Advanced glycation end products-induced insulin resistance involves repression of skeletal muscle GLUT4 expression

Little is known about advanced glycation end products (AGEs) participation in glucose homeostasis, a process in which skeletal muscle glucose transporter GLUT4 (Scl2a4 gene) plays a key role. This study investigated (1) the in vivo and in vitro effects of AGEs on Slc2a4/GLUT4 expression in skeletal muscle of healthy rats, and (2) the potential involvement of endoplasmic reticulum and inflammatory stress in the observed regulations. For in vivo analysis, rats were treated with advanced glycated rat albumin (AGE-albumin) for 12 weeks; for in vitro analysis, soleus muscles from normal rats were incubated with bovine AGE-albumin for 2.5 to 7.5 hours. In vivo, AGE-albumin induced whole-body insulin resistance; decreased (~30%) Slc2a4 mRNA and GLUT4 protein content; and increased (~30%) the nuclear content of nuclear factor NF-kappa-B p50 subunit (NFKB1), and cellular content of 78 kDa glucose-regulated protein (GRP78). In vitro, incubation with AGE-albumin decreased (~50%) the Slc2a4/GLUT4 content; and increased cellular content of GRP78/94, phosphorylated-IKK-alpha/beta, nuclear content of NFKB1 and RELA, and the nuclear protein binding into Slc2a4 promoter NFKB-binding site. The data reveal that AGEs impair glucose homeostasis in non-diabetic states of increased AGEs concentration; an effect that involves activation of endoplasmic reticulum- and inflammatory-stress and repression of Slc2a4/GLUT4 expression.


Results
Treatment of rats with AGE-albumin induced insulin resistance. As previously described 16 , twelveweek treatment with AGE-albumin reduced the glucose decay constant rate (kITT) during the insulin tolerance test (Fig. 1D); however, the treatment did not alter rat body mass gain, blood glucose or plasma insulin concentrations ( Fig. 1A-C). Our hypothesis that this insulin resistance involves a direct effect of AGE-albumin in soleus muscle was supported by the detection of reduced expression of Ddost gene (Fig. 1E). Besides, participation of reduced skeletal muscle glucose disposal in the whole-body insulin resistance was evinced by the repression of Slc2a4 mRNA and GLUT4 protein expression (Fig. 1F).

Treatment of rats with AGE-albumin induced markers of reticulum endoplasmic and inflammatory stress in skeletal muscle.
In vivo treatment with AGE-albumin increased the expression of GRP78 chaperone ( Fig. 2A), revealing ER stress activation. Phosphorylation of IKKA and IKKB (Fig. 2B) and IKBA and IKBB content (Fig. 2C) were unchanged; however nuclear content of NFKB1 (Fig. 3) was increased, indicating the activation of a fundamental inflammatory pathway.
In vitro treatment of rat skeletal muscle with AGE-albumin repressed GLUT4 expression. To confirm a direct effect of AGE-albumin in skeletal muscle, soleus muscles from untreated healthy rats were incubated with AGE-or C-albumin. After 2.5-hour incubation, AGE-albumin decreased (50%) the Slc2a4 mRNA expression (Fig. 4A); but this period was not enough to decrease the total amount of cellular GLUT4 protein content. Nevertheless, extension of the incubation period to 5 and 7.5 hours ( Fig. 4B and C) induced a progressive decrease in GLUT4 expression, which reached a 25% reduction (P < 0.05) after 7.5-hour challenge with AGE-albumin.
Considering that AGE-albumin treatment might also impair GLUT4 vesicles translocation, we tested the glucose uptake in muscles incubated for 2.5 hours, when GLUT4 content was still unaltered. The results revealed that neither the basal (2.017 ± 0.185 vs 2.15 ± 0.21 μmol/g tissue, control-vs AGE-albumin, respectively; n = 6; P = 0.979) nor the insulin-stimulated (4.80 ± 0.326 vs 4.62 ± 0.213 μmol/g tissue, control-vs AGE-albumin, respectively; n = 6; P = 0.948) glucose uptake was altered by the presence of AGE-albumin, although the positive effect of insulin (P < 0.001) was observed in both conditions. In vitro treatment of rat skeletal muscle with AGE-albumin induced endoplasmic reticulum and inflammatory stress markers. In order to specify the role of AGE-albumin, the expression of AGE receptors was investigated. The results show that Ager mRNA expression increased in muscles incubated (2.5 hours) with AGE-albumin (Fig. 5A).
Two-and-a-half-hour incubation of muscles with AGE-albumin increased the expression of chaperones GRP78 and GRP94 proteins (Fig. 5B), revealing the activation of UPR. Besides, the results show that phosphorylation of IKKA and IKKB (Fig. 5C) increased, whereas the cellular content of IKBA and IKBB decreased (Fig. 5D), revealing the rapid activation of inflammatory pathway.
In vitro treatment of rat skeletal muscle with AGE-albumin increased nuclear protein binding into NFKB-binding site of Slc2a4 gene promoter. Firstly, we confirmed that the NFKB pathway activation culminates in increased nuclear content of both NFKB1 and RELA subunits of NFKB ( Fig. 6A-C).
Once established that 2.5-hour incubation of muscle with AGE-albumin activated the NFKB inflammatory pathway, culminating with nuclear translocation of NFKB1 and RELA proteins, we investigated the nuclear Figure 1. Treatment of rats with advanced glycated albumin induces insulin resistance and represses Slc2a4/ GLUT4 expression. Body weight gain (A), blood glucose (B), plasma insulin (C), blood glucose decay and constant rate (kITT) during insulin tolerance test (D), skeletal muscle expression of advanced glycosylation end product-specific receptor (Ager), and of dolichyl-diphosphooligosaccharide-protein glycosyltransferase non-catalytic subunit (Ddost) mRNAs (E), solute carrier family 2 member 4 (Slc2a4) mRNA, and solute carrier family 2 facilitated glucose transporter member 4 (GLUT4) protein (F) were measured in rats chronically treated with advanced glycated-(AGE; black bars; closed circles) or control-(C; white bars; open circles) rat albumin. Representative autoradiograms and Ponceau S staining of respective lanes, used as protein loading control, are shown in (F). Data are mean ± SEM of 4 (panel D) or 7 (panels A-C,E and F) animals. Means were compared by unpaired two-tailed t test. *P < 0.05 and **P < 0.01 vs control-albumin.

Discussion
AGEs are prevalent in DM and represent one of the basis for the development of long-term DM complications 17 . Albumin, the most abundant plasmatic protein, when modified by glycation displays an important role in cellular and tissue damage 18 . In the present study, we investigated whether glycated-albumin, regardless of a high glucose milieu, could modulate GLUT4 expression in skeletal muscle, which represents the latest downstream step for insulin-induced glucose disposal. In healthy subjects, skeletal muscle accounts for up to ~80% of glucose disposal under insulin-stimulated conditions, playing a fundamental role in glycemic homeostasis 19 . Regarding that, reduced GLUT4 expression in skeletal muscle was extensively reported in experimental models of type 2 DM (T2DM) in mice [20][21][22][23][24] . In human T2DM, although some pioneering studies have failed to detect reduced SLC2A4/ GLUT4 expression in skeletal muscle [25][26][27] , this was firstly reported by Dohm and colleagues 28 , and further definitely confirmed by studies employing more sensitive analyses of GLUT4 quantification 8,29,30 . Furthermore, studies that have been investigated epigenetic regulation of Slc2a4 gene in muscles from T2DM patients have now given attention for gene repression in this condition 30,31 . According to the hypothesis that AGEs contribute to glycemic control impairment, reduced insulin-induced glucose uptake by skeletal muscle of animals fed a high AGE content diet was already reported 32,33 . Besides, by intraperitoneal injection of AGE-albumin, we have recently reported that AGEs impair whole-body insulin sensitivity 16 , which now can be ascribed to a reduction in skeletal muscle Slc2a4/GLUT4 expression.
In order to clarify the mechanisms potentially involved in this GLUT4 regulation, we firstly measured the expression of AGEs receptors Ager and Ddost. Treatment with AGE-albumin reduced the expression of Ddost, a regulation already described in mice subjected to chronic ingestion of oral AGEs 32 . Cellular effects of AGEs include a convergent activation of oxidative, ER and inflammatory stress 12 . Accordingly, AGE-albumin treated rats showed increased cellular GRP78 and nuclear NFKB1 content, respective markers of ER and inflammation stress, which can be responsible for impaired Slc2a4/GLUT4 expression.
In in vitro incubations, a direct effect of AGE-albumin in soleus muscle was demonstrated by the Slc2a4 mRNA reduction after 2.5-hour AGE-albumin incubation, which reflected on decreased GLUT4 protein 2.5 hours later. Increased expression of Slc2a4 mRNA and GLUT4 protein has been observed as soon as 30 min to 120 min, respectively, after an enhancer stimulus [33][34][35] . However, decreasing effects have been described to occur later on; especially for GLUT4 protein, since the repressor effect depends not only on the transcriptional/translational inhibition, but also on the mRNA/protein half-life.
Decreased GLUT4 content was reported in muscle after 3-hour incubation with tumor necrosis factor alpha 36 , and in L6 muscle cells after 16-hour culture with linoleic or oleic fatty acids 37 . Here, GLUT4 protein reduction was detected 5 hours after muscle incubation with AGE-albumin, and that became more evident after 7.5 hours. These results can explain the AGE-induced reduction of glucose uptake described in L6 and in C2C12 cells after 8-hour AGEs treatment 3,38 .
Although the powerful effect of GLUT4 reduced expression in the AGE-induced impairment in muscle glucose disposal, we cannot discard the participation of impaired GLUT4 storage vesicles translocation to the plasma membrane. Reduced activity of some steps of insulin signaling pathway has been reported to occur in response to AGEs overload both in vivo 32,33 and in vitro 36 . Considering that, we measured the 2-deoxy-D-glucose (2DG) Figure 3. Treatment of rats with advanced glycated albumin activates nuclear factor NF-kappa-B in skeletal muscle. Nuclear factor NF-kappa-B p50 protein (NFKB1) (A,B) and nuclear factor NF-kappa-B p65 protein (RELA) (A,C) contents were measured in cytosolic and nuclear subcellular fractions from skeletal muscles of rats chronically treated with advanced glycated albumin-(AGE; black bars) or control-(C; white bars) rat albumin. In (A), representative autoradiograms and Ponceau S staining of respective lanes, used as protein loading control, are shown. Data are mean ± SEM of 7 animals. Means were compared by unpaired two-tailed t test. *P < 0.05 vs control-albumin. uptake in muscles incubated in vitro with control-and AGE-albumin for 2.5 hours, a time point in which the total cellular GLUT4 content was not altered. No differences were observed in both basal and insulin-stimulated conditions, although the expected positive effect of insulin was clearly observed. This result indicates that, at least for 2.5 hours, the insulin-mediated traffic of GLUT4 storage vesicles was unaffected by AGEs, reinforcing the important role of Slc2a4/GLUT4 repressed expression.
In vitro incubation of soleus muscle confirmed that AGE-albumin can directly and rapidly activate ER and inflammatory stress, an effect that is probably related to the increased expression of the AGE receptor gene Ager. Muscle incubation with AGE-albumin for 2.5 hours increased GRP78 and GRP94 chaperones, evincing the initial activation of the UPR. Increased protein content of ER chaperone GRP78 has been observed in several tissues of 4-week high-AGEs fed mice 39 , but there is no report of this regulation in skeletal muscle. Curiously, the in vivo treatment with AGE-albumin did not alter GRP94; however, in lead-and polycystic ovary syndrome-induced UPR, the GRP78 increase was reported to be more significant than the GRP94 increase 40,41 .
As a direct effect and/or as an UPR-related effect, AGEs can induce inflammatory stress. In vitro, AGE-albumin induced a clear activation of the canonical NFKB pathway in muscle; an effect that culminated with increased nuclear content of NFKB1 and RELA proteins. Besides, electrophoretic mobility assay revealed an increased nuclear protein binding activity into a Slc2a4 promoter NFKB-binding site. NFKB-mediated repression of Slc24 expression was proposed to be an inflammatory effect of tumor necrosis factor alpha in adipocytes several years ago 42 . Only recently the NFKB repressor effect on Slc2a4 transcription was finally confirmed 13 , in both adipose and muscle tissues. Furthermore, the repressor effect involves both NFKB1 (p50) and RELA (p65) proteins acting as a heterodimer 13 . Thus, the present data clearly show that AGEs, in vitro, repress Slc2a4/GLUT4 expression by a NFKB-mediated pathway. Although NFKB is a powerful repressor of Slc2a4 transcription 13 , proposed to mediate AGEs effect in the present study, we cannot discard the possibility that AGEs also reduce the transcriptional activity of some Slc2a4 enhancer.
The present data reveal that chronically administered AGE-albumin, regardless of a hyperglycemic condition, is able to impair glycemic homeostasis, by activating ER and inflammatory stress in skeletal muscle, which culminates with repression of Slc2a4/GLUT4 expression. This effect was also observed in vitro, in muscles from normal rats incubated with AGE-albumin for a few hours, in which the ER and inflammatory stress lead to increased NFKB1 and RELA binding activity into the Slc2a4 promoter, thus explaining its gene transcription repression. These data reveal that AGEs may worsen glycemic control in diabetic subjects and impair glycemic homeostasis in non-diabetic states of increased AGEs concentration; and that involves an ER-and inflammatory-mediated repression of Slc2a4/GLUT4 expression.

Figure 5.
In vitro treatment of rat skeletal muscle with advanced glycated albumin activates endoplasmic reticulum stress and inflammatory pathways. Advanced glycosylation end product-specific receptor (Ager) and dolichyl-diphosphooligosaccharide-protein glycosyltransferase non catalytic subunit (Ddost) mRNAs (A); 78 kDa glucose-regulated protein (GRP78) and 94 kDa glucose-regulated protein (GRP94) (B); phosphorylated inhibitor of nuclear factor kappa-B kinase subunits alpha (IKKA) and beta (IKKB) (C), and nuclear factor kappa-B inhibitors alpha (IKBA) and beta (IBB) (D) were measured in muscles of untreated rats incubated with advanced glycated-(left muscle; AGE; black bars) or control-(right muscle; C; white bars) bovine albumin for 2.5 hours. On the left side of panels B, C and D, representative autoradiograms and Ponceau S staining of respective lanes, used as protein loading control, are shown. Data are mean ± SEM of 6 (p-IKKA, p-IKKB and IKKA); 7 (Ager, Ddost and IKKB); or 8 (GRP78 and GRP94) animals. Means were compared by paired twotailed t test. *P < 0.05, **P < 0.01 and ***P < 0.001vs control-albumin. Figure 6. In vitro treatment of rat skeletal muscle with advanced glycated albumin increases nuclear content of NFKB1 and RELA proteins, and the nuclear proteins binding into Slc2a4 promoter NFKB-binding site. Nuclear factor NF-kappa-B p50 protein (NFKB1) (A,B) and nuclear factor NF-kappa-B p65 protein (RELA) (A,C) proteins were measured in cytosolic and nuclear subcellular fractions. Nuclear protein binding into NFKB-binding site of Slc2a4 gene promoter was analyzed by electrophoretic mobility shift assay and revealed two protein/DNA complexes a and b (D), which were separately quantified (E). Muscles were incubated for 2.5 hours with advanced glycated-(AGE, black bars) or control-(C, white bars) bovine albumin. In (A), representative autoradiograms and Ponceau S staining of respective lanes, used as protein loading control, are shown; and were cropped from different membranes. Data are mean ± SEM of 7 animals. Means were compared by paired two-tailed t test. *P < 0.05 vs control-albumin.

Material and Methods
Advanced glycation of albumin. The advanced glycation of rat (A6414; Sigma-Aldrich, Saint Louis, Missouri, USA) and bovine (A6003; Sigma-Aldrich, Saint Louis, Missouri, USA) albumin was performed in vitro by incubating albumin with freshly prepared 10 mM glycolaldehyde (Sigma Chemical Co., St. Louis, MO, USA) solution in phosphate buffer (PBS) at 37 °C, in a shaking water bath under N 2 atmosfere, in the dark. Control albumin was incubated with PBS alone. Samples were extensively dialyzed against PBS and kept frozen at −80 °C until experiments. The amount of endotoxins was <50 pg endotoxin/mL as determined by the chromogenic Limulus amebocyte assay (Falmouth, MA, USA) (data not shown). Carboxymethyllysine determined by ELISA was 12.6 times greater in rat AGE-albumin as compared to C-albumin. In addition, carboxymethyllysine and pyrraline amounts (mmol/mmol of lysine) were determined by liquid chromatography-mass spectrometry being highly superior in glycated samples as compared to C, as previously described 43 . Animals. The in vivo effect of AGEs was investigated in four-week old male Wistar rats obtained from de Central Animal Facility of the University of São Paulo Medical School, and housed in controlled environment (12-h light/dark cycle), with chow diet and water ad libitum. Animals were randomized into two groups receiving daily intraperitoneal (i.p.) injections of 20 mg/kg/day of rat control-(C) or AGE-albumin 44 for 12 weeks. At the end of week 12, animals were anesthetized via i.p. injection with sodium thiopental (60 mg/kg, Cristália, São Paulo, Brazil), and subjected to an insulin tolerance test or to blood (inferior vena cava) and soleus muscle (left and right) collection. Blood was processed for glucose and insulin concentration measurement 45 . The muscles were immediately frozen and stored at −80 °C for further analyses. This experimental protocol was approved by the Institutional Care and Research Advisory Committee (CAPPesq HC-FMUSP #002/14).
The in vitro effect of AGEs was investigated in soleus muscle harvested from untreated healthy control 65-to 75-day-old male Wistar rats (180 to 200 g body weight), obtained from the Animal Center of the Institute of Biomedical Sciences, University of São Paulo. Animals were housed under controlled conditions as described above. After i.p. anesthesia with thiopental sodium (60 mg/kg, Cristália, São Paulo, Brazil), left and right soleus muscles were harvested for the in vitro study. The experimental protocol was approved by the Ethical Committee for Animal Research of the Institute of Biomedical Sciences of the University of São Paulo (protocol #124/134/2).
All procedures performed on animals were in accordance with the relevant guidelines and regulations.
Insulin tolerance test (ITT). ITT was performed as previously described 45 . Tail blood samples were col- In vitro muscle incubation. Both right and left soleus muscles were gently dissected to preserve the integrity of tendons. One tendon was fixed into a horizontal metallic support, whereas the other was connected to an isometric transducer by a pulley (TBM-4F, World Precision Instruments INC., Sarasota, FL, USA). Muscle length was adjusted to produce maximal twitch tension (~3 g). The muscles were immersed in 100 mL Krebs-Heinseleit buffer, pH 7.4, containing 8 mM D-glucose and 1 mg/mL of control-(right muscle) or AGE-(left muscle) bovine albumin. Muscles were incubated at 37 °C, continuously oxygenated with 95% O 2 : 5% CO 2 , for 2.5 hours, with buffer replacement every 1.25 hours. For GLUT4 protein analysis, incubation time was extended to 5 and to 7.5 hours, with buffer replacement every 2.5 hours. At the end of the incubation periods, muscles were immediately frozen and stored at −80 °C for further analyses.
Electrophoretic mobility shift assay (EMSA). Nuclear proteins for EMSA were extracted from muscles incubated in vitro, and EMSA was performed as previously described 13 . The oligonucleotide used as probe corresponds to the -134/-113 sequence of the mouse Slc2a4 gene, which was previously confirmed to bind NFKB1 and RELA using nuclear proteins from rat L6 muscle cells 13 . EMSA performed with this probe revealed two protein/ DNA complexes (A and B) in rat muscle cells, and competition assays confirmed the presence of both NFKB1 and RELA in these complexes 13 .
Statistical Analyses. Comparison of results from rats treated or not with rat AGE-albumin in vivo was performed by unpaired two-tailed t test, after confirmation that the variances were not significantly different. Comparison of results from muscles incubated (left) or not (right) with bovine AGE-albumin was performed by paired two-tailed t test. Glucose uptake was analyzed by one-way ANOVA. Differences were considered significant when P < 0.05, and the number of samples is informed in the legends.

Data availability.
There is no restriction on the availability of materials and data.