The biological mechanism by which obesity predisposes to insulin resistance is unclear. One hypothesis is that larger adipose cells disturb metabolism via increased lipolysis. While studies have demonstrated that cell size increases in proportion to BMI, it has not been clearly shown that adipose cell size, independent of BMI, is associated with insulin resistance. The aim of this study was to test this widely held assumption by comparing adipose cell size distribution in 28 equally obese, otherwise healthy individuals who represented extreme ends of the spectrum of insulin sensitivity, as defined by the modified insulin suppression test.

Subcutaneous periumbilical adipose tissue biopsy samples were fixed in osmium tetroxide and passed through the Beckman Coulter Multisizer to obtain cell size distributions. Insulin sensitivity was quantified by the modified insulin suppression test. Quantitative real-time PCR for adipose cell differentiation genes was performed for 11 subjects.

All individuals exhibited a bimodal cell size distribution. Contrary to expectations, the mean diameter of the larger cells was not significantly different between the insulin-sensitive and insulin-resistant individuals. Moreover, insulin resistance was associated with a higher ratio of small to large cells (1.66 ± 1.03 vs 0.94 ± 0.50, p = 0.01). Similar cell size distributions were observed for isolated adipose cells. The real-time PCR results showed two- to threefold lower expression of genes encoding markers of adipose cell differentiation (peroxisome proliferator-activated receptor γ1 [PPARγ1], PPARγ2, GLUT4, adiponectin, sterol receptor element binding protein 1c) in insulin-resistant compared with insulin-sensitive individuals.

These results suggest that after controlling for obesity, insulin resistance is associated with an expanded population of small adipose cells and decreased expression of differentiation markers, suggesting that impairment in adipose cell differentiation may contribute to obesity-associated insulin resistance.