present data that support the hypothesis that the increase in insulin sensitivity, while off recombinant methionyl human leptin, can be attributed to the rapid expansion of the adipose tissue mass. Furthermore, these authors cite the positive effects of leptin therapy on insulin sensitivity of lipodystrophic patients and of patients with mutations of the insulin receptor. As discussed in our article (7), there are limitations to those examples. In lipodystrophic patients there is marked insulin resistance, which is absent in leptin deficiency; on the contrary, prior to treatment our patients did not exhibit pronounced insulin resistance, even in the context of morbid obesity (4). In mutations of the insulin receptor there is defective signaling, which again is absent in leptin deficiency. Therefore, those examples cannot be directly compared with our data. We agree that our study does not allow us to conclude that leptin therapy by itself has a direct effect, worsening insulin sensitivity; as we discussed extensively, such an effect could be indirect through other actions of leptin. Additionally, as commented extensively in our paper, rapid weight gain is a confounding factor that needs to be minimized in future studies. Nevertheless, the uniqueness of our patients (complete absence of leptin and of leptin resistance and normal insulin sensitivity) allows us to suggest that leptin may worsen insulin sensitivity either directly or indirectly. The adipoinsular axis has many complex interactions that need to be elucidated (1). It is possible that leptin may lead to changes in insulin sensitivity but not to overall glucose homeostasis. In clinical research studies, the weight of leptindeficient patients should ideally be kept stable during the onand off-leptin euglycemic hyperinsulinemic clamps. However, it is unethical to allow excessive weight regain after leptin therapy is temporarily discontinued. Alternatively, one could restrict the patients to a rigid isocaloric diet at all times. However, it is virtually impossible for leptin-deficient patients to maintain such a diet while off leptin. The limitations of clinical research in living patients are such that not all possible variables can be controlled for at the same time. Given the constraints of patient-oriented investigation, we still believe that our results provide some of the clearest evidence that in humans leptin worsens insulin resistance, and we stand by our comment that hyperleptinemia may mediate the increased insulin resistance of obesity. Others have independently come to the same conclusion (2). We agree that further studies are needed to clarify the acute and chronic effects of leptin on insulin sensitivity. To determine whether the rapid increase in adipose tissue mass is responsible for the increase in insulin sensitivity, we plan to determine glucose uptake by performing euglycemic hyperinsulinemic clamps with radiolabeled glucose. In addition, patients will be advised to maintain isocaloric diets as much as possible. By continuing to investigate the effects of leptin on insulin sensitivity, we may be able to elucidate the roles of leptin on the adipoinsular axis. As a last word, the regulation of human glucose homeostasis is of fundamental relevance to the survival of our species. Therefore, the hallmark of such regulation is extensive redundancy, as evidenced by a vast network of signals that determine overall glucose utilization (3, 5). The specific contribution of each signal to glucose homeostasis is not necessarily fixed and immutable but may vary across time and across individuals or groups of individuals with different disorders in the context of their overall metabolic milieu.
Read full abstract