The diacylglycerol and protein kinase C pathways are not involvedin insulin signalling in primary rat hepatocytes Irmelin Probst1, Ulrich Beuers2, Birgit Drabent1, Kirsten Unthan-Fechner
Trang 1The diacylglycerol and protein kinase C pathways are not involved
in insulin signalling in primary rat hepatocytes
Irmelin Probst1, Ulrich Beuers2, Birgit Drabent1, Kirsten Unthan-Fechner1and Peter Bu¨tikofer3
1
Institut fu¨r Biochemie und Molekulare Zellbiologie, Georg-August – Universita¨t Go¨ttingen, Germany;2Medizinische Klinik II-Großhadern, Ludwig-Maximilians-Universita¨t Mu¨nchen, Germany;3Institut fu¨r Biochemie und Molekularbiologie,
Universita¨t Bern, Switzerland
Diacylglycerol (DAG) and protein kinase C (PKC) isoforms
have been implicated in insulin signalling in muscle and fat
cells We evaluated the involvement of DAG and PKC in the
action of insulin in adult rat hepatocytes cultured with
dexa-methasone, but in the absence of serum, for 48 h Our
results show that although insulin stimulated glycolysis and
glycogen synthesis, it had no effect on DAG mass or
molecular species composition Epidermal growth factor
showed the expected insulin-mimetic effect on glycolysis,
whereas ATP and exogenous phospholipase C acted as
antagonists and abolished the insulin signal Similarly to
insulin, epidermal growth factor had no effect on DAG mass
or molecular species composition In contrast, both ATP
and phospholipase C induced a prominent increase in
sev-eral DAG molecular species, including 18:0/20:4, 18:0/20:5,
18:0/22:5 and a decrease in 18:1/18:1 These changes were
paralleled by an increase in phospholipase D activity, which
was absent in insulin-treated cells By immunoblotting or by measuring PKC activity, we found that neither insulin nor ATP translocated the PKCa, -d, -e or -f isoforms from the cytosol to the membrane in cells cultured for six or 48 h Similarly, insulin had no effect on immunoprecipitable PKCf Suppression of the glycogenic insulin signal by phorbol 12-myristate 13-acetate, but not by ATP, could be completely alleviated by bisindolylmaleimide Finally, insu-lin showed no effect on DAG mass or translocation of PKC isoforms in the perfused liver, although it reduced the glu-cagon-stimulated glucose output by 75% Together these results indicate that phospholipases C and D or multiple PKC isoforms are not involved in the hepatic insulin signal chain
Keywords: hepatocytes; insulin; ATP; diacylglycerol mole-cular species; protein kinase C
Among the three major insulin-sensitive organs, i.e liver,
muscle and fat tissue, the liver plays a key role in the
regulation of blood glucose homeostasis by channelling
excess glucose into glycogen after food uptake and by
producing glucose through glycogenolysis and
gluconeo-genesis in the states of hunger and starvation Insulin, the
dominant hormone of the absorptive phase, acts via
receptor-mediated tyrosine phosphorylation of insulin
receptor substrates (IRSs) Two well established signalling
cascades are initiated when adaptor proteins are recruited
to the IRSs through their src homology 2 domains (a) the
growth factor receptor binding protein activates the ras/
mitogen-activated protein kinase pathway and (b)
phosphatidylinositol 3-kinase activates the protein kin-ase B/glycogen synthkin-ase kinkin-ase-3 cascade Recent data suggest that a third signalling pathway, downstream of phosphatidylinositol 3-kinase, may also be involved: phos-pholipase D (PLD)-dependent generation of phosphatidic acid (PA) and diacylglycerol (DAG), with subsequent activation of DAG-insensitive atypical protein kinase C (PKC) isozymes such as f and k, as well as activation of DAG-sensitive PKC isozymes [1–3] These studies, which were performed on muscle and fat cells, showed insulin-dependent increases in lipid mediator concentrations [4–7] and translocation and activation of various PKC isoforms [6–13], suggesting their probable involvement in insulin action [8,10,14,15]
In contrast, the available data on hepatic systems are scarce, controversial and have been obtained using primary adult rat hepatocyte suspensions and cultures, and different hepatoma cell lines, as model systems In hepatocyte suspensions, insulin provoked increases in DAG mass [16,17], whereas activation of PLD was demonstrated by two groups [17,18], but not by another group [19] Similarly, activation of PKC was demonstrated, in two reports, in both cytosolic and membrane fractions of crude extracts [16,20], but not in a third [21] Furthermore, activation of atypical PKCf was demonstrated in hepatocytes cultured without glucocorticoid for 3 days [22], whereas two other reports showed enhanced translocation of the d isoform in different hepatoma cell lines [23,24]
Correspondence to I Probst, Institut fu¨r Biochemie und Molekulare
Zellbiologie, Humboldtallee 23, 37073 Go¨ttingen, Germany.
Fax: + 49 551 395960, Tel.: + 49 551 395961,
E-mail: iprobst@gwdg.de
Abbreviations: DAG, diacylglycerol; EGF, epidermal growth factor;
IRSs, insulin receptor substrates; ODN, oligodesoxynucleotides;
PA, phosphatidic acid; PMA, phorbol 12-myristate 13-acetate;
PKC, protein kinase C; PLC, phospholipase C; PLD,
phospholipase D; TGF-a, transforming growth factor-a.
Enzymes: phospholipase C (EC 3.1.4.3); phospholipase D
(EC 3.1.4.4); protein kinase C (EC 2.7.1.37).
(Received 25 April 2003, revised 26 August 2003,
accepted 25 September 2003)
Trang 2The aim of the present work was to study the possible
involvement of lipid signalling and PKC during hepatic
insulin action in a differentiated model for the adult organ,
the primary adult rat hepatocyte cultured serum-free with
dexamethasone This system shows high insulin sensitivity
and responsiveness towards a multitude of
insulin-depend-ent parameters [25–27] The effects of insulin were compared
with those of epidermal growth factor (EGF), ATP and
exogenous phospholipase C (PLC)
Materials and methods
Materials
Enzymes, M199 medium, collagenase A and the
transfec-tion agent DOSPER were from Roche Molecular
Biochem-icals (Mannheim, Germany) Bovine insulin was from Serva
(Heidelberg, Germany) Bisindolylmaleimide I and protein
G–agarose were from Calbiochem (Bad Soden, Germany)
Phorbol 12-myristate 13-acetate (PMA), rottlerin, PLC
from Clostridium perfringens, IGEPAL and dexamethasone
were from Sigma (Taufkirchen, Germany) A stock solution
of PMA (10 mM) was made in dimethylsulfoxide; before
use it was diluted 1 : 100 in M199 medium containing
0.2% (w/v) bovine serum albumin D-[U-14C]Glucose,
[32P]ATP[cP], [9,10-3H]myristic acid, [9,10-3H]palmitic acid
and the Renaissance Western blot chemiluminescence
reagent were from New England Nuclear (Dreieich,
Ger-many) The DAG quantification test kit and the PKC
enzyme assay system were purchased from Amersham
(Braunschweig, Germany) The PKCf isoenzyme-specific
pseudosubstrate [Ser159]PKC-e-[153–164]-NH2 was from
Bachem (Heidelberg, Germany) Silica gel 60 TLC plates
with concentration zones were from Merck (Hannover,
Germany) Whatman P-81 paper was from Herolab
(Wies-loch, Germany) Rabbit anti-PKC peptide Igs, anti-a, -b, -c,
-d, -e, and -f for immunoblotting were obtained from Gibco
(Grand Island, NY, USA) Rabbit anti-PKCf for
immu-noprecipitation and activity determination, and the
anti-PKCf blocking peptide, were from Santa Cruz (Heidelberg,
Germany) PKCf antisense oligodesoxynucleotides were
from Biognostik (Goettingen, Germany) and cytofectin
from Eurogentec/Glen Research (Ko¨ln, Germany)
Cell culture
Hepatocytes from fed male Wistar rats (of weight
180–250 g) were isolated by recirculating collagenase
per-fusion in situ, purified by centrifugation through Percoll and
cu ltu red in M199 mediu m on 6-cm plastic dishes [28] For
the first 3 h, medium contained 4% newborn calf serum,
1 nMinsulin and 0.1 lM dexamethasone Serum was then
omitted and the cells were cultured for the next 4 or 43 h
with 1 nMinsulin and 0.1 lMdexamethasone Medium was
changed at 22 h The gas atmosphere contained CO2/O2/N2
(5 : 17 : 78)
Cell experiments
After 4 or 46 h of continuous culture, dishes were washed
twice and incubated in M199 (2.5 mL per dish) After 1 h
the medium was replaced with M199 containing 2 m
lactate (2 mL per dish) For the determination of glycogen synthesis and glycolysis, the medium was supplemented with [14C]glucose (30 kBq per dish) After a 30-min prein-cubation, zero-time samples were taken and the experiment was started by the addition of agonists to the dishes Inhibitors were added 10 min before the agonists The incubation was terminated by rapidly aspirating the medium and immersing the dishes in liquid N2
Glycolysis and glycogen synthesis Glycolysis was determined by the rate of lactate release into the culture supernatant Labelled glucose was separ-ated from labelled lactate by chromatography of 100 lL
of culture supernatant on Dowex 1· 8 (formate form), as outlined previously [25] The rate of glycogen synthesis was determined by extracting and quantifying the
14C-labelled glycogen from one 6 cm dish, as described previously [27]
Liver perfusion Rat livers were perfused in situ, via the portal vein, with Krebs-Henseleit bicarbonate buffer, pH 7.4 (5 mMglucose,
2 mM lactate, 0.2 mMpyruvate; 95% O2/5% CO2; 37C; constant flow without recirculation, 5.5–6 mLÆmin)1Æg)1of liver) Experiments were performed between 09.00 h and 11.00 h; preperfusion lasted for 20 min before the onset of sampling from the inferior vena cava Liver samples were taken from the front lobe at 35 min
Lipid extraction and separation by TLC Hepatocytes from one 6-cm dish were scraped into 2 mL
of methanol and transferred into a glass tube Chloroform (1 mL) was added and lipids were extracted for 10 min at
4C Subsequently, 1 mL of chloroform and 1.7 mL of
1MNaCl were added under vigorous mixing After 5 min, samples were centrifuged (400 g for 5 min), the aqueous phase was discarded and the organic phase dried under N2
at room temperature Lipids were concentrated in the V-shaped tip of the tube by repetitive solvent evaporation and resuspension using diminishing volumes of chloro-form Dried lipids were stored under N2 at )20 C For separation of lipids by TLC, extracts were redissolved in
50 lL of chloroform/methanol (2 : 1, v/v) and applied in a 1-cm zone on a 20· 20 cm2silica gel plate with concen-tration zone The solvent system used for separation of DAG was heptan/diisopropylether/acetic acid (60 : 40 : 8, v/v/v) Alkylacyl and alk-1-enylacyl subclasses co-migrate
on this TLC system with the DAG species For the determination of PLD activity in hepatocytes, 0.3% (v/v) butanol was added to cells incubated in the presence or absence of insulin or ATP After lipid extraction, phos-phatidylbutanol (formed by PLD-mediated phosphatidyl-transfer onto butanol) was separated by TLC using ethyl acetate/isooctane/acetic acid/H2O (130 : 20 : 30 : 100, v/v/ v/v) as the solvent system [17] Phosphatidylbutanol was quantified by a procedure that chars saturated and unsaturated lipids equally [29] followed by densitometry using authentic phosphatidylbutanol, prepared as des-cribed previously [30], as a standard
Trang 3Determination of DAG mass
DAG content was quantified radioenzymatically by
incu-bating aliquots of the lipid extract with DAG kinase and
[32P]ATP[cP], as described by Preiss et al [31] The manu
-facturer’s instructions for the commercially available DAG
test kit were followed.32P-labelled PA was purified using
chloroform/methanol/acetic acid (65 : 15 : 5, v/v/v) as a
solvent system and quantified with a Storm 860
phosphoi-mager (Pharmacia, Freiburg, Germany)
Analysis of DAG molecular species
Hepatocytes from one 10-cm dish were extracted as outlined
above After drying the lipid extract under nitrogen, DAGs
were extracted with ether and immediately benzoylated, as
described by Blank et al [32] Diradylglycerobenzoates were
separated into their subclasses (diacyl, alkylacyl, and
alk-1-enylacyl types) by TLC using benzene/hexane/ether
(50 : 45 : 4, v/v/v) as a solvent system, and the individual
molecular species were separated by HPLC using an
octadecyl reverse-phase column in acetonitrile/isopropanol
(80 : 20, v/v) as the mobile phase Individual peaks were
quantified by measuring absorbance at 230 nm To identify
individual molecular species, representative samples were
analysed by combined HPLC/MS [33] using the
instrumen-tation described in Bu¨tikofer et al [34] Briefly, after the UV
detector, methanol/0.2M aqueous ammonium acetate
(10 : 90; v/v) was added via a T-connector, and the total
flow was introduced through a thermospray interface into a
Finnigan MAT model TSQ70 mass spectrometer The
M + NH4+ ions of the diradylglycerobezoates were
monitored by selected ion recording The positional
distri-bution of the fatty acyl and fatty alcohol chains of individual
molecular species was not determined The inclusion of the
antioxidant, butylated hydroxytoluene, in the different
solvents was found not to be necessary
PKC activity in cytosol and membranes of crude extracts
Hepatocytes from two 6-cm dishes were homogenized in
500 lL of lysis buffer (20 mM Hepes, pH 7.5, 250 mM
sucrose, 1 mM EGTA, 1 mM sodium vanadate, 1 mM
sodium pyrophosphate, 1 mMNaF, 20 lgÆmL)1leupeptin,
20 lgÆmL)1aprotinin, 1 mMphenylmethanesulfonyl
fluor-ide, 20 mM 2-mercaptoethanol) and centrifuged at
100 000 g for 30 min Supernatant and membrane fraction
were diluted with lysis buffer (without sucrose and EGTA)
and 8–14 lg of protein from each fraction was assayed for
the ability to phosphorylate a synthetic EGF-receptor
peptide (RKRTLRRL) The Amersham assay contained
25 lL of sample, 25 mM Tris/HCl, pH 7.5, 34 lgÆmL)1
phosphatidylserine, 2.7 lgÆmL)1 PMA, 102 lM receptor
peptide, 3.4 mM dithiothreitol, 1.36 mM calcium acetate,
109 lM ATP, and 6.5 mM MgCl2 in a total volume of
55 lL For the specific measurement of the atypical PKC
isoenzyme-f, Ca2+was omitted from the assay, 80 lgÆmL)1
phosphatidylserine was substituted for the kit lipid reagent
and the kit peptide substrate was replaced by 50 lM
[Ser159]PKC-e-[153–164]-NH2 Assays were conducted in
the presence or absence of the substrate for 3–9 min at
30C and stopped with 0.3M phosphoric acid Aliquots
were spotted on P-81 filter paper, washed three times with
75 mMphosphoric acid and counted
Immunoblotting of PKC isoforms Hepatocyte preparation and immunoblotting were per-formed exactly as described previously [35] Samples from the perfused liver (200 mg) were homogenized in 1 mL of lysis bu ffer by sonication (5· 10 s) and centrifuged at
8000 g for 3 min; the supernatant was processed as outlined previously [35] The bands of PKC isoforms were identified by (a) comparison with molecular mass markers run on each gel, (b) comparison with the bands
of a rat brain cytosol sample rich in all relevant PKC isoforms run on each gel, (c) PMA-induced PKC translocation from the cytosol to the membrane fraction (except for the nonmobile f-isoform; samples of control and PMA-treated cells were run on each gel for compar-ison), and (d) comparison of bands after incubation of a membrane blot with buffer in the presence or absence of
an antigen (PKC isoform) of the respective PKC antibody The bands on the immunoblots at about
80 000 molecular mass, representing PKC isoforms a, d and f, and at 90 000 molecular mass, representing PKC isoform e, were quantified by densitometry
Immunoprecipitation and activity assay of PKCf Hepatocytes from one 6 cm dish were homogenized in
500 lL of PKC lysis buffer (see above) supplemented with 0.5% IGEPAL (Nonidet P-40) and 1% Triton X-100 The lysate was sonicated for 10 s and centrifuged for 20 min at
20 000 g after 30 min of incubation at 4C Supernatants (200 ll, 1 mg of protein) were incubated under mild agitation for 4 h at 4C with 5 lg of anti-PKCf, which had been coupled to protein G–agarose (30 lL of agarose in NaCl/Pi, 1 h, 4C) Immobilized immune complexes were recovered by centrifugation, washed three times with complete lysis buffer and twice in kinase buffer (50 mM Tris pH 7.5, 10 mMMgCl2, 1 mMsodium vanadate, 1 mM dithiothreitol, 10 lgÆmL)1leupeptin, 10 lgÆmL)1aprotinin, 0.2 mM phenylmethanesulfonyl fluoride) Kinase buffer (25 lL) was added to the beads and the enzyme was assayed in a total volume of 50 lL containing 80 lgÆmL)1 phosphatidylserine, 50 lM[32P]ATP[cP] (15 kBq per assay) and 50 lM PKC e-peptide Enzyme activity showed time linearity for at least 15 min Assays were conducted for
10 min and processed, as described for PKC, in crude extracts Immunoblot analysis showed that insulin treat-ment of the cells did not alter the amount of PKCf in the immunoprecipitate
Results
Hepatocytes used in the present study were routinely cultured serum-free in the presence of 0.1 lM dexametha-sone for 46 h In each subsequent short-term experiment, measurement of lipid mediators or PKC was always paralleled by the determination of the physiological action
of insulin on glucose metabolism ATP and exogenous PLC, which both stimulate DAG formation [36–38], were used as positive controls In addition, EGF, an insulin-mimetic as
Trang 4well as an insulin-antagonistic factor [39–41], was included
in some experiments
Metabolic effects
We found that the addition of insulin to our primary rat
hepatocyte cultures stimulated glycolysis 4.5-fold, with a
50% effective dose (ED50) of 0.3 nM, whereas EGF
increased glycolysis twofold, with an ED50of 0.5 ngÆmL)1
(Fig 1B) Similar results have been reported before for
other hepatocyte culture systems [27,41] Furthermore, transforming growth factor-a (TGF-a) completely mim-icked EGF action in the lower concentration range (0.1–3 ngÆmL)1); however, it elicited an extrastimulatory response (+30%) at higher concentrations (Fig 1B) In contrast to its known inhibitory action on glycogen synthase [36], PLC was insulin-mimetic at low concentrations (0.1–3 mUÆmL)1) and stimulated glycolysis by up to 3.5-fold (Fig 1A) However, at concentrations of
> 5 mUÆmL)1, the effect of PLC diminished with increas-ing concentrations At these higher concentrations, PLC strongly antagonized the action of insulin As reported previously [42], ATP (> 10 lM) inhibited basal and insulin-activated glycolysis (results not shown)
Furthermore, we observed that the addition of insulin stimulated glycogen synthesis ninefold, whereas ATP, PLC and EGF/TGF-a severely inhibited both basal and insulin-activated rates of glycogen synthesis (Fig 2) These findings are in good agreement with previous reports using other hepatocyte culture systems [28,37,40,43] Unexpectedly, however, EGF and TGF-a were found to be insulin-mimetic at a low concentration (1 ngÆmL)1) (Fig 2) In some experiments, cells were cultured for only 6 h; these cells were less insulin-responsive (as shown by a threefold increase of glycogenesis)
Determination of DAG mass and DAG molecular species Increases in DAG and PA, through insulin-dependent activation of PLC and PLD, have been reported previously for rat hepatocytes [16–18] In contrast to these studies, we found no increase in DAG mass when the cells were stimulated with 1–100 nM insulin, either in 6-h cultures (data not shown) or in 48 h cultures (Fig 3) Similarly, the addition of 10 ngÆmL)1EGF or 10 ngÆmL)1TGF-a also showed no effect As shown previously [36,37], ATP and PLC are capable of rapidly elevating the level of DAG In agreement with these reports, we found that the addition of
100 lM ATP doubled DAG mass within 5 min; interest-ingly, the presence of PLC increased DAG mass at both insulin-mimetic (5 mUÆmL)1) and insulin-antagonistic (100 mUÆmL)1) concentrations (Fig 3)
It has been previously shown that the addition of tritium-labelled fatty acids to hepatocytes results in the incorpor-ation of label into the phospholipid fraction [17,18]; subsequent addition of insulin led to an increase in the production of [3H]DAG and [3H]PA We investigated such
a possible mechanism by labelling cells from 24 h to 46 h of culture with 110 kBqÆmL)1of [3H]myristate or [3 H]palmi-tate and determined the amount of radioactivity recovered
in the DAG fraction Again, our results showed no differences between cells incubated in the presence or absence of insulin (results not shown)
To study whether the observed increase in DAG mass after stimulation of rat hepatocytes with ATP or PLC was specific for certain molecular species, diacyl, alkylacyl and alk-1-enylacyl subclasses were separated and their molecular species composition was determined by combined HPLC/
MS The results in Table 1 show a typical molecular species composition of the diacylglycerol subclass from untreated hepatocytes; the corresponding HPLC trace is shown
in Fig 4A Control and agonist-stimulated hepatocytes
Fig 1 Insulin-mimetic effects of epidermal growth factor (EGF),
transforming growth factor a (TGF-a) and phospholipase C (PLC) on
glycolysis Hepatocytes were cultured for 46 h in the presence of 1 n M
insulin and 0.1 l M dexamethasone Subsequently, they were washed
free of hormones and incubated for 30 min in M199 medium
con-taining 0.1 l M dexamethasone and 2 m M lactate before the agonists
were added [ 14 C]Lactate production from 5 m M [ 14 C]glucose was
measured for 2 h Data represent mean values ± SD from three
dif-ferent hepatocyte preparations.
Trang 5contained almost exclusively diacyl-type molecular species (> 98% of total species) The HPLC profile, and thus the composition of DAG species, was not altered when cultures were treated with insulin (100 nM), EGF (10 ngÆmL)1)
or TGF-a (10 ngÆmL)1), for various periods of time (0.5–60 min) at different cell densities (results not shown) These results are entirely consistent with our observation that insulin, EGF and TGF-a have no effect on DAG levels
in primary rat hepatocytes
In contrast, 100 lM ATP and 100 mUÆmL)1 PLC showed a dramatic change in the HPLC profile (Fig 4B) Relative increases were seen for peak 11 (18:0/20:5), peak 17 (18:0/22:5) and peak 18 (18:0/20:4), whereas peak 20 (18:1/ 18:1) was reduced (Fig 5) The most prominent effect was a 3.9-fold enrichment of the species 18:0/20:4 (peak 18), which was observed for both agonists
Our results clearly contrast those of Baldini and cowork-ers who showed an insulin-dependent increase in DAG and
PA in hepatocytes [17,18] However, their studies were carried out either with hepatocyte suspensions or with cells cultured for 24 h in the absence of dexamethasone and insulin, but in the presence of 10% (v/v) fetal bovine serum
We therefore investigated the effect of insulin on DAG molecular species composition using their culture condi-tions In agreement with their results, we found that in cells cultured for 24 h, insulin provoked the elevation of two molecular species of DAG (18:0/20:4 and 18:0/20:5), while one species was decreased (18:1/18:1) Thus, insulin indeed mimicked the effects of ATP and PLC although the changes were smaller, i.e 30–50% of the ATP responses (results not shown) However, when we studied the metabolic insulin responsiveness of the cells cultured under these steroid-free conditions, we found that the activation of glycogen
Fig 2 Modulation of basal and insulin-stimulated glycogen synthesis by epidermal growth factor (EGF),transforming growth factor a (TGF-a),ATP and phospholipase C (PLC) Hepatocytes were cultured as described in the legend to Fig 1 Incorporation of [14C]glucose into glycogen was measured for 2 h Data represent mean values ± SD from four to seven different hepatocyte preparations.
Fig 3 Total cellular diacylglycerol (DAG) mass of hepatocytes after
treatment with ATP,phospholipase C (PLC),insulin and epidermal
growth factor (EGF)/transforming growth factor a (TGF-a)
Hepato-cytes were cultured as described in the legend to Fig 1 The DAG
content was quantified in lipid extracts using the DAG kinase assay.
Data represent mean values ± SD from three to five different
hepatocyte preparations.
Trang 6synthesis was severely reduced by 90% compared to cells
cultured with dexamethasone (Fig 2)
Measurement of PLD activity
A possible involvement of PLD in insulin signalling was
investigated in cells using our serum-free culture
condi-tions, in the presence of dexamethasone, by determining
transphosphatidylation activity with 0.3% butanol as the
acceptor [17] We found that cell exposure to insulin in
the presence of butanol did not increase the formation
of phosphatidylbutanol As reported previously [19],
transphosphatidylation was, however, five- to 10-fold
enhanced in the presence of ATP (positive control, data not shown)
Translocation of PKC Rat hepatocytes in culture expressed PKC isoforms a, d, e and f The a-isoform was mainly associated with the cytosolic fraction, and the d-, e- and f-isoforms were approximately equally distributed between the cytosol and membrane fraction (Table 2) In control experiments, the conventional cPKCa and the novel nPKCs d and e, but not the atypical aPKCf, were translocated to the membrane fraction by the phorbol ester, PMA (Table 2) These results are in good agreement with a previous study [35] Neither insulin nor ATP were able to translocate any of the isoforms within 1–15 min after agonist addition (Table 2)
Measurement of PKC activity
In a first series of experiments, PKC activity was determined
as overall activity in cytosol and membranes using the EGF-receptor peptide as a non isoform-specific substrate and the PKCe pseudosubstrate as a preferred substrate for PKCf Translocation of the PKC by PMA from the cytosol to the membrane was clearly demonstrated by the cytosolic decrease and membranous increase of enzyme activity (Table 3); in contrast, insulin showed no effect on PKC activity
In a second series of experiments, PKCf was immuno-precipitated and its activity was determined in precipitates from cells treated with or without insulin for 1–15 min We were unable to detect an insulin-dependent increase in the activity of the immunoprecipitated enzyme, which agrees with the inability of insulin to translocate PKCf
Inhibitor studies Stimulation of glycogen synthesis by insulin could not be inhibited by the relatively selective PKC inhibitor bis-indolylmaleimide I, which predominantly inhibits conven-tional and novel isoforms, i.e the a-, d- and e-isoforms (Fig 6) Owing to its isoform specificity, the inhibitor completely alleviated the insulin-antagonistic effect of PMA, which is mediated via DAG-dependent PKC isoforms In contrast, bisindolylmaleimide I was unable to revert the ATP-mediated blockade of the insulin signal (Fig 6) Selective inhibition of PKCd by the inhibitor rottlerin (5–10 lM) was also without effect on insulin signalling (data not shown)
Finally, we tried to inhibit insulin signalling by transfect-ing hepatocytes with antisense oligodesoxynucleotides (ODN) targeted against PKCf We found that cells transfected with 2.5 lgÆmL)1 cytofectin and 0.125 nM fluorescent ODN, or with 2–10 lgÆmL)1 DOSPER and 0.5–2.5 lM fluorescent ODN, showed up to 80% fluores-cent nuclei, and the amount of PKCf was reduced slightly (< 30%) after 3 days of culture when PKCf antisense ODN was added It should be noted, however, that both cell vitality (measured by the release of lactate dehydrogenase) and insulin signalling (measured as glycogen synthesis) were significantly decreased by the transfection agents as well as
by the control ODN alone (results not shown) Thus,
Table 1 Diacylglycerol molecular species composition of rat
hepato-cytes cultured for 48 h Diacylglycerols were analysed as
diacyl-glycerobenzoate derivates by combined HPLC/MS Individual
molecular species are listed in order of their elution from the HPLC
column (Fig 4A) Values are mean ± SD of three determinations
from a typical experiment.
Peak no Molecular species Composition (%)
3 16:1, 20:4 1.2 ± 0.1
4 18:2, 20:4
+ 16:0, 20:5 a
4.4 ± 0.4
5 16:2, 18:2 5.1 ± 0.2
6 + 7 18:2, 18:2
+ 18:1, 18:3a + 16:1, 20:3 a
1.7 ± 0.2
8 + 9 16:1, 18:2
+ 16:0, 22:6a + 14:0, 14:0 a
2.6 ± 0.3
10 16:1, 16:1
+ 14:0, 18:2a + 14:0, 16:1 a
4.4 ± 0.2
11 16:1, 22:4
+ 18:0, 20:5a
2.4 ± 0.2
12 16:0, 20:4 3.8 ± 0.2
13 18:1, 18:2
+ 18:0, 22:6a + 16:1, 18:1a
12.5 ± 0.3
14 16:0, 18:2
+ 18:0, 18:3a
8.8 ± 0.5
15 16:0, 16:1 0.7 ± 0.1
17 18:0, 22:5 4.1 ± 0.6
18 18:0, 20:4 4.3 ± 0.9
19 17:0, 18:2
+ 16:1, 17:0 a
3.6 ± 0.7
20 18:1, 18:1 13.6 ± 0.8
21 16:0, 18:1
+ 18:0, 18:2 a
14.3 ± 0.4
22 16:0, 16:0
+ 18:0, 22:4 a
1.4 ± 0.1
25 18:0, 18:1 2.5 ± 0.01
26 16:0, 18:0 3.8 ± 0.3
27 18:0, 18:0 1.0 ± 0.4
a
These species co-elute from the HPLC column.
Trang 7although this method has been successfully applied to
down-regulate specific PKC isoforms and to study PKC
involvement in signal transduction in other cell systems
previously [44], it seems to not (yet) be applicable to primary
hepatocytes
Insulin effects in the perfused liver
The effects of insulin on DAG mass and PKC isoform
translocation were examined in the intact organ to exclude
the possibility that the data obtained with hepatocytes were
restricted to the isolated cell system The anti-glucagon
action of insulin was chosen to demonstrate the hormone’s
metabolic activity The perfused liver received 50 pM
glucagon for 5–10 min; this first bolus served as an internal
metabolic vitality control From 30 to 35 min, the liver
received no agonist (basal control), 1 lM PMA (positive
control for PKC translocation), a second bolus of 50 p
glucagon or a staggered infusion of 10 nM insulin (25–35 min) and 50 pM glucagon (Fig 7B,C) The anti-glucagon effect of insulin was demonstrated by a 75% reduction of the glucagon-stimulated glucose output PMA alone stimulated glucose production (data not shown) [45]
Of all agonists used, only PMA translocated PKC isoforms
a, d and e (Fig 7A) Differences in DAG mass (lgÆmg)1of protein) were not observed between control liver (8.3) and livers treated with glucagon (7.9), insulin/glucagon (8.2), or PMA (8.2, n¼ 3 for all treatments)
Discussion
In muscle and fat tissue, lipid messengers such as DAG and PA, as well as DAG-dependent and -independent PKC isoforms, have recently been proposed to play a role
in the insulin signal leading to activation of glucose uptake [1–3]
Fig 4 HPLC profile of diacylglycerol (DAG) molecular species of control and ATP-stimulated hepatocytes Hepatocytes cultured for 2 days were incubated for 5 min with M199 as vehicle (A) or 100 l M ATP (B), and the molecular species of DAG were analysed as described in the Materials and methods Data represent mean values ± SD of three determinations from a representative experiment of 10.
Trang 8In contrast, in liver preparations these novel insulin
signalling pathways have been poorly studied and the
available data are confusing and controversial [16–22] The
results presented in this report were obtained using (a)
the highly insulin-sensitive in vitro liver system of cultured
hepatocytes and (b) the perfused liver, and speak clearly against an involvement of phospholipases and PKC isoforms in hepatic insulin signalling, for the following reasons First, we found that the addition of insulin to rat hepatocytes did not increase DAG mass or change the
Fig 5 Changes in hepatocyte diacylglycerol (DAG) molecular species composition in response to ATP and phospholipase C (PLC) stimulation Hepatocytes cultured for 2 days were incubated with vehicle (control, s), 100 l M ATP (d), or 5 (n) or 100 (m) mUÆmL)1PLC, and the molecular species of DAG were analysed as described in the Materials and methods The figure shows time-dependent changes of four DAG species expressed
as percentages of total DAG Data represent the mean values ± SD from four to six different hepatocyte preparations.
Table 2 Effect of 4b-phorbol 12-myristate 13-acetate (PMA),insulin and ATP on the distribution of protein kinase C (PKC) isoforms Hepatocytes cultured for 6 h and 48 h were incubated with 0.1 l M PMA, 10 n M insulin or 0.1 l M ATP for 5 min, and subsequently homogenized and separated into cytosol and particulate membrane fraction The membrane-bound fraction of the PKC isoforms is expressed as the percentage of the total (membrane + cytosol) signal from immunoblots Results are given as mean values ± SD from five to eight experiments using different hepatocyte preparations.
Agonist Culture
Percentage of membrane-bound PKC
Control 6 h 19.5 ± 7.7 51.5 ± 11.1 41.5 ± 5.8 41.2 ± 7.7
48 h 13.3 ± 11.4 37.7 ± 5.9 40.1 ± 4.0 41.7 ± 11.7 PMA 6 h 39.7 ± 11.0* 77.3 ± 10.0* 66.5 ± 12.2* 43.5 ± 7.0
48 h 44.0 ± 12.3* 67.9 ± 9.7* 54.7 ± 11.1* 37.5 ± 16.6 Insulin 6 h 20.0 ± 8.0 52.8 ± 9.6 41.6 ± 12.1 38.8 ± 2.3
48 h 20.7 ± 15.3 39.2 ± 13.3 35.6 ± 7.4 36.7 ± 10.1 ATP 6 h 27.8 ± 4.9 41.2 ± 9.6 58.8 ± 8.7 39.0 ± 9.9
48 h 19.0 ± 7.3 37.1 ± 7.4 42.7 ± 10.4 36.7 ± 7.9
* P < 0.05 vs control.
Trang 9DAG molecular species composition Second, an
involvement of PLD could not be demonstrated as
insulin-stimulated hepatocytes showed no evidence for
transphosphatidylation activity Third, we found no
evidence of translocation of PKC isoforms from the cytosol
to the membrane fraction after stimulation of hepatocytes
with insulin Fourth, insulin-stimulated cells showed no
increase in membrane-bound PKC activity and did not increase the activity of immunoprecipitated PKCf Fifth, the action of insulin on glycogen synthesis was not abolished
by the specific PKC inhibitor, bisindolylmaleimide, whereas
it completely reversed the insulin-antagonistic effect of PMA Sixth, insulin did not alter DAG mass and PKC isozyme distribution in the perfused liver
Our results are in good agreement with two previous reports showing a lack of PLD [19] and PKC [21] activation upon stimulation of rat hepatocytes with insulin In contrast, they clearly contradict several other recent studies showing insulin-mediated activation of PLD and PKC activities in hepatocyte suspensions and cultures [16–18,20,22] We suggest that this controversy may be a result of the use of different cell systems: hepatocytes in suspension often show reduced insulin responsiveness, whereas primary cultured cells can easily lose their insulin sensitivity when cultured without dexamethasone There is ample evidence that, for a number of insulin-sensitive metabolic parameters, hormone responsiveness is only retained when the cells are cultured long term in the presence of a glucocorticoid [26] Interestingly, the reports showing insulin-dependent increases in DAG, and activa-tion of PLD and/or PKC, all used cell suspensions or glucocorticoid-deprived cultures [16–18,20,22], whereas the hepatocytes used in this report were cultured in the presence
of dexamethasone A clear example of how dramatically the results may change, depending on the culture conditions, was obtained when we incubated hepatocytes in the absence
of dexamethasone; this led to insulin-dependent increases in DAG molecular species rich in stearate and arachidonate, which agree with Baldini’s data for steroid-deprived cells [17,18] However, our parallel observation, that the cells cultured under these conditions showed a dramatic reduc-tion of insulin-stimulated glycogen synthesis, casts serious doubts on the validity of these steroid-free cultures A similar controversy also exists concerning the mechanism of action of EGF in hepatocytes A review of the literature shows that EGF-dependent phospholipase activation, and
Table 3 Determination of protein kinase C (PKC) activity in crude extracts (pmolÆmin-1Æmg-1 of protein) and PKCf immunoprecipitates (pmolÆmin -1 Æmg -1
of lysate protein) Hepatocytes cultured for 6 h and 48 h were exposed to vehicle, 0.1 l M phorbol 12-myristate 13-acetate (PMA)
or 10 n M insulin for 10 min Data represent mean values ± SD from three different hepatocyte preparations.
Treatment Culture
Protein kinase C activity Crude extracts
PKCf-immunoprecipitate Cytosol Membrane
48 h 141.5 ± 33 c 249.1 ± 40 c ND
48 h 220.0 ± 16 171.7 ± 29 ND
48 h 165.5 ± 35 188.3 ± 40 2.26 ± 0.14
48 h 146.0 ± 38 199.7 ± 29 2.13 ± 0.21
a Assay with Ca 2+ and the epidermal growth factor-receptor peptide (Amersham test kit) as substrate b Assay without Ca 2+ and with peptide-e as substrate ND, not determined,cP < 0.05,dP < 0.005.
Fig 6 Sensitivity of insulin-,4b-phorbol 12-myristate 13-acetate
(PMA)- and ATP-modulated glycogen synthesis to bisindolylmaleimide.
Hepatocytes cultured for 48 h were incubated with the agonists and
the inhibitor for 2 h Data represent the mean values ± SD from three
different hepatocyte preparations.
Trang 10increases in DAG, PA, inositoltrisphosphate and cytosolic
calcium, were detected to various degrees when hepatocyte
suspensions or glucocorticoid-free hepatocyte cultures were
used [36,39,41,46,47] In contrast, in our
dexamethasone-treated cultures, EGF had no effect on DAG levels, which is
in agreement with the results of Dajani et al [38] who u sed
similar culture conditions Working with hepatocyte
sus-pensions and cultured cells, Nojiri & Hoek [47] pointed out
that EGF-induced inositoltrisphosphate formation was
effectively reduced by actin rearrangement, which occurs
during the transition of the cells from the suspended to the
cultured state As dexamethasone is known to retain
cuboidal hepatocyte morphology in cultures and to
influence actin polymerization [48], the differences in insulin
signalling (and also in EGF signalling) observed between
steroid-treated and untreated cultures might well reflect the differences in cytoskeletal cell architecture and thus point to
a major regulatory role of actin fibers in the propagation of hormone and growth factor signals The recent finding that focal adhesion kinase regulates protein kinase B, glycogen synthase kinase-3 and glycogen synthase, in an insulin-dependent manner [49], supports the hypothesis of cross-talk between insulin and integrin-signalling pathways The lack of an insu lin-elicited increase in DAG, shown here for dexamethasone-treated hepatocytes and for the perfused liver, excludes the involvement of conventional and novel PKCs, but not that of atypical PKCf in signal transduction Our results indicate that PKCf is not involved
in the activation of glycogen synthesis by insulin This finding is in good agreement with a previous report showing that a specific inhibitor of PKCf had no effect on the activation of glycogen synthase, although the authors observed the insulin-dependent activation of PKCf in their glucocorticoid-deprived hepatocyte cultures [22] Recent observations also showed insulin-mediated activation of PKCs in hepatoma cell lines [23,24,50] However, in our view, these data do not support a role for PKCs in the adult hepatic insulin signalling cascade because hepatoma cells are in an abnormal proliferative state
Recently, doubts have been raised regarding whether atypical PKCs are indeed involved in glucose transport in L6 myotubes [51] and in 3T3-L1 adipocytes transiently transfected with wild-type or mutant PKCk and f [52] Thus, activation of atypical PKCs by insulin might depend
on cell differentiation status (via culture conditions), and PKC isoforms may indirectly modulate insulin action by interfering with enzyme compartmentalization and associ-ation with the cytoskeleton
Acknowledgments
We are very grateful to Frank Rhode for his expert help with liver perfusions and we thank Dr Ralf Wimmer for the measurements of PKC distribution in liver tissue samples This work was supported by grants from the Swiss National Science Foundation (to P.B.) and the Deutsche Forschungsgemeinschaft (to I.P and U.B.).
References
1 Formisano, P & Beguinot, F (2001) The role of protein kinase C isoforms in insulin action J Endocrinol Invest 24, 460–467.
2 Farese, R.V (2002) Function and dysfunction of aPKC isoforms for glucose transport in insulin-sensitive and insulin-resistant states Am J Physiol Endocrinol Metab 283, E1–E11.
3 Lorenzo, M., Teruel, T., Hernandez, R., Kayali, A.G & Webster, N.J (2002) PLCgamma participates in insulin stimulation of glucose uptake through activation of PKCzeta in brown adipo-cytes Exp Cell Res 278, 146–157.
4 Boggs, K.P., Farese, R.V & Buse, M.G (1991) Insulin adminis-tration in vivo increases 1,2-diacylglycerol in rat skeletal muscle Endocrinology 128, 636–638.
5 Hoffman, J.M., Standaert, M.L., Nair, G.P & Farese, R.V (1991) Differential effects of pertussis toxin on insulin-stimulated phos-phatidylcholine hydrolysis and glycerolipid synthesis de novo Studies in BC3H-1 myocytes and rat adipocytes Biochemistry 30, 3315–3322.
6 Yamada, K., Standaert, M.L., Yu B., Mischak, H., Cooper, D.R.
& Farese, R.V (1994) Insulin-like effects of sodium orthovanadate
Fig 7 Insulin effects on glucose metabolism and protein kinase C
(PKC) distribution in the perfused liver After the first glucagon bolus,
livers were further perfused without agonist (C, control), with glucagon
(Ggn) (B), with insulin/glucagon (Ins/Ggn) (C) or with 1 l M phorbol
12-myristate 13-acetate (PMA; shown only for PKC distribution) (A).
Liver samples of the front large lobe were sampled at 35 min Data
represent the mean values from three different perfusions for each
agonist *P < 0.05 were indicated (A vs control) and for the insulin
values from 34 to 41 min (C vs glucagon).