Reactive oxygen species and bacterial biofilms in diabetic wound healing

Abstract

Chronic wounds are a common and debilitating complication for the diabetic population. It is challenging to study the development of chronic wounds in human patients; by the time it is clear that a wound is chronic, the early phases of wound healing have passed and can no longer be studied. Because of this limitation, mouse models have been employed to better understand the early phases of chronic wound formation. In the past few years, a series of reports have highlighted the importance of reactive oxygen species and bacterial biofilms in the development of chronic wounds in diabetics. We review these recent findings and discuss mouse models that are being utilized to enhance our understanding of these potentially important contributors to chronic wound formation in diabetic patients.

Chronic Wounds Are a Major Public Health Concern

Chronic wounds are wounds that do not follow the normal healing progression and persist, usually for more than 3 mo. Chronic wounds are a burden for both affected patients and the healthcare system. Most chronic wounds can be assigned to one of three categories: diabetic foot ulcers, ulcers resulting from arterial or venous insufficiency, or pressure ulcers (56). We focus here on diabetic foot ulcers, which are the most common cause of nontraumatic major amputation around the world and the most costly type of wound (25, 61, 67, 84). An estimated 67% of amputations occur in diabetic patients (73), and diabetic foot ulcers precede an estimated 80% of diabetic lower extremity amputations (32). Today, 382 million people have been diagnosed with diabetes, and 25% of these patients will develop a foot ulcer in their lifetime (32, 69). One-quarter of these ulcers will not heal, and up to 28% will result in an amputation (32, 69). It is estimated that an amputation due to diabetes occurs every 30 s worldwide (26). If a chronic wound leads to amputation, the prognosis for the patient is poor. Five-year mortality rates after the development of a foot ulcer are between 43 and 55%, which is higher than some common cancers like melanoma, lymphoma, and cancer of the testis, breast, prostate, uterus, larynx, and cervix (65, 70). Furthermore, the numbers of patients with diabetic foot ulcers are increasing. By 2025, more than 500 million people are expected to have diabetes, 125 million diabetics will develop foot ulcers, and 20 million diabetics will undergo amputations (18).

Patients with lower limb chronic wounds can experience impaired mobility, time lost from work, and adverse effects on finances. In a study assessing the psychological implications of leg ulcers, 68% of affected patients reported a negative emotional impact on their lives, including feelings of fear, social isolation, anger, depression, and negative self-image (64). In addition to the effects on individual patients, diabetic foot ulcers are also a significant drain on healthcare funds. In the US alone, the cost for the treatment of foot diseases in patients with diabetes is at least $6 billion annually (32). For all of these reasons, chronic diabetic wounds are an important public health problem.

Normal Wound Healing

On a cellular level, chronic wounds occur when the normal wound healing process is derailed. A complex series of events involving many cell types (neutrophils, monocytes, endothelial cells, keratinocytes, and fibroblasts) is activated by the presence of an acute wound (30, 77). The stages of wound healing are hemostasis, inflammation, proliferation, and remodeling. The first stage is to re-establish hemostasis at the wound site by formation of a network of fibrin fibrils generated via thrombin-mediated cleavage of fibrinogen (78). In the next stage, neutrophils are recruited to the wound site to destroy pathogenic organisms. Macrophages arrive at the wound site and engulf debris and dead cells. This inflammatory phase ends when inflammatory cells are actively removed, likely via apoptosis. In the following phase, keratinocytes proliferate to reform the epidermis. Fibroblasts migrate to the wound site and are stimulated to proliferate by platelet-derived growth factor (PDGF) and other growth factors (16, 33, 83, 92). During this phase, fibroblasts and their derived myofibroblasts break down the fibrin clots by secreting plasminogen activators that convert plasminogen to plasmin (49, 96), which cleaves fibrin to dissolve the clot. Fibroblasts and myofibroblasts also secrete collagen and other extracellular matrix proteins that form granulation tissue. The formation of granulation tissue is accompanied by the development of new blood vessels, re-epithelialization of the wound, and contraction of the granulation tissue to bring the edges of the wound closer together (2, 30, 34, 74, 92). In a final phase, fibroblasts, macrophages and endothelial cells secrete extracellular matrix proteins and matrix metalloproteinases that remodel the granulation tissue, replacing the collagen type III matrix with a stronger, more organized collagen type I matrix (19).

Chronic Wounds Fail to Progress through Timely Repair

While normal wound healing proceeds via a well-orchestrated series of events moving through the four sequential, temporary, and overlapping phases of hemostasis, inflammation, proliferation, and remodeling, chronic wounds stall before reaching the later phases. It has not been straightforward to define chronic wounds in humans or mice. In 1994, Lazarus et al. (43) defined a chronic wound as one that fails to progress through a timely sequence of repair. Chronic wounds have also been defined as wounds that exhibit defective regulation of one or more steps in the wound healing process (46, 56). In particular, it has been noted that chronic wounds often stall in the inflammatory phase (46, 56). Other definitions of chronic wounds focus on the amount of time they require to heal. In humans, wounds are considered chronic if they have not healed for 3 mo (46). In mice, some argue that if a wound fails to heal by 26 days, it should be considered chronic (5, 18).

Challenges in Studying Chronic Wounds

When chronic wounds are the subject of investigation, researchers face significant challenges. Different types of wounds in diabetics may have different etiologies. Determining the root cause of a wound is complex because the patients often have multiple different comorbidities. Another major challenge in wound healing research is the difficulty of obtaining samples from the early stages of healing in a chronic wound in humans. When a patient presents for treatment of a chronic wound, the early phases of its wound healing have passed and can no longer be studied.

For these reasons, some researchers have turned to mouse models to complement studies on wound healing in humans. Use of genetically inbred mice can provide a large number of subjects with a consistent genotype for studying wound healing that is unachievable in humans. The ability to manipulate the mouse genome provides opportunities to study the effects of a single gene on wound healing. Furthermore, with mice, the early events after a wound is introduced can be studied under precise temporal control, with sufficient numbers of mice to draw statistically significant conclusions.

It is important to note that mouse models also have significant limitations. Mouse models do not accurately simulate many aspects of the condition in diabetic wounds of humans. In general, animal models are limited in their ability to simulate processes that are associated with aging, which is a major factor in chronic wounds in humans. For chronic wounds, characteristics of aging such as neuropathy, peripheral arterial disease, or venous insufficiency (54, 56) are likely to play an important role and are not well simulated in current mouse models. In addition, the wound healing process is different in mice than humans. Mouse skin has a thin muscle layer, the panniculus carnosus layer, under the dermis. In humans, and other higher primates, there are only remnants of panniculus carnosus muscle, primarily in the neck where it is called the platysma muscle. After injury in a mouse, this muscle layer produces rapid wound contraction that promotes healing (100). In humans, wounds heal through re-epithelialization and granulation tissue formation without the benefit of contraction (54). In humans, contraction is slow and incomplete. For these reasons, mouse models of wound healing have limitations in their ability to provide insight into the human condition. We describe here recent studies using both human and mouse models designed to better understand the distinction between chronic and acute wounds.

Chronic Wounds Are Stalled in the Inflammatory Phase

Chronic wounds have been characterized as being stalled in the inflammatory phase of wound healing, which is based on the observation of higher levels of certain types of immune cells. In one study, patients with both venous insufficiency and diabetic chronic wounds had a lower ratio of CD4+ to CD8+ T cells due to low levels of CD4+ T cells, higher numbers of macrophages, and more B cells and plasma cells than patients with properly healing acute wounds (45). Chronic wounds were characterized by prolonged and persistent T cell and macrophage infiltration (45). In another study of nonhealing venous leg ulcers, inflammatory cells, B cells, and macrophages were present in high numbers (98). In yet another study, Rosner and colleagues (72) found a statistically significant increase in the mean numbers of macrophages and neutrophils from the distant area toward the center of the ulcer in venous leg ulcers. At the center and the edge of the ulcer, macrophages were abundant, comprising 63% (center) and 53% (edge) of the cells. In the distant area, T cells were predominant and were largely CD8+ (72). B cells were rare (72). One limitation of this study was that wounds healing normally were not included. Moore and colleagues (52), in another study, found that at the wound margin of chronic wounds, CD45+ leucocytes were largely lymphocytes, and T lymphocytes were more prevalent than B lymphocytes. These investigators also found CD68+ macrophages present in all layers of the dermis at the wound margin, but found that these macrophages lacked activation markers CD16 (Fc γ III receptor) and CD35 (C3b receptor) (52). In this study as well, no normally healing wounds were included for comparison. The reason for the increased infiltration of immune cells into chronic wounds is not clear, but chronic wounds have also been found to have higher levels of proinflammatory growth factors and cytokines (3, 24, 27, 82, 89) that may recruit these inflammatory cells. These inflammatory factors and cytokines have been hypothesized to reflect complications from metabolic dysfunction in diabetic patients (20).

Chronic Wounds Have Higher Levels of Proteinases

Chronic wounds often contain an accumulation of devitalized tissue (7), which may reflect higher levels of proteases such as matrix metalloproteinases and gelatinases that degrade extracellular matrix proteins (101), as well as lower levels of the molecules that inhibit these enzymes (11, 24, 44, 59, 101). Elevated levels of matrix metalloproteinase activity can contribute to the degradation of the extracellular matrix, growth factors, and their receptors (7, 42, 90, 97).

Chronic Wounds Contain Senescent Fibroblasts

Another characteristic of chronic wounds is poor infiltration of fibroblasts and blood vessels into the wound area (19). The fibroblasts that are present in venous ulcers have been reported to display the morphological characteristics of senescent cells (19, 24, 50, 51, 56). The granulation tissue in chronic wounds usually contains fewer fibroblasts and less collagen than in wounds that heal well (41, 48, 86). Extracellular matrix deposition and remodeling are often compromised (19). Fibroblasts isolated from chronic wounds have been shown to have diminished migratory capacity (7), reduced proliferative capacity in response to stimulation with PDGF (1), diminished responsiveness to transforming growth factor (TGF)-β (60), and altered production of extracellular matrix proteins (23). Thus, the fitness and function of fibroblasts in chronic wounds may be jeopardized, or the wound environment may affect their altered ability to coordinate wound healing.

Treatments for Chronic Wounds

There are several options for treatment of diabetic foot ulcers. Standard of care includes the debridement of necrotic tissue [a process in which necrotic and senescent tissue as well as foreign and infected material are removed from the wound, and healthy tissue is cut to stimulate the healing process (7)], mechanical off-loading of pressure areas (66, 95), improving the regulation of blood glucose, infection control, wound dressings that promote a moist environment to assist healing (95), and revascularization. When wounds do not respond to standard care, advanced wound care modalities can be incorporated into the treatment plans. Some of these treatments involve reintroducing peptide growth factors that promote wound healing. Recombinant PDGF (79, 94), EGF, VEGF, and TGF-β have been reported to promote wound healing (87). Other treatment products are designed to reduce the high levels of matrix metalloproteinases that are thought to inhibit healing (40, 76, 85). Biological skin equivalents are also available, and these can take the form of an acellular scaffold that provides a matrix for cells, a combination of cells and a scaffold, or skin equivalents that are created from single-layered or bilayered cultured skin constructs, or membranes from other parts of the body such as the amnion (21, 47, 68, 88, 103). Despite the availability of multiple different wound healing treatments, many patients do not heal, making improvements in our management of wounds a critical public health need.

Elevated Reactive Oxygen Species in Diabetic Wounds

In the last decade, there has been increased attention to the possibility that reactive oxygen species (ROS) are part of the mechanism through which some wounds exhibit a delay in healing. In this context, ROS include radical and nonradical molecules that can be formed from oxygen, such as superoxide anion, hydroxyl radicals, singlet oxygen, and hydrogen peroxide. A hypothesis that excessive ROS can delay wound healing was proposed in 2005 for venous leg ulcers (98). The argument was based on the importance of neutrophils for defense against bacteria. In the initial response to a fresh wound, neutrophils release an oxidative burst of ROS generated by the plasma membrane-localized NADPH oxidase complex. Wlaschek and Scharffetter-Kochanek (98) reasoned that patients with venous ulcers exhibit an extended inflammatory phase with more neutrophils in the region for longer periods of time than individuals with rapidly healing wounds. They hypothesized that excess neutrophils and macrophages present in chronic wounds would result in excess levels of ROS. The result, they argued, would be an imbalance between the pro-oxidants and antioxidants in the wound, which leads to oxidative stress (98). Some studies have provided data consistent with this hypothesis. Superoxide released by neutrophils in chronic wounds has been reported to be increased 170% compared with acute wounds (98). Elevated ROS in the wound environment has been hypothesized to lead to perpetuation of inflammation, activation of proteolytic pathways, and tissue damage (98).

Because ROS are difficult to measure, some researchers have measured the byproducts of the reaction of ROS with more stable molecules to assess ROS levels in chronic wounds. Levels of allantoin and uric acid, stable oxidation products, were fivefold higher in fluid from chronic wounds than fluid from acute wounds (36). Measurement of an oxidized prostaglandin, 8-isoprostane, revealed higher levels of 8-isoprostane in fluid from chronic wounds than acute wound fluid (102). Patients with chronic wounds have also been shown to have low levels of the antioxidant molecules that are important for detoxifying ROS. In particular, chronic wounds have been reported to have lower levels of natural antioxidants vitamins A and E (71) and glutathione (53). These data support the notion that there is a high level of ROS in diabetic chronic wounds.

Physiological Effects of High ROS

Elevated levels of ROS in chronic wounds may have one of many deleterious effects. ROS have been reported to damage endothelial cells, thus preventing angiogenesis necessary for proper wound healing (28). ROS have also been reported to induce upregulation of ICAM-1, which results in extravasation of activated leukocytes (28). Elevated levels of ROS have also been reported to inhibit keratinocyte migration and thus delay re-epithelialization of wounds (57). In addition, ROS can inactivate the tissue inhibitors of metalloproteinases and thereby promote a highly proteolytic microenvironment in chronic wounds (81). In fibroblasts, matrix metalloproteinase expression is induced after exposure to singlet oxygen (99), hydrogen peroxide (8, 91), or hydroxyl radicals (9).

Chronic Wounds Are Commonly Infected with Bacteria

Further contributing to the altered microenvironment in chronic wounds is the fact that they are frequently colonized by bacterial species that impair wound healing (6, 35). Some of these species can form biofilms, which are complex microenvironments consisting of bacteria and polymers the bacteria secrete such as polysaccharides, proteins, lipids, and nucleic acids (22). The bacteria within biofilms are notoriously resistant to antibiotics (22, 80, 104). Studies in human patients have shown that chronic wounds are more likely to contain biofilms than acute wounds. In one study, James and colleagues (35) found that 60% of the chronic wounds studied contained biofilm-forming bacteria, compared with just 6% of acute wounds. Bacterial biofilms in the chronic wounds were often found to be characterized by diverse microbial communities including anaerobic bacteria. In another study, Bjarnsholt and colleagues (4) proposed a similar hypothesis that chronic wounds result from a failure to eradicate pathogens. They analyzed chronic wounds and found microcolonies with the capacity to become biofilms. They hypothesized that formation of biofilms protects these bacteria from being killed by neutrophils in chronic wounds. Subsequent studies have continued to find that wound debridement samples contain high levels of bacteria, with Staphylococcus aureus being the most common type (58). The exact role of biofilms in chronic wounds is not yet clear. One study found that biofilm-forming bacteria and soluble products from these bacteria reduce keratinocyte viability in vitro and decrease the ability of keratinocytes to close an in vitro wound (39), suggesting one possible mechanism for their association with chronic wounds.

db/db Mouse as a Model for Wound Healing in Diabetics

Challenges associated with obtaining early chronic wound samples from human diabetic patients have led to the creation of mouse models as a tool to better understand the early process of chronic wound formation. One frequently used model is the db/db mouse. Db/db mice have an inactivating mutation in the gene encoding the leptin receptor (13). These mice develop insulin resistance, obesity, and diabetes with hyperglycemia (13, 14).

Wound healing in db/db mice is delayed compared with wild-type controls (93). Normal healing in wild-type mice is associated with an increase and a subsequent decrease in the levels of the neutrophil chemoattractant macrophage inflammatory protein (MIP)-2 and the macrophage chemoattractant monocyte chemoattractant protein (MCP)-1. In contrast, db/db mice have a large and more sustained increase in the levels of both MIP-2 and MCP-1, which results in elevated numbers of polymorphonuclear cells and macrophages in the wound tissue (93). This sustained cytokine release may contribute to the delayed wound healing observed in these animals (93).

It should be noted that the db/db mouse has significant drawbacks as a model to simulate human diabetes. Human diabetics do not typically have a mutation in the leptin receptor, and therefore the genetic basis for the diabetic phenotype in these mice may be different from that in humans. In addition, the db/db mice are obese. If wounds are introduced without being splinted, an inconsistent expansion of the wounds can occur due to the altered skin tension. A rabbit model for diabetes also exists in which islet cells die in response to treatment with streptozotocin, simulating Type 1 diabetes. However, since Type 2 diabetes is much more common in human patients, findings in a rabbit model of Type 1 diabetes may not be relevant for the much larger human population of Type 2 diabetic patients.

High Levels of ROS and Biofilms in the db/db Mouse Model

The db/db mouse model has been used to study in more depth the role of ROS and biofilms in delayed wound healing in diabetics. Mudge and colleagues (53) found that wounds introduced into a db/db mouse have low levels of the antioxidant glutathione compared with wounds introduced into control mice. Furthermore, treating wounds in db/db mice with glutathione resulted in significantly expedited wound healing, demonstrating the critical role of glutathione and antioxidants in determining wound healing rates in mice with simulated diabetes (53).

Dhall et al. (18) of the Martins-Green laboratory found that superoxide dismutase (SOD) activity and levels of hydrogen peroxide (the product of SOD) are higher in db/db mice compared with controls. The activities of catalase and glutathione peroxidase, antioxidant enzymes that convert hydrogen peroxide to water, were lower in db/db mice than controls (18). The results are consistent with the presence of oxidative stress in db/db mice.

The Martins-Green laboratory (18) went on to inhibit catalase and glutathione peroxidase and thereby create high levels of ROS in the wounds. This treatment caused the wounds in db/db mice to form exudate and become chronic, in some cases lasting 100 days. db/db mice treated with inhibitors of catalase and glutathione peroxidase were more likely to develop spontaneous microbial infections. Microbial communities in wounds in db/db mice evolved from containing species that do not produce biofilms to containing biofilm-producing species (18). Treating the wounds with antioxidants, α-tocopherol and N-acetylcysteine, improved wound healing dramatically (18). When the wounds were treated with antioxidants, the bacteria in the wounds were less likely to generate biofilms and were more sensitive to antibiotic treatment. Granulation tissue was formed with thicker collagen fibers (18). The bacteria in the db/db mice treated with antioxidant inhibitors develop complex communities of bacteria including bacteria that are similar to those in human skin such as Staphylococcus epidermidis and other coagulase-negative staphylococci (29). The results support an important role for a combination of bacteria that create biofilms and ROS in promoting wounds that fail to heal in diabetic mice.

ROS and Biofilms in LIGHT^−/− Mice

The laboratory of Dr. Martins-Green also reached a similar conclusion based on studies in a different model system of delayed wound healing. Petreaca et al. (63) of the Martins-Green laboratory discovered that VEGF induces macrophage apoptosis via a tumor necrosis factor superfamily member called TNFSF14 or LIGHT (Homologous to Lymphotoxins, exhibits Inducible expression, and competes with herpes simplex virus Glycoprotein D for Herpes virus entry mediator, a receptor expressed by T lymphocytes). Inhibition of LIGHT with an anti-LIGHT neutralizing antibody prevented VEGF-induced apoptotic death of cultured macrophages. The authors hypothesized that LIGHT promotes macrophage apoptosis during wound healing. They further hypothesized that mice with genetic inactivation of LIGHT would not be able to execute the series of events that normally occur during wound healing and would experience chronic wounds. To test their hypotheses, the authors generated genetically engineered mice with inactivation of TNFSF14/LIGHT (62). Wounds introduced into LIGHT^−/− mice do exhibit several characteristics that are similar to those of chronic wounds in human patients. The LIGHT^−/− mice have higher production of cytokines than control mice, in particular CXCL8, MCP-1/CCL2, and IP-10/CXCL10, which are expected to recruit neutrophils, macrophages and T-lymphocytes. In wounds in LIGHT^−/− mice, neutrophils appeared within the first few hours after wounding and the number of neutrophils was significantly elevated compared with wounds introduced into wild-type mice. Neutrophils remained within the wound until day 9 in wounds introduced into LIGHT^−/− mice, while in wounds introduced in control mice, the levels of neutrophils had declined by day 3. Macrophages were also more abundant in wounds in LIGHT^−/− mice, but only on day 7. T cells were consistently more abundant in the wounds from the LIGHT^−/− mice. A subset of the wounds in the LIGHT^−/− mice failed to heal even after a prolonged period of time. These wounds showed signs of infection; they became crusty, oozed fluids, and opened and closed repeatedly. Taken together, the data indicate that the LIGHT^−/− model shares some similarities to the situation in chronic wounds in diabetic patients.

With this new mouse model in hand, Dhall et al. (17) of the Martin-Greens laboratory found that wounds developed in LIGHT^−/− mice have higher levels of hydrogen peroxide than wounds created in wild-type mice. These higher levels of reactive species in wounds generated in LIGHT^−/− mice were associated with increased levels of oxidized macromolecules such as lipid peroxides, and increased apoptosis and necrosis in the LIGHT^−/− mice compared with wounds in wild-type controls.

The authors tested the importance of ROS for wound healing by inhibiting antioxidant enzymes and introducing biofilm-forming bacteria (17). Wounds were reproducibly converted to a chronic state in LIGHT^−/− mice by a combination of inhibition of the antioxidant enzymes catalase and glutathione peroxidase immediately postwounding, combined with subsequent application of the biofilm forming bacterium S. epidermidis. If any one of these three components was eliminated, the wounds did not consistently convert to a chronic state. Thus, the LIGHT^−/− mouse model, like the db/db model, supports a role for a combination of high ROS and the presence of biofilm-producing bacteria in the development of chronic wounds.

Limitations of the Mouse Models

The availability of this new mouse model raises the question whether the model is likely to be informative about the wound healing process in diabetic patients. It is important to note that all mouse models are limited in their ability to model the human condition. In this case, the mice have a genetic abnormality in LIGHT rather than diabetes and, therefore, do not have some of the important characteristics of diabetic patients. Many diabetic patients are overweight. Because the LIGHT^−/− mice are normal weight, they will not simulate some important aspects of diabetes. Diabetic patients also have high levels of glucose in their blood, and the LIGHT^−/− mice do not share this property. This can be a critically important difference because hyperglycemia itself can lead to abnormal leucocyte function, such as decreased phagocytosis (37). Additionally, while LIGHT expression was inhibited in this mouse model, recent studies have shown that plasma levels of LIGHT are actually increased in Type 2 diabetes and obesity (15, 31). Also of importance, the wounds that are generated in the db/db and LIGHT^−/− mice are on the dorsum, a position in which they will not experience pressure. These wounds will therefore not recapitulate the situation for most diabetic ulcers that are largely on the patients' legs and feet and will consequently bear weight. Whether a wound is on a portion of the body that bears weight is likely an important determinant of healing rates, as relieving the pressure from chronic wounds is an effective approach to therapy (66, 95).

ROS and Biofilms in Diabetic Mice: an Alternative Model of Low Levels of ROS

In addition to the db/db and LIGHT^−/− models, other models for diabetic wound healing are also being put forward. The TallyHo mouse has been developed to study diabetes in mice. The TallyHo mouse has polygenic mutations that cause hyperglycemia, hyperinsulinemia, weight gain, and impaired glucose tolerance (38). Because diabetes in human patients is mostly polygenic, the TallyHo mouse model is considered more similar to human Type 2 diabetes than models of streptozotocin-induced diabetes or diabetes induced by a single mutation such as db/db (38). The TallyHo mouse has been demonstrated to exhibit delayed wound healing for incisional wounds, excisional wounds, and ischemia-reperfusion injury (10).

Nguyen and colleagues (55) studied the Tallyho model for mouse diabetes. They introduced S. aureus into wounds in C57BL/6 control mice and TallyHo mice and found the TallyHo mice had higher levels of bacteria than the controls. The TallyHo wounds into which S. aureus had been introduced, but not wounds in control mice with introduced S. aureus, became purulent and inflamed. The presence of S. aureus delayed wound healing for both diabetic and nondiabetic mice, but the effect was especially strong in the diabetic mice. The authors monitored myeloperoxidase activity as an indicator of the neutrophil oxidative burst. They found that sterile wounds from diabetic and wild-type mice showed modest myeloperoxidase activity. Inoculated wounds in wild-type mice had increased myeloperoxidase activity compared with noninoculated controls, while the induction in inoculated, diabetic mice was lower than in the control mice. The lack of neutrophil oxidative burst may be an important contributor to the increased bacterial burden in the diabetic wounds. The authors conclude, in contrast to the studies from the Martins-Green laboratory described above, that an impaired ability to respond to bacteria with an oxidative burst represents a potential mechanism underlying chronic wounds in diabetics.

Outstanding Questions for the Field

These studies raise numerous important questions for future research. One future question will be whether the LIGHT^−/− model can be valuable for understanding chronic wounds in diabetic patients, given that the LIGHT gene is not directly related to diabetes and LIGHT levels are actually high in diabetics. Furthermore, the most important downstream regulators of the phenotype in the LIGHT^−/− mice are not known, as in addition to a reported effect on macrophage apoptosis, according to the authors, LIGHT^−/− mice also have increased levels of forkhead box protein A1 (FOXA1), cytochrome p450 2E1 (CYP2E1), and Toll-like receptor 6 (TLR6) (17). More generally, can mouse models that result in immune dysregulation and a stall in the inflammatory phase of wound healing be a valuable tool for understanding the processes that occur in diabetic patients?

Another important outstanding question for the field will be to determine which model of ROS in diabetic foot ulcers is more relevant. Is there an increase in harmful ROS in the wounded diabetic foot as a result of the higher levels of immune cells with the potential to generate ROS? Or alternatively, do high levels of glucose in diabetic patients inhibit the ability of immune cells in diabetic chronic wounds to become activated and perform a ROS burst, thus leading to lower levels of ROS and a failure to clear pathogens? Or, do each of these models occur in different subsets of patients?

Additional open questions for the field involve the development of bacterial biofilms. For instance, do these biofilms inhibit wound healing and if so, how? Are there strategies to limit biofilm production that can benefit patients with biofilm-containing chronic wounds? Would lowering ROS levels with antioxidants protect against biofilm formation in human patients? Or would lowering ROS allow bacteria to flourish and impede healing?

Finally, it will be important for the diabetes field to determine how we can best use mouse models to understand the course of wounds in diabetic patients. For instance, can mouse models be developed that allow scientists to rigorously investigate the role of weight, diet, or exercise in wound healing? The availability of additional mouse models that allow scientists to dissect the effects of different diabetic complications on wound healing could potentially develop into a valuable tool for researchers.

Given that up to 40% of diabetic patients die within 1 yr of amputation (12, 75), effective treatment and prevention modalities for patients with diabetic foot ulcers are desperately needed for an aging population. Improving our knowledge of the underlying biology of chronic wounds could result in evolved standards of care, potentially reducing the incidence of chronic wounds and their impact on patients and healthcare systems.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

A.N., A.M.A., E.R.Z., L.H., and H.A.C. edited and revised manuscript; A.N., A.M.A., E.R.Z., L.H., and H.A.C. approved final version of manuscript; H.A.C. drafted manuscript.