Diabetic osteopathy is an important but generally neglected complication of diabetes (1). The importance of diabetic osteopathy is illustrated by patients with T1D having a twofold greater risk of any fracture and up to a sixfold greater risk of hip fracture than people without diabetes (2). In people with T2D, the risks of hip and foot fracture are increased by 1.8-fold and 2.7-fold, respectively, compared with the risk to people who do not have diabetes (3). However, the mechanisms underlying increased bone fragility in diabetes are complex and still incompletely understood (4).
Regular physical exercise was shown in humans and rodents to prevent and improve T2D, mainly by improving insulin sensitivity in muscle and other tissues. Exercise also has beneficial effects on bone health, as it increases bone mass in animal models and in women with osteoporosis (5). The mechanisms by which exercise affects glucose metabolism and bone remodeling are not fully defined. This apparent coupling between bone and muscle has been thought to be mainly dependent on mechanotransduction, which implies that bone formation is triggered by the mechanical forces applied to muscle. However, the existence of a bone–muscle endocrine axis has also been proposed to explain the concomitant positive affect of exercise on bone, muscle, and insulin sensitivity (6). This hypothesis has led to the identification of several exercise-regulated endocrine factors produced by either bone or muscle and implicated in various physiological functions. At least two bone-derived hormones, or osteokines, have been shown to couple exercise to the regulation of glucose metabolism or to muscle function: undercarboxylated osteocalcin (ucOCN) and lipocalcin-2. ucOCN circulating levels increase following exercise in mice and humans (7,8), and this hormone is necessary and sufficient to increase muscle function during exercise in mice (8). In addition, ucOCN improves insulin sensitivity and energy expenditure and can prevent insulin resistance in rodents (9). The lipocalin-2 levels also increase in the blood following exercise, but this osteokine acts mainly on the brain to restrict food intake (10). Among the many muscle-derived factors, or myokines, that have been identified over the past decade, two are known to affect both bone and energy metabolism: interleukin-6 (IL-6) and irisin. Muscle-derived IL-6 rises sharply in the circulation during exercise and promotes lipolysis in the fat tissue and gluconeogenesis in the liver (11). Interestingly, IL-6 also has an effect on bone by promoting bone formation and bone resorption (12,13).
Irisin is another myokine released in response to exercise that is generated by the proteolytic cleavage of a membrane-bound precursor named fibronectin-type III domain-containing 5 (FNDC5) (14). Initially described as a hormone promoting the “browning” of the white adipose tissue, thereby stimulating thermogenesis and energy expenditure, irisin has since been shown to affect other tissues, including bone (15). However, the precise action of irisin on bone cells remains uncertain, as multiple studies presented conflicting results. Some groups reported that recombinant irisin promotes bone formation by osteoblasts in culture, leading to increased bone density in vivo (16,17), while another group found that irisin stimulates osteoclast formation and bone resorption (18). Results obtained with loss-of-function models are also contradictory. Global inactivation of the Fndc5 gene in mice did not affect bone density under normal conditions but prevented the bone loss associated with ovariectomy by reducing the number of bone-resorbing osteoclasts (19). In contrast, conditional inactivation of Fndc5 in osteoblast progenitors resulted in a low-bone-mass phenotype associated with decreased numbers of osteoblasts and increased numbers of osteoclasts (20). Obviously, additional work is needed to understand the role of irisin in bones under normal and pathological conditions. In addition, two important questions have not been addressed in previous studies: Does exercise prevent diabetes osteopathy through irisin? If so, what is the source of irisin affecting bone cells?
The study by Behera et al. (21), published in this issue of Diabetes, sheds new light on the function of irisin in bones in the context of diabetes and may help answer these questions. Here, the authors take advantage of mice in which T2D is recapitulated by a combination of high-fat-diet feeding and the partial destruction of their pancreatic β-cells through injections of streptozotocin. They then tested if exercise (i.e., 60 min of treadmill running 5 times per week) could reverse the bone loss associated with this pathological condition in an irisin-dependent manner. First, they show that exercise reverses both hyperglycemia and the low bone mass associated with diabetes in their model. They also found that T2D was associated with a reduced circulating level of irisin, while, as expected from previous studies, exercise could reverse this trend. A key finding of this study is the identification of bone as a major source of circulating irisin in response to exercise. Expression of Fndc5 is induced in bone and in bone marrow mesenchymal cell cultures following exercise, and knockdown of Fndc5 in the femurs through lentiviral delivery of specific shRNAs completely abolishes the rise in serum irisin associated with exercise. In bone marrow mesenchymal cell cultures, irisin reduces the level of pyroptosis, a specialized cell death program triggered by a proinflammatory signal, including the IL-1β/inflammasome pathway (22). Mechanistically, the authors identify a single miRNA, miR-150, whose expression is increased in bone marrow mesenchymal cells derived from the T2D mice but not in cultures derived from T2D animals who have exercised. This miRNA appears to target and downregulate directly the Fndc5 mRNA, explaining why the irisin level is reduced in T2D mice and how exercise can reverse this trend. Importantly, silencing of miR-150 in T2D mice increased serum irisin levels in the blood and normalized bone formation and bone resorption parameters. Altogether, the data presented in this study suggest that bone-derived irisin mediates, at least in part, the positive effect of exercise on bone and insulin sensitivity (Fig. 1).
The importance of this study resides in the fact that it provides additional evidence for a beneficial effect of irisin on bone formation and a negative effect on bone resorption. It also identified irisin as a key mediator of the positive effect of exercise on bone density under diabetic conditions. These findings are in line with earlier studies showing that recombinant irisin promotes osteoblast function and inhibits osteoclast differentiation (16,17) and that, conversely, inactivation of Fcnd5 in osteoblasts causes reduction in osteoblastic bone formation and increased bone resorption (20). Importantly, this new study also suggests that the rise in circulating irisin following exercise is, for the most part, originating from bone, not from muscle, at least in the context of T2D.
Several questions remain unanswered regarding the role and regulation of irisin in bone. The exact cell type producing irisin in bone in response to exercise remain to be identified, since the evidence provided here is based on a nonspecific knockdown of Fcnd5 with lentiviral particles injected in the bone marrow cavity. Additional experiments involving cell-specific inactivation of Fcnd5 in bone mesenchymal stem cells, osteocytes, or osteoclasts would help answer this question. Of note, a previous study in which Fcnd5 was specifically inactivated in the osteoblast lineage using Osterix-Cre suggests that between 10% and 15% of circulating irisin originates from these bone cells (20). Nevertheless, these data call into question the myokine status of irisin, since they suggest this hormone can be released by tissues other than muscle. It is also not clear from the current study if the circulating level of irisin is relevant or not to bone health. Indeed, if bone cells produce irisin, it most likely acts in a paracrine manner on the osteoblasts or mesenchymal progenitors present in the bone marrow. Another unanswered question is by which mechanism does exercise suppress miR-150 expression in bone cells, thereby increasing Fcnd5 expression? Does it involve another myokine, such as IL-6, produced in response to exercise and signaling to osteoblastic cells in the bone? Although αVβ5 integrin was shown to be a receptor for irisin in white adipose tissue and osteoclasts, the current study did not determine if this was also the case in osteoblast precursors. Finally, in their study Behera et al. (21) show that exercise and recombinant irisin increase the expression of osteocalcin at the mRNA and protein levels in bone, resulting in higher levels of circulating bioactive ucOCN. As mentioned above, this osteokine, produced in response to exercise (7,8), promotes insulin sensitivity and reduces blood glucose in prediabetic mouse models (23). These data therefore raise the intriguing possibility that the beneficial impact of irisin on glucose metabolism is in part indirectly mediated by ucOCN (Fig. 1). In conclusion, this study revealed a new role of bone-derived irisin in protecting bone in the context of diabetes and suggests it is an attractive therapeutic approach for diabetic osteopathy.
See accompanying article, p. 2777.
Funding. This work was supported by funding from the Canada Research Chair program, Diabetes Canada, and the Canadian Institutes of Health Research (PJT-175025 and PJT-159534).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.