Báo cáo khoa học: Use of lithium and SB-415286 to explore the role of glycogen synthase kinase-3 in the regulation of glucose transport and glycogen synthase pdf

Using phospho-specific antibodies to screen for the phosphorylation and hence activation status of these signalling molecules we observed that, unlike insulin, neither Li or SB-415286 had

Use of lithium and SB-415286 to explore the role of glycogen

synthase kinase-3 in the regulation of glucose transport

and glycogen synthase

Katrina MacAulay1, Eric Hajduch1, Anne S Blair1, Matthew P Coghlan2*, Stephen A Smith2

and Harinder S Hundal1

1

Division of Molecular Physiology, Faculty of Life Sciences, MSI/WTB Complex, University of Dundee, UK;2GlaxoSmithKline, Harlow, UK

Glycogen synthase kinase 3 (GSK3) is inactivated by insulin

and lithium and, like insulin, Li also activates glycogen

synthase (GS) via inhibition of GSK3 Li also mimics

insu-lin’s ability to stimulate glucose transport (GT), an

obser-vation that has led to the suggestion that GSK3 may

coordinate hormonal increases in GT and glycogen

synthe-sis Here we have used Li and SB-415286, a selective GSK3

inhibitor, to establish the importance of GSK3 in the

hor-monal activation of GT in terms of its effect on GS in L6

myotubes and 3T3-L1 adipocytes Insulin, Li and SB-415286

all induced a significant inhibition of GSK3, which was

associated with a marked dephosphorylation and activation

of GS In L6 myotubes, SB-415286 induced a much greater

activation of GS (6.8-fold) compared to that elicited by

insulin (4.2-fold) or Li (4-fold) In adipocytes, insulin, Li and

SB-415286 all caused a comparable activation of GS despite

a substantial differentiation-linked reduction in GSK3

expression (
85%) indicating that GSK3 remains an important determinant of GS activation in fat cells Whilst

Li and SB-415286 both inhibit GSK3 in muscle and fat cells, only Li stimulated GT This increase in GT was not sensitive

to inhibitors of PI3-kinase, MAP kinase or mTOR, but was suppressed by the p38 MAP kinase inhibitor, SB-203580 Consistent with this, phosphorylation of p38 MAP kinase induced by Li correlated with its stimulatory effect on GT Our findings support a crucial role for GSK3 in the regula-tion of GS, but based on the differential effects of Li and SB-415286, it is unlikely that acute inhibition of GSK3 contributes towards the rapid stimulation of GT by insulin

in muscle and fat cells

Keywords: adipocyte; muscle; GSK-3; insulin; p38 MAP kinase

One of the major physiological effects of insulin is to

promote the uptake, metabolism and storage of glucose in

adipose tissue and skeletal muscle [1] The hormonal

regulation of these cellular processes is initiated by the

binding of insulin to its receptor and activation of the

receptor kinase, which tyrosine phosphorylates intracellular

target substrates, in particular insulin receptor substrate 1

(IRS-1) and its relatives IRS-2 and IRS-3 [2–5] Of the

numerous IRS binding proteins, the serine/lipid kinase phosphoinositide 3-kinase (PI3K) has been implicated strongly as a component of the signalling cascade that stimulates glucose transport and glycogen synthesis [6–9] Another important component of this cascade is protein kinase B (PKB), which lies downstream of PI3K and whose activation is dependent upon phosphorylation of two key amino acid residues, Thr308 and Ser473 [10,11] 3-Phosphoinositide-dependent kinase (PDK1) phosphory-lates Thr308 [12,13], whereas phosphorylation of Ser473 is thought to be mediated by a separate, as yet unidentified, upstream kinase that has been tentatively called PDK2 [14] Activated PKB has been shown to induce the translocation of GLUT4 to the cell surface and stimulate glucose transport in muscle and fat cells [15], whereas it phosphorylates and inhibits glycogen synthase kinase-3 (GSK3) [14] GSK3 is one of several kinases that phosphorylate glycogen synthase (GS), an event that helps

to maintain the enzyme in an inactive state [16] In order to stimulate glycogen synthesis, insulin has to suppress phosphorylation and simultaneously promote the dephos-phorylation of GS via activation of glycogen-associated protein phosphatase 1 (PP1G) The greatest decrease in bound phosphate on GS has been shown to occur at sites 3a, 3b, 3c and 4 [17], which are target sites for GSK3

Correspondence to H S Hundal, Division of Molecular Physiology,

MSI/WTB Complex, University of Dundee, Dundee, DD1 5EH, UK.

Fax: + 44 1382 345507, Tel.: + 44 1382 344969,

E-mail: h.s.hundal@dundee.ac.uk

Abbreviations: GS, glycogen synthase; GSK3, glycogen synthase

kinase-3; HBS, Hepes buffered saline; HRP, horse-radish peroxidase;

IRS, insulin receptor substrate; MAPK, mitogen activated protein

kinase; a-MEM, a-minimal essential media; PDK,

3-phospho-inositide-dependent kinase; PI3K, phosphoinositide 3-kinase;

PKB, protein kinase B; PP1G, protein phosphatase 1.

*Present address: AstraZeneca, Cardiovascular and Gastrointestinal

Research Area, Mereside, Alderley Park, Macclesfield, Cheshire,

SK10 4TG, UK.

(Received 15 May 2003, revised 17 July 2003,

accepted 31 July 2003)

Thus, in addition to control by allosteric regulators the

activation status of GSK3 is likely to be a key determinant

of GS activity

Whilst there is considerable evidence implicating GSK3

in the regulation of glycogen metabolism there are

conflicting reports in the literature as to whether the kinase

also participates in the hormonal activation of glucose

transport Lithium (Li) is a widely used inhibitor of GSK3

and studies using this ion have shown that it can exert

insulin-like effects on both glycogen synthesis and glucose

uptake in insulin sensitive tissues [18–22] These

observa-tions raise the possibility that GSK3 may help to

coordi-nate increases in glucose uptake and glycogen synthesis

allowing for more effective
channelling of glucose into

glycogen in response to insulin However, one of the

potential difficulties in interpreting data from studies that

utilize Li as an inhibitor of GSK3 is that the ion also

affects the activity of a number of other molecules such as

casein kinase-2 and mitogen activated protein kinase

(MAPK)-2 [23] as well as enzymes involved in the

metabolism of glucose [24] It is difficult therefore to

exclude the possibility that the observed stimulatory effects

of Li on glucose transport may be mediated by a

mechanism that is independent of GSK3 Indeed, in

3T3-L1 adipocytes the expression of a constitutively active

form of GSK3 has no significant effect on insulin

stimulated translocation of the GLUT4 glucose transporter

and glucose transport [25] However, the value of these

findings is unclear given that the importance of GSK3 in

the regulation of glycogen metabolism in this cell type

remains poorly defined Brady et al have shown that

GSK3 activity is reduced substantially during

differenti-ation of 3T3-L1 adipocytes and have suggested that the

primary mechanism by which insulin stimulates GS in

mature adipocytes is through activation of PP1G rather

than inactivation of GSK3 [26]

In an attempt to establish the importance of GSK3 in the

acute regulation of glucose transport in terms of its

regulatory control of GS in muscle and fat cells we have

investigated the effects of Li and the anilinomaleimide,

SB-415286, a potent and highly selective inhibitor of GSK3

(Ki¼ 31 nM) [27] We demonstrate here that whilst

expres-sion of GSK3 declines substantially during differentiation of

3T3-L1 adipocytes, both Li and SB-415286 promote

activation of GS to a level comparable, if not greater, than

that elicited by insulin Furthermore, whilst both Li and

SB-415286 inactivate GSK3, our data indicate that only Li

acutely stimulates glucose transport in L6 myotubes and

3T3-L1 adipocytes

Materials and methods

Cell culture

L6 muscle cells were cultured to myotubes as described

previously [28] in a-minimal essential media (aMEM)

containing 2% (v/v) foetal bovine serum and 1% (v/v)

antimycotic/antibiotic solution (100 UÆmL)1 penicillin,

100 lgÆmL)1 streptomycin, 250 ngÆmL)1 amphotericin B)

at 37°C with 5% CO2 3T3-L1 fibroblasts (provided by

H Green, Department of Cell Biology, Harvard Medical

School, Boston, MA, USA) were differentiated into

adipocytes as described previously [29,30] Cells were cultured in 10-cm dishes for lysate preparation and in 6-well plates for glucose uptake assays Differentiated muscle cells or adipocytes were serum starved for 5 h and

3 h, respectively, before addition of appropriate reagents for times and at concentrations indicated in the figure legends

Preparation of cell lysates L6 myotubes and 3T3-L1 adipocytes were serum starved as described above Plates were washed three times with 0.9% (w/v) ice-cold saline Two-hundred lL of lysis buffer (50 mMTris pH 7.4, 0.27Msucrose, 1 mM Na-orthovana-date pH 10, 1 mM EDTA, 1 mM EGTA, 10 mM Na b-glycerophosphate, 50 mMNaF, 5 mMNa pyrophosphate, 1% (w/v) Triton X-100, 0.1% (v/v) 2-mercaptoethanol, 0.1 lMmicrocystin-LR and protease inhibitors) was added Cells were scraped off the plates using a rubber policeman and homogenized by passing through a 26-gauge hyper-dermic needle prior to centrifugation (15000 g, 4°C for

10 min) and stored at)20 °C

Glucose uptake L6 myotubes or 3T3-L1 adipocytes were serum starved as described above and incubated with Li, wortmannin, SB-415286, SB-203580, PD-98059, rapamycin, sucrose and insulin at times and concentrations indicated in figure legends Cells were washed three times with warm Hepes-buffered saline (HBS; 140 mMNaCl, 20 mMHepes, 5 mM KCl, 2.5 mM MgSO4, 1 mM CaCl2, pH 7.4) Glucose uptake was assayed by incubation of 2-deoxy-[3

H]-D-glucose (1 lCiÆmL)1, 26.2 CiÆmmol)1) for 10 min as described previously [28,31] Nonspecific binding was deter-mined by quantifying cell-associated radioactivity in the presence of 10 lMcytochalasin B Radioactive medium was aspirated prior to washing adherent cells three times with 0.9% (w/v) ice-cold saline Cells were subsequently lysed in

50 mM NaOH and radioactivity quantified using a Beck-man LS 6000IC scintillation counter Protein concentration

in cell lysates was determined using the Bradford reagent as described previously [32]

Glycogen synthase The activity of GS was assayed as described previously [31] Briefly, assay buffer (67 mMTris pH 7.5, 5 mM dithiothre-itol, 89 mM UDP-glucose, 6.7 mM EDTA, 13 mgÆmL)1 glycogen, 1 lCi per assay uridine diphospho-[6-3H]-D -glucose) was added to 45 lL cell lysate in the presence and absence of 20 mMglucose-6-phosphate After a 30-min incubation at 37°C the reaction was terminated by spotting the reaction mixture onto 31ETCHR Whatman filter paper (Whatman, Maidstone, UK) and washed three times in 66% (v/v) ethanol for 20 min Filters were finally washed in acetone and air dried before incorporation of glucose from uridine diphospho-[6-3H]-D-glucose into glycogen was quantitated using a Beckman LS6000IC scintillation coun-ter GS activity was expressed as a ratio of the activity in the absence of glucose-6-phosphate over that in the presence of the allosteric activator

Fifty lg of cell lysate protein was subjected to SDS/PAGE

on a 10% resolving gel as described previously [28]

Separated proteins were transferred onto nitrocellulose

membranes, which were subsequently blocked using NaCl/

Tris containing 0.1% (v/v) Tween 20 and 5% (w/v) milk

protein Membranes were probed with antibodies against

the phosphorylated forms of p70S6K (1 : 1000), PKB

(1 : 1000), p42/44 MAPK (1 : 1000), p38 MAP kinase

(1 : 1000), GSK-3a/b (1 : 1000) all from New England

Biolabs or to GS phosphorylated at the GSK3 (site 3)

epitope (1 : 500) or with antibodies against native PKBa

(1 : 1000) or GSK3a and b (1 : 1000) Antibodies to PKBa,

GSK3a, GSK3b and phospho GS were a gift from the

Division of Signal Transduction and Therapy, University of

Dundee cMyc antibodies were from Sigma The membranes

were washed three times in NaCl/Tris/0.1% Tween 20 (v/v)

for 15 min prior to incubation with horseradish peroxidase

(HRP) anti-rabbit IgG (1 : 1000), HRP anti-mouse IgG

(1 : 1000) or HRP anti-sheep/goat IgG (1 : 500, all from

Sigma) Protein bands on nitrocellulose were visualized using

enhanced chemiluminescence by exposure to Konica

Medi-cal Film (Konica Corporation, Hohenbrunn, Germany)

GSK3 assay

L6 myotubes were deprived of serum for 4 h in a-MEM and

washed twice with warm HBS Cells were incubated

subsequently at 37°C in HBS/25 mM D-glucose for 1 h

During the last hour insulin and or wortmannin were added

at times and concentrations indicated in the figure legends

prior to cell lysis Myotubes were extracted from 10-cm

dishes using ice-cold lysis buffer GSK3a, GSKb or

myc-tagged GSK3S9Awere immunoprecipitated from 100 lg cell

lysate and incubated with or without 25 mUÆmL)1PP2A1

prior to assay using phospho-GS peptide-1 as substrate [33]

Cell transfection

L6 cells were transfected with pSG5 vector, which encodes

resistance to G418 sulphate cDNA encoding myc-tagged

GSK-3b in which serine 9 was mutated to an alanine was

subcloned into the pSG5 vector Phosphorylation of the

serine 9 site on GSK-3b is considered important for its

inactivation by insulin and thus mutation of this site to an

alanine (S9A) renders the kinase constitutively active

Con-trol cells were transfected with the empty vector lacking the

GSK-3bS9AcDNA L6 cells transfected with GSK3S9Awere

cultured as described earlier, but with the addition of

0.8 mgÆmL)1G418 sulphate to the media at all stages to

select for transformed cells Transfected cells were used for

analysis of glucose uptake and GSK3 activity as described

earlier

Statistical analyses

For multiple comparisons statistical analysis was performed

using one-way analysis of variance (ANOVA) followed by a

Newman–Keuls post-test Data analysis was performed

usingGRAPHPAD PRISMsoftware and considered statistically

significant at P values < 0.05

Results and discussion

Effects of insulin, Li and SB-415286 on GSK3 activity

As an initial starting point for our studies we investigated the effects of insulin, Li and the maleimide, SB-415286, on GSK3 activity from L6 myotubes Insulin caused a significant inactivation (by up to 40%) of both GSK3 isoforms, which was blocked by prior treatment of cells with the PI3-kinase inhibitor wortmannin (Fig 1A) Because both Li and SB-415286 inhibit GSK3 by competitively blocking Mg and ATP binding, respectively [27,34], it was not possible to directly determine the effect of these inhibitors on cellular GSK3 activity in vivo However, both

Li (50 mM) and SB-415286 (50 lM) induced a substantial suppression of immunoprecipitated GSK3 activity when they were included in the in vitro kinase assay by 73% and 97%, respectively Identical results were also obtained with SB-216763 (data not shown), a structurally unrelated maleimide, which, like SB-415286, also exhibits selectively

Fig 1 Effects of insulin and wortmannin on GSK3 a and b activity in L6 myotubes and relative abundance of GSK3 isoforms in L6and 3T3-L1-adipocytes (A) L6 myotubes were pretreated for 10 min with either

100 n M insulin alone or with 100 n M wortmannin for 15 min before exposing cells to insulin Following these incubations cells were lysed and GSK3 a or b immunoprecipitated for analysis of kinase activity GSK3 activity was expressed as a re-activation ratio (i.e GSK3 activity measured without PP2A 1 treatment divided by GSK3 activity after PP2A 1 treatment) Values are the mean ± SEM for three experiments carried out in duplicate The asterisk signifies a statistically significant change from the untreated sample (P < 0.01) (B) Lysates (50 lg protein) from L6 myoblasts, L6 myotubes, 3T3-L1 fibroblasts and 3T3-L1 adipocytes were immunoblotted using antibodies against GSK3a and b.

for GSK3 [27] An attempt was made to assess the effects of

insulin on GSK3 activity in 3T3-L1 adipocytes, but proved

technically difficult as kinase activity in immunoprecipitates

from unstimulated fat cells was found to be extremely low

To establish why this may be so we immunoblotted lysates

from 3T3-L1 preadipocytes, fully differentiated 3T3-L1

adipocytes as well as L6 myoblasts and myotubes with

antibodies against GSK3a and b Fig 1B shows that whilst

preadipocytes express both GSK3 isoforms, the abundance

of the b isoform declines by
85%, whereas that of the a

isoform is virtually undetectable in differentiated adipocytes

In contrast, such a loss was not observed during

differenti-ation of L6 muscle cells, which, if anything, showed a

marginal increase in GSK3 abundance during

differenti-ation Our inability to detect GSK3 activity in differentiated

adipocytes is at odds with the study of Orena et al [21] in

which the authors reported the presence of significant GSK3

activity The reasons for this discrepancy are unclear, but

the assay protocol used in the present study relied upon

measuring kinase activity in GSK3 immunoprecipitates,

whereas that of Orena et al utilized whole cell extracts to

monitor phosphorylation of a primed GSK3 peptide

substrate [21] It is conceivable that this technical difference

may give rise to the apparent discrepancy between the two

studies Nevertheless, it should be stressed that the marked

decline in GSK3 expression that we observe in 3T3-L1

adipocytes is fully consistent with previous data showing that

GSK3 activity diminishes substantially during adipogenesis

of 3T3-L1 adipocytes [26] thereby helping to explain the low

immunoprecipitable activity that we observe in our hands

Effects of insulin, Li and SB-415286 on signalling

elements implicated in the regulation of GSK3 and GS

To further understand the effects of Li and SB-415286 on

cell signalling events we assessed their effects and that of

insulin on p70S6K, PKB, p42/p44 MAP kinases, GSK3 and

GS PKB is considered to be the upstream inactivator of

GSK3 in vivo [35], but evidence also exists showing that the

latter can be targeted by p70S6K and the classical MAP

kinase pathway in response to nutrients and certain growth factors [33,36,37] Using phospho-specific antibodies to screen for the phosphorylation and hence activation status

of these signalling molecules we observed that, unlike insulin, neither Li or SB-415286 had any detectable effect on the phosphorylation of PKBSer473, p70S6K, p42/p44 MAP kinases or GSK3 in L6 myotubes or 3T3-L1 adipocytes (Fig 2) It is noteworthy that in L6 myotubes insulin induces phosphorylation of both GSK3 a and b, whereas in 3T3-L1 adipocytes we observed only a single phospho-band that correlates with that of GSK3b The lack of an equivalent GSK3a phospho-signal is consistent with the virtual absence of this isoform in our 3T3-L1 adipocytes (Fig 1B) As indicated earlier, SB-415286 potently inhibits GSK3 by an ATP competitive mechanism [27] Conse-quently, this compound did not affect insulin’s ability to induce phosphorylation of GSK3 or that of PKB, p70S6K and the p42/p44 MAP kinases in response to insulin (Fig 2) In unstimulated cells, GS is phosphorylated on site

3 by GSK3 and indeed a phospho-antibody directed against site 3 confirmed that this was the case in muscle and fat cells (Fig 2) The adipocyte data suggests that despite the substantial reduction in GSK3 activity and expression that occurs during differentiation of 3T3-L1 cells ([26] and Fig 1B), sufficient GSK3 activity remains in these cells to induce phosphorylation of GS on site 3 GS phosphoryla-tion fell significantly upon treating both muscle and fat cells with insulin, and was undetectable following incubation of either cell type with Li or SB-415286 (Fig 2) As site 3 phosphorylation can be taken as a downstream read out of GSK3 activity the observation that both Li and SB-415286 induce a complete abolition of GS phosphorylation on this site reflects that both agents cause a far greater inhibition of GSK3 than that elicited by insulin

Regulation of GS activity

To establish the importance of GSK3 inhibition on GS activity we monitored the effects of insulin, Li, SB-415286 and wortmannin (a PI3K inhibitor) on the incorporation of

Fig 2 Representative immunoblots showing the effects of insulin, SB-415286and lithium on the phosphorylation status of key signalling molecules (A) L6 myotubes and (B) 3T3-L1 adipocytes were pretreated for 60 min with

50 l M SB-415286 or 50 m M lithium prior to a 10-min incubation of cells with 100 n M insulin Lysates (50 lg protein) were immunoblotted using phospho-specific antibodies against p70S6K, PKB, p42/44 MAP kinases, GSK3a/b and GS Equal loading of cell lysate protein was determined by probing with an antibody to native PKBa The blots are representative from up to four separate experiments.

labelled UDP-glucose into glycogen in the absence and

presence of glucose-6-phosphate (the allosteric activator of

GS) Insulin stimulated GS activity in both L6 myotubes

and 3T3-L1 adipocytes by 4.2- and 2.5-fold, respectively

(Fig 3A) This stimulation was reduced significantly in both

cell lines by wortmannin, suggesting that activation of PI3K

precedes that of GS This proposition is consistent with the

observation that the inhibition of GSK3 (mediated by PKB)

and the activation of PP1 in response to insulin are both

PI3K-dependent processes in L6 muscle cells and 3T3-L1

adipocytes [10,26] GS was also activated by Li and

SB-415286 in both cell types, but, unlike insulin, activation of

the enzyme in response to these stimuli was not sensitive to

wortmannin (Fig 3) This finding is compatible with the

suggestion that Li and SB-415286 target GSK3 directly and

that inhibition of the kinase by these agents does not rely

upon activation of upstream signalling molecules, such as

PI3K or PKB (Fig 2) Collectively, these findings suggest

strongly that targeted inactivation of GSK3, using either Li

or SB-415286, is sufficient to induce activation of GS to a level similar or greater than that by insulin Since the activity

of GS depends on the relative activities of GS kinases and PP1G, inhibiting GSK3 (one of the principal GS kinases) will shift the balance towards dephosphorylation and activation of GS The notion that GSK3 is critical for glycogen metabolism is strengthened further by our finding that despite the significant decline in GSK3 abundance during differentiation of 3T3-L1 adipocytes, selective inhi-bition of this kinase, using SB-415286, appears to mimic the hormonal activation of GS in this cell type Consequently, whilst PP1G is likely to play a significant role in the hormonal activation of GS in 3T3-L1 adipocytes, the importance of GSK3 in the insulin-mediated regulation of this enzyme in fat cells should not be readily discounted [26] Moreover, it is also noteworthy that an analysis of the GS activity ratio in 3T3-L1 preadipocytes reveals that in unstimulated cells, basal GS activity was
80% lower than that measured in differentiated adipocytes This lower GS activity is fully concordant with the much higher level of GSK3 expression that prevails in preadipocytes

Is GSK3 a regulator of glucose transport

in insulin-responsive cells?

The potential involvement of GSK3 in the regulation of glucose transport remains unclear at present Two recent studies have suggested that acute inhibition of GSK3 using

Li or long-term suppression of the kinase using inhibitors that exhibit selectivity towards GSK3, enhance glucose uptake in muscle and fat cells [21,38] In contrast, another study expressing a constitutively active form of GSK3 reported no significant changes in insulin-stimulated glucose uptake or GLUT4 translocation, although a slight reduc-tion in basal glucose uptake was noted [25] In an attempt to clarify this matter we investigated the effects of both Li and SB-415286 on basal and insulin-stimulated glucose uptake

in L6 myotubes and 3T3-L1 adipocytes Fig 4 shows that insulin enhances glucose uptake in both muscle and fat cells

by 2- and 3.4-fold, respectively When both cell types were exposed to Li, at a concentration that inhibits GSK3, glucose uptake was stimulated by
1.8 fold (L6 myotubes) and 2.6 fold (3T3-L1 adipocytes) (Fig 4A and B) In contrast, however, incubation of muscle and fat cells with

50 lM SB-415286, circumstances during which there is a substantial inhibition of GSK3 (based on analysis of site 3

GS phosphorylation, Fig 2) and an attendant activation of

GS (Fig 3), did not elicit any change in basal or insulin-stimulated glucose uptake (Fig 4) Since both Li and

SB-415286 inhibit GSK3, but only one of these stimulates glucose uptake the findings imply that GSK3 may not have any significant regulatory input into the acute activation of glucose transport by insulin in our experimental system These observations are, to some extent, consistent with the recent work of Henriksen et al [39] who reported that whilst acute inhibition of GSK3 with Li enhanced glucose uptake in skeletal muscle of lean Zucker rats, inhibition of the kinase using a selective organic inhibitor (CT 98014) had no stimulatory or insulin potentiating effect on glucose uptake This inhibitor also failed to stimulate glucose transport in skeletal muscle of Zucker diabetic rats, but

Fig 3 Effects of insulin, wortmannin, lithium and SB-415286on GS

activity (A) L6 myotubes and (B) 3T3-L1 adipocytes/preadipocytes

were pretreated for 5 min with 100 n M wortmannin prior to treatment

with 100 n M insulin (10 min), 50 m M lithium (60 min) and 50 l M

SB-415286 (60 min) Glycogen synthase activity was determined by

assaying incorporation of glucose from uridine diphospho-[6- 3 H]- D

-glucose into glycogen and expressed as a ratio of the activity in the

absence divided by that in the presence of glucose-6-phosphate Values

are the mean ± SEM for three experiments each carried out in

duplicate.

interestingly potentiated the effects of insulin on muscle

glucose uptake in these animals This potentiation was

associated with an increase in sarcolemmal GLUT4 content

following insulin-treatment of muscle Precisely how

inhi-bition of GSK3 under these circumstances leads to an

increase in cell surface GLUT4 still remains poorly defined

However, given that GSK3 activity is thought to be

enhanced in insulin-resistant muscle and the kinase has

been implicated in down-regulating insulin signalling via its

ability to serine phosphorylate IRS1 [40], it is possible that

inhibition of GSK3 potentiates insulin signalling at the level

of proteins such as IRS1 This possibility is supported by the

observations of Nikoulina et al [38] who found that whilst

acute inhibition of GSK3 had no stimulatory effect on glucose uptake in cultured human muscle cells, sustained inhibition of GSK3 (over 96 h) led to an increase in both basal and insulin-stimulated sugar uptake This adaptive increase in glucose uptake could not be linked to alterations

in cellular GLUT4 expression, but was associated with changes in the abundance of both IRS1 and GSK3, which were elevated and repressed, respectively Whether induction of IRS1 is sufficient to elicit the increase in glucose uptake reported by Nikoulina et al [38] remains unclear at present, given that phosphorylation of PKB/Akt,

a kinase implicated in the hormonal regulation of glucose transport [15], was unaffected by prolonged inhibition of GSK3

To assess whether SB-415286 may have an insulin sensitizing effect in our muscle cell system we compared the effects of the maleimide on the phosphorylation of PKB and GSK3 and upon the stimulation of glucose transport in response to a submaximal and maximal insulin concentra-tion Fig 5 shows that insulin induces phosphorylation of both PKB and GSK3, and modestly stimulates glucose uptake at submaximal concentrations, although the responses were clearly lower than that observed in response

Fig 4 Effects of insulin, SB-415286and lithium on glucose transport.

(A) L6 myotubes and (B) 3T3-L1 adipocytes were pretreated with

50 l M SB-415286 or 50 m M lithium for 60 min prior to a 30-min

sti-mulation with 100 n M insulin and analysis of 2-deoxyglucose uptake.

Values are the mean ± SEM for three experiments carried out in

triplicate, asterisks signify statistically significant changes from the

untreated sample (P < 0.01).

Fig 5 Effects of SB-415286on the phosphorylation of PKB and GSK3 and the stimulation of glucose transport induced by submaximal and maximal insulin treatments in L6muscle cells L6 myotubes were pre-incubated with 50 l M SB-415286 for 60 min prior to incubation with insulin (1 n M or 100 n M ) for 10 min (for phospho-blots) or for 30 min (uptake assays) Cells were lysed and 50 lg lysate protein was immunoblotted using phospho-specific antibodies against PKB and GSK3a/b As a loading control, lysates were immunoblotted with an antibody to native PKBa The blots are representative from up to three separate experiments Alternatively cells following insulin treatment were assayed for 2-deoxyglucose uptake Values are the mean ± SEM for three experiments carried out in triplicate, asterisks signify statis-tically significant changes from the untreated sample (P < 0.01).

to a maximally effective insulin dose Pre-incubating L6 cells

with SB-415286 did not enhance the phosphorylation of

either kinase nor did it increase sugar uptake in response to a

submaximal insulin dose (Fig 5) These findings are not

entirely out of line with work from rodent studies showing

that whilst GSK3 inhibition improves insulin responsiveness

in muscle of insulin resistant animals it had no insulin

potentiating effect in skeletal muscle of lean animals [39,41]

To assess whether chronic inhibition of GSK3 modifies

glucose uptake in muscle cells, we incubated L6 myotubes

chronically with SB-415286 prior to analysis of basal and

insulin-stimulated glucose uptake However, it proved

tech-nically difficult to extend the incubation period beyond 24 h

as the integrity and plate-adherent properties of terminally

differentiated myotubes was severely compromised

Never-theless, we observed that sustained exposure of L6 myotubes

to 50 lMSB-415286 for 24 h led to a small, but significant

enhancement in basal glucose uptake (basal untreated,

32.9 ± 2.8 pmolÆmin)1 per mg protein)1; basal treated

47.5 ± 4.4 pmolÆmin)1 per mg protein)1, values are

mean ± SEM from three observations) However,

irres-pective of whether cells were exposed to SB-415286, we did

not observe any potentiation in insulin stimulated glucose

uptake (insulin treatment alone, 56.7 ± 2.2 pmolÆmin)1per

mg protein)1; insulin + SB-415286, 54.7 ± 6.1 pmolÆmin)1

per mg protein)1, values are mean ± SEM from three

observations) The precise mechanism underlying the

observed increase in basal glucose uptake remains poorly understood, but it is plausible that changes in the cellular expression of proteins regulating this process may contribute

to this phenomena as reported by Nikoulina et al [38]

An important question that emerges from these studies concerns the mechanism by which Li stimulates glucose transport in muscle and fat cells To gain some insight into this issue we subsequently monitored the effects of a number

of inhibitors that target PI3K, the MAP kinase pathway, p38 MAP kinase and mTOR on Li-stimulated glucose uptake in L6 myotubes In line with previous work [28], Fig 6 shows that wortmannin (a PI3K inhibitor) suppresses basal glucose uptake by
50% and induces a complete inhibition of insulin-stimulated glucose transport This latter finding is in full agreement with the widely accepted belief that PI3K plays a critical role in the hormonal regulation of glucose transport [1] However, despite the fall

in basal glucose uptake the net stimulation in glucose uptake elicited by Li was largely unaffected by wortmannin implying that PI3K was not involved in this regulatory response Similar analyses, using PD-98059 and rapamycin, excluded the involvement of the classical MAP kinase pathway and mTOR, respectively (Fig 6A) However, the acute stimulation of glucose uptake by Li was virtually abolished in the presence of SB-203580, which inhibits p38 MAP kinase [31] Interestingly, whilst SB-203580 blocked Li-stimulated glucose transport it had no effect on the ion’s

Fig 6 Effects of wortmannin, SB-203580, PD-98059 and rapamycin on insulin and lithium (Li)-stimulated glucose uptake and GS activation (A) L6 myotubes were pretreated for 5 min with 100 n M wortmannin, 10 l M SB-203580, 10 l M PD-98059 or 10 l M rapamycin prior to cell stimulation with 100 n M insulin for 30 min or 50 m M Li for 60 min Inhibitors were present throughout the period of incubation with insulin and Li At the end

of these incubation periods 2-deoxyglucose was assayed as described in Materials and methods Values are the mean ± SEM for three experiments each performed in triplicate, asterisks signify statistically significant changes from the untreated sample (P < 0.01), whereas the double asterisk indicates a significant change from the wortmannin-treated sample (P < 0.01) (B) For GS activity, cells were treated with 50 m M Li or 10 l M

SB-203580 for 60 min or with 100 n M insulin for 10 min prior to assaying incorporation of glucose from uridine diphospho-[6- 3 H]- D -glucose into glycogen and expressed as a ratio of the activity in the absence divided by that in the presence of glucose-6-phosphate Values are the mean ± SEM for three experiments each carried out in duplicate, single asterisks signify statistically significant changes from the untreated sample, whereas the double asterisk signifies a significant change to the wortmannin-treatment alone (P < 0.01).

ability to induce a stimulation of GS (presumably via

inhibition of GSK3) in muscle cells (Fig 6B) This latter

finding adds further support to the argument that the

increase in glucose uptake elicited by Li is likely to be

mediated by a mechanism that is distinct from that used to stimulate GS

The observation that SB-203580 suppresses Li-stimulated glucose uptake implies that Li stimulates the p38 MAP kinase pathway The notion that Li activates this stress signalling pathway is not unprecedented Li has been shown

to acutely activate p38 MAP kinase in a human intestinal epithelial cell line, HT-29, and promote the transcription of the interleukin-8 gene [42] In line with previous studies, Fig 7A shows that Li induced the phosphorylation/activa-tion of p38 MAP kinase in L6 muscle cells and that, like the stimulation of glucose uptake, this was suppressed by SB-203580, but not by wortmannin Fig 7B shows that phosphorylation of p38 MAPK was induced by Li in a dose-dependent manner with maximal phosphorylation being induced in response to 50 mMLi At this concentra-tion, Li also maximally stimulated glucose uptake in muscle cells (Fig 7B) It is conceivable that the use of Li at the high concentrations that are used typically to inhibit GSK3 may stimulate glucose transport as a result of an increase in extracellular osmolarity However, the finding that equi-valent concentrations of sucrose fail to elicit any significant increase in glucose uptake would tend to negate this possibility (Fig 7B)

To further investigate whether inactivation of GSK3 has any regulatory input into the stimulation of glucose uptake

by insulin we expressed a constitutively active Myc-tagged form of GSK3b in L6 cells in which serine 9 was mutated to

an alanine (S9A) Immunoprecipitation and immunoblot-ting using GSK3b or Myc antibodies confirmed the

Fig 7 Li induces p38 MAPK phosphorylation and stimulates glucose

uptake in L6myotubes in a dose-dependent manner (A) L6 myotubes

were stimulated with 100 n M insulin for 10 min or pretreated with

100 n M wortmannin or 10 l M SB-203580 for 5 min prior to a 60-min

incubation with 50 m M Li At the end of this incubation cells were

lysed and 50 lg of lysate protein immunoblotted using antibodies

against phospho-p38 MAP kinase The same blot was reprobed with

an antibody to native PKB to establish equal loading of protein in the

different sample lanes (B) L6 myotubes were incubated with 20 m M ,

50 m M or 100 m M Li or sucrose for 60 min At the end of this

incu-bation cells were lysed and 50 lg lysate protein were resolved by SDS/

PAGE and immunoblotted with a phospho-specific antibody against

p38 MAPK Alternatively, at the end of the 60 min incubation cells

were used for assaying glucose uptake Values are mean ± SEM from

three experiments each performed in triplicate, asterisks signify

sta-tistically significant changes from the appropriate sucrose treatment

(P < 0.01) The immunoblots are representative from three similar

experiments.

Fig 8 Effects of insulin on glucose uptake in L6cells expressing a constitutively active GSK3 S9A L6 cells were transfected with myc-tagged GSK3S9Awhich was immunoprecipitated using antibodies to either c-myc or GSK3b and the immunoprecipitate probed with the reciprocal antibody L6 cells transfected with the empty expression vector (L6-EV) were used as a control L6-EV or GSK3S9Aexpressing cells were incubated with 100 n M insulin for 30 min prior to assaying glucose uptake The uptake values are mean ± SEM for three experiments, each conducted in triplicate Asterisks signify a significant change from the respective basal value (P < 0.05).

expression of GSK3S9Ain L6 cells Whilst insulin

inacti-vated GSK3b from L6 cells transfected with the empty

expression vector by 46 ± 5% (mean ± SEM of four

experimental observations) the hormone failed to induce

any inhibition of the kinase when immunoprecipitated from

GSK3S9Aexpressing cells (data not shown) Consistent with

this observation, insulin did not stimulate GS in cells

expressing the GSK3S9Amutant [activity ratios (±

glucose-6-phosphate) for GS in the absence and presence of insulin

in control cells were 0.029 ± 0.01 (basal), 0.67 ± 0.01

(insulin), and in GSK3S9Aexpressing cells were 0.06 ± 0.03

(basal), 0.05 ± 0.01 (insulin)] Nevertheless, when

GSK3S9Aexpressing cells were stimulated with insulin and

glucose uptake assayed we observed no significant

differ-ences in sugar uptake compared with cells transfected with

the empty vector (Fig 8) This observation is consistent

with our pharmacological data and is in line with previous

work by Summers et al who reported that whilst expression

of a GSK3S9Ain 3T3-L1 adipocytes reduced basal glucose

uptake slightly it failed to influence insulin’s ability to

acutely stimulate glucose uptake or GLUT4 translocation

in this cell line [25]

In summary, we have shown that suppressing GSK3

activity in L6 myotubes and 3T3-L1 adipocytes, using Li or

SB-415286, is capable of stimulating GS to a level that is

comparable to that observed in response to insulin

However, whilst clearly important for the hormonal

regu-lation of GS, our data does not support a role for GSK3 in

the acute regulation of glucose transport based on (a) the

differential effects of Li and SB-415286 on hexose uptake

and (b) the inability of a constitutively active GSK3 to

modulate insulin-stimulated glucose uptake Nevertheless,

given that inhibition of GSK3 (using Li or SB-415286)

appears to be sufficient for inducing activation of GS in

muscle and fat cells, and that inhibition of the kinase

potentiates insulin action in muscle of insulin-resistant rats

[39], and that prolonged GSK3 inhibition not only enhances

basal glucose uptake but elevates IRS1 expression [38]

suggests that long-term manipulation of GSK3 may be of

therapeutic value in improving glucose utilization and

sensitivity of muscle and adipose tissue to insulin

Acknowledgements

We are grateful to our colleagues in the MRC Protein Phosphorylation

Unit and the DSTT for providing some of the reagents used in this

study We also thank D J Powell for technical help and useful

discussions This work was supported by the MRC, BBSRC, Diabetes

and Wellness Research Foundation, Diabetes UK and

GlaxoSmith-Kline K M is supported by a BBSRC studentship and A B was

supported by a MRC-CASE studentship.

References

1 Litherland, G.J., Hajduch, E & Hundal, H.S (2001) Intracellular

signalling mechanisms regulating glucose transport in

insulin-sensitive tissues Mol Memb Biol 18, 195–204.

2 White, M.F & Kahn, C.R (1994) The insulin signaling system.

J Biol Chem 269, 1–4.

3 Araki, E., Lipes, M.A., Patti, M.E., Bruning, J.C., Haag, B.,

Johnson, R.S & Kahn, C.R (1994) Alternative pathway of insulin

signaling in mice with targeted disruption of the IRS-1 gene.

Nature 372, 186–190.

4 Hansen, P.A., Corbett, J.A & Holloszy, J.O (1997) Phorbol esters stimulate muscle glucose transport by a mechanism distinct from the insulin and hypoxia pathways Am J Physiol 36, E28–E36.

5 Kaburagi, Y., Satoh, S., Tamemoto, H., Yamamoto-Honda, R., Tobe, K., Veki, K., Yamauchi, T., Kono-Sugita, E., Sekihara, H., Aizawa, S., Cushman, S.W., Akanuma, Y., Yazaki, Y & Kadowaki, T (1997) Role of insulin receptor substrate-1 and pp60

in the regulation of insulin-induced glucose transport and glut4 translocation in primary adipocytes J Biol Chem 272, 25839– 25844.

6 Cheatham, B., Vlahos, C.J., Cheatham, L., Wang, L., Blenis, J & Kahn, C.R (1994) Phosphatylinositol 3-kinase activation is required for insulin stimulation off pp70, s6 kinase, DNA synth-esis, and glucose transporter translocation Mol Cell Biol 14, 4902–4911.

7 Tsakiridis, T., McDowell, H.E., Walker, T., Downes, C.P., Hundal, H.S., Vranic, M & Klip, A (1995) Multiple roles of phosphatidylinositol 3-kinase in regulation of glucose-transport, amino-acid-transport, and glucose transporters in L6 skeletal-muscle cells Endocrinology 136, 4315–4322.

8 Quon, M.J., Chen, H., Ing, B.L., Liu, M.L., Zarnowski, M.J., Yonezawa, K., Kasuga, M., Cushman, S.W & Taylor, S.I (1995) Roles of 1-phosphatidylinositol 3-kinase and ras in regulating translocation of glut4 in transfected rat adipose-cells Mol Cell Biol 15, 5403–5411.

9 Yang, J., Clarke, J.F., Ester, C.J., Young, P.W., Kasuga, M & Holman, G.D (1996) Phosphatidylinositol 3-kinase acts at an intracellular membrane site to enhance GLUT4 exocytosis in 3T3-L1 cells Biochem J 313, 125–131.

10 Cross, D.A.E., Alessi, D.R., Cohen, P., Andjelkovic, M & Hemmings, B.A (1995) Inhibition of glycogen synthase kinase-3

by insulin-mediated protein kinase B Nature 378, 785–789.

11 Alessi, D.R., Andjelkovic, M., Caudwell, B., Cron, P., Morrice, N., Cohen, P & Hemmings, B.A (1996) Mechanism of activation

of protein kinase B by insulin and IGF-1 EMBO J 15, 6541– 6551.

12 Alessi, D.R., James, S.R., Downes, C.P., Holmes, A.B., Gaffney, P.R.J., Reese, C.B & Cohen, P (1997) Characterization of a 3-phosphoinositide-dependent protein kinase which phosphory-lates and activates protein kinase Ba Curr Biol 7, 261–269.

13 Stokoe, D., Stephens, L.R., Copeland, T., Gaffney, P.R.J., Reese, C.B., Painter, G.F., Holmes, A.B., McCormick, F & Hawkins, P.T (1997) Dual role of phosphatidylinositol-3,4,5-trisphosphate

in the activation of protein kinase B Science 277, 567–570.

14 Cohen, P., Alessi, D.R & Cross, D.A.E (1997) The tenth datta lecture – PDK1, one of the missing links in insulin signal trans-duction? FEBS Lett 410, 3–10.

15 Hajduch, E., Litherland, G.J & Hundal, H.S (2001) Protein kinase B: a key regulator of glucose transport? FEBS Lett 492, 199–203.

16 Lawrence, J.C & Roach, P.J (1997) New insights into the role and mechanism of glycogen synthase activation by insulin Diabetes 46, 541–547.

17 Parker, P.J., Caudwell, F.B & Cohen, P (1983) Glycogen syn-thase from rabbit skeletal muscle – effects of insulin on the state of phosphorylation of the 7 phosphoserine residues in vivo Eur J Biochem 130, 227–243.

18 Cheng, K., Creacy, S & Larner, J (1983)
Insulin-like effects

of lithium ion on isolated rat adipocytes I Stimulation of glycogenesis beyond glucose transport Mol Cell Biochem 56, 177–182.

19 Chen, X., McMahon, E.G & Gulve, E.A (1998) Stimulatory effect of lithium on glucose transport in rat adipocytes is not mediated by elevation of IP1 Am J Physiol 275, E272–E277.

20 Tabata, I., Schluter, J., Gulve, E.A & Holloszy, J.O (1994)

Lithium increases susceptibility of muscle glucose transport to

stimulation by various agents Diabetes 43, 903–907.

21 Orena, S.J., Torchia, A.J & Garofalo, R.S (2000) Inhibition of

glycogen-synthase kinase 3 stimulates glycogen synthase and

glucose transport by distinct mechanisms in 3T3-L1 adipocytes.

J Biol Chem 275, 15765–15772.

22 Furnsinn, C., Noe, C., Herdlicka, R., Roden, M., Nowotny, P.,

Leighton, B & Waldhausl, W (1997) More marked stimulation

by lithium than insulin of the glycogenic pathway in rat skeletal

muscle Am J Physiol 273, E514–E520.

23 Cohen, P (2001) The role of protein phosphorylation in human

health and disease The Sir Hans Krebs Medal Lecture Eur.

J Biochem 268, 5001–5010.

24 Bosch, F., Rodriguez-Gil, J.E., Hatzoglou, M., Gomez-Foix,

A.M & Hanson, R.W (1992) Lithium inhibits hepatic

gluco-neogenesis and phosphoenolpyruvate carboxykinase gene

expression J Biol Chem 267, 2888–2893.

25 Summers, S.A., Kao, A.W., Kohn, A.D., Backus, G.S., Roth,

R.A., Pessin, J.E & Birnbaum, M.J (1999) The role of glycogen

synthase kinase 3 beta in insulin-stimulated glucose metabolism.

J Biol Chem 274, 17934–17940.

26 Brady, M.J., Bourbonais, F.J & Saltiel, A.R (1998) The

activa-tion of glycogen synthase by insulin switches from kinase

inhibi-tion to phosphatase activainhibi-tion during adipogenesis in 3T3-L1

cells J Biol Chem 273, 14063–14066.

27 Coghlan, M.P., Culbert, A.A., Cross, D.A.E., Holder, J.C.,

Yates, J.W., Pearce, N.J., Rausch, O.L., Murphy, G.J., Carter,

P.S., Roxbee Cox, L., Mills, D., Brown, M.J., Haigh, D., Ward,

R.W., Smith, D.G., Murray, K.J & Reith, A.D (2000) Selective

small molecule inhibitors of glycogen synthase kinase-3 modulate

glycogen metabolsim and gene transcription Chem Biol 24, 1–11.

28 Hajduch, E., Alessi, D.R., Hemmings, B.A & Hundal, H.S.

(1998) Constitutive activation of protein kinase Ba (PKBa) by

membrane targeting promotes glucose and System A amino acid

transport, protein synthesis and GSK3 inactivation in L6 muscle

cells Diabetes 47, 1006–1013.

29 Frost, S.C & Lane, M.D (1985) Evidence for the involvement of

vicinal sulfhydryl groups in insulin- activated hexose transport by

3T3-L1 adipocytes J Biol Chem 260, 2646–2652.

30 Green, H & Kehinde, O (1975) An established preadipose cell

line and its differentiation in culture II Factors affecting the

adipose conversion Cell 5, 19–27.

31 Blair, A.S., Hajduch, E., Litherland, G.J & Hundal, H.S (1999)

Regulation of glucose transport and glycogen synthesis in L6

muscle cells during oxidative stress: Evidence for cross-talk

between the insulin and SAPK2/p38 MAP kinase signalling

pathways J Biol Chem 274, 36293–36299.

32 Bradford, M.M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding Anal Biochem 72, 248–254.

33 Cross, D.A.E., Alessi, D.R., Vandenheede, J.R., McDowell, H.E., Hundal, H.S & Cohen, P (1994) The inhibition of glycogen synthase kinase-3 by insulin or IGF-1 in the rat skeletal muscle cell line L6 by wortmannin, but not rapamycin: evidence that wort-mannin blocks activation of the MAP kinase pathway in L6 cells between Ras and Raf Biochem J 303, 21–26.

34 Ryves, W.J & Harwood, A.J (2001) Lithium inhibits glycogen synthase kinase-3 by competition for magnesium Biochem Bio-phys Res Commun 280, 720–725.

35 Lawlor, M.A & Alessi, D.R (2001) PKB/Akt: a key mediator of cell proliferation, survival and insulin responses? J Cell Sci 114, 2903–2910.

36 Peyrollier, K., Hajduch, E., Blair, A.S., Hyde, R & Hundal, H.S (2000) L -Leucine availability regulates phosphatidylinositol 3-kinase, p70, S6 kinase and glycogen synthase kinase-3 activity in L6 muscle cells: evidence for the involvement of the mammalian target of rapamycin (mTOR) pathway in the L -leucine-induced up-regulation of System A amino acid transport Biochem J 350, 361–368.

37 Armstrong, J.L., Bonavaud, S.M., Toole, B.J & Yeaman, S.J (2001) Regulation of glycogen synthesis by amino acids in cultured human muscle cells J Biol Chem 276, 952–956.

38 Nikoulina, S.E., Ciaraldi, T.P., Mudaliar, S., Carter, L., Johnson,

K & Henry, R.R (2002) Inhibition of glycogen synthase kinase 3 improves insulin action and glucose metabolism in human skeletal muscle Diabetes 51, 2190–2198.

39 Henriksen, E.J., Kinnick, T.R., Teachey, M.K., O’Keefe, M.P., Ring, D., Johnson, K.W & Harrison, S.D (2003) Modulation of muscle insulin resistance by selective inhibition of GSK-3 in Zucker diabetic fatty rats Am J Physiol 284, E892–E900.

40 Eldar-Finkelman, H & Krebs, E.G (1997) Phosphorylation of insulin receptor substrate 1 by glycogen synthase kinase 3 impairs insulin action Proc Natl Acad Sci USA 94, 9660–9664.

41 Ring, D.B., Johnson, K.W., Henriksen, E.J., Nuss, J.M., Goff, D., Kinnick, T.R., Ma, S.T., Reeder, J.W., Samuels, I., Slabiak, T., Wagman, A.S., Hammond, M.E & Harrison, S.D (2003) Selec-tive glycogen synthase kinase 3 inhibitors potentiate insulin acti-vation of glucose transport and utilization in vitro and in vivo Diabetes 52, 588–595.

42 Nemeth, Z.H., Deitch, E.A., Szabo, C., Fekete, Z., Hauser, C.J & Hasko, G (2002) Lithium induces NF-kappa B activation and interleukin-8 production in human intestinal epithelial cells J Biol Chem 277, 7713–7719.