 coupled assays were performed as
described. Data (means ^ SEM) were obtained from a minimum of
six independent experiments.
0.10 (^ 0.04)
0.032 (^ 0.009)
0.09 (^ 0.02)
2.45 (^ 0.15)
n
n
n
n
ˆ
ˆ
ˆ
ˆ
5
5
3
3
Effect of ATP, NE and serotonin on cAMP synthesis. Primary
astrocyte cultures were treated for 10 min with ATP (100 mM),
serotonin (5-HT; 10 mM) or norepinephrine (NE; 10 mM), and cAMP
synthesis was determined as previously described. ATP and serotonin
did not induce cAMP production, while NE signi®cantly stimulated
cAMP production ( p , 0.05 with paired t-test).
could possibly be regulated by signaling components
activated by ATP.
PKC and cAMP formation are not involved in the
inhibition of cRaf-1
We investigated the activation of the cAMP/PKA pathway,
which has been shown to inhibit cRaf-1 (Wu et al. 1993).
However, inhibition by this pathway appears unlikely
because ATP did not increase the concentrations of cAMP
in astrocytes (Table 1 and Lenz et al. 2000).
Because ATP activates PKCd (Neary et al. 1999a), and
since PKC can phosphorylate and inhibit the FGFR1
(Gillespie et al. 1995), we considered the possibility that
PKC was mediating the ATP-induced inhibitory effect on
cRaf-1 activation, either by inhibiting FGFR1 or any of the
signaling components that link this receptor to cRaf-1.
Therefore, we inhibited or down-regulated PKC and asked
whether such treatment would relieve the inhibitory effect
of ATP on cRaf-1. However, inhibition of PKC with
GF109203X (Toullec et al. 1991) or Ro31±8220 (Davis
et al. 1989) or down-regulation with PMA did not relieve
the inhibitory effect of ATP on cRaf-1 (Table 2). Although
inhibition of PKC or down-regulation did not prevent
inhibition of FGF-induced activation of cRaf-1 by ATP,
both inhibitors and down-regulation blocked ATP-induced
stimulation of ERK (Table 2), a ®nding consistent with
previous studies (Neary et al. 1999a). However, the
stimulation of cRaf-1 induced by FGF-2 was not affected
by these treatments (Table 2). In addition, the PKC
inhibitors blocked the stimulation of cRaf-1 induced by
phorbol esters (data not shown).
Extracellular ATP also inhibits the activation of cRaf-1
by EGF and PDGF
Growth factors activate receptor tyrosine kinases (RTKs),
which contain a speci®c set of tyrosine residues whose
phosphorylation is responsible for the signal transduction
mechanisms activated by a RTK. Some RTKs also have
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
Purinergic inhibition of MEK activators 1005
Table 2 Effect of PKC inhibition on ERK and cRaf-1 activity induced by ATP and/or FGF-2
cRaf-1
Kinase inhibition\treatment
ERK
ATP
None
GF109203X
Ro31±8220
PMA DR
2.19
1.16
1.29
1.33
ATP
(Š^
(Š^
(Š^
(Š^
0.09)
0.08)
0.28)
0.25)
0.84
1.00
0.73
1.12
FGF-2
(Š^
(Š^
(Š^
(Š^
0.26)
0.06)
0.42)
0.16)
4.11
4.45
3.63
3.52
(Š^
(Š^
(Š^
(Š^
ATP 1 FGF-2
0.71)
0.85)
0.40)
1.36)
0.98
1.54
1.05
0.52
(Š^
(Š^
(Š^
(Š^
0.23)
0.64)
0.24)
0.19)
Cells were preincubated with 5 mM of the PKC inhibitors GF109203X or Ro 31-8220 for 20 min or with 100 nM PMA overnight (PMA DR) and then
incubated with ATP and/or FGF-2 for 5 min. ERK assays and coupled assays for cRaf-1 were performed. ANOVA analysis indicated that PKC
blockade or down-regulation signi®cantly inhibited ERK activation induced by ATP ( p , 0.001) but did not affect cRaf-1 stimulation by FGF-2
( p . 0.05) or the inhibition of FGF-2-induced cRaf-1 activation by ATP ( p . 0.05). Experiments were repeated at least twice.
other phosphorylation sites involved in their regulation
(Gillespie et al. 1995). Considering the differences among
the growth factor receptors, both in the coupled signal
transduction pathways and in their regulation, we tested
whether ATP could also inhibit cRaf-1 activation induced by
other growth factors. As shown in Fig. 4, cRaf-1 activity
induced by both EGF and PDGF was inhibited by ATP to a
similar extent when compared to the inhibition of cRaf-1
activity induced by FGF-2.
Extracellular ATP inhibits the activation of MEKK1 and
MEKK2 by FGF-2
Several regulatory mechanisms are speci®c for a particular
MEK1 activator, as is the case of the cAMP/PKA pathway,
which inhibits cRaf-1 (Wu et al. 1993), but does not inhibit
B-Raf (Vossler et al. 1997). Other MEK1 activators, like
MEKK 1, 2 and 3 have low homology on the catalytic
domain, and no homology on the regulatory domains with
the Raf family, although they have MEK1 and Ras as a
common substrate and activator, respectively (Russel et al.
1995; Blank et al. 1996). Therefore, considering these
Fig. 4 Effect of ATP on the activation of cRaf-1 by FGF-2, EGF and
PDGF. Cells were treated with ATP, the indicated growth factors
(GFs), FGF-2 (25 ng/mL), EGF (10 ng/mL) or PDGF (2.5 ng/mL), or
a combination of ATP and the GFs (A 1 GF) for 5 min and coupled
assays for cRaf-1 were performed. Data (mean ^ SEM) were
obtained from a minimum of three experiments. ANOVA test
#
p , 0.001.
differences between the Raf and MEKK families, the effect
of ATP on these MEK1 activators might indicate whether
the mechanism acts directly on cRaf-1 or on an upstream
and shared component. We found that ATP also inhibited
the activation of MEKK1 and MEKK2 by FGF-2 (Fig. 5),
thus pointing to a mechanism that is common to these
MEK1 activators.
Serotonin, another GPCR agonist, does not block
FGF-2-induced cRaf-1 activity
In view of the fact that ATP activates the ERK pathway
through P2Y receptors (King et al. 1996; Ralevic and
Burnstock 1998; Lenz et al. 2000) that are members of the
G protein-coupled receptor (GPCR) superfamily, we examined the action of two other GPCR agonists that stimulate
ERK, serotonin and norepinephrine (NE), on the activation
of cRaf-1 by FGF-2. Serotonin was chosen because, like
ATP, it did not induce cAMP formation in our astrocyte
cultures (Table 1). However, unlike ATP, no inhibition of
FGF-2-induced cRaf-1 activity was observed on exposure of
Fig. 5 Effect of ATP and FGF-2 on different MEK activators. Cells
were treated with ATP and/or FGF-2 for 5 min and coupled assays
were performed using the indicated antibodies for the immunoprecipitation. As indicated by the supplier (Santa Cruz Biotechnology),
MEKK2 antibody has some cross-reactivity with MEKK3. Data
(mean ^ SEM) were obtained from a minimum of 3 experiments.
#
ANOVA test p , 0.001, **p , 0.01 and *p , 0.05.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
1006 G. Lenz et al.
Table 3 Effect of NE and serotonin on FGF-2-induced cRaf-1 and
ERK activity
cRaf-1
ERK
Kinase
Basal
Basal
NE
5-HT
FGF-2
Basal
FGF-2
4.11 (Š^ 0.71)
2.28 (Š^ 0.11)
0.87 (Š^ 0.25) 1.37 (Š^ 0.46) 1.79 (Š^ 0.16) 2.36 (Š^ 0.24)
1.24 (Š^ 0.50) 3.56 (Š^ 0.13) 2.22 (Š^ 0.19) 2.63 (Š^ 0.18)
Cells were treated with FGF-2 and/or norepinpehrine (NE; 10 mM) or
serotonin (5-HT; 10 mM) for 10 min. cRaf-1-coupled assays and ERK
assays were performed. Data are mean ^ SEM. ANOVA tests indicate
that NE inhibited the FGF-2 induced cRaf-1 activity, while serotonin
had no effect on this activity.
the cells to serotonin (Table 3). By contrast, NE blocked
FGF-2-induced cRaf-1 activity (Table 3). This result was
expected, however, since NE increases cAMP synthesis in
astrocyte cultures (24 times the control; Table 1) and
increased synthesis of cAMP has been shown to inhibit
cRaf-1 in astrocytes (Kurino et al. 1996).
Effects of ATP plus FGF-2 on the activities of MEK and
ERK
Signal transduction components located upstream of MEK,
including cRaf-1, integrate and transmit signals to ERK,
which, together with the kinases activated by it, is the main
effector of the ERK cascade. It was important therefore to
determine how the inhibition seen at the level of cRaf-1
affected the activation of MEK and ERK. It should be noted
that both ATP and FGF-2 can activate MEK and ERK,
thereby complicating the study of the contribution of each
signaling molecule on these kinases. The cotreatment of
ATP 1 FGF-2 induced MEK and ERK activities that
resembled the activities seen with FGF-2 alone, except for
some differences in the kinetics of activation (Figs 6a and
b). At 1 and 2 min, MEK activity appeared to be slightly
higher with ATP 1 FGF-2 than with FGF-2 alone, but this
difference was not statistically signi®cant. In the case of
ERK, activity was signi®cantly greater at 1 and 2 min with
ATP alone or with ATP 1 FGF-2 compared to FGF-2
alone, whereas at 15 and 60 min ERK activity was
signi®cantly greater with FGF-2 alone or with
ATP 1 FGF-2 compared to ATP alone (Fig. 6b). Thus,
the activity of ERK induced by ATP 1 FGF-2 followed the
activity induced by ATP alone for the ®rst 2 min and after
15 min followed the activity induced by FGF-2 alone,
thereby resulting in an early and sustained activation of
ERK.
Discussion
Fig. 6 Time courses of the activation of the ERK cascade components. Cell were treated as in Fig. 1 for the indicated times, and the
activity of MEK (a) and ERK (b) were measured. ANOVA tests
show that the MEK activation by FGF-2 was not different from
A 1 F at any time. ERK activation by FGF-2 was different
from A 1 F at 1 min ( p , 0.05) and 2 min ( p , 0.01) and ERK
activation by ATP was different from A 1 F at 15 min ( p , 0.05)
and 60 min ( p , 0.001). Data (means ^ SEM) were obtained from
a minimum of two experiments. A, ATP; X, FGF-2; K, A 1 F.
The pathway by which FGF-2 and other polypeptide growth
factors activate the ERK cascade is well described. After
activation of the FGF-2 receptor, a series of protein±protein
interactions leads to the activation of Ras which then
recruits cRaf-1, the ®rst protein kinase in the ERK cascade
(Seger and Krebs 1995; Kouhara et al. 1997). The pathway
by which extracellular ATP activates the ERK cascade in
astrocyte cultures is not as well described, but recent studies
have shown that PLD and a calcium-independent PKC
isoform are upstream of ERK (Neary et al. 1999a).
However, in contrast to the FGF-2 pathway, the ATP
pathway in rat cortical astrocytes does not involve the
recruitment of cRaf-1, B-Raf, or MEKK1/2 (Lenz et al.
2000), apparently using an as yet unknown MEK activator.
Both FGF-2 and ATP induce proliferation of astrocyte
cultures through a mechanism dependent on the MEK/ERK
cascade (Neary et al. 1999a). Moreover, when astrocytes are
treated with ATP 1 FGF-2, the proliferation is increased
dramatically and synergistically (Neary et al. 1994).
Surprisingly, we have now found that cotreatment with
ATP leads to the inhibition of FGF-2-induced activation of
cRaf-1, an effect that was maximal at 5 min and sustained
for at least 1 h (Figs 1 and 3).
Rat cortical astrocytes express purine-preferring and
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
Purinergic inhibition of MEK activators 1007
Fig. 7 Scheme of the inhibitory cross-talk
of ATP activated signal transduction pathways on the activation of cRaf-1 by FGF-2
and other growth factors.
purine/pyrimidine-preferring P2Y receptor subtypes (Lenz
et al. 2000). The similarity of the concentration±response
curves for the inhibitory effect of ATP on cRaf-1 activation
and the activation of ERK by ATP (Fig. 2) suggests that the
activation of ERK and the inhibition of cRaf-1 are mediated
by the same P2Y receptor, or by purinergic receptors with
very similar af®nities. The time course of the activation of
components of the cascade also demonstrates some interesting features of the cross-talk. Since the inhibition of
cRaf-1 activation was maximal before ERK reached its
maximal activation, inhibition very likely occurred directly
from an early ATP-activated signal transduction component
to FGF-2-induced pathways at or above cRaf-1 and not
involving protein expression or negative feedback, as for
example the phosphorylation of SOS1 by ERK (Por®ri and
McCormick 1996). Further, the concentration-response and
the kinetics of inhibition suggest that direct interaction of
ATP with FGF-2 or the FGF-2 receptor is unlikely.
In an effort to identify the pathway by which stimulation
of ATP receptors leads to inhibition of cRaf-1 activation, we
tested some known mechanisms of inhibition involving
protein kinases. PKC is activated by ATP in astrocytes
(Neary et al. 1999a) and phosphorylates and inhibits the
FGFR1 (Gillespie et al. 1995). As shown in Table 2, PKC
blockers did not reverse the inhibition of cRaf-1 induced by
ATP. Accordingly, these inhibitors did not affect FGF-2
induced stimulation of cRaf-1 but were effective in
inhibiting ERK activation by ATP, as previously described
(Neary 1996; Neary et al. 1999a, 1999b). This suggests that
the signal transduction pathway(s) activated by ATP which
inhibit(s) cRaf-1 and which activate ERK bifurcate at a
point above the activation of PKC (Fig. 7). Although PKA is
well described as an inhibitor of cRaf-1 (Wu et al. 1993), it
is reasonable to suggest that this kinase is not involved
because ATP does not increase cAMP levels in astrocytes
(Table 1, and Lenz et al. 2000).
Several mechanisms for the inhibition of cRaf-1 have
been described in the literature, which are either dependent
on phosphorylation, as is the case of Akt (Zimmermann and
Moelling 1999; Guan et al. 2000) or dependent on
protein±protein interactions, as seen with the Raf-kinase
inhibitor protein (RKIP) (Yeung et al. 1999). BLAST
sequence alignments of the Raf family with the MEKK
family indicated that none of the sequences corresponding
to the main inhibitory phosphorylation sites in Raf are
conserved in MEKK1, and MEKK2 (BLAST search ±
Swissprot ± Sequence Retrieval System). This suggests that
direct phosphorylation of the MEK1 activators is probably
not involved in the inhibitory mechanism activated by ATP,
because ATP inhibited the effect of FGF-2 on both Raf and
the MEK1 activators, MEKK1 and MEKK2 (Fig. 5).
Furthermore, since the activation of cRaf-1 by three
different growth factors was inhibited by ATP, this suggests
that the inhibitory mechanism acts on a component that is
activated by FGF-2, EGF and PDGF and that can activate
cRaf-1, MEKK1 and MEKK2 (Figs 4, 5 and 7).
Ras is known to activate cRaf-1 and MEKK1 (Russel et al.
1995). Thus Ras, together with its regulators, represent
putative molecular targets of the inhibitory cross-talk (box
in Fig. 7). Signal transduction components that appear to
inhibit Ras, such as the kinase suppressor of Ras (KSR)
which was shown to activate the ERK cascade, only induces
inhibition when overexpressed (Cacace et al. 1999). Another
putative Ras inhibitor, the Ras suppressor (Rsu-1), blocks
the transformation induced by Ras, seemingly by inhibiting
the JNK pathway; in contrast, the ERK cascade is activated
by Rsu-1 (Masuelli and Cutler 1996).
GPCR agonists other than ATP that activate ERK could
shed light on the signaling events that are responsible for
the inhibition of cRaf-1. For this, we used serotonin,
which shows a remarkable resemblance with ATP-induced
signaling, since both (a) activate ERK in a manner
independent of cRaf-1 (Fig. 1 and Table 3); (b) do not
induce cAMP production (Table 1); and (c) induce an
increase in intracellular Ca21 with the initial peak dependent
on internal Ca21 and the sustained phase dependent on
external Ca21 (Neary et al. 1991; Jalonen et al. 1997;
Hagberg et al. 1998). Considering that ATP inhibits cRaf-1
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
1008 G. Lenz et al.
activation by FGF-2, whereas serotonin does not, this
suggests that the shared signaling events, like the increase in
intracellular Ca21 and the lack of increase of cAMP, are
probably not involved in the inhibitory cross-talk. Another
GPCR agonist, norepinephrine, increased cAMP and inhibited cRaf-1 activity (Tables 1 and 3). The norepinephrineinduced inhibition was probably mediated by cAMP/PKA,
because increased cAMP is known to inhibit cRaf-1 in
astrocytes (Wu et al. 1993; Kurino et al. 1996).
Because activation of Raf leads to stimulation of ERK, it
was important to analyze the effect of the inhibitory crosstalk on ERK. As shown in Fig. 6, ERK activity induced by
ATP 1 FGF-2 presented fast kinetics of activation, just like
ERK stimulated with ATP, but a slow deactivation, similar
to that seen with FGF-2. Except for the early time points, the
maximal ERK activity stimulated by ATP 1 FGF-2 was
only slightly larger than the maximal activity induced by
FGF-2, certainly much less that would be expected if the
activities induced by these two transmitters were additive.
Nonetheless, although ATP inhibited FGF-2-induced cRaf-1
activation, the combined presence of extracellular ATP and
FGF-2 allowed for an early and sustained activation of ERK.
This sustained activity may be due to the regulation of
mechanisms responsible for the deactivation of the ERK
cascade, such as the phosphatase 2 A, which can be
inhibited by growth factors (Heriche et al. 1997).
The ERK cascade is involved in the signaling of several
cellular events. Therefore, kinetics and intensity of signaling
of the ERK cascade are important ways of differentiating
signals. For example, sustained kinetics induces differentiation while transient kinetics induces proliferation in PC12
cells (Marshall 1995). Recently, several studies have shown
that the relationship between cRaf-1 signal intensity and
progress through the cell cycle presents an inverted U shape,
with intermediate intensities stimulating proliferation and
extremely low and high activities blocking the progression
through the cell cycle (Sewing et al. 1997; Woods et al.
1997). These complex relationships between the ERK
cascade activity and cellular events might in part explain
how the inhibitory cross-talk described in the present work
could be related to a signaling mechanism that leads to the
synergistic effect of ATP and FGF-2 on proliferation. For
example, if cRaf-1 activity stimulated by FGF-2 is too high
to stimulate a maximal proliferation, then a shift towards
lower activity, as induced by ATP, could induce an increase
in proliferation. On the other hand, since PDGF and EGF do
not present any synergism in proliferation with ATP (Neary
et al. 1994), but gave similar results when compared to
FGF-2 on the inhibitory cross-talk, the inhibitory effect
described in the present article may not be directly related to
the synergism.
In summary, we have demonstrated that signaling by
extracellular ATP in astrocytes stimulates an inhibitory
pathway towards cRaf-1, a key signaling component
recruited by polypeptide growth factors. In addition, we
present evidence that this inhibitory pathway acts on signal
transduction elements that are upstream of cRaf-1 (and
MEKK 1 and 2), below the growth factor receptor, and
independent of PKC or cAMP. Considering the importance
of extracellular ATP and FGF-2 in astrocyte proliferation
such as in reactive gliosis, the elucidation of the molecular
events involved in the interaction of the signal transduction
pathways activated by ATP and FGF-2 may represent an
important step towards the understanding of astrocyte
responses to brain injury.
Acknowledgements
We are grateful to Dr Peter Dunkley for discussion and Drs C. I.
Tasca and D. O. Souza for their help with cAMP determinations. This work was supported by the Department of Veterans
Affairs, NIH, PRONEX-FINEP, and Department of Bio®sica,
UFRGS, Brazil.
References
Bhalla U. S. and Iyengar R. (1999) Emergent properties of networks of
biological signaling pathways. Science 283, 381±387.
Blank J. L., Gerwins P., Elliot E. M., Sather S. and Johnson G. L.
(1996) Molecular cloning of MEKK 2 and 3. J. Biol. Chem. 271,
5361±5368.
Burnstock G. (1990) Overview: purinergic mechanisms. Ann. NY Acad.
Sci. 603, 1±18.
Burnstock G. (1997) The past, present and future of purine nucleotides
as signalling molecules. Neuropharmacology. 36, 1127±1139.
Cacace A. M., Michaud N. R., Therrien M., Mathes K., Copeland T.,
Rubin G. M. and Morrison D. K. (1999) Identi®cation of
constitutive and ras-inducible phosphorylation sites of KSR:
implications for 14±323 binding, mitogen-activated protein
kinase binding, and KSR overexpression. Mol. Cell. Biol. 19,
229±240.
Davis P. D., Hill C. H., Keeck E., Lawton G., Nixon J. S.,
Sedgwick A. D., Wadsworth J., Westmacott D. and
Wilkinson S. E. (1989) Potent selective inhibitors of protein
kinase C. FEBS Lett. 259, 61±63.
Deng G., Matute C., Kumar C. K., Fogarty D. J. and Miledi R. (1998)
Cloning and expression of a P2y purinoceptor from the adult
bovine corpus callosum. Neurobiol. Dis. 5, 259±270.
Eclancher F., Kehrli P., Labourdette G. and Sensenbrenner M. (1996)
Basic ®broblast growth factor (bFGF) injection activates the glial
reaction in the injured adult rat brain. Brain Res. 737, 201±214.
Farnsworth C. L., Freshney N. W., Rosen L. B., Ghost A.,
Greenberg M. E. and Feig L. A. (1995) Calcium activation of
Ras mediated by neuronal exchange factor Ras-GRF. Nature 376,
524±527.
Finklestein S. P., Apostolides P. J., Caday C. G., Prosser J., Phillips M. F.
and Klagsburn M. (1988) Increased basic ®broblast growth factor
(bFGF) immunoreactivity at the site of focal brain wounds. Brain
Res. 460, 253±259.
Franke H., Krugel U. and Illes P. (1999) P2 receptor-mediated
proliferative effects on astrocytes in vivo. Glia 28, 190±200.
Gillespie L. L., Chen G. and Paterno G. D. (1995) Cloning of FGFR
splice variant from Xenopus embryos that lacks a PKC site is
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009
Purinergic inhibition of MEK activators 1009
important for the regulation of receptor activity. J. Biol. Chem.
270, 22758±22763.
Guan K. L., Figueroa C., Brtva T. R., Zhu T., Taylor J., Barber T. D.
and Vojtek A. B. (2000) Negative regulation of the serine/
threonine kinase B-Raf by Akt. J. Biol. Chem. 275, 27354±27359.
Hagberg G. B., Blomstrand F., Nilsson M., Tamir H. and Hansson E.
(1998) Stimulation of 5-HT2A receptors on astrocytes in primary
culture opens voltage-independent Ca21 channels. Neurochem.
Int. 32, 153±162.
Heriche J. K., Lebrin F., Rabilloud T., Leroy D., Chambaz E. M. and
Goldberg Y. (1997) Regulation of protein phosphatase 2A by
direct interaction with casein kinase 2a. Science 276, 952±955.
Jalonen T. O., Margraf R. R., Wielt D. B., Charniga C. J., Linne M. L.
and Kimelberg H. K. (1997) Serotonin induces inward potassium
and calcium currents in rat cortical astrocytes. Brain Res. 758,
69±82.
King B. F., Neary J. T., Zhu Q., Wang S., Norenberg M. D. and
Burnstock G. (1996) P2 purinoceptors in rat cortical astrocytes:
expression, calcium-imaging and signalling studies. Neuroscience
74, 1187±1196.
Kouhara H., Hadari Y. R., Spivak-Kroizman T., Schilling J.,
Bar-Sagi D., Lax I. and Schlessinger J. (1997) A lipid-anchored
Grb2-binding protein that links FGF-receptor activation to the
Ras/MAPK signaling pathway. Cell 89, 693±702.
Kurino M., Fukunaga K., Ushio Y. and Miyamoto E. (1996) cAMP
inhibits activation of MAPK and cell proliferation in response to
growth factors in cultured rat cortical astrocytes. J. Neurochem.
67, 2246±2255.
Kyriakis J. M., Force T. L., Rapp U. R., Bonventre J. V. and
Avruch J. (1992) Raf-1 activates MAP kinase kinase. Nature 358,
417±421.
Kyriakis J. M., Force T. L., Rapp U. R., Bonventre J. V. and Avruch J.
(1993) Mitogen regulation of c-Raf-1 protein kinase activity
toward mitogen-activated protein kinase kinase. J. Biol. Chem.
268, 16009±16019.
Lenz G., Gottfried C., Luo Z., Avruch J., Rodnight R., Nie W. J.,
Kang Y. and Neary J. T. (2000) P2Y purinergic receptor subtypes
recruit different MEK activators in astrocytes. Br. J. Pharmacol.
129, 927±936.
Luo Z., Tzivion G., Belshaw P. J., Vavvas D., Marshall M. M. and
Avruch J. (1996) Oligomerization activates c-Raf-1 through a
Ras-dependent mechanism. Nature 383, 181±185.
Marshall C. J. (1995) Speci®city of receptor tyrosine kinase signaling:
transient versus sustained extracellular signal-regulated kinase
activation. Cell 80, 179±185.
Masuelli L. and Cutler M. L. (1996) Increased expression of the Ras
suppressor Rsu-1 enhances Erk-2 activation and inhibits Jun
kinase activation. Mol. Cell. Biol. 16, 5466±5476.
Mattingly R. R. and Macara I. G. (1996) Phosphorylation-dependent
activation of the Ras-GRF/CDC25Mn exchange factor by muscarinic receptors and G-protein bg subunits. Nature 382,
268±272.
Neary J. T. (1996) Trophic action of extracellular ATP on astrocytes:
synergistic interactions with FGF and underlying signal transduction mechanisms. Ciba Found. Symp. 198, 130±141.
Neary J. T. and Zhu Q. (1994) Signaling by ATP receptors in astrocytes.
Neuroreport 5, 1617±1620.
Neary J. T., Laskey R., van Breemen C., Blicharska J.,
Norenberg L. O. B. and Norenberg M. D. (1991) ATP-evoked
calcium signal stimulates protein phosphorylation/dephosphorylation in astrocytes. Brain Res. 566, 89±94.
Neary J. T., Whittemore S. R., Zhu Q. and Norenberg M. D. (1994)
Synergistic activation of DNA synthesis in astrocytes by FGF and
extracellular ATP. J. Neurochem. 63, 490±494.
Neary J. T., Rathbone M. P., Cattabeni F., Abbracchio M. P. and
Burnstock G. (1996) Trophic actions of extracellular nucleotides
and nucleosides on glial and neuronal cells. Trends Neurosci. 19,
13±18.
Neary J. T., Kang Y., Bu Y., Yu E., Akong K. and Peters C. M. (1999a)
Mitogenic signaling by ATP/P2Y purinergic receptors in astrocytes: involvement of a calcium-independent PKC ERK pathway
distinct from the PI-PLC-calcium pathway. J. Neurosci. 19,
4211±4220.
Neary J. T., McCarthy M., Cornell-Bell A. and Kang Y. (1999b)
Trophic signaling pathways activated by purinergic receptors in
rat and human astroglia. Prog. Brain Res. 120, 323±332.
Por®ri E. and McCormick F. (1996) Regulation of EGF receptor
signaling by phosphorylation of the Ras exchange factor hSOS1.
J. Biol. Chem. 271, 5871±5877.
Ralevic V. and Burnstock G. (1998) Receptors for purines and
pyrimidines. Pharmacol. Rev. 50, 413±492.
Russel M., Lange-Carter C. A. and Johnson G. L. (1995) Direct
interaction between Ras and the kinase domain of mitogenactivated protein kinase kinase kinase (MEKK1). J. Biol. Chem.
270, 11757±11760.
Schlessinger T. K., Fanger G. R., Yujiri T. and Johnson G. L. (1998)
The tao of MEKK. Front. Biosci. 3, d1181±1186.
SchoÈnwasser D. C., Marais R. M., Marshall C. J. and Parker P. J. (1998)
Activation of the mitogen-activated protein kinase/ERK pathway
by conventional, novel, and atypical protein kinase C isotypes.
Mol. Cell. Biol. 18, 790±798.
Seger R. and Krebs E. G. (1995) The MAPK Signaling Cascade.
FASEB J. 9, 726±735.
Sewing A., Wiseman B., Lloyd A. C. and Land H. (1997) High-intensity Raf signal causes cell cycle arrest mediated by p21Cip1.
Mol. Cell. Biol. 17, 5588±5597.
Suhasini M., Li H., Lohmann S. M., Boss G. R. and Pilz R. B. (1998)
Cyclic-GMP-dependent protein kinase inhibits the Ras/mitogenactivated protein kinase pathway. Mol. Cell. Biol. 18, 6983±6994.
Toullec D., Pianetti P., Coste H., Bellevergue P., Grand-Perret T.,
Ajakane M., Baudet V., Boissin P., Boursier E. and Loriolle F.
(1991) The bisindolylmaleimide GF 109203X is a potent and
selective inhibitor of PKC. J. Biol. Chem. 266, 15771±15781.
Vossler M. R., Yao H., York R. D., Pan M. G., Rim C. S. and
Stork P. J. S. (1997) cAMP activates MAP kinase and Elk-1
through a B-Raf- and Rap1-dependent pathway. Cell 89, 73±82.
Webb T. E., Simon J. and Barnard E. A. (1998) Regional distribution of
[35S]2 0 -deoxy 5 0 -O-(1-thio) ATP binding sites and the P2Y1
messenger RNA within the chick brain. Neuroscience 84,
825±837.
Wieraszko A., Goldsmith G. and Seyfreid T. N. (1989) Stimulationdependent release of ATP from hippocampal slices. Brain Res.
485, 244±250.
Woods D., Parry D., Cherwinski H., Bosch E., Lees E. and
McMahon M. (1997) Ras-induced proliferation or cell cycle
arrest is determined by the level of cRaf-1 activity with arrest
mediated by p21Cip1. Mol. Cell. Biol. 17, 5598±5611.
Wu J., Dent P., Jelinek T., Wolfman A., Weber M. J. and Sturgill T. W.
Inhibition of the EGF-activated MAPK (1993) Signaling
pathway by adenosine 3 0 ,5 0 -monophosphate. Science 262,
1065±1069.
Yeung K., Seitz T., Li S., Janosch P., McFerran B., Kaiser C., Feek F.,
Katsanakisk K. D., Rose D. W., Mischak H., Sedivy J. M. and
Kolch W. (1999) Suppression of Raf-1 kinase activity and MAP
kinase signalling by RKIP. Nature 401, 173±177.
Zimmermann S. and Moelling K. (1999) Phosphorylation and
regulation of Raf by Akt (protein kinase B). Science 286,
1741±1744.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 1001±1009