MKK6/3 and p38 MAPK Pathway Activation Is Not Necessary for Insulin-induced Glucose Uptake but Regulates Glucose Transporter Expression*

p38 mitogen-activated protein kinase (MAPK), which is situated downstream of MAPK kinase (MKK) 6 and MKK3, is activated by mitogenic or stress-inducing stimuli, as well as by insulin. To clarify the role of the MKK6/3-p38 MAPK pathway in the regulation of glucose transport, dominant negative p38 MAPK and MKK6 mutants and constitutively active MKK6 and MKK3 mutants were overexpressed in 3T3-L1 adipocytes and L6 myotubes using an adenovirus-mediated transfection procedure. Constitutively active MKK6/3 mutants up-regulated GLUT1 expression and down-regulated GLUT4 expression, thereby significantly increasing basal glucose transport but diminishing transport induced by insulin. Similar effects were elicited by chronic (24 h) exposure to tumor necrosis factor α, interleukin-1β, or 200 mm sorbitol, all activate the MKK6/3-p38 MAPK pathway. SB203580, a specific p38 MAPK inhibitor, attenuated these effects, further confirming that both MMK6 and MMK3 act via p38 MAPK, whereas they had no effect on the increase in glucose transport induced by a constitutively active MAPK kinase 1 (MEK1) mutant or by myristoylated Akt. In addition, suppression of p38 MAPK activation by overexpression of a dominant negative p38 MAPK or MKK6 mutant did not diminish insulin-induced glucose uptake by 3T3-L1 adipocytes. It is thus apparent that activation of p38 MAPK is not essential for insulin-induced increases in glucose uptake. Rather, p38 MAPK activation leads to a marked down-regulation of insulin-induced glucose uptake via GLUT4, which may underlie cellular stress-induced insulin resistance caused by tumor necrosis factor α and other factors.

Virtually all mammalian cells utilize glucose as a major energy source and thus possess facilitative glucose transporters on their cell surfaces. The facilitative glucose transporter gene family encodes at least six isoforms (GLUT1-6) with varying tissue distributions, subcellular localizations, and ki-netics for glucose uptake (1,2). Among them, GLUT4, in which expression is strictly limited to muscle and fat cells, resides in an intracellular compartment under basal conditions and moves to the cell surface in response to stimulation by insulin and other factors. In contrast, the ubiquitously expressed GLUT1 is mainly located on the cell surface, irrespective of stimulation, and is thus considered to be mainly involved in the maintenance of basal glucose uptake into the cells.
The mechanisms whereby signal-transducing molecules regulate the activity of glucose transporters have been extensively studied. At present, it is clear that activation of phosphatidylinositol 3-kinase is essential for insulin-induced GLUT4 translocation to the plasma membrane and increased glucose uptake (3)(4)(5)(6), although the mediators downstream of phosphatidylinositol 3-kinase remain controversial (7)(8)(9). In addition, activation of classical mitogen-activated protein kinase (MAPK) 1 (also termed extracellular signal-regulated kinase (ERK)), which plays a central role in cellular transformation, reportedly up-regulates GLUT1 expression, thereby augmenting glucose transport (10,11).
Recent studies have revealed the existence of at least three independent MAPK pathways (12)(13)(14). Besides ERK, two novel MAPKs have been identified and designated p38 MAPK (15,16) and stress-activated protein kinase (also known as c-Jun N-terminal kinase (JNK)) (8,17). Among the mitogenic factors and stress-inducing stimuli that mediate activation of these enzymes, insulin is reported to activate p38 MAPK (18,19). Interestingly, SB203580, a specific inhibitor of p38 MAPK, inhibits insulin-stimulated increases in glucose transport in both 3T3-L1 adipocytes and L6 myotubes without affecting insulin-stimulated GLUT4 translocation to the cell surface (20). It has therefore been hypothesized that p38 MAPK activation is involved in an insulin-induced enhancement of intrinsic GLUT4 activity on the cell surface. In addition, p38 MAPK is also activated in response to tumor necrosis factor (TNF) ␣, interleukin 1 (IL-1), and hyperosmotic shock (21)(22)(23), reportedly leading to insulin resistance (24 -27).
In that context, our study was undertaken to clarify the role of the p38 MAPK pathway in the regulation of glucose trans-port. To modulate p38 MAPK activity, dominant negative p38 MAPK and MAPK kinase (MKK) mutants as well as constitutively active MKK mutants were overexpressed in 3T3-L1 adipocytes and L6 myotubes using an adenovirus-mediated transfection system. In addition, the effects induced by TNF␣, IL-1␤, and 200 mM sorbitol were examined in 3T3-L1 adipocytes. We show that MKK6-and MMK3-mediated activation of p38 MAPK significantly alters expression of GLUT1 and GLUT4, increasing basal glucose transport and reducing insulin responsiveness.

EXPERIMENTAL PROCEDURES
Materials-3-Isobutyl-1-methylxanthine and 2-deoxy-D-glucose were purchased from Wako Bioproducts. Enhanced chemiluminescence (ECL) detection system was from Amersham Pharmacia Biotech. TNF␣ and IL-1␤ were from Genzyme Corp. Inhibitors wortmannin and SB203580 were from Sigma and Calbiochem, respectively. [␥-32 P]UTP was from ICN. All other reagents from commercial sources were of analytical grade.
Cell Culture-3T3-L1 fibroblasts were initially maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% donor calf serum (Life Technologies, Inc.) under an atmosphere of 90% air, 10% CO 2 at 37°C. Differentiation was induced 2 days after the cells reached confluence by replacing their normal culture medium with DMEM supplemented with 0.5 mM 3-isobutyl-1-methylxanthine, 4 g/ml dexamethasone (Sigma), and 10% fetal bovine serum (Life Technologies, Inc.) for 48 h. Thereafter, the cells were incubated for an additional 4 -10 days in DMEM supplemented with 10% fetal bovine serum; the medium was changed every other day. With this protocol, more than 90% of the cells expressed the adipocyte phenotype (3).
L6 myoblasts were initially grown in DMEM supplemented with 10% fetal calf serum under an atmosphere of 95% air, 5% CO 2 at 37°C, after which they were seeded onto 10-cm plastic culture dishes at a density of 3,000 cells/cm 2 . To promote their fusion into myotubes, they were rendered quiescent by incubation for 10 days in DMEM containing 2% serum. Myotube formation was assessed according to the percentage of nuclei present in multinucleated myotubes. With this protocol, 80 -90% of myoblasts were fused into myotubes.
Gene Transduction-To obtain recombinant adenoviruses, the expression cosmid cassette pAdexCAwt was ligated with the cDNA encoding LacZ from Escherichia coli or one of the aforementioned mutants, respectively, after which homologous recombination between the recombinant cosmid cassette and its parental virus genome was carried out as described previously (33). Infection of 3T3-L1 (5, 34) and L6 cells (35) with the indicated adenoviruses was carried out as described previously. Assessing triglyceride content of the infected cells confirmed that overexpression of the MAPK family mutants did not affect differentiation of 3T3-L1 cells into adipocytes (data not shown).
Immunoblotting-3T3-L1 adipocytes in a 12-well tissue culture dishes were lysed, boiled in Laemmli buffer containing 10 mM dithiothreitol, and subjected to SDS-polyacrylamide gel electrophoresis (PAGE). Immunoblotting was performed using an ECL system according to the manufacturer's instructions.
p38 MAPK Assay-Adipocytes were serum-deprived for 3 h before being exposed to 100 nM insulin for 5 min. p38 MAPK activity was assayed using a specific kit (New England Biolabs) according to the manufacturer's instructions. Briefly, the cells were lysed, immunoprecipitated with immobilized anti-phospho-p38 MAPK (Thr 180 /Tyr 182 ) antibody, and immunoblotted with anti-phospho-ATF2 (Thr 71 ) antibody, a natural substrate of p38 MAPK.
Subcellular Fractionation-Fractionation of subcellular membranes from 3T3-L1 adipocytes was carried out essentially as described previously (36). Aliquots of subcellular membrane fractions containing equal amounts of protein were subjected to SDS-PAGE, followed by immunoblotting using anti-GLUT1 and anti-GLUT4 antisera and an ECL detection system.
RNA Extraction-Total cell RNA was isolated from 3T3-L1 adipocytes using an Isogen RNA isolation kit (Nippon Gene, Tokyo, Japan). RNA concentrations were estimated based on absorbance at 260 nm, and 10 g of RNA from each sample were used for the RNase protection assay described below.
Preparation of Riboprobes-Riboprobes were synthesized as de- scribed previously (37). To obtain mouse GLUT1 and GLUT4 cDNAs, polymerase chain reaction was performed based on the reported sequences using mouse cDNA libraries. The amplified fragments, which corresponded to nucleotides 790 -939 of mouse GLUT1 cDNA and 790 -946 of mouse GLUT4 cDNA, respectively, were subcloned into a TA vector. These resultant plasmids were then linearized with BamHI and HindIII, respectively, and used for in vitro transcription.
RNase Protection Assay-RNase protection assays were carried out according to the manufacturer's instructions (RPA III; Ambion, Austin, TX). Pooled samples of 10 g of total RNA from adipocytes were hybridized with the riboprobes for GLUT1 and GLUT4. After treatment with RNase, the protected fragments were resolved on 5% polyacrylamide-urea gels and subjected to autoradiography, and the band intensities were determined using a Molecular Imager GS-525.
Statistical Analysis-Results were expressed as means Ϯ S.E. Comparisons were made using unpaired Student's t test. Values of p Ͻ 0.05 were considered statistically significant.
Effects of p38 MAPK and MKK6 Mutants on p38 MAPK Activity in 3T3-L1 Adipocytes in the Presence and Absence of Insulin-p38 MAPK, which is activated by insulin-mediated dual phosphorylation on threonine 180 and tyrosine 182 in the TGY motif, phosphorylates a number of downstream targets, including ATF2. We overexpressed p38 MAPK-AF, AA-MKK6, EE-MKK6, and EE-MKK3 in 3T3-L1 adipocytes, after which p38 MAPK activities were measured by immunoblotting antiphospho-p38 MAPK (Thr 180 /Tyr 182 ) immunoprecipitates with anti-phospho-ATF2 (Thr 71 ) antibody (Fig. 2). As shown previously, insulin did indeed activate p38 MAPK in control 3T3-L1 adipocytes overexpressing LacZ. By contrast, overexpression of EE-MKK6 led to marked activation of p38 MAPK irrespective of insulin stimulation, whereas overexpression of p38 MAPK-AF or AA-MKK6 strongly inhibited insulin-induced activation of p38 MAPK. EE-MKK3 also activated p38 MAPK; however, the magnitude of the enhancement was smaller than seen with EE-MKK6 (data not shown).
Effect of Dominant Negative p38 MAPK and MKK6 Mutants on Glucose Transport in 3T3-L1 Adipocytes- Fig. 3A shows the basal and insulin-stimulated 2-deoxy[ 3 H]glucose uptake into 3T3-L1 adipocytes overexpressing LacZ or dominant negative p38 MAPK (p38-AF) and MKK6 (AA-MKK6) mutants. In each case, insulin markedly enhanced glucose uptake. Moreover, overexpression of either p38-AF or AA-MKK6 significantly enhanced the insulin-induced increases in glucose uptake, with the latter having the greater degree. Based on these results, it seems clear that p38 MAPK activation is not essential for insulin-induced increases in glucose uptake.
On the other hand, overexpression of a constitutively active form of MKK6 (EE-MKK6) led to an ϳ30-fold increase in glucose uptake by 3T3-L1 adipocytes, even in the absence of insulin, and insulin stimulation further increased glucose uptake by ϳ50% (Fig. 3B). Pretreatment for 15 min with 100 nM of wortmannin, a phosphatidylinositol 3-kinase inhibitor, abolished the insulin-induced increase in glucose uptake by cells overexpressing either LacZ or EE-MKK6, although wortmannin had no effect on the increase in glucose uptake induced by EE-MKK6.
As reported previously, a membrane-targeted, constitutively active form of Akt (myr-Akt) also increased glucose transport in both 3T3-L1 adipocytes and L6 myotubes (Figs. 4, A and B).
Effect of Overexpressing EE-MKK6 or LASDSE-MEK1 on the Subcellular Distribution of Glucose Transporter in 3T3-L1 Adipocytes-To understand better the mechanism underlying the increase in glucose transport induced by MKK6 and MEK1, subcellular fractionation of 3T3-L1 adipocytes was performed, and the GLUT1 and GLUT4 contents of the plasma membrane (PM) and low density microsome (LDM) fractions were analyzed. We found that EE-MKK6 and LASDSE-MEK1 markedly increased levels of GLUT1 protein in both the PM and LDM fractions (Fig. 5, A and B), whereas levels of GLUT4 protein were markedly diminished (Fig. 5, C and D).
Effect of Overexpressing EE-MKK6 or LASDSE-MEK1 on Expression of Glucose Transporter mRNA in 3T3-L1 Adipocytes-By using an RNase protection assay, we found that, as compared with control (LacZ), expression of GLUT1 mRNA was significantly increased, whereas that of GLUT4 mRNA was significantly diminished in 3T3-L1 adipocytes overexpressing either EE-MKK6 or LASDSE-MEK1 (Fig. 6).
Time-dependent Effect of SB203580 on Stimulated Glucose Transport in 3T3-L1 Adipocytes Overexpressing EE-MKK6 -To confirm whether or not MKK6-induced changes in glucose transport and GLUT1/4 expression were mediated via p38 MAPK, the effects of SB203580, a specific p38 MAPK inhibitor, were examined. As shown in Fig. 7, SB203580 time-dependently inhibited the increase in glucose uptake induced by overexpression of EE-MKK6 in 3T3-L1 adipocytes; an almost complete blockade was achieved when cells were preincubated with SB203580 for 48 h.
By contrast, preincubation with SB203580 had no significant effect on the increase in glucose uptake induced by overexpression of myr-Akt.
Effects of SB203580 on Glucose Transport and Expression of GLUT1/4 in 3T3-L1 Adipocytes Overexpressing Constitutively Active MKK Mutants-Preincubation for 24 h with 10 M SB203580 significantly inhibited the increase in glucose transport normally induced by overexpression of EE-MKK6 and EE-MKK3 in 3T3-L1 adipocytes but not that induced by LASDSE-MEK1 (Fig. 8). Moreover, preincubation with SB203580 restored expression of both GLUT1 and GLUT4 protein to control levels in 3T3-L1 adipocytes overexpressing EE-MKK3 or EE-MKK6 (Fig. 9). By contrast, the effects of overexpressing LASDSE-MEK1 on expression of GLUT1 and GLUT4 protein were unaffected by SB203580.

DISCUSSION
Glucose transport is known to be affected by cell differentiation and oncogenic transformation, as well as by various hormones and growth factors. Such modulation has often been attributed to altered expression of one or more glucose transporter isoforms (i.e. GLUT1-6) (1, 2) or to modulation of glucose transporters' (particularly GLUT4) capacity to translocate to the plasma membrane (38 -40). The current study was undertaken to understand better the role of MKK6/3-p38 MAPK activation on the regulation of glucose transport and glucose transporter expression.
Activation of p38 MAPK Is Not Essential for Insulin-induced Increases in Glucose Uptake-It was recently reported that treatment with SB203580, a specific p38 MAPK antagonist, inhibited insulin-induced glucose uptake by both 3T3-L1 adipocytes and L6 myotubes, without inhibiting translocation of GLUT1 or GLUT4 (20). Consequently, it was hypothesized that activation of p38 MAPK plays a key role in the insulin-induced enhancement of intrinsic GLUT4 activity on the cell surface. On the other hand, p38 MAPK is also known to be activated by stressful stimuli, such as TNF␣, IL-1, and hyperosmotic shock (21)(22)(23), which appear to induce insulin resistance (41)(42)(43). To clarify whether or not activation of p38 MAPK is indeed essential for insulin-induced increases in glucose transport, we overexpressed dominant negative p38 MAPK and MKK6 mutants, which markedly suppressed endogenous p38 MAPK activity but significantly enhanced insulin-induced glucose uptake by 3T3-L1 adipocytes. We also observed that 10 M SB203580 almost completely blocked insulin-induced activation of p38 MAPK but that the insulin-induced increase in glucose uptake was reduced by only 30% (data not shown). A higher concentration of SB203580 (100 M) inhibited the insulin-induced increase in glucose uptake by more than 90%; however, at this concentration, SB203580 also significantly suppressed the activity of Akt kinase (data not shown). It therefore appears that the inhibition of insulin-induced increases in glucose uptake by SB203580 was not mediated by an effect on p38 MAPK activity but by an effect on some other molecule(s), perhaps Akt or GLUT4 itself. Furthermore, it appears unlikely that activation of p38 MAPK plays a significant role in the regulation of the insulin-induced increases in glucose transport mediated by GLUT4.
Activation of p38 MAPK Leads to Up-regulation of GLUT1 and Down-regulation of GLUT4 -We found that the overexpression of constitutively active MKK6 or MKK3 mutants markedly increased expression of GLUT1 protein in both 3T3-L1 adipocytes and L6 myotubes, whereas expression of GLUT4 was diminished. These effects were reflected by a significant increase in basal glucose transport and diminished insulin-induced glucose transport, respectively. That SB203580 inhibited MMK6/3-induced alterations in glucose uptake and expression of GLUT1/4 makes it highly likely that p38 MAPK plays a central role in the regulation of glucose metabolism.
Most interesting, overexpression of constitutively active MEK1, which induces chronic activation of ERK, yielded the same result as did overexpression of constitutively active MKK6 and MKK3. Previously, up-regulation of GLUT1 by oncogenes fps, src, and ras or by 12-O-tetradecanoylphorbol-13acetate was shown to underlie the increased glucose transport related to cellular transformation and/or proliferation (10,11). The GLUT1 gene contains two enhancer elements as follows: one in the 5Ј region contains one serum-response element, two 12-O-tetradecanoylphorbol-13-acetate-response elements (TREs), one cyclic AMP-response element, and three GC boxes; and the other, located in the second intron, contains one cyclic AMP-response element and two TREs (44). Both of these enhancers are involved in the augmented GLUT1 gene transcription induced by ras and v-src, by serum and by plateletderived growth factor (11,44,45). Since transcription activator protein-1 is a target of ERK1/2 (46 -48), the association of activator protein-1 with TRE may contribute to the activation of the MEK1-ERK pathway by acting at the 5Ј region enhancer(s) of the GLUT1 gene to up-regulate its expression and thus increase glucose transport. Indeed, this phenomenon may be central to the cellular transformation and/or proliferation mediated by activation MEK1-ERK pathway.
MKK6/3 and MKK7 activate p38 MAPK and JNK, respectively (31, 49 -52). In 3T3-L1 adipocytes, activation of JNK appears to induce increased synthesis of glycogen (18), and in the present study, we found that MKK6/3, but not MKK7, up-regulated expression of GLUT1 mRNA. Thus, the regulation of GLUT1 gene transcription may not be mediated entirely by activator protein-1 but by a more complex mechanism involving other factors.
On the other hand, the GLUT4 gene, which contains 11 exons and 10 introns, possesses a promoter region containing a weak "TATA" box sequence homology, a CCAAT box to which CCAAT/enhancer binding protein likely binds (53), four potential binding sites for nuclear transcription factor Sp1 (54), a skeletal muscle-specific activation domain (55). Together, these domains play a crucial role in the up-regulation of GLUT4 expression during adipose and muscle cell differentiation. In that respect, we found no significant alteration in cellular triglyceride content in 3T3-L1 adipocytes overexpressing the various MAPK family mutants studied; thus the down-regulation of GLUT4 expression in 3T3-L1 adipocytes overexpressing constitutively active forms of MKK6/3 and MEK1 cannot be attributed to the dedifferentiation from adipocytes to fibroblasts.
In many cases, expression of GLUT1 and GLUT4 are regulated oppositely, but whether changes in the expression of one affect the expression of the other remains unknown. If such interdependence exists, down-regulation of GLUT4 might be induced by the elevated intracellular glucose concentration mediated by up-regulation of GLUT1. However, there is reportedly no significant alteration in the GLUT4 expression in the skeletal muscle of transgenic mice overexpressing GLUT1 (56). We therefore suggest that expression of GLUT1 and GLUT4 is regulated independently, although further study will be necessary to resolve this issue definitively.
Involvement of p38 MAPK Activation in the Insulin Resistance Induced by TNF␣, IL-1␤, and Hyperosmolarity-Several studies have shown that chronically exposing adipocytes to TNF␣ or oxidant stress markedly decreases their GLUT4 content as a result of decreased GLUT4 gene transcription and a reduced half-life of its mRNA (26,42,57,58). Conversely, chronic exposure to TNF␣ increases expression of GLUT1 (27,58,59), although a contradictory result was obtained in one study (60). In the present study, we found that prolonged (24 h) exposure to TNF␣, IL-1␤, or hyperosmolarity increased glucose transport, increased GLUT1 expression, and decreased GLUT4 expression in a manner similar to that seen in cells overexpressing constitutively active MKK6. These phenomena were all attenuated by SB203580, implicating p38 MAPK in their occurrence.
Exposure to TNF␣, IL-1␤, and hyperosmolarity is also known to lead to insulin resistance (24 -27). Consistent with the aforementioned findings, down-regulation of GLUT4 and up-regulation of GLUT1 are considered to be components of the mechanism underlying such insulin resistance, although reduced expression of insulin receptor substrate-1 (61) and inhibition of insulin-induced tyrosine phosphorylation of insulin receptor may also be involved in TNF␣-induced insulin resistance (41,62).
In summary, activation of the MKK6/3-p38 MAPK pathway markedly enhances glucose transport by up-regulating GLUT1 expression, irrespective of the stimulus. This effect may help cells meet the increased energetic demands incurred when they are under stress, perhaps determining whether cells survive or succumb to apoptosis, a process in which the MKK6/3-p38 MAPK pathway also plays a role. In addition, although further study will be required to clarify whether the MKK6/3-p38 MAPK pathway is indeed activated under conditions associated with insulin resistance (e.g. obesity), we anticipate that p38 MAPK activation will prove to be a key component of stress-induced insulin resistance in fat and muscle cells.