Dissociation between the insulin-sensitizing effect of rosiglitazone and its effect on hepatic and intestinal lipoprotein production

Context: Despite its potent, well-documented insulin-sensitizing effects, rosiglitazone (RSG) does not effectively ameliorate the hypertriglyceridemia of insulin-resistant or diabetic individuals and has even been shown to slightly but significantly increase triglyceride-rich lipoproteins (TRL) in some studies. The mechanism of this effect is currently not known. Objective: We investigated the effect of RSG treatment on TRL metabolism. Design: This was a 12-wk, single-sequence, cross-over study of rosiglitazone vs. placebo for 6 wk. Participants: Participants included 17 nondiabetic men with a broad range of insulin sensitivity. Intervention: Intervention included rosiglitazone 8 mg/d vs. placebo for 6 wk. Main Outcome Measure: TRL metabolism (concentration, production and catabolic rates) was assessed in a constant fed state with a 12-h primed constant infusion of [D3](L)-leucine and multi-compartmental modeling. Results: RSG treatment resulted in significant insulin sensitization with no change in body weight. Fasting plasma triglyceride (TG) concentration, however, was higher with RSG vs. placebo (P = 0.0006), as were fasting and fed TRL-TG, TRL-apoB-48, and TRL-apoB-100 (fed TRL-apoB-48: 0.93 +/- 0.08 vs. 0.76 +/- 0.07 mg/dl, P = 0.017, and fed TRL-apoB-100: 15.57 +/- 0.90 vs. 13.71 +/- 1.27 mg/dl, P = 0.029). This small but significant increase in plasma TRL concentration was explained by a tendency for RSG to increase TRL production and reduce particle clearance, as indicated by the significantly increased production to clearance ratios for both apoB-48-containing (0.43 +/- 0.03 vs. 0.34 +/- 0.03, P = 0.048) and apoB-100-containing (7.0 +/- 0.4 vs. 6.2 +/- 0.6, P = 0.029) TRL. Conclusion: These data indicate dissociation between the insulin-sensitizing effects of RSG and absence of anticipated reductions in production rates of apoB-100- and apoB-48-containing-TRL particles, which may explain the absence of TG lowering seen in humans treated with this agent.