Macrophage Dysregulation and Impaired Skin Wound Healing in Diabetes

1Skin wound healing involves distinct but overlapping phases of hemostasis, inflammation, proliferation, and remodeling. Immediately following injury, platelet aggregation is important to achieve coagulation and hemostasis in the wound. The inflammatory phase is induced by pro-inflammatory mediators released by injured tissues, and is critical for controlling infection, clearing necrotic debris, and induction of the wound healing process. Next, the proliferative or tissue formation phase involves proliferation of a number of cell types to form a provisional connective tissue matrix, new blood vessels, and epithelial closure. Finally, during the remodeling phase the newly formed tissues are remodeled to improve their integrity. A diverse set of cells such as platelets, mast cells, neutrophils, monocytes (Mo), macrophages (Mϕ), lymphocytes, keratinocytes, fibroblasts, and endothelial cells all contribute to the process of skin wound healing.

2Among all cell types Mo/Mϕ play critical roles in each phase of wound repair through host defense, tissue debridement and cell regulatory functions. Studies using LysM-Cre/DTR genetically modified mice that allow for inducible depletion of Mo/Mϕ by diphtheria toxin (DT) administration provide strong evidence that these cells are required for normal wound healing, promoting angiogenesis, collagen deposition, and closure.

3Properly regulated numbers and phenotypes of Mo/Mϕ are crucial for efficient wound repair, and the dysregulation of either may lead to impaired wound healing. For example, increased numbers of wound Mo/Mϕ have been shown to be associated with impaired wound healing in diabetes. Similarly, an impaired transition from pro-inflammatory into pro-healing wound Mo/Mϕ phenotypes and reduced phagocytic ability contribute to chronic inflammation and impaired wound healing in diabetes. This brief review considers the origin, heterogeneity and function of wound Mϕ during normal wound healing followed by discussion of how dysregulation of numbers and phenotypes of wound Mϕ may lead to impaired diabetic wound healing. The review also highlights the possible links between altered bone marrow myelopoiesis, wound macrophage dysfunction and impaired wound healing, and finally highlights gaps in the current literature, whose filling could lead to new therapeutic interventions for diabetic wounds.

4Skin wound Mϕ originate both from tissue resident Mϕ and infiltrating Mo with significantly larger contribution from the latter. Dermal Mϕ are likely early responders to skin wounding via recognition of damage associated molecular pattern (DAMP) molecules or pathogen associated molecular pattern (PAMP) molecules. These tissue-resident Mϕ originate from yolk sac but are replenished by fetal liver-derived Mo in the embryo and by bone marrow Mo after birth. The major functions of these Mϕ are maintenance of skin homeostasis and integrity, tissue repair, and stress response. In addition, Langerhans cells, which are epidermal dendritic cells but share Mϕ markers such as MHC-II, F4/80 and CD14 also play important roles in wound healing. Langerhans cells originate both from the yolk sac during primitive hematopoiesis and fetal liver-derived Mo during definitive hematopoiesis. However, in contrast to dermal Mϕ, Langerhans cells are maintained by self-replication without any replenishment from bone marrow monocyte pool.

5Skin wounding induces a rapid, large infiltration of inflammatory Mo (CCR2+Ly6C+) into wounds followed by conversion of the Mo into Mϕ (Ly6C–F4/80+) as healing progresses. Blood Mo are thought to be the main source of wound Mo/Mϕ and a rapid decrease in CD11b+CD115+Ly6Chi blood Mo 4–6 h post wounding correlates in time with the increase of inflammatory Mo in skin wound Mo. After infiltrating wounds, novel recent findings demonstrate that inflammatory Mo/Mϕ (Ly6ChiF4/80–/lo) proliferate rapidly peaking on day 6 post-wounding. In contrast, the majority of mature wound Mϕ (Ly6C–F4/80+) remain at resting G0 phase indicating that proliferation of infiltrating inflammatory Mo followed by their differentiation into mature Mϕ results in wound Mϕ expansion. In addition, several studies have demonstrated that bone marrow-derived Mo contribute to skin wound Mϕ and that similar to other tissue injuries such as myocardial infarction and hindlimb ischemia, skin wounding also promotes bone marrow monopoiesis in mice. However, unlike myocardial infarction, skin wounding-induced monopoiesis in bone marrow occurs independently of IL-1R1 signaling. Altogether, these data clearly suggest that there is a communication between skin wounding and bone marrow for increased Mo production which may be critical for normal wound healing.

6During the early phases of healing, wound Mo/Mϕ exhibit a pro-inflammatory or “classically activated” M1-like phenotype, which gradually transitions into a healing-associated or “alternatively activated” M2-like phenotype; it is well documented that such transition of phenotypes is essential for normal wound healing. In addition to the M1/M2 or related classification schemes, other phenotypically distinct Mϕ subsets have been reported to play important roles during skin wound healing. For example, dermal Mϕ identified as CD64+, MERTK+, and CCR2–/low influence wound healing by their highly phagocytic nature. A recent study has demonstrated two distinct Mϕ subsets in skin wounds which differ in both function and origin and are distinguishable by surface CX3CR1 staining. CX3CR1hi Mϕ were derived from tissue resident Mϕ and were predominantly alternatively activated, whereas CX3CR1–/lo wound Mϕ were derived from recruited Mo and exhibited both classical and alternative activation states. Another subset of tissue resident Mϕ known as skin trans endothelial radio-resistant anti-inflammatory Mϕ (STREAM) were found to be located in perivascular regions and constitutively express an anti-inflammatory transcriptional profile. Interestingly, these Mϕ were resistant to polarization toward inflammatory phenotypes under inflammatory stimuli, hence appearing to be critical for tissue repair and regeneration. Similarly, another report described CD11b+F4/80+CD206+CD301b+ wound Mϕ to be critical for reparative mechanism which are increased during the proliferative phase of wound healing. These data suggest that there are functionally distinct Mϕ subsets in skin wounds which play critical roles at different stages of wound healing, however, how the reported Mϕ subsets may be related to each other, how each is regulated and their precise roles in healing remain to be determined.

7Tissue-resident dermal Mϕ are likely among the earliest responders to skin injury, helping to induce the inflammatory response via release of hydrogen peroxide resulting in recruitment of blood neutrophils and monocytes. During early wound healing, Mo/Mϕ help to clear the wound of contaminating microbes as well as apoptotic neutrophils and cellular debris via phagocytosis. The importance of Mϕ in such wound debridement is supported by studies targeting macrophage peroxisome proliferator-activated receptor γ (PPARγ), which plays a role in efferocytosis. PPARγ KO mice exhibit increased accumulation of apoptotic neutrophils in wounds and impaired wound healing, indicating impaired clearance of apoptotic cells. Further, treatment with a PPARγ agonist reduced accumulation of apoptotic neutrophils in wounds and improved healing.

8Another important function of Mϕ is to regulate the activity of other wound cells via the production and release of many different cytokines and growth factors. Early after tissue injury, Mϕ release numerous inflammatory mediators including IL-1β, TNF-α, IL-6, and others to amplify the inflammatory response. In addition, wound Mo/Mϕ are an important source of growth factors such as VEGF, which is critical for angiogenesis and tissue growth. Later in the healing process, Mϕ secrete other growth factors such as TGF-β, FGF, and IGF-1 that induce cell proliferation and protein synthesis which are critical for healing. Finally, Mϕ have been shown to be involved in collagen degradation during tissue remodeling phase of repair. A study that depleted Mϕ in LysM-Cre/DTR mice at different stages of wound healing supports the notion that Mϕ change functions throughout the healing process – loss of Mϕ during early stages of healing leads to reduced epithelialization, granulation tissue formation and wound contraction whereas Mϕ depletion during the mid-phase abrogates transition of wound tissues from regeneration to maturation phase. Collectively, these studies demonstrate that Mϕ play diverse roles throughout each stage of wound healing and thus are integral components of wound repair.

9Diabetes is a metabolic disorder leading to low-grade systemic inflammation which is known to have a significant impact on the immune system. Dysregulated metabolic pathways and host immune response contribute to impairments in each phase of wound healing, ultimately causing delayed wound healing in diabetes. Along with various other factors, alterations in both the number and phenotype of wound Mo/Mϕ likely contribute to impaired wound healing in diabetes. The number of infiltrating Mo is found to be higher in the wounds of leptin receptor mutant (Leprdb) db/db type 2 diabetic mice early after wounding. Following such increased monocyte accumulation, the macrophage subsets (Ly6C+F4/80+ and Ly6C–F4/80+) are also increased significantly in db/db mouse wounds at later time points suggesting a persistent Mo/Mϕ response in diabetic wounds. Further, a recent study has shown that the proportion of early wound Mϕ differentiated from infiltrating bone marrow-derived monocytes in db/db diabetic mice are increased whereas wound Mϕ derived from tissue resident Mϕ remain unaltered as compared to non-diabetic wounds. Impaired wound healing in high-fat-diet induced obese (DIO) pre-diabetic mice is also associated with persistent accumulation of inflammatory Mo/Mϕ in non-healing wounds. Prolonged infiltration of blood Ly6Chi Mo is thought to be responsible for the sustained accumulation of inflammatory Mo/Mϕ in the skin wounds of DIO mice. Such extended infiltration of inflammatory Mo may contribute to the observed defect in the transition from Ly6Chi into Ly6Clo Mo/Mϕ phenotypes (Figure 1A). In contrast to these reports, there are also reports of decreased numbers of wound Mo/Mϕ early after wounding in diabetic mice attributed in part to an early impairment of chemotaxis into the wound. The discrepancy in wound macrophage numbers between studies could result from technical differences in the assessment of wound Mϕ and deserves further study.

10Wound Mo/Mϕ in diabetic mice persistently express high levels of M1-like Mϕ markers such as NOS2, TNF-α, IL-1β, MMP9, and low levels of M2-like Mϕ markers such as Arginase 1, CD206, CD36. Diabetic wound-derived Mϕ also exhibit decreased expression of pro-healing factors such as IGF-1, TGF-β and VEGF. Similar to diabetic mice, wound biopsies from human diabetic foot ulcers also show increased proportion of M1-like Mϕ (CD68+ and IL-1β+) and decreased proportion of M2-like Mϕ (CD163+, CD206+, and Arginase-1+), respectively (Figure 1B). The sustained pro-inflammatory phenotype of diabetic wound Mϕ likely helps to drive a persistent pro-inflammatory microenvironment in diabetic wounds characterized by increased levels of IL-1β, IFN-γ, TNF-α, and IL-12 as well as decreased levels of pro-healing factors such as IGF-1, TGF-β1, VEGF, and IL-10. Together, these reports indicate that dysregulation of Mϕ polarization likely plays critical roles in impaired diabetic wound healing in both mice and humans.

11There is also consensus in the literature that wound Mo/Mϕ in diabetic mice and humans show impaired phagocytosis that may contribute to impaired wound healing. Wound Mϕ in diabetic mice exhibit reduced phagocytosis resulting in increased accumulation of apoptotic cells in the wounds and a sustained pro-inflammatory microenvironment. In addition, reduced PPAR-γ expression in diabetic wound Mϕ and improved wound healing in diabetic mice with topical wound treatment with PPAR-γ agonist is consistent with the idea that PPAR-γ-mediated macrophage clearance of apoptotic wound neutrophils may play an important role in wound healing. In another report, antibody mediated topical inhibition of receptor for advanced glycation end products (RAGE) signaling showed reduced number of neutrophils in diabetic wounds in association with enhanced phagocytosis by Mϕ and improved wound healing further supporting the notion that increased neutrophil accumulation in diabetic wounds results from reduced phagocytic ability of Mϕ which is closely associated with impaired wound healing in diabetes (Figure 1C). Together, these reports suggest that decreased efferocytosis by wound Mϕ is critical for impaired wound healing in diabetes.

12Dysregulated Mϕ activity in diabetic wounds impairs processes critical for normal wound healing. For example, Mϕ are known to play important roles in neovascularization during wound healing, and decreased production of VEGF-A as well as reduced VEGFR1 signaling by diabetic wound Mϕ likely contribute to impaired angiogenesis in diabetic wounds. Mϕ likely also play important roles in the impaired maturation and remodeling of the vasculature in diabetic wounds. Altogether, the available data suggest that dysregulation of wound Mo/Mϕ both in terms of numbers and phenotypes plays critical roles in impaired diabetic wound healing.

13Several reports have shown that diabetes increases monopoiesis in the bone marrow of STZ-induced and Akita (Ins2Akita) type 1 diabetic mice where pancreatic β cells are destroyed by toxic effects of STZ or misfolded insulin, DIO pre-diabetic mice, and db/db type 2 diabetic mice at steady state. STZ-induced diabetes and DIO mice also exhibit extramedullary myelopoiesis in the spleen resulting in increased number of splenic Mo during homeostasis. Furthermore, diabetic patients have been shown to display elevated number of circulating inflammatory Mo (CD14+CD16–) indicating diabetes-induced enhanced monopoiesis in human. However, skin wounding-induced Mo expansion in mouse bone marrow is not further augmented by diabetes. Collectively, these data suggest increased steady-state monopoiesis in diabetes likely contributes to enhanced wound Mo/Mϕ and impaired wound healing (Figure 2A).

14Increased monopoiesis in diabetes is associated with modification of the hematopoietic stem and progenitor cell (HSPC) compartment. Recently, bone marrow-derived myeloid progenitors (LK, Lin–cKit+Sca-1– cells) from db/db mice have been shown to be intrinsically modified to produce increased number of Mo upon IL-1β or M-CSF challenge in vitro corroborating enhanced potential of diabetic myeloid progenitors to Mo production. Alterations of the HSPC niche, which regulates HSPC maintenance, mobilization and differentiation, likely play an important role in the alterations of HSPC phenotypes in diabetes. Increased level of sympathetic nerves in the HSC niche is believed to induce the alterations in niche components in diabetes.

15Importantly, increased IL-1R1 signaling in myeloid restricted bone marrow progenitors may be responsible for increased monopoiesis in obese and pre-diabetic mice. In these studies, IL-1β produced by adipose Mϕ is thought to mediate communication between adipose tissue and bone marrow. However, diabetic db/db recipient mice transplanted with Il1r1–/– or WT donor cells showed no difference in myeloid cell output indicating that IL-1R1 signaling is likely not involved in diabetes-associated increased myelopoiesis in the db/db mouse model. These seemingly disparate roles of IL-1R1 signaling in increased myelopoiesis in different animal models of pre-diabetes and diabetes deserve further study.

16In addition, RAGE has also been shown to be associated with increased monopoiesis in other models of diabetes such as STZ-induced and Akita diabetic mice. Neutrophil-derived S100A8/A9 mediated activation of NF-kB-dependent RAGE signaling in common myeloid progenitors (CMPs) which in turn induces GMP proliferation via growth factors leads to increased monopoiesis in these diabetic mice. Taken together, these data suggest that IL1-R1 and RAGE signaling may act as intrinsic drivers to promote monopoiesis in pre-diabetic and diabetic mice.

17Mϕ phenotypes in diabetic wounds can be altered both through local effects mediated by the wound microenvironment and through epigenetic modifications that may occur in bone marrow progenitors that are passed down to macrophage progeny. There is strong evidence in the literature supporting the importance of the local microenvironment in determining macrophage phenotypes. First, the ability of diabetic wound conditioned medium to induce M1-like phenotypes in bone marrow-derived Mϕ in vitro supports the hypothesis that diabetic wound microenvironment may play important roles in determining Mϕ phenotypes. Studies showing that local modification of diabetic wound microenvironment by blocking IL-1β or RAGE, or inhibiting inflammasome pharmacologically can shift wound Mϕ toward pro-healing phenotypes further support the notion that wound microenvironment has significant effect on dysregulation of Mϕ phenotypes in diabetic wounds and that sustained activation of NLRP3 (NLR family, pyrin domain-containing 3) inflammasome or of the RAGE pathway in wound Mo/Mϕ is likely involved (Figure 2B). Further, the level of MCP-1 is found to be higher in the wounds of DIO mice and antibody mediated inhibition of this chemokine during the inflammatory phase of wound healing normalizes the number of inflammatory wound Mo/Mϕ and improves wound healing. Together, these data suggest that wound microenvironment may play important roles in dysregulated Mϕ functions in diabetic wounds.

18Intrinsic modifications of bone marrow progenitors that are passed down to Mo/Mϕ may also contribute to dysregulated Mϕ phenotypes in diabetic wounds. For example, differential responses to classical and alternate activation in vitro, amplified pro-inflammatory phenotypes and sustained potential to produce inflammatory tissue Mϕ observed in bone marrow-derived cells from different diabetic mouse models suggest that diabetes functionally alters progenitors in the bone marrow which may lead to dysregulated Mϕ phenotypes in diabetic wounds. Additional studies indicated that epigenetic modifications of HSPCs may be involved in such phenotypic alterations in diabetic bone marrow progenitors. For example, epigenetic modification of HSPCs by means of decreased repressive histone methylation mark H3K27me3 at the IL-12 gene promoter has been shown to be passed down to wound Mϕ in DIO mice resulting in increased pro-inflammatory wound Mϕ and impaired wound healing. In addition, transplantation of diabetic hematopoietic stem cells (HSCs) from db/db into WT mice has been shown to cause delayed wound healing in association with sustained accumulation of M1-like wound Mϕ indicating that dysregulation of diabetic wound Mϕ phenotypes happens via a HSC autonomous mechanism. Epigenetic modification of myeloid lineage associated genes such as Notch1, PU.1 and Klf4 by DNA methyltransferase 1 (Dnmt1) mediated hypermethylation may induce such alterations in diabetic HSCs (Figure 2A). Altogether, these data suggest both wound microenvironment-mediated alteration of wound Mϕ and epigenetic modification of HSPCs as potential mechanisms of dysregulated Mϕ phenotypes in diabetic wounds.

19In summary, both sustained increases in the number of wound Mo/Mϕ and dysregulation of their phenotype, caused both by intrinsic alterations in bone marrow progenitors and by a pro-inflammatory wound microenvironment, lead to impaired wound healing in diabetes. Improved understanding of factors that regulate wound Mo/Mϕ numbers and phenotype have led to new therapeutic interventions attempting to normalize the Mϕ response in mice targeting the NLRP3 inflammasome/IL-1β and RAGE pathways; these findings await translation to humans. Another approach for normalizing Mϕ phenotypes in non-healing wounds could be altering epigenetic modifications of genes associated with dysregulated Mϕ phenotype. However, more comprehensive knowledge on epigenetic changes that drive persistent inflammation in diabetic wound Mo/Mϕ will be useful to specifically target relevant genes.

20Studies showing that diabetes-associated increased monopoiesis at steady-state contributes to increased Mϕ accumulation in diabetic wounds suggest the possibility that targeting monopoiesis may help normalizing wound Mϕ accumulation and improve diabetic wound healing. However, much remains to be learned about the regulation of monopoiesis during wound healing.

21Lastly, several studies have highlighted the importance of Mo subsets in health and disease. Heterogenous Mo/Mϕ populations have also been identified in various tissues such as lung, aorta, and heart under diseased conditions. However, our knowledge of the heterogeneity of skin wound Mϕ is lacking, especially during impaired healing. Multiplex analyses such as single cell RNA-sequencing, multiparameter flowcytometry, and imaging mass cytometry will be helpful to acquire knowledge on Mϕ subsets present during wound healing, and how these are regulated during normal and impaired wound healing which may, in turn, provide insight into new approaches for manipulating inflammation and improving healing.

22PB and TK have made a substantial, direct and intellectual contribution to the work. Both authors contributed to the article and approved the submitted version.

23The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.