For nearly the past decade we have been studying the relationship between insulin resistance and adipocyte size in a number of settings. The key ideas we have distilled from these investigations include: 1. Adipose cells vary widely in size from about 20 microns in diameter to well over 100 microns, befitting their unique role as cells that can expand and contract to take up and release lipid as needed. 2. The distributions are not typical unimodal Gaussian distributions but typically have a Gaussian-like peak of large cells and an exponential-like tail of small cells. We have interpreted the Gaussian peak as representing mature adipocytes and the tail as newer cells in the process of growing as they take up lipid. Dynamical models by others in the lab (see reports by Vipul Periwal, LBM) have confirmed that a hypothesized growth process incorporating newly recruited cells emerging with diameter around 20 microns and growing according to a size-dependent growth law naturally generates such distributions. A distinct nadir between the left tail and the right peak is found if cells are assume to grow slowly until they reach a threshold diameter of, say, 30 - 60 microns and then grow rapidly. In some individuals, both human and rodent, an additional peak of cells with intermediate diameter can be observed, which may even move to the right with time, suggesting a bolus of cells recruited close together in time that grows and joins the peak of mature cells. 3. We have correlated the characteristics of the size distributions, chiefly the fraction of small cells and the typical size of the large cells, with insulin resistance or sensitivity in a variety of study populations. Our initial study, in collaboration with the McLaughlin lab at Stanford University, examined moderately obese subjects (BMI near 30 kg/m2) who were insulin sensitive (IS) or insulin resistant (IR) and found that the insulin resistant group had an increased proportion of small cells. This was interpreted as a signature of impaired adipocyte development, which could result in impaired lipid storage capacity and lead to spillover of lipid to other organs poorly equipped to handle the fat load. Such spillover, or ectopic fat has been proposed to cause insulin resistance in muscle and liver and impaired insulin secretion in the pancreas. The study noted a trend toward larger large cells among the IR subjects, but this did not reach statistical significance. One would expect larger large cells given a smaller proportion of large cells if BMI and total fat mass are the same between the groups, as they were by design. A follow-up study with a larger group of subjects and a broader range of BMI has confirmed the increased proportion of small cells as well as larger large cells (see Ref. # 1 of the 2014 report). The increased size was also reported by us in Ref. # 1 of the 2012 report. A considerable body of evidence from other studies supports the notion that large cells are intrinsically less efficient at storing lipid than small cells (again, see report of Periwal, LBM, for a theoretical view of this). The study described in Reference # 1 of this report showed that enhancement of insulin sensitivity by treatment with the insulin sensitizing drug rosiglitazone, of the thiazolidenedione (TZD) class, was associated with the enlargement of existing large adipose cells (hypertrophy) and recruitment of new small adipose cells in T2D patients (hyperplasia) studied longitudinally over three months. In contrast, the cross-sectional studies described above showed that insulin sensitivity is associated negatively with hyperplasia and hypertrophy. We suggest that the apparent paradox can be resolved by the principle that impaired fat storage capacity leads to insulin resistance, whereas enhanced storage capacity results in improved insulin sensitivity. This principle can be extended to cover the case of weight loss, which would improve insulin sensitivity by reducing the fat load. Reference # 1 draws attention to the possibility that adipocytes can enlarge for either harmful or beneficial reasons. In insulin-resistant subjects, large cell size may result from impaired capacity to recruit new small cells or expand existing small cells. TZD treatment, in contrast, may increase the size of the large cells by reducing lipolysis. We further propose that in metabolically healthy individuals, hypertrophy and hyperplasia are linked, with mature adipocytes generating a signal to recruit new small cells when they become too large. This link may be disrupted in individuals with hypertrophic obesity, contributing to insulin resistance, and repaired by TZD treatment.
