Obesity-induced inflammatory changes in adipose tissue (original) (raw)
Earlier transcriptional profiling experiments performed in adipose tissue from lean and obese animal models (8), or following treatment of obese animals with members of the thiazolidinedione class of antidiabetic agents (9), have pointed to striking regulation of a large repertoire of inflammatory genes in adipose tissue. Two articles in this issue of the JCI illustrate that macrophage infiltration into adipose tissue in obesity could be integral to these inflammatory changes (10, 11). Both of these studies have principally emerged from large-scale gene-expression analysis. Xu and colleagues directly compared gene expression in multiple tissues between five obese mouse models and their lean controls (10). Weisberg et al. chose to profile gene expression in the white adipose tissue of mice of varying degrees of obesity, in order to identify correlations between gene expression and the degree of obesity (11). In support of earlier studies, both approaches indicate that the largest class of genes significantly regulated in obesity consists of macrophage and inflammatory genes in white adipose tissue.
In these articles, both groups considered the possibility that the cellular source of these inflammatory changes may not be only adipocytes, but also reticuloendothelial cells present in adipose tissue or perhaps adipocyte precursors (10, 11). The studies to address this possibility led to the observations that some inflammatory responses took place outside of adipocytes in obesity, and that significantly higher numbers of macrophages infiltrated into the expanding adipose tissue. Coming to this conclusion, however, was not easy, precisely due to the similarities of macrophages to adipocytes and preadipocytes, as mentioned above. To resolve the critical question of accurate identification of the macrophages, the investigators performed a series of elegant and rigorous experiments. Xu et al. began by separating white adipose tissue into a buoyant fraction containing adipocytes and a pelleted stromal-vascular fraction consisting of nonadipocyte cells in the adipose tissue, such as preadipocytes and macrophages. They found that several inflammatory genes chosen from their microarray analysis were in fact expressed predominantly in the stromal-vascular fraction (10). The challenge was then to distinguish between adipocyte precursors and macrophages in this fraction. Histological examination revealed that clusters of small, nucleated cells were present in obese, but not in lean, adipose tissue, and that these clusters became larger and more numerous as the animals aged and gained weight. Staining for F4/80, a macrophage-specific antigen, was also prominent in the obese fat pads as compared with those of lean mice. Experiments in cultured clonal preadipocytes confirmed that the same inflammatory genes identified in the stromal-vascular fraction of white adipose tissue were not expressed in these cells. However, upon stimulation with TNF-α, an inflammatory cytokine that is overexpressed in obesity and contributes to insulin resistance, preadipocytes did express some inflammatory genes, suggesting that they too might mount a similar inflammatory response under select conditions.
Weisberg and colleagues also stained for the F4/80 antigen and found that the percentage of F4/80+ cells correlated positively with average adipocyte size (11). To confirm that the F4/80+ cells were in fact macrophages, the authors first used FACS to separate F4/80+ cells from F4/80– cells in the stromal-vascular fraction of ob/ob mouse fat pads. The F4/80+ cells were found to express several other macrophage-specific markers. Taking advantage of the fact that macrophages originate in the bone marrow, the authors subsequently lethally irradiated mice expressing the CD45.2 leukocyte marker and transplanted bone marrow from CD45.1-expressing mice. After placing the mice on a high-fat diet, the authors observed that 85% of the F4/80+ cells in the adipose tissue were in fact carrying the donor marker and, thus, were bone marrow derived. Furthermore, in Csf1op/op mice, which lack macrophages, the presence of F4/80+ cells was greatly reduced as compared with that in wild-type mice. However, these animals were not studied under obese conditions. Importantly, Weisberg et al. also provide evidence that macrophage infiltration of adipose tissue is characteristic of human obesity, by determining that both BMI and average adipocyte size were significant predictors of macrophage accumulation in adipose tissue, as assessed by the percentage of CD68+ cells.