Mechanisms involved in tumor necrosis factor-alpha induction of insulin resistance and its reversal by thiazolidinedione(s).

Insulin resistance (IR) remains one of the major pathogenic mechanisms for non-insulin-dependent type 2 diabetes mellitus. We have previously modelled IR in H-411E liver cells in culture. In past experiments, we used both labeled glucose uptake, lipogenesis, and stimulation of calmodulin gene expression to quantify the ability of the antidiabetic drugs (pioglitazone and metformin) to reverse tumor necrosis factor-alpha (TNF-alpha)-induced IR in these insulin-treated cells. In these current experiments, H-411E liver cells were rendered IR by a combination of TNF-alpha and insulin. In other experiments, the ability of C2 ceramide (Cer) to inhibit insulin action and induce IR was assessed as well as the phospholipase C inhibitor D609 to reverse IR induced by these TNF-alpha-like agents. C2 Cer, like TNF-alpha, inhibited insulin action. D609 reversed TNF-alpha induced--and to a lesser extent, C2 Cer-induced--IR. At selected times, the cells were also treated with troglitazone (TRG) in 2 groups: (1) 1-time exposure and (2) chronic exposure followed by acute exposure. TRG concentrations ranged from 0.015 to 15.0 micromol/L. Our data demonstrate a powerful effect of TRG in reducing IR and restoring insulin sensitivity in TNF-alpha-treated H-411E cells. Furthermore, pretreatment with TRG, reflecting chronic exposure, as in human clinical use, was more potent than 1-time acute exposure. These data support the efficacy of using thiazolidinediones (TRG) in human type 2 diabetes, and support the use of this cell culture model to further study the effects of thiazolidinediones on TNF-alpha-induced insulin resistance.

Members of the Sp transcription factor family regulate rat calmodulin gene expression.

We have previously demonstrated that insulin positively regulates transcription of the rat calmodulin (CaM) I gene and that both basal and insulin stimulation of this gene are critically dependent on Sp1. Furthermore, a 392 bp CaM promoter was stimulated by insulin equal to the full promoter but lost activity with deletion of any of the three Sp1 sites (Solomon SS, Palazzolo MR, Takahashi T, Raghow R. Endocrinology 1997;138:5052-5054). Herein we document that Sp1 preferentially binds to the upstream sites Sp1(2) and Sp1(3) but not Sp1(1). Furthermore, gel-mobility super-shift assays demonstrate that both Sp1 and Sp3 protein are found in these complexes. When pPac-Spl, pPac-Sp3, pPac-USp3, and pPac-Sp4 were cotransfected with rCaM 1-392 promoter into Drosophila SL2 cells and challenged with 10,000 microU/mL insulin, we discovered that (1) Sp1 enhanced both basal and insulin-stimulated CaM I gene expression; (2) USp3, a "long" form of the Sp3 molecule, had a stimulatory effect on CaM I gene expression; (3) Sp1 or USp3 is involved in mediating insulin-stimulation of the CaM I gene in SL2 cells; and (4) Sp3, a "short" form of the Sp3 molecule, and Sp4 inhibited Spl-stimulated and insulin-stimulated Sp1-mediated CaM I gene expression. Together these data corroborate and extend our previous observations on Sp1 and elucidate that other members of the Sp family of transcription factors may also be involved in regulating the activity of the CaM promoter.

Transcription factor Sp1 is necessary for basal calmodulin gene transcription and for its selective stimulation by insulin.

Insulin positively regulates transcription of rat calmodulin (CaM) I gene and activates the low Km cyclic AMP (cAMP) phosphodiesterase (PDE). To elucidate the mechanism of transcriptional regulation, rat hepatoma (H-411E) cells were transfected with DNA constructs containing the putative CaM promoters coupled to a luciferase reporter and challenged with insulin. Activation of the full length 1835 bp rat CaM I promoter containing all three Sp1 sites or truncated promoters with combinations of one to three of the Sp1 sites was studied in Sp1 deficient Drosophilia SL2 cells and in SL2 cells co-transfected with an Sp1 expression vector and re-challenged with insulin. Our results demonstrate that Sp1 is obligatory for basal activation of the CaM promoter. The maximal insulin stimulation of CaM promoter is elicited only if it contains at least two Sp1 sites.

Insulin stimulates rat calmodulin I gene transcription through activation of Sp1.

We have shown previously that insulin positively regulates transcription of the rat calmodulin (CaM) I gene. This activation occurs concomitantly with the activation of the low-Km adenosine 3':5'-cyclic phosphate phosphodiesterase (PDE), which appears to be coregulated with CaM. Rat hepatoma H-411E cells were transfected with plasmids containing various lengths of the putative CaM promoter coupled to a luciferease reporter and were challenged with insulin. We demonstrate that insulin-stimulated transcription of CaM I gene is mediated by a 392-bp 5'-flanking region of the CaM I gene, encompassing 185 bp downstream and 207 bp upstream of the start site of transcription. The CaM I promoter contains three potential Sp1 sites, located at -114 through -109 [(3), +], -77 through -72 [(2), -] and at +53 through +58 [(1), +]. The gel mobility shift assays demonstrated that nuclear protein(s) associate with all three sp1 sites. We present data demonstrating the relative importance of the three Sp1 sites for the insulin effect: prCaM I 1835, 3.8x, delta 1081; prCaM I 392, 5.3x, delta 1055; prCaM I 180, 3.7x, delta 462; prCaM I 237, 1.6x, delta 478; prCaM I 139, 2.6x, delta 182; prCaM I 130, 2.1x, delta 194; and prCaM I 1463, negligible activity. In summary, the maximal insulin stimulation of CaM gene expression is seen when the promoter region contains at least two Sp1 sites.

Identification of specific sites in the TNF-alpha molecule promoting insulin resistance in H-411E cells.

Data from a number of laboratories support a potential role for tumor necrosis factor-alpha (TNF-alpha) in the loss of insulin sensitivity and the pathogenesis of insulin resistance (IR) in diabetic animal models and human patients. We designed experiments to establish a dose-response relationship for TNF-alpha and IR in H-411E cells in culture. IR was measured by inhibition of the ability of graded amounts of insulin to stimulate expression of calmodulin (CaM) mRNA in these cells. This was assessed by autoradiographs of Northern blot(s) of CaM mRNA probed with labeled oligonucleotide cDNA for rat CaM. We found that TNF-alpha at 0.1, 1.0, and 10.0 ng/ml opposed 10,000 microU/ml insulin (i.e., %IR = 20%, 67%, and 88%, respectively). At 1.0 ng/ml TNF-alpha, insulin at the concentration of 1000 microU/ml (0.006 micromol/L) stimulated CaM mRNA at a 41% level and at 10,000 microU/ml (0.06 micromol/L) at a 63% level. Furthermore, oligopeptide TNF-alpha homologs (at 1000 x the molar concentration of TNF-alpha) TNF-alpha 69-100 and TNF-alpha 133-157 conferred 66% and 101% IR, respectively, while all other peptide fragments of TNF-alpha were essentially without effect. Studies done with both monoclonal and polyclonal antibodies to the TNF-alpha receptor demonstrated blocking activity by polyclonal but not by monoclonal anti TNF-alpha receptor antibody. This supports the concept that the activity of the peptide fragments occurs through the TNF-alpha receptor and not through nonspecific translocation across the plasma membrane. These data suggest that the epitopes on TNF-alpha that mediate IR reside in two regions of the molecule spanning amino acid residues 69-100 and 133-157.

Insulin-stimulated calmodulin gene expression in rat H-411E cells can be selectively blocked by antisense oligonucleotides.

Reduced expression of calmodulin (CaM) and decreased activity of low Km cyclic AMP (cAMP) phosphodiesterase (PDE) are associated with uncontrolled diabetes. This condition can be readily mimicked in hepatocytes cultivated in insulin-depleted medium (Solomon, et al J. Lab. Clin. Med. in press, 1994). To investigate the relationship between CaM and low Km cAMP PDE gene expression in response to insulin, we specifically blocked expression of the three CaM genes by antisense oligonucleotides under insulin-deficient and -sufficient conditions in a rat hepatoma cell line, H-411E. We observed that both the low Km cAMP PDE activity and the steady state levels of CaM mRNA were increased in response to insulin by 50 and 100%, respectively. When antisense oligonucleotide to CaM I, II or III was added to the cultures, only CaM I antisense oligonucleotide blocked insulin stimulation of both CaM I mRNA and protein with concommittant marked inhibition of insulin's expected stimulation of low Km cAMP PDE. Furthermore, in another experiment utilizing both antisense and oligonucleotide probes specific for CaM I, II, or III together, only CaM I mRNA expression was blocked. We conclude that H-411E cells respond to insulin by appropriate increases in CaM transcripts. Furthermore, the stimulatory effect of insulin on both CaM synthesis and activation of low Km cAMP PDE could be blocked by antisense to CaM I, but not II or III genes. Therefore, in addition to the above conclusions, H-411E hepatoma cells appear to be an excellent in vitro system to explore the molecular mechanisms by which CaM and low Km cAMP PDE genes are regulated in the diabetic state.

Regulation of calmodulin gene expression by insulin is both transcriptional and post-transcriptional.

Previously we demonstrated that in streptozotocin-induced or spontaneously diabetic BB rats (BB-SDR), low-Km cyclic AMP (cAMP), phosphodiesterase (PDE), and calmodulin (CaM) are decreased. Isolated fat cells of diabetic animals synthesized less CaM and contained reduced levels of CaM transcripts (Solomon SS, Palazzolo MR, Green SA, Raghow R. Biochem Biophys Res Commun 1990; 168: 1007-12). Treatment of diabetic animals with insulin restores CaM transcripts to normal. RNA was extracted from isolated hepatocytes from BB-SDR rats in primary tissue culture treated with insulin (from 2.8 x 10(4) to 1.4 x 10(6) microU/ml) for 48 hours, was immobilized on nitrocellulose, and was sequentially hybridized with radiolabeled probes for CaM, actin, and tubulin. Insulin stimulates steady state levels of mRNA for calmodulin > actin > tubulin. Furthermore, decreased steady state levels of CaM mRNA in hepatocytes from diabetic animals are restored to normal levels with in vitro insulin incubation. Data from nuclear transcription run-on assays demonstrate that insulin stimulates transcription of mRNA CaM by 80%. In addition, we observed RNA degradation in the untreated diabetic but not insulin-treated liver. These data support transcriptional as well as post-transcriptional effects of insulin on CaM mRNA. We postulate that in uncontrolled diabetes, elevations in levels of cAMP in tissue result in part from decreased activity of the apparently co-regulated PDE and CaM and that PDE inactivation in diabetes results from both insulin insufficiency and CaM down-regulation.

Effects of sequence and timing of hormonal additions on adipose tissue: activation of low-Km cyclic adenosine monophosphate phosphodiesterase.

Epinephrine (EPI) is lipolytic and insulin (INS) antilipolytic in the isolated fat cell (IFC). We have previously demonstrated that in a perifusion system the antilipolytic action of INS is more powerful when IFC are exposed to INS before EPI. In contrast to their opposite effects on lipolysis, both INS and EPI stimulate low-Km cyclic adenosine monophosphate (cAMP) phosphodiesterase (PDE) in adipose tissue. In view of these observations, we decided to determine the effects of sequential addition of EPI and INS on stimulation of PDE from rat adipose tissue. Using previously published methods, the effects of INS and EPI on PDE were assessed alone, together with INS followed by EPI, and then with EPI followed by INS. The resulting data demonstrate that EPI and INS individually both stimulate PDE (P less than .001); EPI plus INS together stimulate PDE minimally compared with EPI or INS alone (P less than .001); when adipose tissue is included with INS first, then followed by EPI, activation of PDE is much less than INS or EPI alone (P less than .001); and when adipose tissue is stimulated by EPI then INS, there is no activation of PDE, different from EPI or INS alone (P less than .001). In conclusion, in perifused IFC, INS and EPI always oppose each other. In studies using activation of PDE, EPI and INS each stimulate PDE, but INS opposes EPI when incubated simultaneously. When adipose tissue is incubated first with INS followed by EPI, PDE is activated. In contrast, when the reverse order is applied, no activation of PDE is observed.(ABSTRACT TRUNCATED AT 250 WORDS)

Expression of calmodulin gene is down-regulated in diabetic BB rats.

In past studies, we have demonstrated that in streptozotocin-induced diabetic or spontaneously diabetic (BB) animal models, low Km cAMP phosphodiesterase and calmodulin are decreased while a low MW inhibitor of calmodulin is increased. To extend these studies, we have determined the rate of [35S]-methionine incorporation into calmodulin in isolated fat cells from these diabetic animals, i.e. streptozotocin-induced diabetic and the BB rats, spontaneous diabetic rat, non-diabetic rat, and control. We found markedly decreased rates of synthesis of calmodulin in the fully diabetic BB rat. In order to investigate the mechanism of the reduced calmodulin biosynthesis, we probed poly A+ mRNA from control and diabetic rat livers with a calmodulin specific anti-sense oligonucleotide probe and found that the fully diabetic animals, streptozotocin-induced diabetic and genetically diabetic BB, contained markedly reduced levels of calmodulin transcripts. Thus, both calmodulin protein and its putative mRNA are decreased in diabetic rat liver. We believe that in uncontrolled diabetes, the observed elevation in the levels of cyclic AMP in plasma and tissue results in part from decreased activity of phosphodiesterase. The insulin-sensitive phosphodiesterase appears to be regulated by calmodulin. We hypothesize that cyclic AMP phosphodiesterase inactivation in diabetes results in part from insulin insufficiency and to a less well-defined genetic lesion leading to calmodulin down-regulation.

Activation of cyclic AMP phosphodiesterase by phorbol and protein kinase C pathway: differences in normal and diabetic tissue.

Diabetes mellitus is associated with high levels of adenosine 3',5'-cyclic monophosphate in tissue and plasma. Diabetes inhibits and insulin stimulates and restores low Km adenosine 3',5'-cyclic monophosphate phosphodiesterase activity. We recently reported that phorbol ester, a tumor promoting agent known to act through protein kinase C also stimulates phosphodiesterase. Here, we address the issue of whether or not the activation of phosphodiesterase by insulin and phorbol ester is different in streptozotocin diabetic adipose tissue. Rat adipose tissue was incubated with insulin, phorbol ester or other known components or effectors of the protein kinase C pathway, i.e. 1,2 dioleoyl-glycerol, 1- oleoyl, 2- acetylglycerol, Ca(++)-Ionophore A 23187, and nifedipine. After incubation, preparation and assay of adenosine 3',5'-cyclic monophosphate phosphodiesterase was made. As in previous data streptozotocin-diabetes inhibits basal phosphodiesterase by about 50% (P less than .02); insulin and phorbol ester each stimulate phosphodiesterase, in streptozotocin-diabetes less than normal (P less than .025); nifedipine inhibits phorbol stimulated phosphodiesterase in streptozotocin-diabetes but not normal (P less than .001); and nifedipine inhibits insulin stimulated phosphodiesterase in normal (84%) and diabetic (97%) (P less than .005). In normal and diabetic tissue, diacyl glycerol and oleoyl-acyl glycerol stimulate phosphodiesterase, are augmented by ionophore and inhibited by nifedipine. In addition 32P incorporation studies and measurements of tyrosine kinase activity are presented which support these differences between normal and diabetic. In summary then, these data suggest common pathways of activation for low Km adenosine 3',5'-cyclic monophosphate phosphodiesterase by insulin and phorbol ester; imply a relationship between two second messenger systems, phosphoinositides and adenosine 3',5'-cyclic monophosphate; and demonstrate a difference in activation of phosphodiesterase between normal and diabetic adipose tissue.

Inhibitor of calmodulin and cAMP phosphodiesterase activity in BB rats.

Diabetes mellitus in humans is associated with increased plasma and tissue levels of cAMP and decreased cAMP phosphodiesterase (PDE) activity. Calmodulin (CM) is a low-molecular-weight protein essential for activation of cAMP PDE. The inhibitor (INH) is a low-molecular-weight substance that inhibits the activity of CM in multiple systems, including PDE. Spontaneously diabetic BB rats (SDR) and their nondiabetic littermates (NDR) were used in this study. Holtzman rats were rendered diabetic by streptozocin (STZ). STZ-induced diabetic rats (STZ-DR) and BB rats were studied with and without the benefit of insulin therapy. Calmodulin was assayed both by bioassay and by specific radioimmunoassay. The inhibitor was bioassayed by its ability to inhibit CM-activated PDE. Results showed that both spontaneous and STZ-induced diabetes are associated with a decrease in activity of the low-Michaelis constant (Km) cAMP PDE in the liver (39%, SDR; 70% STZ-DR). Calmodulin activity was also decreased in the livers of both animals (13%, SDR; 68%, STZ-DR). Similar data were obtained for NDRs. The inhibitor, on the other hand, was increased in the livers of untreated SDRs and STZ-DRs (155%, SDR; 125%, STZ-DR). No change was noted for NDRs. All these changes were restored toward normal after treatment with insulin. These data suggest that in diabetes the defect in the cAMP PDE-CM-INH system is demonstrated in both an environmental model, as illustrated by STZ-DRs, and a genetic model, as shown by SDRs and NDRs. The inhibitor activity, however, is not changed significantly in NDRs. We speculate that the inhibitor activity plays a role in dictating whether the genetic NDR will or will not become clinically diabetic.

Spontaneous diabetic BB rat: studies of cyclic adenosine 3',5'-monophosphate phosphodiesterase and calmodulin.

Uncontrolled diabetes in man is associated with increased plasma and tissue levels of cAMP and decreased cAMP phosphodiesterase (PDE) activity. Spontaneously diabetic BB rats (SDR) were used in these experiments. Specific tissues (i.e. liver and epididymal fat) were studied without therapeutic insulin. Another group of normal animals were rendered diabetic by streptozotocin (STZ) and killed without benefit of insulin therapy. Calmodulin (CM), a small molecular weight protein essential for activation of specific cAMP PDE was assayed. STZ diabetes is associated with a decrease (58%) in CM biological activity and in immunoreactive CM in fat (69%) and liver (13%) tissues. Similarly, SDR rats and the nondiabetic genetic controls (NDR) demonstrate decreased CM bioactivity in fat (76% and 56%, respectively) and decreased CM immunoreactivity in liver (68% and 74%, respectively) compared to normal control rats. In addition, maximum velocity (Vmax) of the low Michaelis-Menten constant (Km) cAMP PDE is decreased in SDR animals, as compared to controls in both fat (42%) and liver (39%) tissues. Similar data are presented for NDR animals. STZ diabetes is also associated with a reduction in Vmax of the low Km cAMP PDE in both liver (70%) and fat (70%) tissues. These changes found in the NDR animals suggests that the diabetic defect may be under dual regulation: genetic and environmental.