(Received for publication, October 22, 1996, and in revised form, April 7, 1997)
From the Institut National de la Santé et de la Recherche Médicale, Faculté de Médecine Nice 06107 Cedex 2, France
Phosphatidylinositol (PI) 3-kinase is activated by various growth factors such as PDGF (platelet-derived growth factor) and insulin. The aim of the present study was to determine whether PDGF could modulate insulin activation of PI 3-kinase in 3T3-L1 adipocytes. When cells were preincubated for 5-15 min with PDGF, PI 3-kinase activity associated to insulin receptor substrate 1 (IRS 1) in response to insulin was decreased, due to reduced association of the PI 3-kinase p85 subunit with IRS 1. In addition, following this PDGF pretreatment, the tyrosine phosphorylation of IRS 1 in response to insulin and its electrophoretic mobility were diminished. The change in the mobility of IRS 1 could be attributed to PDGF-induced serine/threonine phosphorylation of the protein which was partly inhibited by PI 3-kinase inhibitors. By contrast, epidermal growth factor, which does not stimulate PI 3-kinase, had no effect on the association of PI 3-kinase with IRS 1 in response to insulin. This series of results indicates that the PDGF-induced serine/threonine phosphorylation of IRS 1 could be due to activation of PI 3-kinase pathway. Furthermore, this phosphorylation of IRS 1 is associated with a decrease in its tyrosine phosphorylation by insulin and in its association with the p85 subunit of PI 3-kinase. This study suggests that a cross-talk exists between the different pathways stimulated by PDGF and insulin in intact cells.
Phosphatidylinositol (PI) 3-kinase1 is a common element of the signaling pathway of a large number of tyrosine kinase receptors. PI 3-kinase is a heterodimer consisting of an 85-kDa regulatory subunit (p85) containing one Src homology 3 (SH3) domain and two Src homology 2 (SH2) domains (1-3) and an 110-kDa catalytic subunit (4). The catalytic subunit phosphorylates inositol lipids at the D-3 position of the inositol ring and has been shown to possess a serine kinase activity (5). By contrast, the p85 regulatory subunit functions as an adaptor which, via its SH2 domains, links PI 3-kinase to tyrosine-phosphorylated proteins such as autophosphorylated tyrosine kinase receptors (6). This association leads to the stimulation of the kinase activities of the p110 subunit (6, 7). Since PI 3-kinase is activated by a large range of peptide growth factors, this enzyme activity appears to be implicated in various cellular responses including promotion of cell growth, regulation of cell differentiation, and metabolism (for review see Ref. 6). Despite this, each growth factor triggers distinct and specific biological responses in each particular cell type. In the present study, we looked at the effects of a prior stimulation by platelet-derived growth factor (PDGF) on the further ability of insulin to activate PI 3-kinase. We took advantage of the 3T3-L1 adipocytes where PI 3-kinase can be activated by both insulin and PDGF but not by EGF (8, 9). When PDGF stimulates tyrosine phosphorylation of its receptor, the p85 regulatory subunit of PI 3-kinase associates to phosphorylated Tyr-Xaa-Xaa-Met motifs of the PDGF receptor (10, 11). In contrast, insulin activates the tyrosine kinase activity of its receptor that subsequently phosphorylates insulin receptor substrate 1 (IRS 1) on YMXM motifs allowing binding of p85 to IRS 1 and PI 3-kinase activation (12-14). Our data show that pretreatment with PDGF decreases the ability of insulin to phosphorylate IRS 1 on tyrosine residues and consequently decreases both the amount of the p85 subunit and the PI 3-kinase activity associated with IRS 1. These results demonstrate that, in 3T3-L1 adipocytes, a cross-talk exists between the pathways stimulated by PDGF and insulin. This phenomenon could play a role in the regulation of the cell responses to growth factors and may explain the modulation of the cellular responses to different stimuli.
Antibodies to IRS 1 were obtained from a rabbit injected with a peptide corresponding to the sequence comprising amino acids 489-507 of the protein and used at 1/100 dilution for immunoblotting. Antibodies used for immunoblotting of the p85 subunit of the PI 3-kinase and of phosphotyrosine-containing proteins were from UBI (Lake Placid, NY). Antibodies to phosphotyrosine used in immunoprecipitation assays were obtained after injection of a rabbit with phosphotyrosine coupled to bovine immunoglobulins.
MaterialsInsulin was a gift from Lilly (Paris, France). PDGF-BB was from Pepro Tech. Inc. (Rocky Hill, NJ). Bovine serum albumin was from Intergen Co. (Purchase, NY). All other biochemicals were from Sigma or Serva (Heidelberg, Germany). Radiochemicals were from ICN Pharmaceuticals (Orsay, France).
Cell Culture3T3-L1 fibroblasts were cultured in DMEM, 10% fetal calf serum and induced to differentiate into adipocytes as described (8). 3T3-L1 adipocytes were used between 8 and 12 days after initiation of differentiation, when more than 95% of the cells exhibited an adipocyte-like phenotype. Sixteen hours before each experiment, the cells were changed to serum-free DMEM supplemented with 0.5% (w/v) bovine serum albumin.
Measurement of PI 3-Kinase Activity3T3-L1 adipocytes were
pretreated at 37 °C without or with PDGF (50 ng/ml) or EGF (100 nM) for different periods. Then the cells were stimulated
or not for 5 min with insulin (100 nM). The cells were
solubilized for 40 min at 4 °C in 700 µl of buffer A (20 mM Tris, pH 7.4, 137 mM NaCl, 100 mM NaF, 10 mM EDTA, 2 mM
Na3VO4, 10 mM pyrophosphate, 1 mM PMSF, 100 units/ml aprotinin) containing 1% Nonidet
P-40 (v/v). Lysates were centrifuged for 10 min at 13,000 × g. Supernatants were incubated for 2 h at 4 °C with
antibodies to IRS 1 or to phosphotyrosine coupled to protein A-Sepharose beads. Immune pellets were washed twice with each of the
three following buffers: (a) phosphate-buffered saline containing 1% Nonidet P-40 and 200 µM
Na3VO4; (b) 100 mM Tris, pH 7.4, 0.5 M LiCl, 200 µM
Na3VO4, and (c) 10 mM
Tris, pH 7.4, 100 mM NaCl, 1 mM EDTA, 200 µM Na3VO4. Bead-associated PI
3-kinase activity was assayed from the phosphorylation of PI in the
presence of [-32P]ATP as described previously (15).
The reaction products were separated by thin layer chromatography.
Quantification was performed after autoradiography by Cerenkov counting
of the spot corresponding to phosphatidylinositol 3-phosphate.
Cells were pretreated with PDGF (50 ng/ml) for different periods and then stimulated with insulin (100 nM) for 5 min. When indicated, cells were treated with 100 nM wortmannin or with 50 µM LY294002 for 20 min prior to PDGF stimulation. Then the cells were solubilized in buffer A containing 1% Nonidet P40. Cell lysates were incubated for 2 h with antibodies to IRS 1 preadsorbed on protein A-Sepharose. Pellets were washed as described above and were treated with Laemmli buffer, boiled for 10 min, and proteins analyzed by SDS-PAGE with a 7.5% acrylamide gel. Proteins were transferred to a polyvinylidene difluoride (PVDF) sheet. The sheet was incubated with blocking buffer (phosphate-buffered saline, 5% bovine serum albumin, w/v) for 2 h at room temperature and then overnight at 4 °C with antibodies to IRS 1, to the p85 subunit of PI 3-kinase, or to phosphotyrosine. Membranes were washed three times (10 min each) in phosphate-buffered saline containing 1% Nonidet P-40. When antibodies to phosphotyrosine were used, a further incubation was performed for 1 h at room temperature with rabbit anti-mouse immunoglobulins. Finally, sheets were incubated for 1 h at room temperature with 125I-protein A (5 × 105 cpm/ml blocking buffer) and washed as above. Blots were submitted to autoradiography.
32P-Orthophosphate Labeling of 3T3-L1 Adipocytes3T3-L1 adipocytes (one 100-mm dish/condition) in phosphate-free DMEM supplemented with 0.5% bovine serum albumin (w/v) were labeled for 2.5 h at 37 °C with [32P]orthophosphate (1 mCi/5 ml). When indicated, cells were treated with 100 nM wortmannin for 20 min prior to a 15- or 60-min stimulation with 50 ng/ml PDGF. Then the cells were stimulated or not for 5 min with 100 nM insulin. Cells were washed with ice-cold buffer A and solubilized for 40 min at 4 °C in 800 µl of buffer A supplemented with 1% Nonidet P-40 (v/v). Samples were centrifuged at 13,000 × g for 10 min. Supernatants were immunoprecipitated for 2 h at 4 °C with antibodies to IRS 1 preadsorbed on protein A-Sepharose. The immune pellets were washed as described previously, treated with Laemmli buffer, boiled for 10 min, and separated on a 7.5% acrylamide SDS-PAGE. The gel was dried and autoradiographed.
We first determined the time
course of PI 3-kinase activation by PDGF in 3T3-L1 adipocytes. Cells
were incubated for various periods with 50 ng/ml PDGF. PI 3-kinase
activity was then measured in immunoprecipitates obtained with
antibodies to phosphotyrosine (Fig. 1).
PDGF maximally stimulated the PI 3-kinase activity between 5 and 15 min
of stimulation. Then the activity progressively decreased to 60 min.
This time course of PI 3-kinase activation paralleled the tyrosine
phosphorylation of the PDGF receptor (Fig. 1) suggesting that PI
3-kinase activity measured in antiphosphotyrosine immunoprecipitates was due to the association of the enzyme with the
tyrosine-phosphorylated PDGF receptors. By contrast, EGF did not induce
change in the PI 3-kinase activity in 3T3-L1 adipocytes but was able to
activate mitogen-activated protein kinase (data not shown).
Effect of PDGF Pretreatment on Insulin-induced PI 3-Kinase Activity Associated with IRS 1
Since PI 3-kinase associates to activated
PDGF receptors, we investigated whether a pretreatment of the cells by
PDGF could interfere with a subsequent activation of PI 3-kinase by
insulin. To test this hypothesis, 3T3-L1 adipocytes were stimulated
with PDGF for different periods, before a subsequent 5-min insulin stimulation. PI 3-kinase activity was then measured in
immunoprecipitates obtained with antibodies to IRS 1 (Fig.
2). Insulin alone markedly increased PI
3-kinase activity associated to IRS 1 (15- ± 3-fold increase in four
different experiments). When the cells were pretreated with PDGF for
5-15 min before insulin stimulation, PI 3-kinase activity associated
to IRS 1 decreased. Maximal inhibition (32%) was observed after 15 min
of PDGF pretreatment. No inhibition was observed at 30 min of PDGF
pretreatment and PI 3-kinase activity decreased again by 30% after 60 min pretreatment. Different from the results observed with PDGF, EGF
pretreatment of the cells did not modify the level of PI 3-kinase
activity associated to IRS 1 in response to a subsequent insulin
stimulation.
Effect of PDGF Pretreatment on Insulin-induced Tyrosine Phosphorylation of IRS 1 and Its Association with PI 3-Kinase
The
previous experiments indicate that the PI 3-kinase activity associated
with IRS 1 following insulin stimulation is diminished by a PDGF
pretreatment. We then wanted to determine whether it reflected a
diminution of the amount of the enzyme associated with IRS 1 or an
inhibition of the PI 3-kinase activity itself. 3T3-L1 adipocytes were
pretreated with PDGF for different periods before insulin stimulation
(5 min). Proteins were solubilized and immunoprecipitated with
antibodies to IRS 1. The immunoprecipitated proteins were analyzed by
SDS-PAGE and transferred to PVDF membranes. The PI 3-kinase associated
with IRS 1 was detected by antibodies to its p85 subunit (Fig.
3A). In the absence of PDGF
pretreatment, insulin induced the association of the PI 3-kinase p85
subunit with IRS 1. The pretreatment with PDGF for 5-60 min markedly
reduced the amount of the p85 associated to IRS 1. In parallel, the
tyrosine phosphorylation of proteins was visualized with an antibody to phosphotyrosine. The insulin-induced autophosphorylation of the -subunit of its receptor was not affected by PDGF (data not shown). By contrast, a PDGF pretreatment induced a decrease in the
insulin-induced tyrosine phosphorylation of IRS 1 (Fig. 3B).
It should be noted that PDGF did not induce any tyrosine
phosphorylation of IRS 1 (data not shown and Ref. 8). The decrease in
insulin-induced IRS 1 tyrosine phosphorylation was not due to a
modification in the amount of IRS 1 present in the immune pellets since
it was similar in all conditions, indicating that PDGF pretreatment
modified neither the total cellular amount of IRS 1 nor the ability of the antibody to immunoprecipitate IRS 1 (Fig. 3C). PDGF
induced an increase in the apparent molecular weight of IRS 1, which
was similar to that induced by insulin (Fig. 3C).
Effect of PDGF Treatment on IRS 1 Phosphorylation
To test
whether the slower IRS 1 electrophoretic migration in PDGF-treated
cells was due to a change in its phosphorylation state, 3T3-L1
adipocytes were labeled with [32P]orthophosphate. Cells
were then stimulated for 15 or 60 min without or with PDGF (50 ng/ml),
before insulin treatment (100 nM, 5 min). At the end of the
incubation, the proteins were solubilized and subjected to
immunoprecipitation with antibodies to IRS 1. Immunoprecipitated
proteins were analyzed by SDS-PAGE followed by autoradiography (Fig.
4). A low level of IRS 1 phosphorylation was observed in basal conditions. Both insulin and PDGF induced the
incorporation of [32P]orthophosphate into IRS 1 (2.5- and
2-fold increase, respectively). Since insulin (Fig. 3), but not PDGF
(data not shown and Ref. 8), induced the tyrosine phosphorylation of
IRS 1, these results suggest that the phosphorylation occurred mostly
on tyrosine residues in the presence of insulin and on serine/threonine
residues following PDGF stimulation. Furthermore, the phosphorylations
induced by insulin and PDGF were not additive since an acute insulin
stimulation after a PDGF pretreatment did not markedly increase
32P incorporation into IRS 1. These results indicate that
PDGF treatment could induce the phosphorylation of IRS 1 in intact
cells on serine/threonine residues and decreased its ability to be
tyrosine-phosphorylated by insulin and thus to bind the p85 subunit of
PI 3-kinase.
Effect of PI 3-Kinase Inhibitors on PDGF-induced IRS 1 Phosphorylation
We then looked at the possibility that PI
3-kinase was involved in the phosphorylation of IRS 1 in response to
PDGF. Indeed, PI 3-kinase is activated in response to PDGF (Refs. 8 and
9 and Fig. 1), and PI 3-kinase exerts a serine kinase activity toward IRS 1 in vitro (16-18). Cells were incubated with two
pharmacological inhibitors of the enzyme, i.e. with 100 nM wortmannin or 50 µM LY294002, for 20 min
prior to any incubation with PDGF or insulin (Fig.
5). IRS 1 tyrosine phosphorylation was
then followed by immunoblotting with antibodies to phosphotyrosine as
described previously. Pretreatment of the cells with wortmannin
enhanced the tyrosine phosphorylation of IRS 1 induced by insulin
alone, as also reported (19), whereas LY294002 was without effect. Pretreatment with LY294002 and wortmannin blocked the inhibitory effect
of PDGF on the tyrosine phosphorylation of IRS 1 induced by insulin,
and concomitantly, they prevented the electrophoretic mobility shift
induced by PDGF (Fig. 5). When cells were labeled with
[32P]orthophosphate as described above, wortmannin
markedly inhibited, by 80%, the serine/threonine phosphorylation
induced by PDGF (Fig. 6).
The present study was designed to look at a possible modulation of
insulin action by growth factors, focusing at the level of PI 3-kinase.
PI 3-kinase is involved in multiple signaling pathways (20), such as
those activated by insulin and PDGF. Activation of this enzyme results
from the association of the p85 subunit of the PI 3-kinase to
tyrosine-phosphorylated IRS 1 in response to insulin (12-14, 21) or
directly to tyrosine-phosphorylated PDGF receptor in response to PDGF
(10, 11). Here we have taken advantage of 3T3-L1 adipocytes, which
express insulin, PDGF, and EGF receptors, to show that a pretreatment
of the cells with PDGF, but not with EGF, decreased the PI 3-kinase
activity and the amount of p85 subunit associated with IRS 1 in
response to a subsequent insulin stimulation. This effect was rapid,
since it occurred after a very short PDGF treatment (5-15 min).
Moreover, in PDGF-treated cells, the insulin-induced tyrosine
phosphorylation of IRS 1 was reduced, and this could explain the lower
amount of p85 subunit associated with IRS 1 in response to insulin. The
decrease in IRS 1 tyrosine phosphorylation was not due to an inhibition
of insulin receptor tyrosine kinase activity since PDGF treatment did
not modify the insulin-induced autophosphorylation of the receptor. By
contrast, PDGF treatment lowered the electrophoretic mobility of IRS 1, the likely consequence of a phosphorylation of IRS 1 following PDGF
treatment. Since PDGF did not induce any tyrosine phosphorylation of
IRS 1 (data not shown and Ref. 8), the shift can only result from an
increase in the serine/threonine phosphorylation of IRS 1 following
PDGF treatment. These results are reminiscent of previous studies using
okadaic acid (22) or tumor necrosis factor- (23, 24) agents which
simultaneously increase IRS 1 serine/threonine phosphorylation and
decrease insulin-stimulated IRS 1 tyrosine phosphorylation. It is thus
tempting to think that the stimulation of IRS 1 serine/threonine
phosphorylation induced by PDGF treatment was responsible for the
inhibition of insulin-induced IRS 1 tyrosine phosphorylation. The
mechanism by which the serine/threonine phosphorylation of IRS 1 could
modulate its tyrosine phosphorylation is not well defined. In the case
of tumor necrosis factor-
treatment, it was proposed that the
phosphorylated form of IRS 1 could act as an inhibitor of the insulin
receptor tyrosine kinase activity (24). This explanation seems unlikely
in the present study since insulin receptor autophosphorylation was not
altered by a PDGF pretreatment (data not shown). When cells were
treated with okadaic acid, the tyrosine kinase activity of the insulin
receptor was not changed, but the serine/threonine-phosphorylated form
of IRS 1 had a decreased ability to interact with the insulin receptor (22). Recently, it has been proposed that the phosphorylation of the
serine residues of IRS 1, in the proximity of the tyrosine residues
involved in the binding of PI 3-kinase, could decrease the association
between PI 3-kinase and IRS 1 (25). Such a mechanism could be involved
in the effect we observed in PDGF-treated cells.
Several studies have focused on the identification of kinases involved
in the serine/threonine phosphorylation of IRS 1. Casein kinase II (26)
and the serine kinase activity of PI 3-kinase (16-18) have been shown
to phosphorylate IRS 1 in vitro. A series of data argue for
a role of PI 3-kinase in the phosphorylation of IRS 1 following
PDGF stimulation. First, EGF, which is unable to activate PI 3-kinase
in 3T3-L1 adipocytes but which activates multiple other signaling
pathways such as the mitogen-activated protein kinase kinase cascade,
does not modify the insulin-induced PI 3-kinase activity associated
with IRS 1. Second, the serine/threonine phosphorylation of IRS 1 induced by PDGF is partly inhibited by two pharmacological inhibitors
of PI 3-kinase as indicated by the reversal of the mobility shift of
IRS 1 or by the cell orthophosphate labeling experiments. These data
thus suggest that PI 3-kinase can induce a serine/threonine
phosphorylation of IRS 1 not only in vitro in
immunoprecipitates obtained from insulin-stimulated cells (16-18) but
also in cells stimulated by PDGF. It is also possible that the
phosphorylation of IRS 1 observed after a PDGF treatment results from
the activation of another serine kinase(s) activated by the PI 3-kinase
pathway. Indeed, novel serine/threonine kinases, targets of PI
3-kinase, such as the protein kinase B (27-29) and atypical protein
kinase C and -
(30, 31) are activated by PDGF. Serine kinase
activities, independent of the PI 3-kinase pathway, could also
participate in the phosphorylation of IRS 1 since the PDGF-induced
phosphorylation of IRS 1 was not completely blocked by wortmannin.
Further studies are necessary to identify the enzyme implicated in the
serine/threonine phosphorylation of IRS 1.
The decrease in insulin-induced PI 3-kinase activity associated with
IRS 1 was the likely consequence of the reduced IRS 1 tyrosine
phosphorylation at early time points of PDGF pretreatment. However,
other mechanisms could be involved in such an effect. Indeed, it has
been shown that the p85 subunit becomes tyrosine-phosphorylated after a
PDGF stimulation (32, 33) and that this tyrosine phosphorylation decreased its ability to associate with tyrosine-containing proteins such as the activated PDGF receptor. Such an hypothesis is unlikely in
our case since we never observed such a tyrosine phosphorylation of p85
in our cell line. It is also possible that the activated PDGF receptor
binds a large amount of p85 subunit following PDGF stimulation thus
decreasing, by a competitive process, the amount of p85 available to
associate with IRS 1 after insulin stimulation. Indeed, a displacement
of PI 3-kinase from the PDGF receptors toward IRS 1 following insulin
stimulation occurred in a recent study (34) performed with cells
transfected by high levels of insulin receptors while only the
endogenous PDGF receptors were present. The present series of
experiments has been performed in 3T3-L1 adipocytes, expressing only
the endogenous insulin and PDGF receptors. Since it has recently been
shown in NIH 3T3 cells that only 5% of the total pool of p85
associated with the PDGF receptor (33), this hypothesis is not very
plausible to explain our results. It is still unknown why PI 3-kinase
activity associated with IRS 1 returned to maximally insulin-stimulated
values after 30 min of PDGF pretreatment. This occurred although
serine/threonine phosphorylation of IRS 1 by PDGF was maintained for at
least 60 min, concomitant with a constant decrease in the tyrosine
phosphorylation of IRS 1 and of its association with the p85 subunit.
The reason for this observation is still unclear.
Our results clearly show that intercommunications exist between the different growth factors and hormones signaling pathways. Enzymes common to multiple pathways, such as the PI 3-kinase, create the possibility of many different types of interactions among the various pathways. The diversity of these interactions could result in additive, synergistic, or antagonistic effects. The studies of such cross-talks between the different signaling pathways should be new approaches to understand the physiological significance of such events.
We thank T. Grémeaux for technical assistance, C. Minghelli and G. Visciano for illustrations, and M. Cormont for scientific discussion.