A number of studies have established that adipose tissue dysfunction contributes to metabolic dysfunction and type 2 diabetes. Enlargement of adipocyte cell size has been shown to be associated with adipose tissue dysfunction and, in addition, to predict later development of type 2 diabetes in Pima Indians, a population with high propensity for obesity and type 2 diabetes , as well as in a Swedish cohort of middle-aged women . Although obesity is a major risk factor for the development of type 2 diabetes, not all obese individuals become insulin resistant or develop type 2 diabetes. Recently, it was shown by Klöting et al.  that insulin-sensitive, severely obese individuals have smaller adipocyte size compared to an equally obese group matched for age, sex and body fat. This finding was also associated with reduced tissue inflammation and higher insulin-stimulated glucose uptake at least in omental adipose tissue. Furthermore, by comparing non-obese subjects with a known genetic predisposition for either type 2 diabetes or obesity, we have recently shown that for a given amount of body fat individuals with a genetic predisposition for type 2 diabetes had an inappropriate enlargement of their abdominal adipocyte cell size. This difference was evident when they were compared to subjects with a genetic predisposition for obesity or to control subjects lacking a known genetic predisposition . These findings indicate that adipocyte hypertrophy combined with an “obese phenotype” is present in the abdominal adipose tissue long before type 2 diabetes develops and that this is related to insulin resistance rather than obesity per se. Clearly, adipocyte cell size and function are related to whole body insulin sensitivity.
The cohort studied in this report was part of the inter-European EUGENE 2 program (http://www.eugene2.com) . This program focused on carefully phenotyping individuals at risk for type 2 diabetes by virtue of having at least one first degree relative with this condition. This cohort has undergone several genotyping studies and is followed prospectively in order to identify future diabetes development. The individuals included in this study were healthy and non-obese, but they are a high-risk group even though current obesity was not part of the risk profile. Furthermore, they are more insulin-resistant, as a group, than matched control subjects without a family history of type 2 diabetes .
The results from the present study show that markers of adipose tissue dysregulation are present already in these otherwise healthy individuals. Adipose cell size, GLUT 4 protein and mRNA expression as well as circulating levels of adiponectin and RBP4 were all markers for degree of insulin sensitivity. Interestingly, adipose cell size was positively correlated with serum RBP4, which is consistent with previous findings , and inversely with adiponectin levels as well as with GLUT 4 expression. These findings are consistent with the concept that adipose cell expansion, even over the limited range of BMI in this cohort, is associated with insulin resistance as well as markers of a dysregulated adipose tissue measured as low GLUT4 expression and circulating levels of adiponectin and high serum RBP4 levels. The associations also suggest that these molecules may have common gene regulatory sites such as C/EBPalpha although no correlation was found between mRNA levels for this transcription factor and adiponectin or RBP4 levels. It should also be pointed out that we did not measure APM1 mRNA levels since this molecule is subject to important post-transcriptional modifications as well as a regulated secretion pathway , thus making mRNA levels less dependent than total protein secreted and present in the blood. Virtually all clinical studies on the role of adiponectin have also focused on the circulating levels of this protein. It is also noteworthy that while circulating adiponectin levels were only border-line correlated to BMI and serum RBP4 levels not at all, the association with adipose cell size was highly significant. These findings are in line with the results presented by Klöting et al. where circulating adiponectin is decreased and RBP4 increased in equally obese individuals with enlarged adipocytes and reduced insulin sensitivity .
Interestingly, insulin sensitivity was most closely related to waist/hip ratio and adipose cell size while BMI was a poor marker. The lack of correlation with BMI is probably due to the relatively limited range of BMI in this cohort since BMI is well known to be associated with insulin sensitivity in large population samples with different degree of obesity. Thus, this study shows that adipose cell size and adipose tissue distribution are more sensitive parameters over a relatively limited range of BMI in a cohort of non-obese subjects. In addition, these results show that adipose tissue dysregulation does not require obesity per se but rather hypertrophic adipose cells. The hypertrophic characteristics of the adipose tissue is probably due to inability to recruit and/or differentiate preadipocytes which results in excessive lipid deposition in, and enlargement of, pre-existing adipocytes. We recently found that even though precursor cells are available in adipose tissue there seems to be a blockage of the commitment and/or initiation of adipocyte differentiation leading to impaired preadipocyte recruitment in individuals with enlarged fat cells . This is in line with recent findings showing that the number of new adipocytes generated each year is reduced in subjects with adipocyte hypertrophy while the relative death rate is unchanged . What constitutes the blockage is still unknown, but clearly the resulting adipocyte hypertrophy is associated with adipose tissue dysfunction. We here show that this is also related to reduced insulin sensitivity combined with a metabolic risk profile and markers of adipose tissue dysregulation. It is well known that adipose tissue distribution differs between men and women but, interestingly, it has also been shown that the proportion of early-differentiated adipocytes, measured as percentage of PPARgamma expressing cells in the subcutaneous adipose tissue is increased in women when compared to men indicating that there may be important gender-related differences in pre-adipocyte recruitment, proliferation and differentiation potential . Our results show that the correlation between adipocyte cell size and GLUT4 as well as the insulin resistance marker RBP4 are affected by gender. This is also true for the association between adiponectin and GLUT4, both markers of late adipocyte differentiation and function. These findings add further strength to the concept of gender differences in adipocyte differentiation and function. Tchoukalova et al. speculate that a possible mechanism could include gender-specific differences in the micro-environment and/or effects of sex-steroid hormones. Sex steroid-hormones have been shown to influence fat distribution  as well as adipocyte differentiation  and could well be responsible for the gender differences observed. Unfortunately, we did not measure sex-steroids in the present study. However, a recent paper investigating clinical characteristics associated with insulin sensitivity in women with polycystic ovary syndrome (PCOS) showed that the strongest predictors of insulin sensitivity in this group were adipocyte cell size, adiponectin and WHR, while sex steroid-hormones were excluded from the regression model . Further studies are required to elucidate the importance of, and mechanisms behind, these gender-associated differences.
It is well established that enlarged adipose cells leads to infiltration of macrophages and other inflammatory cells, including mast cells [32–34]. The presence of inflammatory cells in the adipose tissue affects the micro-environment and can impair adipocyte differentiation [7, 17]. Indeed, macrophage infiltration in the omental adipose tissue depot, together with circulating adiponectin was found to almost completely explain the degree of insulin sensitivity in severely obese individuals . We previously measured the inflammatory cytokine IL6 in adipose tissue and showed that expression, secretion and, as a consequence, also interstitial levels of this cytokine were increased in the adipose tissue characterized by enlarged fat cells . Thus, inflammation seems to follow adipose cell size enlargement and this is also associated with impaired adipocyte differentiation . Cellular hypoxia has also been implicated in adipose tissue dysregulation in obesity . However, we found no relationship between HIF-1alpha mRNA levels, or VEGF, which is an HIF-1alpha-regulated gene , and adipose cell size or any marker of insulin resistance. This is in agreement with previously reports in obese individuals where the expression in subcutaneous adipose tissue was unrelated to the degree of insulin sensitivity or cell size. In contrast, the expression of HIF-1alpha has been shown to be up-regulated in insulin resistant omental adipose tissue in severe obese individuals . Thus, although HIF-1alpha may play a role in severe obesity, we did not find any association between insulin sensitivity and HIF-1alpha in this small group of individuals with hypertrophic adipocytes.
The results of the present study clearly show that enlarged abdominal adipose cells are associated with reduced systemic insulin sensitivity irrespective of whether obesity is present or not.
A likely reason for the insulin resistance is altered RBP4 and adiponectin levels as well as an inability to store additional lipids in the subcutaneous depot during weight gain. This leads to storage in ectopic sites including visceral depots, liver and muscle which, in turn, further promotes insulin resistance (Reviewed in [15, 37]). Elegant experiments in mouse models have indeed shown that mice overexpressing adiponectin in the subcutaneous adipose tissue become grossly obese with hypercellular adiposity as a consequence of new preadipocyte recruitment and differentiation. The changes associated with this transgene did not impair insulin sensitivity at all . The present results further support the concept that pre-adipocyte recruitment and hypercellular obesity can prevent the development of insulin resistance.
The present study is limited by its small number of subjects and using the quite conservative methods available to correct for multiple testing leaves few significant correlations. However, regardless of these limitations, the results provide important information in a high-risk cohort of first-degree relatives to type 2 diabetic patients showing that a dysregulated adipose tissue occurs early and is associated with insulin resistance. Future studies, such as long-term follow-up studies of the EUGENE 2 cohorts may provide further evidence for this concept as a risk to develop type 2 diabetes as well.