Diabetic foot ulcers (DFUs) are a severe consequence of type I and type II diabetes mellitus (T1DM; T2DM). Obesity and advanced age are primary risk factors for developing T2DM, characterised by sustained hyperglycaemia, which leads to insulin resistance and associated secondary complications, including ischaemia and neuropathy. Interestingly, the presence of neuropathy significantly increased the risk of first-time DFU in both T1DM and T2DM, while other risk factors for DFU are more specific to T1DM (Hangaard et al, 2019). Diabetes in general has a large societal and economic impact, with T2DM patients over 65 costing healthcare providers on average twice as much as the same age group without T2DM (O’Shea et al, 2013). One in four patients with diabetes develop at least one DFU, with prospective studies indicating that 14–26% of these require amputations due to wound infection and osteomyelitis (Guest et al, 2018).
Wound management begins with assessing the patient and wound aetiology, followed by tailored intervention. These often start with debridement, offloading and carefully selected dressings then, if required, move onto more advanced modalities, such as negative pressure wound therapy (Amin and Doupis, 2016). Despite technological advances and improved understanding of diabetic pathophysiology, current treatments almost exclusively address secondary complications of chronicity, rather than the cause. As a result, diabetic ulcers frequently persist or recur, with figures of up to 66% recalcitrance quoted (Hunt, 2011). Accordingly, there remains a clinical need to develop effective therapies that are financially, physiologically and practically viable.
Diabetic skin is predisposed to poor healing
Curiously, changes are observed in diabetic skin prior to ulceration, with up to a third of people with diabetes experiencing a wide range of skin complications (reviewed in Quondamatteo, 2014). Compared to aged skin, where the physiological changes are well documented, the structure and function of diabetic skin has been greatly understudied. Nevertheless, a few key observations have been reported. For example, people with diabetes with high fasting blood glucose are more likely to develop xerosis, and skin blistering (bullosis diabeticorum) is observed in around 0.5% of people with diabetes (reviewed in Quondamatteo, 2014).
The structural stability of diabetic dermis is decreased, with lower elasticity and reduced mean breaking strength, due in part to increased production of advanced glycation end-products (AGEs; Argyropoulos et al, 2016). In diabetes, HbA1c is strongly linked to early glycation (Hb-AGE), a reliable measure of long-term glycaemic control (Turk et al, 1998). AGEs in peripheral tissues occur naturally when the body’s proteins are glycated over time. However, this process is accelerated on a background of sustained hyperglycaemia in diabetes, where years of natural AGE accumulation can occur in a matter of months (Quondamatteo, 2014; Argyropoulos et al, 2016). While hyperglycaemia (high HbA1c) is a major contributor to AGE formation, there will be other environmental factors that modulate the rate of AGE accumulation (e.g. ultraviolet radiation exposure). These AGEs then trigger inflammation-mediated pathways, which induce cellular free-radical release and oxidative stress. This cascade has profound effects in diabetic pathophysiology, leading to vascular and neurological damage that drives nephropathy, retinopathy, neuropathy and atherosclerosis (Huijberts et al, 2008).
In addition to the changes to our own cells, diabetic skin is characterised by microbiota dysbiosis. This includes higher than normal abundance of opportunistic pathogens (e.g. Staphylococcus aureus) in plantar regions (Redel et al, 2013). Indeed, it has recently been shown that low virulent forms of S. aureus preferentially establish infection in diabetic hosts (Tuchscherr et al, 2018), reinforcing the concept that diabetic tissues are intrinsically compromised. Together, these characteristic changes predispose diabetic skin to poor healing by: a) attenuating sensation, which increases the risk of injury; b) altering cellular capacity to repair the injury and; c) modifying tissue properties in favour of pathogenic colonisation.
Cellular basis of wound healing
The skin is a highly regenerative organ that has evolved a tightly regulated series of events to return to a normal state following damage. At the cellular level, the wound healing stages consist of blood clotting, inflammation, re-epithelialisation, angiogenesis, matrix deposition and remodelling (reviewed in Landén et al, 2016). Upon injury, platelets aggregate to prevent blood loss and provide a temporary scaffold of fibrin for wound cells to migrate into. Immune cells, mainly neutrophils and macrophages, then pervade the wound within hours. They remove bacteria and necrotic tissue, releasing soluble signalling factors to communicate with resident skin cells and other infiltrating cells (Singh et al, 2017).
Epidermal keratinocytes respond to these signals, undergoing collective migration to close the wound gap. Endothelial cells also begin to invade the wound, sprouting new vasculature to progress angiogenesis, a process that provides oxygen and nutrients to meet the requirements of the highly metabolic regenerating tissue. Fibroblasts correspondingly proliferate and deposit new matrix to replace the temporary fibrin scaffold (Landén et al, 2016). As healing progresses, inflammation abates, the epidermis reforms and the dermis remodels into a scar.
Diabetic wounds demonstrate impaired cellular behaviours
In acute healing, the molecular and cellular regulation of wound cell behaviours, from epidermal regeneration (Dekoninck and Blanpain, 2019), to extracellular matrix production (Wilkinson et al, 2019c), is well understood. Yet how these processes become dysregulated in the context of diabetic healing remains less well established. In diabetic ulcers, wound edge keratinocytes are known to lose their ability to migrate, instead adopting a hyperproliferative phenotype, marked by abnormal c-myc and beta-catenin expression (Stojadinovic et al, 2005). The epidermal wound edge becomes hyperkeratotic and parakeratotic, with keratinocytes unable to close the open wound. The epidermis surrounding DFUs has also been shown to contain abundant Langerhans cells (Galkowska et al, 2005), which may cause elevated wound matrix metalloproteinase (MMP) levels. MMPs contribute to excessive tissue breakdown, degrading collagen and other dermal proteins (Argyropoulos et al, 2016).
Despite an excessive inflammatory response, diabetic wounds become easily infected. Infiltrating macrophages, neutrophils and T cells have all been shown to be dysfunctional in diabetic tissue (Galkowska et al, 2005; Moura et al, 2017; Wilkinson et al, 2019b). This defective immune cell response adds to an inability to clear infection. Interestingly, antimicrobial peptide production is also altered in DFUs. For example, little to no expression of the cathelicidin antimicrobial peptide, LL-37, is observed in DFU epithelium (Rivas-Santiago et al, 2012). LL-37 is multi-functional, beneficially promoting keratinocyte migration in diabetic mice, as well as inhibiting bacterial growth (Carretero et al, 2008). Thus, damaged defence mechanisms provide an opportunity for pathogen invasion, while simultaneously impairing the function of important wound cells.
Cellular senescence promotes diabetic wound pathology
All cells in the body have a finite lifespan, undergoing programmed cell cycle arrest primarily to prevent the propagation of harmful mutations. Over time, these senescent cells accumulate and contribute to age-related tissue pathology (Wilkinson and Hardman, 2020). The observation that DFUs harbour senescent fibroblasts was made over 20 years ago (Hehenberger et al, 1998). Few studies, however, have investigated the functional implications of these cells in diabetic pathology. We recently extended these findings to demonstrate that diabetic wounds contain senescent macrophages that impair healing (Wilkinson et al, 2019a). Though incapable of dividing, senescent cells remain highly active, secreting a wide range of cytokines, chemokines and proteases, collectively known as the senescence associated secretory phenotype (SASP). This SASP can influence the healing environment by altering the behaviours of local wound cells. SASP factors also appear to promote “bystander senescence” in neighbouring cells (as observed in Wilkinson et al, 2019a), hence, perpetuating cellular senescence and further impairing tissue function.
It is likely that local aspects of the diabetic environment increase the propensity of cells to senesce. DFUs are characterised by tissue hypoxia, a result of decreased tissue perfusion (Amin and Doupis, 2016). This contributes to inflammation and reactive oxygen species production, promoting a hostile tissue environment that impacts on cellular senescence. Hyperglycaemia can also directly and indirectly provoke replicative and stress-induced senescence (Prattichizzo et al, 2018). Indeed, chronic wound fibroblasts display an impaired stress response, rendering them more susceptible to senescence than normal fibroblasts (Wall et al, 2008).
Local bacteria may also augment host cell senescence. For example, pyocyanin, a virulence factor of Pseudomonas aeruginosa found in infected wounds, was shown to induce fibroblast senescence in vitro (Muller et al, 2009). Infection likely exacerbates local inflammation and drives oxidative stress to cause senescence, while persistent infection and immune cell dysfunction will potentiate senescence. These studies emphasise that senescent cell burden may significantly contribute to impaired diabetic wound healing (Figure 1), however, further experimental and clinical investigations are needed to elucidate the molecular, cellular and tissue level impacts of wound senescence.
Manipulating senescence to improve diabetic foot ulcer outcome?
Given the clear mechanistic link between cellular senescence and age-related disease, there has been considerable recent interest in modulating senescence for therapeutic benefit. Senescence-targeted drugs, often referred to as senolytics, are starting to show potential in the clinic. For example, limiting senescent cell burden in the adipose tissue of patients with diabetic kidney disease (Hickson et al, 2019).
Though not specifically directed at senescence, current wound treatments, such as debridement and antimicrobial dressings, will offer a degree of senescent cell reduction. For instance, removing necrotic tissue and alleviating infection will dampen local inflammation and further tissue damage that would otherwise promote senescence in wound cells. Of particular note, the widely used anti-diabetic drug, Metformin, has been suggested as a potential senolytic owing to its hypoglycaemic and anti-SASP effects (Ochoa-Gonzalez et al, 2016). Metformin decreases hyperglycaemia by inhibiting liver gluconeogenesis. The resulting drop in blood glucose levels should directly inhibit AGE production, dampen inflammation and oxidative stress, and prevent subsequent tissue damage. Indeed, diabetic mice treated with Metformin show improved wound closure (Han et al, 2017) and people with diabetes given Metformin experienced fewer DFU-related amputations (Ochoa-Gonzalez et al, 2016). This effect was specific to Metformin, as patients on other anti-diabetic drugs (e.g. insulin, sulfonylurea and glybenclamide) showed no difference in DFU-related amputations.
To suggest using one of the most commonly prescribed drugs for diabetes to treat DFUs may sound paradoxical; however, the reality is that, while Metformin successfully treats early stages of diabetes, long-term adherence is poor (McGovern et al, 2018). Thus, by the time DFUs develop, relatively few patients will still be using Metformin. It is also important to note that almost all senolytic therapy development to date has been in the area of cardiovascular and renal disease. However, senolytics, such as Resveratrol and Quercetin, are now showing promising anti-diabetic effects, which may well extend to clinical endpoints (e.g. DFUs).
Therapeutics with wide-ranging biological actions have traditionally been avoided, due the high risk of off-target effects (side effects). An alternative, possibly more fruitful, approach may be to develop senolytic treatments with local, on-target effects. One option would be to target senescence-linked receptors on wound-relevant cells. For example, we recently demonstrated that the chemokine receptor, CXCR2, potentiates senescence in diabetic wound macrophages (Wilkinson et al, 2019a). Targeting antagonism of CXCR2 accelerated wound repair in diabetic mice by reducing senescence and inflammation. Interestingly, the safety of CXCR2 antagonists has already been evaluated in clinical trials (Kirsten et al, 2015), likely simplifying the route to repurposing for wound applications. Potential senescence-linked therapeutic strategies are summarised in Figure 2.
Conclusion
Hyperglycaemia is known to cause widespread tissue damage, particularly to the vasculature, via AGE production and subsequent local inflammation. From a clinical perspective, long-term effective management of HbA1c is essential to limit early and advanced glycation products. Senescence is appearing as a key cellular process linking this glycation to the widespread tissue pathology observed in diabetes. This has important therapeutic implications. For example, experimental evidence suggests that targeting prescription of and adherence to Metformin (known to have senolytic activity) in patients with active DFUs may directly promote wound healing. Indeed, emerging clinical data point towards the benefit of topical senolytic treatments (Gallelli et al, 2020), with other trials under way.
