Plasma Adiponectin Plays an Important Role in Improving Insulin Resistance With Glimepiride in Elderly Type 2 Diabetic Subjects

The patients consisted of 29 elderly subjects (19 men and 10 women, age 67.8 ± 9.9 years) with type 2 diabetes who, at study entry, showed poor blood glucose control (HbA_1c 8.1 ± 2.5% [mean ± SD]) despite treatment with oral glibenclamide (7.5–10 mg daily) and acarbose (α-glucosidase inhibitor [α-GI]) (300 mg daily). Clinical profiles of the subjects are shown in Table 1. All subjects were recruited from the geriatric outpatient division of the Nagoya University Hospital. Subjects were not included who already had concomitant large vessel disease (e.g., clinical evidence of coronary artery disease) or were on insulin treatment.

All subjects were randomly assigned to either a group whose medication was changed from glibenclamide to glimepiride (Amaryl) at study entry (glimepiride treatment group, n = 17, 11 men and 6 women) or a group whose medication was not changed (control group, n = 12, 8 men and 4 women). The oral dose of glimepiride started from 1 mg daily and increased up to 6 mg daily until a fasting glucose level <120 mg/dl or HbA_1c level <6.5% was achieved. The treatment with α-GI was continued in all patients. The study was approved by the local ethical committee. Written informed consent was obtained from all subjects after they were given a complete description of the study.

Blood and urine were sampled at different times (4, 8, and 12 weeks after the start of glimepiride treatment). The following measurements were examined before and 4, 8, and 12 weeks after the start of treatment: BMI [(body weight)/(height)²], serum total cholesterol, triglyceride, HDL cholesterol, fasting plasma glucose (FPG) and insulin in the morning and 2 h after breakfast, HbA_1c, and C-peptide in 24-h pooled urine (urine CPR). Indexes that are considered to reflect insulin resistance were also measured, including homeostasis model assessment of insulin resistance (HOMA-IR) = FPG (mg/dl) × immunoreactive insulin (IRI) (μU/ml)/405, an insulin secretion index, and homeostasis model assessment of β-cell function (HOMA-β) = 360 × IRI/(FPG-63), an index to reflect the ability of endogenous insulin secretion (15–17). At 4 and 8 weeks after the start of treatment, the following indexes, which are considered to be involved in insulin resistance, were also investigated: plasma PAI-1, plasma 8-epi-PGF2α, plasma TNF-α, and plasma adiponectin. As for seven elderly control subjects with type 2 diabetes, who remained on the combination therapy with glibenclamide and α-GI, plasma concentrations of TNF-α and 8-epi-PGF2α were measured at two time points with an interval of 8 weeks for comparison with the results of the glimepiride treatment group. Enzyme-linked immunosorbent assay kits were used for the measurement of concentrations of 8-epi-PGF2α (Oxis International, Portland, ME), TNF-α (Jimro, Takasaki, Japan), PAI-1 (Mitsubishi Chemical and Medical, Tokyo), and adiponectin (Linco Research, St. Charles, MO).

In 6 of the 17 patients who were randomly selected from the glimepiride treatment group, a hyperinsulinemic-euglycemic clamp was carried out by using an artificial pancreas (STG-22; Nikkiso, Tokyo) (18) before and at 8 weeks after the start of the treatment to assess insulin resistance in the peripheral tissue, especially in the skeletal muscle. The subjects were kept in a fasting state by not having any food intake for >10 h before the trial. Intravenous injection of insulin was made with a constant infusion rate (insulin infusion rate = 1.12 mU · kg⁻¹ · min⁻¹) to suppress the secretion of endogenous insulin. Then, glucose was injected to maintain the fasting blood glucose within the normal range (100–110 mg/dl). When the dose of glucose injection became constant, the constant glucose infusion rate, given as the M value, was recorded as an index to reflect insulin resistance in skeletal muscle. The metabolic clearance rate of glucose (MCR-g) was obtained from the following formula: MCR-g = M/fasting blood glucose × 100. Because the glucose release from the liver into the bloodstream is almost completely inhibited when peripheral insulin concentrations are at the physiological upper limit, the injection volume of glucose and peripheral glucose, mainly taken up by the skeletal muscle, are considered to be in an equilibrium state. Therefore, glucose uptake of the skeletal muscle is estimated from the amount of glucose injected into the body (18,19).

Data are means ± SD, and all variables were normally distributed. The paired Student’s t tests was applied, and a P value of <0.05 was considered significant. Statistical analyses were performed using Stat View software (version 5.01; SAS Institute, Cary, NC).

Clinical profiles of the patients are shown in Tables l and 2. Age, sex, BMI, lipid profiles (total cholesterol, triglycerides, and HDL cholesterol), and HbA_1c did not differ between the group treated with glimepiride and the control group. Eight weeks of treatment with glimepiride significantly reduced HbA_1c and HOMA-IR (Table 1). No significant changes were observed in urine CPR and HOMA-β in the glimepiride treatment group (Table 2), whereas those indexes were increased in the control group (data not shown). Meanwhile, a significant increase in the MCR-g was observed in all subjects, whose insulin resistance was examined by the hyperinsulinemic-euglycemic clamp test (Table 3). The blood glucose levels during fasting and 2 h after breakfast became significantly lower than the baseline levels at 4, 8, and 12 weeks after glimepiride treatment; however, those levels tended to increase in the control group (Table 2).

There was a highly significant elevation in plasma adiponectin concentration by 8 weeks of glimepiride treatment (Fig. 1). A significant decrease in plasma TNF-α concentration was also observed (Table 3) in the glimepiride treatment group. However, plasma adiponectin and TNF-α concentrations did not change in the control group (adiponectin: from 11.5 ± 4.1 to 10.5 ± 4.0 μg/ml; TNF-α: from 0.14 ± 0.10 to <0.05 pg/ml in 8 weeks). After 8 weeks of glimepiride treatment, plasma concentrations of 8-epi-PGF2α and PAI-1 tended to decrease, and the trend did not reach statistical significance (8-epi-PGF2α: P = 0.07; PAI-1: P = 0.06, Table 3), whereas those of the control group did not change over the period of observation.

The results of this study indicate that glimepiride may improve not only glucose metabolism but also insulin resistance in elderly diabetic subjects. They are demonstrated by reductions in HbA_1c and blood glucose levels at fasting and 2 h after breakfast, in concert with a significant increase of MCR-g and a reduction in HOMA-IR. By contrast, the control group did not show change in any of those indexes. BMI and plasma lipid profile tended to improve by glimepiride treatment, not by the control treatment using glibenclamide. These results may suggest that glimepiride not only improves blood glucose metabolism and insulin resistance in peripheral tissues, but also may improve factors related to the insulin resistance syndrome, such as plasma lipid profile and BMI. Eventually, glimepiride may be expected to retard the progression of arteriosclerosis, which is one of the major complications in elderly diabetic subjects.

Alterations of plasma adiponectin concentration and plasma TNF-α concentration in response to glimepiride are intriguing. In both diabetic groups examined, average plasma adiponectin level was significantly lower than that of five nondiabetic elderly subjects (20.6 ± 7.7 μg/ml), which implicates that adiponectin level is decreased in type 2 diabetic subjects, as suggested by Hotta et al. (20). The results of this study are in keeping with those of a previous report (21), showing that blood adiponectin concentration correlates with MCR-g, an index of insulin resistance. It was confirmed that thiazolidine improves insulin resistance by increasing plasma adiponectin (22). Therefore, a similar mechanism may underlie the findings we observed. Recently, attention has been paid to the anti-arteriosclerotic effect of adiponectin in connection with coronary artery disease. Several reports suggest that reduction in plasma adiponectin level may be related to the elevation of insulin resistance (12,23). Adiponectin is a specific plasma glycoprotein, contained in adipose tissue, with various homologies to collagen, complement proteins, and hibernation- associated proteins. In vitro studies have shown anti-atherogenic, anti-inflammatory, and apoptotic effects of adiponectin. The presence of adipocytes is essential for the onset of the normal insulin action. It is noteworthy that secretion factors, such as free fatty acid, TNF-α, and resistin, are involved in insulin resistance resulting from the hypertrophy of adipocytes. In particular, TNF-α is known to induce insulin resistance by inhibiting the activity of insulin receptor tyrosine kinase and the expression and translocation of GLUT4, and adiponectin specifically inhibits the expression of TNF-α among the inflammatory cytokines secreted from macrophages (24–27). Therefore, the observed changes in plasma adipocytokine levels may account for the improvement of insulin resistance by glimepiride.

Reductions of urine CPR and HOMA-β implicate that glimepiride does not stimulate endogenous insulin production, unlike other SU agents. Eight weeks of treatment with glimepiride did not change plasma concentrations of 8-epi-PGF2α, known as a free radical oxidative stress marker, which may suggest that glimepiride has no such short-term effect regarding oxidative stress as the one observed in our previous study using cerivastatin (28). Sampson et al. (29) suggested that acute rather than fasting hyperglycemia may be associated with macrovascular risk in type 2 diabetes by demonstrating a significant rise in plasma 8-epi-PGF2α concentrations when the patients underwent a 75-g glucose tolerance test. Therefore, the same test would have been necessary to assess the effect of glimepiride on reducing oxidative stress because we measured 8-epi-PGF2α concentrations only in fasting conditions. Also, longer periods of observation may be necessary to assess the chronic effect of glimepiride on macrovascular risk.

PAI-1 is considered a risk factor for the development of type 2 diabetes, independent of insulin resistance (14). Observed reduction of PAI-1 by glimepiride, although it did not achieve statistical significance, may implicate a diverse effect of the SU agent other than improvement of insulin sensitivity. The improvement of lipid profile we observed suggests that glimepiride may also have some effect on lipid metabolism. An in vitro experiment investigating oxidative denaturation of LDL in coronary artery endothelial cells has shown that glimepiride inhibits oxidation of LDL in a dose-dependent manner and also lowers plasma LDL (30). To our knowledge, there has been no adequate report conclusively answering the question of whether SU agents have a direct effect on preventing the onset and progression of macrovascular disease. We also found that 8 weeks of glimepiride treatment tended to increase the concentration of plasma NOx (sum of NO₂⁻ and NO₃⁻, data not shown), which implicates that glimepiride may improve nitric oxide availability in vascular endothelial cells, thus retarding the progression of arteriosclerosis when insulin resistance is increased. The extrapancreatic effect of glimepiride needs to be clarified in further investigations.

In summary, the present study demonstrates that glimepiride improves insulin resistance, probably by its extrapancreatic effects, in elderly patients with type 2 diabetes. The mechanism for the improvement may involve a decrease in plasma TNF-α, presumably induced by increased plasma adiponectin. HbA_1c and blood glucose levels were improved, whereas the extrapancreatic β-cell function (HOMA-β) and urine CPR remained unchanged. Thus, the observed improvement of glucose metabolism without stimulating extrapancreatic insulin secretion may indicate the advantage of glimepiride over other SU agents in the management of elderly patients with type 2 diabetes.

This study was supported in part by grant-in-aid number 09470166 from the Japanese Ministry of Education.

We thank W. Adachi, M. Kawamura, and C. Mazoku for technical assistance.