Periodontitis is a disease caused by oral infection associated with polymicrobial dysbiosis and the activation of host immunity (Hajishengallis, 2014). Keystone or keystone-like pathogens, such as P. gingivalis and Tannerella forsythia (T. forsythia), can drive the disruption of periodontal tissue homeostasis and lead to inflammation (Wright et al., 2014; Lamont and Hajishengallis, 2015). Keystone or keystone-like pathogens of periodontitis, predominantly Gram-negative anaerobic or facultative bacteria, are appreciably sensitive to changes in the oxidative environment (Lamont and Hajishengallis, 2015). ROS can disturb the cellular oxidative environment and participate in the killing of keystone pathogens. For instance, a marked increase in ROS generation was observed when leukocytes were treated with P. gingivalis lipopolysaccharide (LPS) or Fusobacterium nucleatum (F. nucleatum) in vitro (Sheikhi et al., 2000; Zhu et al., 2016). Fascinatingly, it has very recently been reported that H₂O₂ is a central determinant of oral polymicrobial synergy (Lamont, 2016). However, several lines of evidence have suggested that periodontal pathogens such as Treponema denticola (T. denticola) have evolved strategies to suppress the induction of ROS (Shin et al., 2013).

Conversely, at basal levels, ROS serve as second messenger particulates in regulating signal transduction, cellular homeostasis, and cell death. For instance, H₂O₂ can trigger defensive inflammatory responses to environmental cues in periodontal tissues through mitogen-activated protein kinase (MAPK) and NF-κB as well as inflammasome activation (Almerich-Silla et al., 2015). Moreover, glucose oxidase, which continuously generates H₂O₂ at relatively low concentrations, could stimulate the proliferation and osteoblastic differentiation of hPDLFs through the upregulation of runt-related transcription factor-2 (Runx2) and osterix (Choe et al., 2012). H₂O₂could also increase the levels of gelatinolytic matrix metalloproteinases (MMPs), enhancing hPDLF migration in an MMP-dependent manner (Cavalla et al., 2015). These findings suggest that ROS participate in the proliferation and differentiation of hPDLFs. However, many studies have reported that H₂O₂ acts predominantly as an inhibitory mediator of cell proliferation and differentiation (Choi et al., 2009). A possible explanation for these contradictory results is that cellular responses to H₂O₂ can differ depending on the concentration of H₂O₂ and the type of cells. For example, Burdon et al. reported that exposure to 1 μM H₂O₂ promoted the proliferation of BHK-21 fibroblasts, while H₂O₂ at 0.5 and 1 mM caused apoptotic cell death (Burdon et al., 1996).

ROS have multifaceted effects, and the function of ROS is determined by the redox state (Zhao et al., 2016). Oxidative stress is induced when ROS are produced in excess of the capacity of the antioxidant system to efficiently counteract their actions, resulting in cytotoxic effects and enhancing periodontal destruction (Nibali and Donos, 2013). The involvement of ROS in the pathogenesis of periodontal diseases is highlighted by the existence of a disturbed redox balance in periodontitis cases. The results of recently published relevant studies have been summarized in Table 1. Data from a few cross-sectional studies have demonstrated low plasma and serum total antioxidant (TAOC) concentrations in periodontitis patients relative to healthy controls (Chapple et al., 2002; Brock et al., 2004; D'aiuto et al., 2010; Baltacioglu et al., 2014a; Thomas et al., 2014; Baser et al., 2015; Patil et al., 2016). Saliva is well recognized as containing a pool of markers for periodontitis (Zhang et al., 2016). Studies have also found similar results regarding salivary TAOC. The salivary TAOC was significantly lower in patients with chronic periodontitis compared with healthy controls (Diab-Ladki et al., 2003; Baltacioglu et al., 2014a; Miricescu et al., 2014; Baser et al., 2015; Zhang T. et al., 2015). Moreover, higher levels of reactive oxygen metabolites and total oxidant status (TOS) were observed in the serum, saliva, and gingival crevicular fluid (GCF) of patients with periodontitis compared with controls (Akalin et al., 2007; D'aiuto et al., 2010; Wei et al., 2010; Baltacioglu et al., 2014a,b). Furthermore, there is a strong negative correlation between salivary TAOC and clinical attachment loss (CAL) in periodontitis patients (Baser et al., 2015). Significant positive correlations were also observed between malondialdehyde (MDA), an LPO product, and TOS levels and periodontal parameters (Akalin et al., 2007). Collectively, these results suggested that reduced TAOC and increased ROS may be risk factors for periodontitis or may be caused by periodontal inflammation. However, it is very difficult to determine whether the change in redox status is the cause or a result of periodontitis. In addition, as summarized in Table 1, decreased levels of specific antioxidants, such as SOD, CAT, and GPX, were observed in periodontitis patients compared with healthy controls (Panjamurthy et al., 2005; Wei et al., 2010; Trivedi et al., 2014; Patil et al., 2016).

Comprehensive reviews on tissue damage caused by ROS have been published (Waddington et al., 2000; Chapple and Matthews, 2007; Nibali and Donos, 2013). To summarize, periodontal tissue damage may arise directly from oxidative stress and indirectly via the activation of cell signaling pathways related to inflammation, apoptosis, and other factors. It has been demonstrated that direct tissue damage caused by ROS can be mediated by (1) the induction of lipid peroxidation and cell membrane destruction (Mashayekhi et al., 2005; Panjamurthy et al., 2005; Tsai et al., 2005; Pradeep et al., 2013), which results in (2) protein denaturation and enzyme deactivation (Nibali and Donos, 2013; Trivedi et al., 2014; Nguyen et al., 2016; Patil et al., 2016), leading to (3) nucleic acid damage (e.g., strand breaks and base pair mutations) and chromosome disruption (Takane et al., 2002) and causing (4) mitochondrial injury and ROS bursts (Battino et al., 1999). Tissue destruction can be assessed by measuring the levels of markers for lipid peroxidation, protein damage, and DNA damage, such as MDA, protein carbonylation and 8-hydroxy-2-deoxyguanosine (8-OHdG) (Sawamoto et al., 2005; Takane et al., 2005; Canakci et al., 2009; Su et al., 2009; Mai et al., 2012; Sezer et al., 2012; Dede et al., 2013; Hendek et al., 2015). The results of relevant published studies have been summarized in Table 2.

A more complex question is how ROS result in periodontal tissue damage by regulating signal transduction and gene transcription, which is described in Figure 1. There are at least four pathways relevant to this topic. First, ROS are able to activate NF-κB, initiating a signaling cascade that regulates inflammatory and immune responses (Morgan and Liu, 2011). Second, ROS are involved in inducing JNK activation, resulting in cell apoptosis (Nakano et al., 2006). Third, ROS are associated with inflammasome activation, leading to pyroptic cell death (Zhou et al., 2010). Fourth, ROS play a critical role in autophagy (Filomeni et al., 2015). This section will focus on evidence for the mechanisms of ROS-mediated activation of NF-κB, JNK, and inflammasomes in periodontitis. The relationship between ROS and autophagy in periodontitis will be described later.

ROS have been reported to both activate and repress NF-κB signaling in studies of different cells and different upstream pathways, and ROS-mediated activation of NF-κB signaling results in the expression of pro-inflammatory cytokines and chemokines (Ozcan et al., 2016). The expression of these cytokines leads to periodontal destruction by triggering inflammatory responses and osteoclastic differentiation (Hans and Hans, 2011; Souza and Lerner, 2013). This effect can be inhibited by exogenously added antioxidants. For example, when intracellular ROS were scavenged during receptor activator for nuclear factor-κB ligand (RANKL)-stimulated osteoclastogenesis, the RANKL-induced activation of NF-κB was abrogated (Nikhil et al., 2015). Furthermore, Thummuri et al. demonstrated that thymoquinone, an antioxidant, could inhibit inflammation-induced ROS generation and the activation of NF-κB in osteoclast precursors (Thummuri et al., 2015).

ROS can also trigger JNK signaling during periodontitis (Wang et al., 2015; Lee et al., 2016). A recent study showed that ROS induced the activation of JNK signaling, which disrupted the periodontal junctional epithelium through the dissociation of E-cadherin (Wang et al., 2015; Lee et al., 2016). Consistent with this result, nicotine-induced ROS generation induced JNK phosphorylation in human gingival fibroblasts (HGFs). Furthermore, constitutive activation of JNK initiated the apoptosis cascade via the caspase-3-dependent pathway (Kang et al., 2011). In contrast to the pro-apoptotic function of JNK in HGFs, another study demonstrated an anti-apoptotic role of JNK in response to bacterial invasion (Wang et al., 2015). This study reported that JNK activation could induce the expression of genes that counter oxidative stress (Cat, Sod2, Prdx3) and apoptosis (Bcl-6) via the activation of the transcription factor forkhead box protein O1 (FoxO1) (Wang et al., 2015). Collectively, these results suggest that the activation of JNK in periodontal cells mediates cell survival, and this function may be condition and cell type dependent.

Another mechanism by which ROS are involved in periodontal pathogenesis is via the activation of inflammasomes. As previously reported, ROS induced the activation of NLRP3 by causing thioredoxin (TRX)-interacting protein (TXNIP) to dissociate from thioredoxin, which may be associated with periodontitis (Schroder et al., 2010; Zhou et al., 2010). Increasing clinical data support this point. Bostanci et al. first reported significantly high levels of NLRP3 as well as enhanced release of pro-inflammatory cytokines (IL-1β and IL-18) in patients with periodontitis compared with healthy controls (Bostanci et al., 2009). Elevated IL-1β and IL-18 levels could contribute to the triggering of periodontal destruction. Consistent with this study, Xue et al. and Huang et al. demonstrated increased levels of NLRP3 in the gingival tissues of periodontitis patients compared with healthy individuals via real-time PCR and immunohistochemistry (Huang et al., 2015; Xue et al., 2015). The activation of NLRP3 inflammasomes via ROS can lead to IL-1β secretion and pyroptosis.

Emerging evidence has indicated that nuclear factor erythroid 2-related factor 2 (Nrf2) plays an important cytoprotective role in oxidative-stress-associated periodontal damage. As a redox-sensitive factor, Nrf2 protects cells against cytotoxic ROS. As illustrated in Figure 1, oxidative stress disrupts critical cysteine residues in Kelch like-ECH-associated protein 1 (Keap1). When Nrf2 is not ubiquitinated, it dissociates from Keap1, translocates into the nucleus and binds to antioxidant response elements (AREs) to initiate the transcription of antioxidant genes, such as heme oxygenase-1 (HO-1), SOD, and CAT (Ma, 2013). Nrf2 knockout exacerbated the loss of periodontal tissues in a mouse model of periodontitis. In addition, an analysis of oral polymorphonuclear neutrophils (oPMNs) and blood PMNs revealed that Nrf2 expression was significantly decreased in patients with severe chronic periodontitis compared with periodontally healthy controls (Sima et al., 2016). Moreover, compared with blood PMNs, thirty Nrf2 pathway-related genes were differentially expressed in oPMNs from chronic periodontitis patients (Sima et al., 2016). These results indicated that Nrf2 and its downstream genes may be involved in the pathological process of periodontitis via their antioxidative effects.

Sirtuins (silent information regulator, Sir2) belong to a conserved family of nicotinamide adenine dinucleotide (NAD)-dependent protein deacylases. There are seven human Sir2 homologs, SIRT1 through SIRT7 (Chalkiadaki and Guarente, 2015). Increasing evidence has indicated that SIRT activation suppresses oxidative stress (Kumar et al., 2017), and the expression of SIRT1 was shown to be increased at the gene and protein levels in LPS-treated human periodontal ligament cells (Park et al., 2012). However, the direct linkage between SIRT and oxidative stress in periodontitis has not been elucidated.

Abundant evidence has demonstrated that periodontitis is highly associated with microbial infection. As an intracellular innate immune defense pathway, autophagy is usually enhanced in infected cells, contributing to antimicrobial defense mechanisms. Autophagy can eliminate intracellular pathogens such as Mycobacterium tuberculosis (M. tuberculosis) (Kim et al., 2012). To avoid lysosomal killing, many pathogens, including Legionella pneumophila (L. pneumophila), have developed strategies to suppress cellular autophagy. However, findings obtained via in vitro experiments in which cultured cells were exposed to bacteria have suggested that periodontal pathogens such as P. gingivalis participate in the induction of autophagy. Belanger et al. found that P. gingivalis trafficked quickly from phagosomes to autophagosomes in human coronary artery endothelial cells (Belanger et al., 2006). This result is consistent with the finding that ROS generated by P. gingivalis contribute to increased levels of LC3 proteins and promoting the conversion of LC3-I to LC3-II (Park et al., 2017). Taken together, these results strongly suggest that the induction of autophagy can facilitate specific periodontal bacterial survival by replication within an autophagosome-like compartment. However, no LC3 lipidation was found when cells were infected with A. actinomycetemcomitans (Blasi et al., 2016), suggesting that the activity of cellular autophagy in response to infection is associated with periodontal bacterial species.

The relationship between autophagy and immunity has been systematically reviewed in the published literature (Levine et al., 2011; Deretic et al., 2013; Shibutani et al., 2015). Here, we mainly focus on the potential immune consequences of autophagy for periodontitis. Autophagy functions as a modulator of classical pattern recognition receptors (PRRs), such as Toll-like receptors (TLRs), and Nod-like receptors (NLRs), regulating the periodontal innate immune response (Deretic et al., 2013; Oh and Lee, 2014). Furthermore, autophagy can suppress the periodontal immune response by inhibiting cytokine secretion. First, autophagy plays a negative role in inflammasome activation and secretion of IL-1β and IL-18. Animal studies have shown that mice lacking LC3B produced higher levels of caspase-1-dependent cytokines than wild-type mice. Similar results were found in LC3B-deficient macrophages (Nakahira et al., 2011). LC3B is a ubiquitin-like protein that participates in autophagosome formation and maturation (Anton et al., 2016). Second, autophagy negatively regulates the secretion of IL-1α. Castillo et al. found that mice lacking Atg5 produced more IL-1α via a ROS-calpain pro-inflammatory pathway (Castillo et al., 2012). It therefore seems reasonable to consider that autophagy might influence periodontal inflammation by regulating both inflammasome-dependent and inflammasome-independent inflammation.

Studies have shown that the inhibition of autophagy in HGFs treated with P. gingivalis LPS induced apoptosis, suggesting a protective role of autophagy (Bullon et al., 2012). To further explore the role of autophagy in periodontitis, a recent study measured the expression of LC3 and observed autophagic vacuoles in periodontal ligament (PDL) tissues from individuals with and without periodontitis. The results showed increased LC3 expression and autophagosome production in inflammatory PDL tissues (An et al., 2016). In addition, co-localization of LC3 and melanoregulin (MREG) was found in gingival epithelial cells isolated from severe periodontal disease-affected individuals, while this effect was absent in cells from healthy or moderately affected individuals (Blasi et al., 2016). As multiple studies have proposed that autophagy may antagonize apoptosis, these results suggest a potential protective role of autophagy in periodontal tissues. However, whether the blocking of autophagy induces apoptosis in periodontal tissues remains unknown.

The activity of mTORC1 is regulated by numerous input signals, such as rapamycin, insulin, and oxidative stress. Studies have found that ROS could influence mTORC1 activity through the tuberous sclerosis complex 1/2 (TSC1/TSC2) heterodimer. Increased levels of ROS activate AMP-activated protein kinase (AMPK), which causes TSC2 phosphorylation and activates the TSC1/TSC2 complex, thus inhibiting mTORC1 and stimulating ULK (an important initiator of the autophagy complex) to induce autophagy (Yu et al., 2010; Zhang et al., 2013; Zhang J. et al., 2015). Conversely, ROS can activate the phosphoinositide-3-kinase (PI3K)-protein kinase B (Akt)-mTORC1 signaling pathway by directly activating PI3K or by regulating the phosphorylation state of Akt, thus inhibiting autophagy induction (Dermit et al., 2016; Su et al., 2017). Stafford et al. found that invasion of P. gingivalis inhibited the mTOR pathway in oral epithelial cells, which was the first reported evidence to suggest a potential role for mTORC1 in periodontitis (Stafford et al., 2013).

As noted above, increased levels of ROS can activate NF-κB, which may result in the upregulation of Atgs, including Beclin 1 (He Z. J. et al., 2017). Moreover, the activation of JNK signaling during oxidative stress leads to the phosphorylation of Bcl-2, which causes Beclin 1 to dissociate from the Vps34 complex and results in the activation of autophagy (Ni et al., 2014). Several studies have shown the relevance of these Atgs in periodontitis. Specifically, there were higher protein expression levels of LC3II/I and Beclin 1, as well as increased transcriptional levels of LC3, Beclin-1, Atg7, and Atg12, in periodontal ligament stem cells isolated from patients with periodontitis compared with healthy individuals (An et al., 2016).

The Atg12-Atg5 conjugate is a ubiquitin-like protein complex that is essential for autophagophore elongation in autophagy (Otomo et al., 2013). A number of studies have provided evidence of the fine-tuning of Atg12-Atg5 in relation to the intracellular redox state (Mai et al., 2012). Pei et al. reported that the levels of Atg12-Atg5 were upregulated in a preodontoblast cell line (mDPC6T cells) after treatment with LPS for 6 h and 12 h but were downregulated after treatment with LPS for 24 h (Pei et al., 2015).

These findings show that autophagy can be induced in response to ROS through two master regulators of autophagosome biogenesis (mTORC1 and Beclin-1) and the Atg5-Atg12 complex, which also plays important roles in autophagosome biogenesis (Figure 2).

Autophagy is also crucial in mitochondrial ROS generation and scavenging, which is predominantly achieved by the release and activation of Nrf2 (Komatsu et al., 2010). Emerging evidence has indicated that Nrf2 and its target genes are crucial for maintaining cellular redox homeostasis in the attenuation of oxidative stress-associated periodontal destruction (Tamaki et al., 2014; Kataoka et al., 2016; Sima et al., 2016).

Collectively, progress in the field of redox regulation in autophagy has provided increasing details of the crosstalk mechanisms between ROS and autophagy. However, there is still no direct evidence demonstrating that the activation/inactivation of autophagy is triggered by redox regulation signaling in periodontitis. Hence, the precise process in periodontal tissues still needs to be elucidated. Whether ROS is an upstream signal of autophagy in periodontitis also requires further investigation.

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