Diabetic Wound Healing in Soft and Hard Oral Tissues

Abstract

There is significant interest in understanding the cellular mechanisms responsible for expedited healing response in various oral tissues and how they are impacted by systemic diseases. Depending upon the types of oral tissue, wound healing may occur by predominantly re-eptihelialization, by re-epithelialization with substantial new connective tissue formation, or by a a combination of both plus new bone formation. As a result, the cells involved differ and are impacted by systemic diaseses in various ways. Diabetes mellitus is a prevalent metabolic disorder that impairs barrier function and healing responses throughout the human body. In the oral cavity, diabetes is a known risk factor for exacerbated periodontal disease and delayed wound healing, which includes both soft and hard tissue components. Here, we review the mechanisms of diabetic oral wound healing, particularly on impaired keratinocyte proliferation and migration, altered level of inflammation, and reduced formation of new connective tissue and bone. In particular, diabetes inhibits the expression of mitogenic growth factors whereas that of pro-inflammatory cytokines is elevated through epigenetic mechanisms. Moreover, hyperglycemia and oxidative stress induced by diabetes prevents the expansion of mesengenic cells that are involved in both soft and hard tissue oral wounds. A better understanding of how diabetes influences the healing processes is crucial for the prevention and treatment of diabetes-associated oral complications.

Introduction

The oral cavity contains a complex mixture of tissue types and cells. Mucosal surfaces that cover muscle and/or small amounts of loose connective tissue include the soft palate, nasopharynx, and buccal mucosa that extends through the floor of the mouth to the tongue (Figure 1). These oral tissues are non-keratinized, contain large amounts of elastin, and harbor epithelium that express different keratins compared to keratinized oral mucosa such as attached gingiva¹. The attached gingiva contains dense connective tissue and is found on the palate and masticatory gingiva that surrounds dentition over alveolar bone. Healing of oral mucosa covering loose connective tissue and muscle is distinct from healing of attached gingiva that covers alveolar bone. Epithelial coverage of excisional wounds in the former is rapid and does not depend upon stromal healing, while for the latter, epithelial closure coincides with substantial new connective tissue formation, which occurs more slowly. In addition, wounds that cover muscle and loose connective tissue heal by epithelial coverage plus contraction while those that cover bone lack a significant contraction component². The oral mucosal epithelium is thicker than the epidermis in skin³. Unlike skin wounds, oral mucosal wounds lack dermal appendages such as hair follicles, sweat glands and sebacious glands. This is significant since the appendages contain stem cells that promote the healing process, which may be particularly important in healing of connective tissue as described below.

Figure 1. Re-epithelialization in diabetic oral wounds.

Left, location of different types of oral mucosa Right graphical diagram of diabetic impact on epithelialization in oral wounds Diabetes causes hyperglycemia and production of advanced glycation end-produce (AGEs) and reactive oxygen species (ROS). This in tum increases recruitment of neutrophil recruitment and redues microbial diversity. oareover, salivary content for pro-healing factors (EGF end histatin) is reduced whereas those that cause destruction (MMP2 and MMP9) are enhanced Another mechanism is FOXO1-mediated overexpression of MMP9 by keratinocytes. The overall outcome is persistent inflammation and an environment that is conducive for keratinocyte apoptosis, coupled with reduced keratinocyte proliferation anti migration

Re-epithelialization in oral wounds

Re-epithelialization of nonkeratinized tissue is faster in oral wounds than in corresponding dermal wounds, whereas re-epithelialization in keratinized attached gingiva is not necessarily faster. Upon wounding, epithelial keratinocytes at the wound edge proliferate and migrate to cover the wound surface. It has been proposed that non-keratinized oral mucosa has a higher proliferation rate than skin epidermis, which may account for the more rapid re-epithelialization⁴. In addition, wounds in nonkeratinized tissue appear to have fewer leukocytes such as neutrophils and mast cells that may lead to less inflammation and more rapid healing⁵. Growth factors that induce keratinocyte migration are produced in the early phases of healing, including epidermal growth factor (EGF), heparin-binding EGF-like growth factor (HB-EGF), transforming growth factor-β (TGF-β), and fibroblast growth factor (FGF)^2,6. It has been proposed that saliva contains growth factors that promote oral wound healing. In a tongue mucosal wound healing model, removal of submandibular salivary glands reduced the salivary EGF content and impaired healing^6,7. Saliva also contains other biologically active growth factors including nerve growth factor (NGF), TGF-β, insulin-like growth factor-1 (IGF-1) and fibroblast growth factor (FGF)⁶. Furthermore, oral mucosal keratinocytes upregulate host response genes more quickly and to a greater extent than dermal keratinocytes, which may contribute to the more rapid healing of nonkeratinized oral mucosal compared to similar dermal wounds¹. In addition to growth factors, saliva contains histatins, which are antimicrobial peptides and have been proposed to promote keratinocyte migration in oral wounds⁸. There is some evidence that the biologically active site of histatins that stimulates keratinocyte migration is distinct from the antimicrobial domain. It should be noted however, that saliva appears to influence large mucosal wounds to a greater extent than small mucosal wounds⁹. Moreover, saliva alone is unlikely to account for the differences in rates of healing between skin and nonkeratinized oral mucosal wounds given that there are several intrinsic differences in the keratinocytes from oral mucosa and skin^10,11.

Acute inflammation is necessary to initiate the recruitment of cells to the wound site and to induce the expression of factors that lead to repair. Although transient inflammation is needed for normal healing to occur, prolonged inflammation limits the repair process. When oral tongue wounds were compared to skin wounds, the expression of pro-inflammatory MMPs, peptidases, cytokines and chemokines was less in the oral wounds, which may reflect differences intrinsic to cells in oral mucosa¹². However, the temporal sequence of factors induced by wounding was similar in mucosal and dermal wounds¹². Reduced cytokine and chemokine expression may facilitate rapid healing in non-keratinized mucosal wounds compared to dermal wounds. An early event in wound healing is the development of a hypoxic environment due to disruption of blood flow. Mucosal wounds produce less hypoxia inducible factor-alpha (HIF-1a) than similar skin wounds¹³. These results suggest that the level of hypoxia in mucosal wounds is less than that in skin and that the response of oral mucosal wounds of nonkeratinized tissue is milder. DiPietro and colleagues have suggested that the differential responses to hypoxia may contribute to differences in rates of healing¹³.

The Impact of Diabetes on Re-Epithelialization of Oral Wounds

Type 1 diabetes mellitus (T1DM) is caused by a loss of beta cells in the pancreas, most frequently due to autoimmune etiology, and results in an insufficient amount of insulin production. Type 2 diabetes mellitus (T2DM) is due to insulin resistance, caused by an inadequate cellular response to normal levels of insulin signaling, combined with beta cell failure to provide a compensatory increase in insulin¹⁴. Oral mucosal wounds heal more slowly in diabetic animals compared to normoglycemic controls in both T1DM and T2DM models^15–15. Contributing factors have consistently been shown reflect the direct effect of high glucose levels, as well as indirect effects caused by increased inflammation, high levels of advanced glycation endproducts (AGEs), and increased formation of reactive oxygen species (ROS)^16,15. These factors inhibit migration of oral and dermal keratinocytes in vitro^17,30. Oral keratinocytes in diabetic wounds in vivo have increased levels of inflammatory cytokines such as TNF and IL-1β and reduced levels of growth factors such as FGF-2 and TGF-β^15,15. The combination of reduced growth factor expression by keratinocytes and increased expression of inflammatory mediators may negatively impact keratinocyte migration.

Saliva has several constituents that are thought to directly promote re-epithelialization, and diabetes has been shown to alter saliva in a way that may be detrimental to re-epithelialization. Diabetes reduces EGF levels in the saliva of humans and mice^18,19, which may contribute to reduced oral wound healing. This conclusion is supported by evidence that supplementation with EGF in the drinking water improves healing of diabetic oral wounds but has little effect on healing of non-diabetic oral wounds⁶. Furthermore, diabetic children have reduced levels of salivary antimicrobial peptides including histatins²⁰. The reduced levels of histatins may negatively impact keratinocyte migration and lessen the effectiveness of the antimicrobial defense in diabetic individuals. In the skin, reduced migration is linked to high levels of MMP-9 in diabetic humans and animal models²¹. In vitro, keratinocyte migration in high glucose is rescued by antibody to MMP-9 and knockdown of the transcription factor, FOXO1²². Interestingly, MMP-2 and MMP-9 are elevated in the saliva of diabetic patients²³, which may impair re-epithelialization.

Diabetes delays both the early and late stages of oral wound healing²⁴. Diabetic wounds have increased inflammation as noted by greater numbers of neutrophils^24,25. Re-epithelialization of gingival wounds is enhanced by TNF inhibitors in diabetic animals but not in normal animals, indicating that the increased levels of TNF in the diabetic oral wounds is problematic whereas the normal level of TNF does not inhibit the healing process²⁶.In skin wounds, diabetes reduces the conversion from a M1 macrophage phenotype to a M2 macrophage phenotype, which may contribute to prolonged inflammation²⁷. Diabetic skin wounds also have a greater senescence-associated secretory phenotype (SASP) with increased inflammatory cytokine expression (e.g. IL-1a, IL-6, IL-8, and CXCL2) and MMP expression²⁸. Moreover, aged mice have a similar increase in the SASP phenotype as diabetic mice, suggesting that there may be similar mechanisms that delay healing with both aging and diabetes²⁸.

Mucosal wounds must deal with the influence of bacteria since the oral cavity has a higher microbial load than skin. Bacterial colonization impairs healing by inducing prolonged inflammation, which intereferes with transition to the next phase of healing, maturation²⁹. Diabetic skin wounds have an impaired host response to bacteria and an increased level of ROS that reduces microbial diversity and increases bacterial colonization³⁰. Bacteria can also have direct effects on keratinocytes by stimulating apoptosis, reducing migration and decreasing proliferation³¹. Bacteria such as P. gingivalis and F. nucleatum can invade oral keratinocytes to interfere with re-epithelialization by down regulating genes that promote proliferation and integrins needed for cell migration³¹. The characteristics of diabetic oral wound healing by re-epithelilization are summarized in Figure 1.

Re-epithelialization Diabetes and FOXO1

As discussed above, diabetes negatively impacts keratinocyte proliferation and migration to interfere with re-epithelialization³². Diabetic conditions impair keratinocyte function, which include high glucose levels, increased formation of advanced glycation end products (AGEs) and reactive oxygen species (ROS). Elevated ROS levels have been shown to interfere with mucosal healing³³ and increase inflammation in gingival epithelium³⁴. The epithelium of diabetic individuals has greater expression of RAGE, which increases inflammation and worsens healing responses^35,36. A unifying explanation that links high glucose, AGEs and ROS formation to diabetes-impaired re-epithelialization is the transcription factor FOXO1. Under normal conditions, FOXO1 improves re-epithelialization by protecting keratinocytes from oxidative stress and by inducing expression of TGF-β, both of which are needed for an appropriate healing response³⁷. In diabetic conditions, high glucose or AGES cause an alteration in FOXO1 binding to the promoter regions of specific genes, most likely due to epigenetic modifications, that result in a diminished capacity to protect against oxidative stress and induce TGF-β1 transcription is significantly reduced³². These molecular changes are observed in vivo and in vitro and in both skin and oral mucosal keratinocytes. On a molecular level, diabetic conditions interfere with FOXO1 induced transcription by reducing FOXO1 binding to consensus response elements on gene promoters such as TGF-β1 and tissue inhibitor of matrix metalloproteinases 1 (TIMP1). In contrast, high glucose and AGEs increase FOXO1-promoter interactions leading to overexpression of factors that reduce re-epithelialization. More specifically, in a microenvironment with high glucose/AGE levels, FOXO1 induces expression of MMP9, serpin peptidase inhibitor clade B member 2 (SerpinB2), chemokine CCL20 and IL-36γ^17,22,32. Interestingly, the negative effect of high glucose on keratinocyte migration is improved by addition of specific inhibitors for MMP9, or CCL20 or IL-36γ or by reducing FOXO1 activity²². The role of FOXO1 in inducing excessive production of these factors is demonstrated in vivo by lineage specific FOXO1 deletion in keratinocytes that reduces high levels of MMP9, SerpinB2, CCL20 and IL-36γ expression²². It should be noted that while high levels of these factors interfere with keratinocyte re-epithelialization, their total absence is also problematic. For example, the total absence of MMP9 is detrimental to keratinocyte migration³⁸ as is absence of IL-36γ³⁹.Thus, impaired re-epithelialization in diabetes is due to overexpression of factors that normally promote an appropriate healing response at physiological concentrations. The expression of FOXO1 provides a mechanistic explanation for how diabetic conditions can lead to overexpression of these factors as shown in Figure 1.

Connective Tissue Healing in Gingival Wounds

As discussed above, keratinized mucosa (eg. attached gingiva) covers the hard palate and alveolar bone, and its healing differs from that of nonkeratinized mucosa by relying on substantial connective tissue healing. Table 1 summarizes the findings from studies that investigate specific mucosal sites. Attached gingiva has a denser connective tissue compartment, which is home to contains fibroblasts, leukocytes, endothelial and other cells cells. Upon wounding, the coordinated healing response often results in complete regeneration of injured gingiva, which contrasts with skin wounds that heal by fibrosis^40–42. Fibroblasts are the principal cell type of connective tissues and are largely responsible for stromal healing by producing extracellular matrix (ECM) proteins such as collagen and proteoglycans. It has been proposed that the unique regenerative response of gingival wounds is attributed to the intrinsic properties of gingival fibroblasts that are distinct from skin fibroblasts. Cultured gingival fibroblasts are more proliferative and express reduced levels of TGF-β1, a pro-fibrogenic cytokine, compared to skin fibroblasts⁴³. Importantly, gingival wounds have higher fibroblast density and increased connective tissue fill compared to skin wounds⁴², which is associated with reduced levels of TGF-β1 and elevated levels of anti-fibrogenic TGF-β3 in oral wounds^44–47. In a similar vein, fetal wounds that heal without scarring also exhibit low levels of TGF-β1 and high TGF-β3 levels^47,48. Moreover, it is proposed that fibroblasts may down regulate immune responses in gingival wounds to facilitate resolution of inflammation and promote healing ^11,42. Animal studies have shown that gingival fibroblasts are derived from neural crest origin (Wnt1-lineage) and can differentiate into tri-lineage mesenchymal cells in vitro⁴⁹, which contrasts with mesoderm-derived skin fibroblasts^49,50. Gingival fibroblasts cultured ex vivo are known as gingival mesenchymal stromal cells (GMSCs). GMSCs exhibit a pro-healing capacity by their anti-inflammatory properties rather than by directly differentiating into mesengenic cells^51–55. It is speculated that gingival fibroblasts may exhibit similar immunomodulatory function in vivo, which remains to be proven. In vitro culture conditions drastically shift the fibroblast transcriptome^56,57. Thus the phenotypic and functional similarities of gingival fibroblasts and GMSCs may not necessarily align in the context of wound healing in vivo. It has been shown however that gingival fibroblasts express toll-like receptors and can secrete pro-inflammatory cytokines such as IL-6 and IL-8 when exposed to bacterial products in vitro and in vivo^58–61. Therefore, gingival fibroblasts, in addition to their traditional role in producing ECM, may exert polarizing effects on inflammation depending on physiological or pathologic condition.

Table 1:

Wound healing characteristics in oral mucosa over loose connective tissue or bone

Mucosal healing over loose connective tissue
Mucosal type (model)	Healing characteristics	Reference
Buccal mucosa (mouse)	Reduced collagen deposition and TGFb1 compared to skin healing	2
Buccal mucosa (human)	Increased epithelial migration and proliferation, rapid resolution of inflammation	11
Buccal mucosa (human)	Increased NETosis and anti-microbial activity	63
Pharyngeal mucosa (human)	Fewer numbers of immune cells in healthy and healed mucosa compared to skin	5
Tongue (mouse)	Enhanced mucosal healing by salivary EGF, and reduced apoptosis and hypoxic response compared to skin wounds	7, 10, 13
Tongue (mouse)	Reduced inflammation	45
Tongue (mouse)	Reduced angiogenesis versus skin wounds	74
Mucosal healing over bone
Mucosal type (model)	Healing characteristics	Reference
Palatal gingiva (pig)	Reduced inflammation, contraction and fibroblast density, improved healing compared to skin wounds	42
Palatal gingiva (mouse)	Reduced inflammation and improved healing by GMSC application	54
Palatal gingiva (mouse)	Keratinocyte-derived VEGF controls angiogenesis in oral wounds	78
Palatal/marginal gingivae (mouse)	Presence of neural-crest-derived gingival MSCs	49, 50
Marginal gingiva (human)	Expresso, of IL-6 and IL-8 by gingival fibroblasts	61
Marginal gingiva (human, mouse)	DEL-1 mediated resolution of inflammation in periodontitis	66
Marginal gingiva (mouse)	Homeostatic defense against masticatory stress by IL-17 driven mechanism	69
Marginal gingiva (mouse)	Protection by gamma-delta T cell-derived amphiregulin in periodontitis model	70

Inflammation in gingival and mucosal wounds is more transient than in skin wounds^11,42. Initial inflammation is characterized by neutrophil infiltration which is necessary to limit bacterial invasion into the wound site via secretion of proinflammatory cytokines, phagocytosis and NETosis^62,63. Clearance of neutrophils occurs via apoptosis⁶⁴, which may be targeted by bacteria to prolong their lifespan and prolong inflammation65. Another mechanism of neutrophil clearance is DEL-1-mediated efferocytosis by macrophages, which in turn reprograms macrophages toward a pro-healing phenotype to facilitate gingival regeneration⁶⁶. Furthermore, lipid mediators such as resolvins and lipoxins have shown to have a pro-resolving effect in gingival inflammation from rabbit models^67,68. Lymphocyte population has also been shown to play an important role in shaping immunity in gingival injury. In mice, induction of mechanical damage to gingiva, either by mastication or external abrasion, stimulates gingival epithelial cells to secrete IL-6, which in turn leads to expansion of Th17 cells to produce defensins and neutrophil chemoattractants⁶⁹. Moreover, gingival γδ T cells have been shown to produce amphiregulin to enhance wound healing in a periodontitis model⁷⁰. It is well established that unresolved inflammation results in impaired wound healing and scarring^71,72. Therefore, transit inflammation may partially contribute to gingival wound healing.

Oral wounds have reduced angiogenic response when compared to skin wounds^73,74, which may be due to excessive angiogenesis of skin wounds that is accompanied by inflammation. Inflammatory signals such as IL-1β and TNF promote neovascularization^75,76 and are linked to aberrant angiogenesis in rheumatoid arthritis⁷⁷. Moreover, oral keratinocytes express reduced levels of VEGF compared to skin keratinocytes⁷⁴, which may be advantageous. VEGF expression by the oral keratinocytes is FOXO1-dependent, and its deletion impairs angiogenesis and wound healing parameters in gingiva⁷⁸.

Diabetic Impact on Cells Involved in Connective Tissue Healing

Diabetes negatively affects multiple cell types that participate in connective tissue wound healing in mucosal surfaces as summarized in Table 2. In both T1DM and T2DM mouse models, new connective tissue fill in palatal gingival wounds is reduced compared to normoglycemic control groups⁷⁹. This is attributed to reduced fibroblast numbers in healing wounds and downregulation of pro-collagen I and III expression⁸⁰. Mechanistically, high glucose levels promote apoptosis and reduce proliferation of gingival fibroblasts^79,81, which mechanistically is due to the impact of greater oxidative stress, excessive activation of FOXO1 and high levels of caspase-3 expression^80,82. High glucose levels also cause abnormal expression of MMP-1 from human gingival fibroblasts in vitro^83,84. Moreover, AGEs upregulate proinflammatory cytokine expression such as IL-6 and IL-8 in human gingival fibroblasts^83,84. Human GMSCs treatment improves wound healing in diabetic murine models^85,86, but it has been shown that GMSCs isolated from diabetic patients exhibit reduced proliferative capacity compared to those from nondiabetic patients⁸⁷. Similarly, bone marrow derived MSCs isolated from T2DM patients exhibit reduced immunomodulatory capacity in suppressing T cell proliferation⁸⁸. Therefore, diabetes affects gingival fibroblasts by reducing their numbers through enhanced apoptosis, increases catabolic events in wounds through upregulation of MMP-1 and down-regulation of TIMP-1, and increases expression of inflammatory cytokines, which altogether lead to suboptimal connective tissue fill in gingival wounds.

Table 2:

Diabetes-induced alterations in wound healing.

Cell or Tissue	Effect of Diabetes	Reference
Epithelium	Elevated ROS	33, 34
Epithelium	Increased expression of receptor for AGE	35
Epithelium	Alteration in downstream genes induced by FOXO1	22
Epithelium	Reduced growth factor expression	17, 78
Epithelium	Increased cytokine expression	15, 17
Epithelium	Reduced re-epithelialization	17, 79
Saliva	Reduced growth factor and increased inflammatory proteins	6, 20, 23
Connective tissue	Increased fibroblast apoptosis	79
Connective tissue	Increased cytokine expression and inflammation	15, 79
Connective tissue	Reduce growth factor expression	102
Connective tissue	Reduced fibroblast proliferation	79
Connective tissue	Increased FOXO1 activity	79
Connective tissue	Increased expression of receptor for AGE	35
Connective tissue	Reduced angiogenesis	78
Connective tissue	Reduced production of connective tissue matrix	78
Gingiva	Increased inflammation and ROS	81
Alveolar bone	Increased bone resorption and osteoclasts	111, 136
Alveolar bone	Reduced bone formation	111, 132, 134, 144–146
Alveolar bone	Increased inflammation	129, 147
Alveolar bone	Increased osteonecrosis	153, 154, 165
Mandibular bone	Reduced bone formation	123, 124
Mandibular bone	Increased postoperative infection risk	123

Diabetes impairs immune cell function and promotes inflammation to delay gingival wound healing. It is well established that diabetic wounds present with persistent inflammation^81,89, particularly with high levels of the pro-inflammatory cytokine TNF. Inhibition of TNF can reverse diabetes-induced apoptosis of fibroblasts and mesenchymal stem cells in skin and osseous wounds^26,90. A major source of TNF is macrophage/monocyte population, and high glucose has been shown to cause epigenetic changes in the promoter region of TNF and MCP-1 to promote persistent inflammation through over-activation of NF-kB^91,92. Moreover, diabetes and a high glucose environment impair phagocytic activity of macrophages that is necessary for clearance of apoptotic bodies and bacteria^93,94. Diabetes has also shown to prime neutrophils to undergo NETosis and cause aberrant inflammation in mice and humans^95,96, which also may potentiate gingival inflammation⁹⁷. Furthermore, metabolic changes in obesity and T2DM have been linked to reduced production of lipid mediators that are important for resolving inflammation in diabetic wounds⁹⁸. Overall, diabetic impact on multiple immune cell types is detrimental for wound healing by interfering with their function and promoting persistent inflammation. This subsequently sets the stage for sub-optimal connective tissue wound healing.

Diabetes can have an opposite effect on angiogenesis depending on various anatomic locations. In diabetic retinopathy and nephropathy, angiogenesis is excessive, and subsequent vascular edema is linked to organ damage^99,100. In diabetic skin wounds and chronic foot ulcers, angiogenesis is minimal, which is associated with delayed healing and reduced antimicrobial activity^99,101. In buccal mucosal wounds of diabetic rats, new blood vessels are more permeable and the numbers are significantly reduced compared to normoglycemic wounds^15,102. It has been shown that TNF inhibition can reverse the number of new blood vessel formation in diabetic osseous wound by restoring VEGF expression¹⁰³, providing a link between persistent inflammation and reduced angiogenesis. Because reduced angiogenesis in gingival wounds can lead to delayed healing⁷⁸, diabetes-mediated reduction in neovascularization may represent an important pathogenic mechanism to alter gingival wound healing. Taken as a whole, it is apparent that diabetes dysregulates the expression of a number of different factors to impair the wound healing process in various cell types as shown in Figure 2.

Figure 2. Connective tissue healing in diabetic oral wounds.

Gingiva is composed of a thick layer of lamina propria overlaying bone. Diabetes-mediated hyperglycemia and ROS are the major driving force for directly induced fibroblast apoptosis and inhibiting proliferation. Moreover, fibroblasts express enhanced levels of pro-inflammatory cytokines IL-6, IL-8 sad MMP-1 in diabetic gingival wounds. Diabetes also interferes with a phagocytic function of macrophages and ttuedir polarisation from pro-inflammatory (MI) to pro-resolving (M2) type. Diabetes-induced epigenetic changes prevent resolution of inflammation. The promoter region of TGFb1 in kerartinocytes are methylated by high glucose whereas that of ITNF in marcphages are acetylated, resulting in reduced anti-inflammatory TGFb1 expression and over-expression of pro-inflammatory TNF. The overall outcome is persistent inflammation, which reduces stromal healing outcome.

Keratinocytes and Connective Tissue Healing

Crosstalk between keratinocytes and fibroblasts is needed for normal connective tissue healing¹⁰⁴. Keratinocytes produce growth factors such as TGF-β, CTGF and VEGFA to stimulate connective tissue healing^17,105–107. TGF-β can directly and indirectly stimulate connective tissue formation; the latter is mediated by connective tissue growth factor (CTGF), also known as cellular communication network factor-2 (CCN-2)105. Application of CTGF in vivo stimulates fibroblast proliferation and deposition of collagen¹⁰⁸. Studies examining human and animal dermal wounds demonstrate that VEGF is produced by keratinocytes during the healing process. The transcription factor FOXO1 activates wound healing responses in keratinocytes including the production of TGF-β1 to promote connective tissue healing in oral wounds¹⁷. FOXO1 ablation in keratinocytes diminishes TGF-β1 expression, which reduces CTGF expression by cells in connective tissue¹⁰⁵. In type 1 diabetic oral wounds, the production of TGF-β1 is reduced. This may be explained by substantially reduced binding of FOXO1 to the TGF-β1 promoter caused by high levels of glucose or AGEs, which reduce TGF-β1 transcriptional activity¹⁷. Lineage specific ablation of FOXO1 reduces VEGF-A expression in keratinocytes in mucosal wounds and delays wound closure and with reduced angiogenesis^77,107.

Molecular Mechanisms Involved in Impaired Diabetic Bone Healing

Dental and periodontal problems are exacerbated by diabetes¹⁰⁹, often resulting in tooth extraction, bone augmentation and restoration with dental implants. These treatment modalities require an appropriate osseous healing reponse that involves concerted activity of osteoclasts, osteoblasts and mesenchymal stem cells (MSCs). It is well established that diabetes has a detrimental impact on bone formation by affecting these cell types. Diabetes uncouples bone remodeling by increasing osteoclast numbers and inhibiting expansion of periodontal fibroblasts and osteoblasts in an experimental periodontitis model^110,111. This is linked to a negative effect of high glucose on the proliferation and osteogenic differentiation of progenitor cells¹¹². The molecular mechanism by which diabetes affects osteoblast progenitors and osteoblasts has been investigated extensively in orthopedic injury models. The early phase of diabetic bone healing is characterized by impaired osteoid matrix production and reduced cellularity, which may be caused by defective MSC recruitment, proliferation and impaired osteogenic differentiation^113,114. Diabetes further suppresses mRNA expression of the Dlx5 and Runx-2 transcription factors, which are responsible for the expression of the osteoblastic phenotype¹¹⁵, suggesting that the low rate of bone formation observed in diabetic animals may be related to deficiencies in the differentiation of MSCs to osteoprogenitor cells and in the osteoblastic cell maturation¹¹⁴. Studies in fracture healing support this hypothesis as diabetes results in impaired bone formation¹¹⁶. Moreover, diabetes leads to aberrant activation of NF-kB and enhances inflammation due in part to the loss of immunomodulatory MSCs¹¹⁷. In turn, persistent inflammation characterized by high levels of TNF reduces MSC expansion through upregulation of FOXO1⁹⁰ and suppression of IHH¹¹⁸ to ultimately impair healing in diabetic fractures. In the oral cavity, animal studies have recently identified genetically traceable MSC populations by a-SMA, Axin2 and Lepr expression^119–121, which reside in the periodontal ligament space between the tooth surface and bone and participate in bone healing after injury. These animal models may be crucial in understanding intraoral MSC and osteoprogenitor behavior necessary for bone healing and how diabetes alters it in vivo.

Guided bone regeneration (GBR) is a surgical procedure that restores deficient bone volume with the use bone graft. A key element is a barrier membrane that separates the bone graft from the soft tissue to exclude soft tissue infiltration that can interfere with bone formation¹²². The effect of experimental diabetes on GBR has been investigated by means of histology and gene expression analyses. GBR can sufficiently regenerate bone even in diabetic rats, but the quantity of bone formed is less than that formed in normoglycemic animals¹²³. Investigation of gene expression profiling has shown that poorly controlled diabetes downregulates pathways related to cell division, energy production and osteogenesis during the proliferative phase of intramembranous bone healing after bone augmentation treatment¹²⁴. Although detailed mechanisms responsible for the reduced cell proliferation rates are still not completely elucidated, it has been suggested that diabetes may suppress the expression of growth factors needed for osteogenesis during the early phase of bone healing^125,126. In support of this diabetes reduces expression of basic fibroblast growth factor (FGF), an important mitogenic factor for MSCs during bone healing¹²⁷. Animal studies indicate that experimentally induced type 1 diabetes diminishes the osteoprogenitor population in bone marrow, possibly due to increased apoptosis via oxidative stress, which in turn, may impair osteoblastogenesis, bone formation, and bone healing¹²⁸. Experimental studies have also demonstrated reduced mRNA and protein expression of platelet derived growth factor (PDGF), TGF-β1, IGF-I and VEGF in femoral fractures of diabetic rats during early osseous healing correlates with reduced cell proliferation^125,126. Diabetes increases TNF in the oral cavity that prolongs detrimental inflammation. FGF-2, TGFβ-1, bone morphogenetic protein-2 (BMP-2), and BMP-6 are suppressed during periodontal bone foramtion, which is rescued when inflammation is reduced with a TNF inhibitor¹²⁹.

The exogenous delivery of IGF-I or rhFGF-2 enhanced the healing of calvarial and periodontal defects in mice and rats^130–132. Preclinical studies have demonstrated that diabetic rats exhibit reduced periodontal bone regeneration, which can be partly rescued with local application of enamel matrix derivative (EMD) that contains a milleu of growth factors^133,134. A recent prospective study in human subjects showed that treatment of periodontal bony defects with enamel matrix derivative, which contains a mixture of undefined growth factors, in diabetic patients, can achieve a regenerative outcome that is comparable to non-diabetic patients after 3 years¹³⁵.

Diabetes has a significant impact on inflammatory responses. Pathways linked to innate and adaptive immune responses are generally altered in poorly controlled diabetics compared to healthy controls. Prolonged inflammation is detrimental to bone formation, in part, because NF-kB activation in response to inflammatory stimuli blocks the expression of bone matrix proteins¹³⁶. Diabetes also interferes with proper immune functions that are necessary for pathogen recognition and removal^137–139. These findings may explain increased risk for infectious complications following GBR treatment in patients with uncontrolled diabetes. Comparable findings were also reported by other authors evaluating the healing of bone defects in animals¹⁴⁰ and humans¹⁴¹. The regulation of osteoblasts and osteoblast precursors is complex and involves the timed expression of genes encoding pro-inflammatory cytokines (Il1α, Il1β, Il-6, TNF), chemokines (Ccl20, Cxcl1, Cxcl2, Cxcl10), chemokine receptors (Ccr2, Ccr5, Ccr6) and cell adhesion molecules (Vcam1 and Icam1) that are needed for migration and activation of these cells^142,143. Prolonged inflammation disrupts the level of expression and timing of these factors.

Taken together, the available data suggest that suppressed differentiation, proliferation and/or bone forming capacity of osteoblastic cells during the critical early healing period appear to play an important role in the pathogenic mechanism underlying poor bone formation in diabetes. In particular, poorly controlled diabetes appears to result in a higher and dysfunctional inflammatory response during the early phase of bone healing which in turn, may be resposnible for the increased risk of infection in subjects with uncontrolled diabetes.

Impact of Diabetes on Tooth Extraction

The healing of bone in extraction sockets following tooth removal in diabetes has been evaluated in a number of animal and human studies. Using T1DM rat model, alveolar bone remodeling following tooth extraction evealed that the quantity and rate of alveolar bone formation was significantly lower in diabetic group compared to controls¹⁴⁴. After a healing period of 10 days, diabetic sockets contained thin and scanty collagen fibers whereas in control groups there were thick collagen fibers that formed a pretrabecular scaffold along the trabeculae¹⁴⁵. Interestingly, there was no evidence of diabetic microangiopathy in the extraction sockets of diabetic animals compared with healthy controls or with insulin-treated animals. Healing of in humans with T2DM showed a deficit in the early stages of healing and in diabetic pigs revealed delayed healing and decreased osteogenic differentiation of MSCs¹⁴⁶. Although substantial bone formation still occurred in diabetic animals there was incomplete mineralization in the center of the socket compared to that of control groups which were completely filled with newly formed bone¹⁴⁶. The observations made in animals were in line with those made in humans. Compared to nondiabetic control group, T2DM patients had delayed bone fill in the extraction sockets at multiple time points post surgery¹⁴⁶ and 54.7% had defective healing manifested by inadequate bone formation. In diabetic mice the extraction sockets consistently have a higher inflammatory profile, with more M1 macrophages and TNF-α expression and less M2 macrophages and PPARγ expression compared to non-diabetic controls¹⁴⁷.

While several studies point to deficits in bone formation after dental extraction in diabetic patients not all reports reach this conclusion^148–151. For example it has been reported that diabetics have increased neutrophil recruitment following tooth extraction but bone healing in diabetic and normal subjects were similar¹⁴⁹. Taken as a whole the overall assessment of healing following tooth extraction in diabetic patients is slower compared to non-diabetic ones, especially in the initial phases¹⁵¹. The available evidence suggests that a) in poorly controlled T1DM bone formation is inhibited, resulting in delayed healing and deficient bone formation; and b) tooth extractions can be performed in well controlled diabetic patients, without major postoperative complications.

Diabetes as a Comorbidity for MRONJ

In the oral cavity, osseous wounds such as extraction socket or exogenous bone grafts eventually heal even under diabetic condition, albeit at a delayed rate. However, major differences exist in bone healing for nondiabetic and diabetic patients that suffer from medication-related osteonecrosis of the jaw (MRONJ). MRONJ may be localized and easily treated or may become widespread and develop into a serious clinical condition characterized by the progressive destruction of the jawbones. It occurs in patients on anti-resorptive medications that have undergone oral or periodontal surgery extraction¹⁵². There is a strong association between diabetes and the incidence of MRONJ^153–157; even though the exact pathogenetic mechanisms remain unclear, several experimental studies have proposed converging pathways. Much like diabetes-mediated delay in extraction socket healing¹⁵⁸, bisphosphonate treatment delays bone formation in extraction sockets in rats¹⁵⁹. There are a number of potential mechanisms that may play a role including diabetes-reduced coupled bone formation, and the impact of diabetes on increased osteocyte and osteoblast apoptosis^160,161. Another possible converging mechanism is altered immune system. Animal and human studies have demonstrated that bisphosphonate treatment impairs neutrophil and macrophage chemotaxis and function to promote immunosuppression^162–164. Studies in mice have also demonstrated a pro-inflammatory response by diabetes and zolendronate treatment in macrophages, which was responsible for MRONJ-like healing after tooth extraction^165,166. The etiology for either disease is multifactorial and complex, and further studies may identify pathways that can be targeted to prevent MRONJ incidents in diabetic patients. The impact of diabetes on intraoral bone healing as discussed above is summarized in Figure 3.

Figure 3. Diabetic impact on various cases of intraoral bone healing.

In the oral cavity, bone healing takes place under different scenarios inch as. left right, periodontal bone regeneration, guided bone regeneration. healing after dental extraction, or pathological healing such as MRONJ. Diabetes significantly delays bone healing in each of these processes. In periodontal bone healing, diabetes enhances osteoclastogenesis while reducing the number of osteoblasts its and periodnotal fibroblasts This is accompanied by persistent inflammation and reduced growth factor expression. In guided bone regeneration, which, restores the hard tissue volume necessary for dental implant placement, diabetes has shown to inhtbit genes associated with cell division and osteogenesis. In exodontia, diabetes delays socket healing through elevation of pro-inflammatory cytokines such as TNF, which may indirectly suppress proper expansion of osteogenic stem cells. Diabetes is a significant co-morbidity for MRONJ. and although exact pathogenesis is unclear, experimental studies have shoun enhanced apoptosis, persistent inflammation and improper recruitment of leukocytes as a converging mechanism.

Conclusions and future directions

Diabetes negatively impacts several aspects of oral and dermal wound healing. When directly compared, the effects of diabetes on both are relatively consistent. Diabetes alters the rate of healing by reducing keratinocyte migration, the production of growth factors by a number of different cell types including keratinocytes, interferes with a number of cellular functions such as cellular proliferation and differentiation and the expression of genes that are essential for forming soft connective tissue and bone matrix. Some of the cellular dysregulation that occurs may be due to alterations in the gene targets of transcription factors such as FOXO1 or over activation of others such as NF-kB. These changes lead to difficulty in down regulating inflammation, increasing oxidative stress and enhancing apoptosis. In the skin these changes may lead to biofilm formation in healing wounds that stopped the repair process and in the mouth, a biofilm may form on surgically exposed bone that prevents healing. In oral wounds that require substantial connective tissue healing, diabetes exerts its deterimental impact on multiple cell types such as gingival fibroblasts, endothelial cells and leukocytes through persistent inflammation and aberrant cell apoptosis. Impaired healing by diabetes can be rescued by controlling inflammation or by genetic inhibition of FOXO1 in experimental models, thus targetting these processes and/or transcription factors with monoclonal antibody or small molecule inhibitor may be a viable method to treat diabetic oral wounds in humans. Similarly, diabetic intraoral bone healing is characterized by reduced expression of growth factors that are necessary for bone formation, therefore the use of recombinant growth factors is an attractive therapeutic option for diabetic patients. Given the difficulty that diabetic patients have in resolving inflammation the optimal treatment may include a therapeutic that facilitates resolution and delivers a growth factor to diabetic wounds.

A mechanistic understanding of oral wound healing is largely based on animal studies. These studies are advantageous by targetting specific genes or factors that may be implicated in different stages of the healing process. However, they may not fully translate to human models as diabetes itself is a multifactorial disease that is often accompanied by other metabolic conditions. Moreover, certain oral diseases in humans such as MRONJ may not be fully replicated in mouse models. Despite these limitations, animal studies have contributed significantly to understanding some detreimental effects of diabetes on oral wound healing, and clinical efforts to supplement growth factors to improve healing process are active underway. Future studies may test a translational value of recombinant growth factor therapy, anti-inflammatory drugs and/or small molecule inhibitors of transcription factors that are implicated in diabetic healing. Any other future studies?