Anand and Chada1 have reported that knockouts of Hmgic prevent adipocyte differentiation and therefore might serve as a target for obesity. Prevention of adipocyte differentiation, however, is destined to exchange obesity for diabetes and suggests an interesting hypothesis: that type II diabetes is the result of the inability of the adipose organ to expand to accommodate excess calories and that type II diabetes in the centrally obese, in spite of their unlikely phenotype, is a form of lipodystrophy. If the adipose organ is unable to accommodate excess energy, the calories are stored in the liver, muscles and, with the development of diabetes, in the blood and urine. This is most clearly seen in acquired total lipodystrophy where a positive energy balance begets diabetes2, and is revealed in obese type II diabetics when, early in their disease, caloric restriction reverses the diabetes.

This is supported by Gavrilova et al.3, who reversed the diabetes in a mouse model of lipodystrophy by surgically implanting adipose tissue. It is supported by Okuno et al.4, who found that thiazolidinediones improved insulin sensitivity in Zucker rats by increasing the number of new adipocytes. It is supported in studies in obese Pima Indians, in which the best correlation with the onset of diabetes is with adipocyte size, suggesting difficulty in differentiating new ones5. It is supported by the observation that insulin sensitivity during overfeeding correlates with the recruitment of new adipocytes6, and that the in vitro yield of newly differentiated adipocytes is greater in lean than in obese subjects.

Adipocytes are terminal cells that when filled to capacity (∼3 μg lipid/cell) are extremely insulin resistant. Therefore, excess energy intake must be balanced by the expansion of the adipose organ. If not, the energy is stored in the liver, muscle and blood and is accompanied by insulin resistance as reported by Randle and colleagues7. Hepatic steatosis and excess lipid in muscle and pancreas is characteristic of obese diabetics8. These abnormal deposits of lipid may lead to the trilogy of defects characteristic of type II diabetes mellitus: over-production of glucose by the liver, muscle insulin resistance and blunted secretion of insulin by the pancreas. Several laboratories now support this, including those of McGarry8 and Unger9. Therefore, any scenario that targets the destruction of adipocytes or prevents their differentiation in the face of excess caloric intake is doomed to exchange obesity for type II diabetes mellitus.

Thirty years ago, Knittle and Hirsch10 suggested that too many adipocytes predisposed to obesity. I suggest that too few adipocytes predisposes to type II diabetes mellitus. This could explain the lower prevalence of type II diabetes in the generalized and more hypercellular obese than in the centrally obese, who for genetic or environmental reasons have lost the ability to accommodate excess energy by differentiating new adipocytes.