HIF-2α blows out the flames of adipose tissue macrophages to keep obesity in a safe zone.

In 1993, Hotamisligil et al. (1) broke new ground by linking immune response and metabolism. A striking observation in this early study was the upregulation of the inflammatory cytokine tumor necrosis factor-a in adipose tissue (AT) of obese and insulin-resistant animals. Although adipocytes were thought to be the major source of inflammatory cytokines in fat, Weisberg et al. (2) later showed that macrophages accumulate in the AT of obese animals, creating a state of low-grade inflammation that could trigger insulin resistance. Since then, the concept of “immunometabolism” has gradually evolved, and many types of immune cells have been shown to contribute to AT inflammation and insulin resistance (3). For instance, macrophages secrete proinflammatory cytokines and other factors, which could impair insulin sensitivity. In light of this evidence, it has been suggested that decreasing inflammation in AT could attenuate the diabetic state. In support of this idea is the fact that anti-inflammatory drugs appear to have provided proof-of-concept that inflammation is linked to the deleterious effects of insulin resistance in animals (4,5). Paradoxically, recent studies suggest that inflammation is also required for maintaining AT homeostasis (6–8). These complex findings emphasize the need for a more nuanced understanding of the role immune cells play in the development of insulin resistance. Adipose tissue macrophages (ATMs) have been categorized into two distinct populations: proinflammatory (M1) and anti-inflammatory (M2). While ATM subpopulations have overlapping inflammatory profiles, it is generally accepted that a polarization of M2 macrophages toward an M1 phenotype occurs in the AT of obese animals. Importantly, the AT microenvironment seems to play a major role in this phenotypical switch (9). One of the signals triggering this ATM polarization is hypoxia, which has been of particular interest in recent years (10). The lack of de novo angiogenesis during the large AT expansion that occurs in obesity leads to a reduced blood flow and increased oxidative stress in this tissue in animals (7). However, whether this hypothesis has application in human AT is still controversial (11). At the molecular level, hypoxia stimulates the transcriptional activity of the hypoxia-inducible factor, HIF-1a. In turn, this factor controls the expression of inducible nitric oxide synthase (iNOS), which is involved in nitric oxide (NO) production (12). Excessive NO production in AT promotes inflammation, thereby contributing to the impairment of insulin sensitivity in obesity (9). HIF-2a is another hypoxia-responsive component of the HIF family that is also expressed in macrophages (13). HIF-1a and HIF-2a have been shown to play opposite roles in the regulation of macrophage function in vitro and in vivo (14). This raises the important question that Choe et al. (15) address in this issue of Diabetes: How do these two isoforms regulate ATM phenotype and its consequence on whole-body metabolism? Choe et al. demonstrate that M2 macrophages express high levels of HIF-2a in the AT of obese mice. It has been recently shown that HIF-1a and HIF-2a antagonize each other in the regulation of NO production, resulting in different effects on the macrophage inflammatory profile (14). iNOS produces NO by metabolizing its substrate, the amino acid L-arginine (16). Arginase 1 (ARG1) is highly expressed in M2 macrophages and competes with iNOS for their common substrate, L-arginine, to produce ornithine and urea (17). Therefore, ARG1 activity can decrease NO production via the limitation of arginine availability (18). Choe et al. provide evidence that HIF-2a could partially prevent obesity-associated inflammatory response and insulin resistance through activation of ARG1 (Fig. 1). Accordingly, in gainand loss-offunction experiments, macrophage HIF-2a prevented