Both rosiglitazone and metformin increase hepatic insulin sensitivity, but their mechanism of action has not been compared in humans. The objective of this study was to compare the effects of rosiglitazone and metformin treatment on liver fat content, hepatic insulin sensitivity, insulin clearance, and gene expression in adipose tissue and serum adiponectin concentrations in type 2 diabetes. A total of 20 drug-naive patients with type 2 diabetes (age 48 ؎ 3 years, fasting plasma glucose 152 ؎ 9 mg/dl, BMI 30.6 ؎ 0.8 kg/m 2 ) were treated in a double-blind randomized fashion with either 8 mg rosiglitazone or 2 g metformin for 16 weeks. Both drugs similarly decreased HbA 1c , insulin, and free fatty acid concentrations. Body weight decreased in the metformin (84 ؎ 4 vs. 82 ؎ 4 kg, P < 0.05) but not the rosiglitazone group. Liver fat (proton spectroscopy) was decreased with rosiglitazone by 51% (15 ؎ 3 vs. 7 ؎ 1%, 0 vs. 16 weeks, P ‫؍‬ 0.003) but not by metformin (13 ؎ 3 to 14 ؎ 3%, NS). Rosiglitazone (16 ؎ 2 vs. 20 ؎ 1 ml ⅐ kg ؊1 ⅐ min ؊1, P ‫؍‬ 0.02) but not metformin increased insulin clearance by 20%. Hepatic insulin sensitivity in the basal state increased similarly in both groups. Insulin-stimulated glucose uptake increased significantly with rosiglitazone but not with metformin. Serum adiponectin concentrations increased by 123% with rosiglitazone but remained unchanged during metformin treatment. The decrease of serum adiponectin concentrations correlated with the decrease in liver fat (r ‫؍‬ ؊0.74, P < 0.001). Rosiglitazone but not metformin significantly increased expression of peroxisome proliferator-activated receptor-␥, adiponectin, and lipoprotein lipase in adipose tissue. In conclusion, rosiglitazone but not metformin decreases liver fat and increases insulin clearance. The decrease in liver fat by rosiglitazone is associated with an increase in serum adiponectin concentrations. Both agents increase hepatic insulin sensitivity, but only rosiglitazone increases peripheral glucose uptake.

To examine whether and how intramyocellular lipid (IMCL) content contributes to interindividual variation in insulin action, we studied 20 healthy men with no family history of type 2 diabetes. IMCL was measured as the resonance of intramyocellular CH 2 protons in lipids/ resonance of CH 3 protons of total creatine (IMCL/Cr T ), using proton magnetic resonance spectroscopy in vastus lateralis muscle. Whole-body insulin sensitivity was measured using a 120-min euglycemic-hyperinsulinemic (insulin infusion rate 40 mU/m 2 ⅐ min) clamp. Muscle biopsies of the vastus lateralis muscle were taken before and 30 min after initiation of the insulin infusion to assess insulin signaling. The subjects were divided into groups with high IMCL (HiIMCL; 9.5 ؎ 0.9 IMCL/Cr T , n ‫؍‬ 10) and low IMCL (LoIMCL; 3.0 ؎ 0.5 IMCL/Cr T , n ‫؍‬ 10), the cut point being median IMCL (6.1 IMCL/Cr T ). The groups were comparable with respect to age (43 ؎ 3 vs. 40 ؎ 3 years, NS, HiIMCL versus LoIMCL), BMI (26 ؎ 1 vs. 26 ؎ 1 kg/m 2 , NS), and maximal oxygen consumption (33 ؎ 2 vs. 36 ؎ 3 ml ⅐ kg ؊1 ⅐ min ؊1 , NS). Whole-body insulin-stimulated glucose uptake was lower in the HiIMCL group (3.0 ؎ 0.4 mg ⅐ kg ؊1 ⅐ min ؊1 ) than the LoIMCL group (5.1 ؎ 0.5 mg ⅐ kg ؊1 ⅐ min ؊1 , P < 0.05). Serum free fatty acid concentrations were comparable basally, but during hyperinsulinemia, they were 35% higher in the HiIMCL group than the LoIMCL group (P < 0.01). Study of insulin signaling indicated that insulin-induced tyrosine phosphorylation of the insulin receptor (IR) was blunted in HiIMCL compared with LoIMCL (57 vs. 142% above basal, P < 0.05), while protein expression of the IR was unaltered. IR substrate-1-associated phosphatidylinositol (PI) 3-kinase activation by insulin was also lower in the HiIMCL group than in the LoIMCL group (49 ؎ 23 vs. 84 ؎ 27% above basal, A bnormal lipid metabolism is a feature of the insulin resistance syndrome (1). In addition to an increase in the total amount of fat, the lipid disturbances include elevated circulating concentrations of triglycerides and free fatty acids (FFAs) (2) and an increase in visceral fat (3). Recently, several studies have shown an association between lipid accumulation in skeletal muscle and insulin resistance (4 -10). In four of these studies, this relation was shown to be caused by intramyocellular rather than extramyocellular lipids, as measured by proton spectroscopy (5-7,11).The causes for intramyocellular lipid (IMCL) accumulation are poorly understood. The first possibility is that IMCL is an innocent bystander and simply reflects overall adiposity. This is not supported by recent experiments in mice lacking subcutaneous fat, the A-ZIP/F-1 mice (12,13). These mice deposit fat intramyocellularly and exhibit severe insulin resistance, which is reversible by fat transplantation and rechanneling of IMCL back to subcutaneous depots. In humans, however, it is less clear whether IMCL is associated with insulin resistance independent of obesity. In 20 Europeans, Forouhi et al. (6) found the rela...

Multiple alterations characterize gene expression in the subcutaneous adipose tissue of patients with HAART-associated lipodystrophy compared with HIV-positive, HAART-treated patients without lipodystrophy. The low expression of transcription factors inhibits adipocyte differentiation. The low expression of PGC-1 may contribute to mitochondrial defects. In addition, IL-6 and CD45 expressions are increased, the latter implying an excessive number of cells of leukocyte origin in lipodystrophic adipose tissue. Mitochondrial injury and an excess of proinflammatory cytokines may lead to increased apoptosis. All these changes may contribute to the loss of subcutaneous fat in HAART-associated lipodystrophy.