When implants are placed, the healing process occurs, such as connective tissue repair, while epithelial cells move horizontally to cover the wound on the connective tissue (Atsuta et al., 2005a; Sculean et al., 2014; Tomasi et al., 2014). Since the implant blocks the growth and horizontal migration of epithelial and connective tissue, they begin to spread along the implant surface at depth and turn back at a certain depth, thus forming a mucosal tissue similar to that of natural teeth, i.e., the pores formed by the implant are gradually sealed by epithelial and connective tissue (Fujii et al., 1998; Gulati et al., 2020).

Histologically, the peri-implant mucosa consists of well-keratinized oral epithelium, sulcular epithelium and a thin barrier epithelium facing the abutment, which corresponds to the around the tooth, called the peri-implant Junctional Epithelium (JE) (Schroeder et al., 1981; Gould et al., 1984; Buser and Brägger, 1989). The height of the peri-implant JE is approximately 2 mm, and the connective tissue below the JE is about 1.5 mm. Thus, the average biological width usually reaches 3.8 mm (Berglundh and Lindhe, 1996). When the biological width of any part of the peri-implant mucosa decreases, marginal bone resorption is generally observed, so the biological width is adjusted to compensate for these changes. The desmosome junctions in JE are fewer than the oral epithelium, so the intercellular gaps are wide and contain many immune cells such as neutrophils (Shimono et al., 2003). The JE is the first line of defense of the titanium or zirconium oxide peri-implant tissue against pathogen invasion. The epithelial tissue attaches to the surface of the implant system via the inner basement membrane and hemidesmosomes (HDs) (Ivanovski and Lee, 2018; Ren et al., 2019). The HDs are mainly located on the root side of the JE, so the implant’s JE is weakly closed (Bartold et al., 2000; Atsuta et al., 2013; Atsuta et al., 2016).

Among the connective tissue surrounding the implant, it is noteworthy that, due to the absence of the osseous layer, there are no vertical fibers inserted vertically into the implant surface, and the gingiva is directly “adapted” to the implant surface (Fujii et al., 1998), indicating that the sealing capacity of this structure is more fragile than that of a natural tooth and more prone to periodontal rupture and subsequent bacterial invasion (Moon et al., 1999). Compared to natural teeth, fewer fibroblasts and mesenchymal cells are usually intermediate to collagen fibers (Al Rezk et al., 2018). The connective tissue is adhered by gingival fibroblasts to the surface of the implant system through adhesive plaque attachment, tight junctions and extracellular matrix connections, and collagen fibers parallel to the implant surface (Bartold et al., 2000; Amberg et al., 2018).

Studies have found an increase in Actinomyces naeslundi and Streptococcus oralis around implants in healthy patients compared to natural teeth (Gürlek et al., 2017). There are still inconsistent and even contradictory conclusions about whether diabetes alters the bacterial composition of the oral cavity (Chapple et al., 2013; Graves et al., 2020). In type 2 diabetic periodontitis, the detection rate of orange complex species in subgingival plaque was significantly higher compared to a healthy periodontal state (Sabancı et al., 2022). A 16s rRNA assay of the subgingival flora of patients with periodontitis showed a significant augment in the relative abundance of Aggregatibacter, Neisseria, Fusobacterium, and Actinomycetes in type 2 diabetic patients (Casarin et al., 2013). Another report indicated that the levels of Porphyromonas gingivalis (P. gingivalis) or Porphyromonas forsythia (P. forsythia) showed an increase (da Cruz et al., 2008). Furthermore, a recent metagenomic shotgun sequencing analysis comparing the subgingival flora of non-diabetic and diabetic patients found that subgingival flora tended to be more abundant in pathogenic species in the T2DM state, regardless of whether the periodontal tissue was healthy, suggesting that T2DM patients face a higher risk of developing periodontitis (Shi et al., 2020).

Similarly, diabetes may also affect peri-implant inflammation by altering the subgingival flora. One study divided patients suffering from peri-implantitis into a systemic health group and a T2DM group (Sabancı et al., 2022). There are no significant differences in microbial species were found in the shallow peri-implant pockets of the two groups (Sabancı et al., 2022). However, Campylobacter rectus (C. rectus), P. gingivalis, Actinobacillus actinomycetemcomitans (A. actinomycetemcomitans) and Tannerella forsythia (T. forsythia) may be reduced in deep pockets around implants by T2DM (Sabancı et al., 2022). An interesting study to identify the microbiota in peri-implantitis pockets by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry showed Neisseria flavescens, Streptococcus constellatus, Slackia exigua, Streptococcus intermedius, Fusobacterium nucleatum, and Gemella morbillorum were the central resident folk detected in the peri-implantitis pockets (Yeh et al., 2019). The above findings suggest that the sequencing identification of bacteria alone may not determine whether they are pathogenic in patients with type 2 diabetes. Non-etheless, the diabetic state produces changes in some microbial species in the peri-implant pocket, and these changes may affect microbial homeostasis, but the exact effects are inconclusive.

Since the discovery of bacterial binding to ECM proteins (e.g., collagen, fibronectin, and laminin) more than 40 years ago (Kuusela, 1978), our understanding of pathogen-host cell interactions has gradually enhanced. The adhesion of bacteria to host tissues is the first decisive step in the infection process (Keller et al., 2015). In Gram-negative bacteria, adhesion to host cells can be achieved by binding to ECM proteins as a tool for pathogen-host contact (Vaca et al., 2020). Even pathogens secrete bacterial proteases to degrade ECM proteins and thus disrupt the barrier (Singh et al., 2012). For example, Treponema denticola produces chymotrypsin-like proteases targeted to degrade Laminin, collagen IV, and fibrinogen, allowing bacteria to invade the basement membrane after ECM degradation (Bamford et al., 2007; Singh et al., 2012). In the oral cavity, JE can express defensins, chemokines and cytokines (Groeger and Meyle, 2019). When bacteria colonize JE, the adhesion of keratinocytes is negatively affected, and laminin and collagen present in JE become binding sites for pathogens (Marre et al., 2020). In conclusion, the difference between hyperglycemia and normoglycemia is mainly in the deep pockets around the implants (Sabancı et al., 2022), and diabetic patients are more susceptible to microbiome changes than normal healthy individuals (Sabancı et al., 2022).

Diabetes increases the physical inflammatory response (Andriankaja et al., 2012; Pacios et al., 2012). Diabetic patients have increased tumor necrosis factor-α (TNF-α), elevated polymorphonuclear leukocyte infiltration, and more bone loss in diabetic rats than in normoglycemic rats (Kang et al., 2012). Injection of equal amounts of bacteria into the soft tissues of diabetic animals caused a more severe inflammatory response than in normal animals (Naguib et al., 2004). These results suggest that diabetic animals have a more dramatic host response to bacterial attacks.

Diabetes augments levels of inflammatory cytokines such as TNF-α, interleukin-1β (IL-1β), interleukin-17 (IL-17), interleukin-23 (IL-23), and interleukin-6 (IL-6) in human periodontal tissue (Bastos et al., 2012; Polak and Shapira, 2018). Increased resistance and permeability of the gingival epithelial cell layer after TNF-α and IL-1β treatment is associated with the development of gingival inflammation (Fujita et al., 2012; Miyagawa et al., 2016; Lagha and Grenier, 2019). Moreover, human gingival fibroblasts (HGF) treated with TNF-α and IL-1β induced increased HGF infiltration (Lv et al., 2020). Enhanced expression of inflammatory cytokines leads to rose vascular permeability and inflammatory cell recruitment (Domingueti et al., 2016) with upregulated receptor activator of nuclear factor kappa-B ligand (RANKL) or decreased osteoprotegerin (OPG) expression, stimulating increased bone resorption (Bastos et al., 2012). The more severe the hyperglycemia, the higher the levels of IL-1β and IL-6 in the oral saliva and the greater the probability of peri-implantitis (Al-Askar et al., 2018; Al-Sowygh et al., 2018; Vissink et al., 2018). And the restraint in anti-inflammatory factors may be one of the reasons for the exacerbation of periodontitis. Regulatory T-cells and M2-type macrophages responsible for anti-inflammation produce anti-inflammatory factors such as interleukin-4 (IL-4), interleukin-10 (IL-10), transforming growth factor-β (TGF-β) and anti-inflammatory lipid mediators, significantly reduced in diabetic complications (Acharya et al., 2017; Van Dyke, 2017).

Diabetes also affects the innate and adaptive immune responses of cells. Neutrophils make up the majority of cells in the gingival sulcus and are an essential component of the host response to tooth-associated biofilms (Hajishengallis, 2014). High glucose stimulates the production of more chemokines, which induce neutrophil recruitment in response to bacterial challenges (Cintra et al., 2014; Manosudprasit et al., 2017; Zheng et al., 2018b), and stimulates neutrophil initiation by increasing protein kinase C (PKC) activity (Karima et al., 2005). In addition, diabetes augments neutrophil activation and reactive oxygen species (ROS) production to increase damage to periodontal tissue (Guerra and Otton, 2011). Macrophages are associated with peri-implant disease. Diabetes promotes the production of IL-1β and TNF-α by macrophages, which may contribute to the enhancement of peri-implant infection (Salvi et al., 1998). Diabetes may increase the polarization of M1 macrophages to upregulate the susceptibility and severity of peri-implantitis (Sreedhar et al., 2017). In addition, higher levels of TNF-α, CC chemokine receptor 5 (CCR5) and CXC chemokine receptor 3 (CXCR3) at peri-implant sites in patients with chronic periodontitis and diabetes suggest a high potential for peri-implant bone loss (Venza et al., 2010). Dendritic cells are also associated with peri-implant disease, and their function may be regulated by diabetes, thereby promoting the disease process. Diabetes may control dendritic cells to alter alveolar bone loss by increasing the production of Th1 or Th17 lymphocytes or decreasing the formation of regulatory T-cells (Santos et al., 2010; Silva et al., 2012; Song et al., 2018).

A fundamental cause of the pathogenesis of peri-implantitis is how diabetes affects the soft tissue barrier of the implant, but it has not received sufficient attention. The sealing structure formed by the soft tissues is an effective barrier to protect the implant from external microorganisms, but the corresponding specific mechanisms remain to be thoroughly investigated.

Matrix Metalloproteinases (MMPs) are a group of protein hydrolases that play an essential role in the turnover of ECM and maintain the balance between remodeling and degradation of ECM (Klein and Bischoff, 2011; Bastos et al., 2017). The substrate of MMP is mainly collagen, but it also includes many other ECM proteins, including fibronectin, laminin, etc. (Ruoslahti, 1996; Stamenkovic, 2003). Regarding the mechanism of neurological damage by cerebral ischemia, MMP-9 destroys neurons by degrading laminin (Lee et al., 2009). MMPs are associated with diabetes-related peri-implantitis. MMP-8 and MMP-9 levels in oral fluids can reflect the disease status of periodontal disease to some extent (Domenyuk et al., 2019). In particular, the level of MMP-8 in oral fluid is upregulated proportionally to the severity of periodontal/peri-implant disease (Sorsa et al., 2000; Kiili et al., 2002; Sorsa et al., 2010; Sexton et al., 2011; Sorsa et al., 2011). MMP-8 in saliva was similar in composition to its counterpart within the gingival sulcus (Gangbar et al., 1990), and a similar pattern of elevated MMP-8 was observed in the gingival sulcus of peri-implantitis to that observed at the site of periodontitis (Ma et al., 2000; Arakawa et al., 2012; Janska et al., 2016). A study reported that patients with both chronic periodontitis and diabetes had much higher levels of MMP-8 and MMP-9 in periodontal tissue compared to patients with chronic periodontitis but systemically healthy patients and healthy controls (Kumar et al., 2006), and there was a trend toward significantly increased MMP-8 in patients transitioning from the absence of both diseases (Hardy et al., 2012). In addition, the excessive release of various MMPs by fibroblasts affected by reactive oxygen species promoted the degradation of connective tissue and bone matrix (Guan et al., 2009; Kaur et al., 2013). The interaction of MMPs with advanced glycosylation end-products (AGEs)/receptors for advanced glycosylation end-products (RAGEs) further contributes to increased inflammation in diabetes. AGEs accumulate and destroy periodontal tissue continuously under hyperglycemic conditions and directly or indirectly mediate intracellular effects through RAGE receptors of epithelial cells, gingival fibroblasts and other cells (Plemmenos and Piperi, 2022). The combination of AGEs and RAGEs in specific cell lines can stimulate the production and activity of MMPs, making the inflammatory response more severe (Ryan et al., 1999; Nesto and Rutter, 2002). AGEs can also influence the production and structure of ECM proteins by affecting the cross-linking of collagen (Snedeker and Gautieri, 2014).

Both oral mucosal and skin wound healing requires four stages: Hemostasis, inflammation, proliferation, and maturation/stromal remodeling. Soft tissue healing after implant placement surgery takes 6–8 weeks, which is much longer than the epithelial healing time for dental surgical wounds (7–14 days) (Salvi et al., 2000; Hämmerle et al., 2014). Wound healing involves re-epithelialization, massive new connective tissue formation and new bone formation (Ko et al., 2021). The diabetic oral wound healing process is characterized by impaired keratinocyte proliferation and migration, altered levels of inflammation, and reduced neo-connective tissue and bone formation (Ko et al., 2021). The factors involved in delayed wound healing in diabetes include hypoxia, fibroblast and epidermal cell dysfunction, impaired angiogenesis, and neovascularization, elevated MMPs, damage from ROS and advanced glycosylation end-products (AGEs), neuropathy and decreased multilevel host immune resistance (Lan et al., 2008; Chrcanovic et al., 2014; Brizeno et al., 2016). Most of these conditions are exacerbated in diabetes mellitus with poor glycemic control and significantly increase the risk of poor wound healing in dental procedures. An in vitro study confirmed that hyperglycemia significantly inhibits the adhesion and proliferation of human gingival fibroblasts on titanium (Liu et al., 2021). Diabetic wounds have more neutrophils, enhanced TNF expression levels, and declined expression of growth factors in wound fluid, which may delay wound re-epithelialization (Siqueira et al., 2010; Romanos et al., 2019). In cutaneous wounds, diabetes reduced the switch from the M1 macrophage phenotype to the M2 macrophage phenotype (Al-Mulla et al., 2011). More pro-inflammatory macrophages implied increased expression of inflammatory cytokines (e.g., IL-1, IL-6, and IL-8) and MMP, which may lead to more prolonged inflammation (Wilkinson et al., 2019). Genes involved in apoptosis, mucosal migration and intercellular communication were reported to be upregulated 4-fold–200-fold when inflammation around the implant increased, which simultaneously slowed the rate of wound healing (Ganesan et al., 2022).

Chronic exposure to high glucose induces cytokine expression and responses to cytokine stimuli in diabetic patients, including the production of ROS (Lim et al., 2017; Zheng et al., 2018b). ROS are a series of molecular oxygen derivatives with pleiotropic properties, classified into radical ROS (e.g., hydroxyl radicals, superoxide anion radicals O₂ ⁻, peroxyl radicals, and alkoxyl radicals) and non-radical ROS (e.g., H₂O₂, organic hydroperoxides, single linear state molecular oxygen, ozone, hypochlorous acid, and hypobromous acid) (Forrester et al., 2018; Hawkins and Davies, 2019; Lushchak and Lushchak, 2021). Physiological levels of ROS regulate various life processes in cells and organs, including cell proliferation, differentiation, migration, and angiogenesis (Niki, 2016). However, when ROS levels exceed physiological concentrations, it will lead to cell growth arrest and death (Niki, 2016; Li et al., 2021). Due to high intracellular glucose levels, the mitochondrial electron transport chain is overactive, inducing the formation of reactive oxygen species and producing and releasing inflammatory factors. Reactive oxygen species can induce apoptosis and damage cellular and stromal DNA and structural components (Yang et al., 2015). Increased mitochondrial reactive oxygen species production exacerbates periodontitis in diabetic patients (Sun et al., 2017).

High glucose levels can lead to the formation of advanced glycosylation end products (AGEs). AGEs accumulate in most tissues of diabetic patients, including kidney, retina, gums, bone and periodontal tissue (Schmidt et al., 1996; Lalla and Papapanou, 2011; Napoli et al., 2017). AGEs bind to receptors for advanced glycosylation end products (RAGEs) and other receptors, activate nuclear factor κB (NF-κB), stimulate the production of reactive oxygen species, and induce the expression of inflammatory cytokines such as IL-6 and TNF-α (Singh et al., 2014; Nonaka et al., 2018). AGEs may inhibit the viability of human gingival fibroblasts and reduce the expression of type I and type III collagen (Ren et al., 2009). Interestingly, one study observed an increased concentration of AGEs in the fluid around the implant in the diabetic group and a significant correlation with the depth of the pocket around the implant (Al-Sowygh et al., 2018). In osteoblasts, excess AGEs promote apoptosis and inhibit osteogenic activity (Alikhani et al., 2007). Inhibition of RAGE reduced TNF-α production, which confirms the role of RAGE receptors in slowing the development of periodontitis (Lalla et al., 2003). A report on periodontitis and systemic disease found that combining AGEs and RAGEs leads to excessive inflammatory response and periodontal tissue destruction in patients with type 2 diabetes (Chapple et al., 2013). The effects of diabetes on dental implants’ soft tissue sealing are shown as follows (Figure 2).