Glucose-Dependent Foxo1 Switch in Healing Wounds: A Shred of Hope for Diabetic Ulcers?

Diabetes mellitus (DM) has been labeled the epidemic of the 21st century, claiming its toll on both quality and duration of life and posing a major economic burden to society (1). By the year 2050, it is predicted that 10% of the population will be diagnosed with DM. Approximately 15% of diabetic patients will develop a nonhealing wound or diabetic foot ulcer. These wounds are the number one cause of lower-limb amputations (2). Several defects have been proposed to explain impaired wound healing in diabetic patients including compromised angiogenesis, inadequate expression of growth factors, excessive cytokine release at the site of injury, reduced epithelial cell migration, and proliferation and defective insulin action in the skin (3–5). However, the specific molecular events leading to impaired wound repair in DM are still unknown. Against this backdrop, it is perhaps not surprising that effective therapeutic strategies to treat these wounds remain limited.

The Foxo1 transcription factor is a member of the evolutionarily conserved, mammalian forkhead box O class (Foxo) family that also includes Foxo3, Foxo4, and Foxo6. These transcription factors play a major role in diverse processes including cellular differentiation, proliferation, apoptosis, stress resistance, and glucose metabolism. Foxos are negatively regulated by the insulin-like growth factor 1 (IGF1)/insulin pathway, and they are activated by oxidative stress through Jun N-terminal kinase signaling. Foxo regulation is orchestrated by numerous post-translational modifications that result in nucleocytoplasmic shuttling (6). In turn, Foxos bind to hundreds of DNA sites at promoter or adjacent regions through a common consensus sequence (7). The Foxos have been implicated in pathways regulating life span and age-associated diseases including cancer and DM. The function of Foxo in response to environmental changes is to maintain homeostasis, and their activity is mainly revealed under stress conditions, such as what is encountered in DM (8). Foxo1 involvement in systemic glucose homeostasis is complex (9); it serves as a mediator and modulator of insulin signaling depending on the cell type and metabolic setting (10), promotes hepatic expression of gluconeogenic and glycogenolytic genes (11), prevents β-cell dedifferentiation (12), and controls energy expenditure through hypothalamic networking (13).

Previous studies have linked Foxo1 to both physiological and pathological processes of epithelial wound healing. Foxo1 expression is required for intact wound healing by promoting keratinocyte migration and upregulation of transforming growth factor β1 (TGF-β1) and reducing oxidative stress (14). However, the optimal expression level of Foxo1 for wound healing is debatable because Foxo1^+/− mice with attenuated Foxo1 expression actually exhibited enhanced wound repair with reduced scarring (15). In pathological healing states such as keloid scars or diabetic skin lesions, Foxo1 expression (15) and activity (16) are increased along with impaired insulin signaling (17,18), and these were reversed with topical insulin administration (18).

In this issue of Diabetes, the newly published article by Xu et al. (19) examines the role of Foxo1 in re-epithelialization of oral mucosal wounds in diabetic and normoglycemic mice. Foxo1 was deleted in keratinocytes using the Cre-lox system, thereby generating experimental (K14.Cre⁺.Foxo1^L/L) and control (K14.Cre⁻.Foxo1^L/L) mice. Type 1 DM was induced by streptozotocin, and mucosal wounds were established by punch biopsy of the dorsal surface of the tongue. In these diabetic mice, mucosal wounds showed significantly delayed healing with reduced cell migration and proliferation. DM enhanced expression of chemokine (C-C motif) ligand 20 (CCL20) and interleukin-36γ (IL-36γ) and increased nuclear translocation of Foxo1. Foxo1 deletion had the opposite effects on wound healing in diabetic versus normoglycemic mice. While it rescued the negative impact of DM, reduced healing time and chemokine expression, and improved epithelial cell migration and proliferation, Foxo1 clearly impaired healing under normoglycemic conditions. Indeed, in nondiabetic mice, Foxo1 deletion reduced TGF-β1 expression in mucosal epithelial cells, a finding that was corroborated in vitro by Foxo1 silencing in primary human mucosal epithelial (PHME) cells. Notably, the reduced migration seen in Foxo1-silenced, TGF-β1−deficient PHME cells was completely reversed by TGF-β1 treatment. Based on these observations, the authors propose that under normal conditions Foxo1 expression is needed in oral mucosal epithelial cells for TGF-β1 expression because it promotes epithelial cell migration and wound healing. In contrast, under hyperglycemic conditions the dominant Foxo1-related effect is increased DNA binding that leads to an inflammatory phenotype. This phenotype was characterized by upregulation of CCL20 and IL-36γ, which resulted in reduced migration, proliferation, and delayed healing (Fig. 1).

Xu et al. (19) address a burning medical issue—the mechanisms that underpin impaired healing of diabetic skin lesions. Although their report focuses on mucosal lesions and not the more common cutaneous lesion in diabetic foot ulcers and it uses a model of type 1 DM that is associated with absolute and not relative insulin deficiency, it makes a noteworthy contribution to the understanding of the molecular mechanisms involved in this common DM complication. Clarifying the apparent dichotomy of Foxo1 activity in normoglycemic versus hyperglycemic/insulin-deficient states may contribute to the development of therapies for complications of DM, possibly through manipulation of Foxo1 cellular localization.

As is most often the case with novel findings, many new questions arise: Is there a glucose threshold for the Foxo1 “switch” vis-à-vis wound healing? Why is Foxo1-dependent TGF-β1 expression less important for diabetic re-epithelialization than for normal wound healing? Finally, as tissue targeting is relatively easy in readily accessible diabetic wounds, can the novel insight into diabetic ulcer formation and healing offered by this article be translated to a local form of therapy? That such an approach may be feasible has been recently suggested by the finding that local delivery of antisense oligodeoxynucleotides to Foxo1 at the skin wound site improved healing (15) in a nondiabetic mouse model. Older clinicians may recall times when insulin was applied locally to diabetic ulcers (20). Because, as shown in the current study, insulin retards nuclear binding of Foxo1, could we refresh this abandoned approach with the aid of a more refined, local Foxo1-related therapy, independent of its systemic effects?