Differential effects of adiposity on peroxisomal proliferator-activated receptor gamma1 and gamma2 messenger ribonucleic acid expression in human adipocytes.

Both genetic and pharmacological studies raise the possibility that a primary increase in the amount or activity of peroxisomal proliferator-activated receptor gamma (PPARgamma) in adipocytes could play a role in common types of human obesity. Using real-time RT-PCR assays we examined the relationship between body mass index (BMI) and PPARgamma isoform expression in freshly isolated human adipocytes. There were no consistent differences in the expression of either PPARgamma1 mRNA or PPARgamma2 mRNA between omental and sc adipocytes. In a group of 17 subjects (BMI range, 17-34 kg/m(2)) there was a strong and highly significant inverse correlation (r = -0.68; P < 0.005) between PPARgamma1 mRNA expression in adipocytes and BMI, whereas no significant relationship was apparent for PPARgamma2. In an independent study PPARgamma1 mRNA levels were decreased (1.1 +/- 0.1 vs. 3.7 +/- 0.8 arbitrary units; P < 0.01) in adipocytes from morbidly obese (BMI, 50.6 +/- 14.1 kg/m(2)) vs. lean (BMI, 21.1 +/- 1.0 kg/m(2)) subjects. In contrast, there was a significant increase in the expression of PPARgamma2 mRNA levels between the morbidly obese and lean groups (1.7 +/- 0.2 vs. 1.1 +/- 0.2 arbitrary units; P < 0.05). Treatment of isolated human adipocytes with TNFalpha resulted in a significant decrease in both PPARgamma1 and PPARgamma2 mRNA levels [40.6 +/- 5.5% relative to control (P = 0.01) and 60.9 +/- 24.8% (P = 0.02) respectively]. The strong inverse relationship between BMI and PPARgamma1 expression in human adipocytes is striking and may represent part of an autoregulatory mechanism restraining the expansion of individual adipocytes in states of positive energy balance. On the other hand, the increase in PPARgamma2 observed in adipocytes of morbidly obese individuals suggests a potential pathogenic effect of this isoform in promoting fat acquisition. Although an autocrine effect of the enhanced TNFalpha secretion seen with increasing obesity might play a role in the changes in PPARgamma1, this would not provide an explanation for the different relationship of PPARgamma2 to adiposity. The significance of the divergent effect of human adiposity on the two isoforms will require a greater understanding of the differential properties of the two isoforms and of the differences in the functions of their respective regulatory elements.