Autophagy can protect host cells from apoptosis when periodontal tissues are subjected to P. gingivalis and its virulence factors. The possible involved mechanism is that the activation of autophagy may suppress apoptosis. For example, compared with healthy people, the expression levels of ATGs such as LC3, Beclin-1, Atg7, and Atg12 are higher in periodontal ligament tissue cells (PDLSCs) of periodontitis patients. However, the apoptotic protein caspase-8 expresses less (An et al., 2016). Coincidentally, another research found that the expression of Atg12 and the conversion of LC3-I to L3-II were increased in human gingival fibroblasts (HGFs) treated with P. gingivalis LPS (Bullon et al., 2012). Pro-inflammatory cytokines, such as TNF-α and IL-1β secretion could be limited by P. gingivalis-induced autophagy (Liu et al., 2018). Inhibiting autophagy with 3-Methyladenine (3-MA) increased the apoptosis rate in infected cells (Bullon et al., 2012), which indicated that autophagy might protect host cells from apoptosis in inflammation (Bullon et al., 2012; Liu et al., 2018).

Besides, autophagy can also reduce the secretion or increases the degradation of inflammatory factors. Chang et al. (Chang et al., 2013) found that butyrate, a kind of metabolic production of P. gingivalis, could impair cell cycle progression and induce inflammation via the generation of ROS in HGFs. However, P. gingivalis LPS treatment could also induce an autophagic response in human cultured keratinocyte cells (HaCaT) via increased ROS (Hagio-Izaki et al., 2018). The expression of LPS binding protein and TLRs increased, resulting in the co-localization of P. gingivalis with autophagosomes (Hagio-Izaki et al., 2018). Antibacterial peptide LL-37 could reduce the number of live P. gingivalis in HaCaT in a dose-dependent manner via xenophagy. Treatment with 3-MA weakened the inhibitory effect of LL-37 on the intracellular P. gingivalis (Yang et al., 2020). The findings indicated that autophagy was beneficial to the defense against the P. gingivalis internalized in HaCaT. Furthermore, inhibiting autophagy could enhance inflammation. Huang et al. (Huang et al., 2020) found high glucose environment could disrupt the autophagy lysosomal pathway (ALP) via a special subunit on the ATPase transmembrane (ATP6V0C), consequently enhancing the expression of IL-1β in HGECs.

Moreover, autophagy also plays a protective role by modulating the proangiogenesis ability to protect cells under the inflammatory environment. Wei et al. (Wei et al., 2018) found that proangiogenesis cytokine, angiogenin (Ang) expression is higher in inflammatory human periodontal ligament stem tissues (PDLSCs) and more tube formation was observed in endothelial cells pretreated with TNF-α and IL-1β. Rapamycin or Beclin-1 overexpression-induced autophagy can upregulate the expression level of Ang in PDLSCs, while knockdown of Beclin-1 downregulates its expression.

Consequently, when periodontal tissue cells are stimulated by P. gingivalis or its virulence factors, autophagy plays a protective role in fighting against the pathogens and decreasing the inflammatory injury. The involved mechanism may be that autophagy can decrease the intracellular P. gingivalis, reduce the secretion of inflammatory factors, inhibit host cell apoptosis death, and promote angiogenesis ability.

As an important indicator for the diagnosis of periodontitis, alveolar bone resorption is largely irreversible (Armitage, 2004). P. gingivalis infection could accelerate the alveolar bone pathological resorption by ROS generation and autophagic activity (Waddington et al., 2000; Tokutomi et al., 2015; Huck et al., 2020). Autophagy plays a dual role in osteoclast differentiation and bone resorption during periodontitis. DeSelm et al. (DeSelm et al., 2011) elucidated that osteoclasts degrade bone by directionally secreting lysosomal enzymes to a complex structure named ruffled border. The ruffled border is the product of the fusion of secretory vesicles to the bone-apposed plasma membrane. Lysosomal fusion with the plasmalemma results in the release of matrix-degrading molecules such as cathepsin K into the extracellular space to digest the bone matrix (DeSelm et al., 2011). However, PE-conjugated LC3 system and ubiquitin systems are essential for generating the osteoclast ruffled border, the secretory function of osteoclasts, and bone resorption (DeSelm et al., 2011). The research found that Syndecan 4 (SDC4), a member of the syndecan family, was highly expressed in the experimental periodontitis rat model (Li J. et al., 2021). While SDC4 silencing inhibits osteoclast differentiation and SDC4 overexpression enhanced osteoclast differentiation and autophagy induced by receptor activator of NF-κB ligand (RANKL). However, inhibiting autophagy with 3-MA abolished osteoclast differentiation which was enhanced by SDC4, indicating that autophagy might promote the osteoclast differentiation during periodontitis (Li J. et al., 2021).

On the other hand, autophagy also plays a negative role in osteoclast differentiation. Song et al. (Song et al., 2019) found that osteoclast differentiation and mRNA expression of osteoclast−specific genes were enhanced in bone marrow macrophages stimulated by IL-17A. Inhibition of autophagy with 3−MA attenuated the IL−17A−induced osteoclast differentiation. Tang et al. (Tong et al., 2018) found that autophagy plays an essential role in the process via AMPK/mTOR/p70S6K signaling pathway. Inhibiting autophagy with chloroquine (CQ) attenuated the OPG’s inhibition of osteoclast differentiation and bone resorption. Besides, autophagy also plays an important role in orthodontic tooth movement (Li Y. et al., 2021). Autophagy activity enhanced on the compression side of the tooth, which was associated with osteoclast recruitment and inflammatory cytokine expression (Li Y. et al., 2021). Treatment with rapamycin, an autophagy promotor, resulted in less bone resorption and tooth movement, suggesting that autophagy downregulated the inflammatory response and osteoclast differentiation during orthodontic tooth movement (Li Y. et al., 2021). To sum up, regulating autophagy activity may be an effective way to protect alveolar bone from excessive resorption in periodontitis.

P. gingivalis can invade antigen-presenting cells (APCs) such as macrophage and dendritic cells (DCs) and elevate autophagy activity. Autophagy can also inhibit inflammation by the elimination of P. gingivalis and degradation of inflammatory factors in APCs.

As for macrophages, Park et al.. (Park et al., 2017) preliminarily elucidated the link among bacteria, autophagy, and inflammation in the model of P. gingivalis infection in THP-1 cells. The infection promoted the secretion of mature IL-1β through activation of NLRP3 inflammasomes and downstream caspase-1. Similarly, Aggregatibacter actinomycetemcomitans (A. actinomycetemcomitans), another important pathogen of periodontitis, can also promote IL-1β maturation through the activation of AIM2 (absent in melanoma 2) inflammasome (Lee et al., 2020). The induction of AIM2 or NLRP3 inflammasomes could trigger autophagosome formation, which is dependent on the presence of the inflammasome sensor (Shi et al., 2012). The inflammasomes could be ubiquitinated and then recruit the ubiquitin-binding protein adaptor p62. With its assistance, the inflammasomes were delivered into autolysosome for degradation (Shi et al., 2012). Besides, both P. gingivalis and A. actinomycetemcomitans-induced autophagy in THP-1 cells inhibited the production of intracellular IL-1β and ROS by promoting bacterial internalization via phagocytosis into the macrophages. This therefore restricted excessive inflammatory response (Park et al., 2017; Lee et al., 2020). Besides, recent studies found two compounds, lipoxin A4 and vitamin D could promote the autophagic activity upon P. gingivalis infection in macrophages (Niu et al., 2020; Zhao et al., 2021). Lipoxin A4 inhibited NLRP3 inflammasome and downstream caspase-1 and IL-1β in P. gingivalis LPS-infected RAW264.7 by promoting autophagy. This action was related to the phosphorylation of NF-κB (Zhao et al., 2021). Calcitriol could significantly enhance the colocalization of P. gingivalis with autophagosome and lysosome markers in U937-derived macrophages, which indicated that the autophagy process could degrade live P. gingivalis (Niu et al., 2020). In summary, the above studies elucidated the role of autophagy in suppressing P. gingivalis-induced inflammation in macrophages. On the one hand, autophagy inhibits the production and promotes the degradation of AIM2 and NLRP3 inflammasomes. On the other hand, autophagy promotes bacterial internalization through phagocytosis and limits excessive inflammatory responses.

The autophagic activity in DCs was different from macrophages upon infection. P. gingivalis recognizes different PRRs on the surface or in the cytoplasm of DCs, which can trigger different responses (Xie et al., 1997; den Dunnen et al., 2009). The PRRs included TLRs and C-type lectin receptors such as DC-SIGN. The activation of TLR2 by FimA fimbriae could inhibit Akt phosphorylation, whereas the activation of DC-SIGN by Mfa-1 fimbriae could promote Akt phosphorylation (El-Awady et al., 2015; Meghil et al., 2019). Phosphorylated Akt could phosphorylate and activate mTORC1, ultimately inhibiting its downstream effector ULK1 to decrease autophagic activity. The interaction between DC-SIGN and Mfa-1 P. gingivalis strains induced lower intracellular killing and higher intracellular content of P. gingivalis. Moreover, Mfa-1 P. gingivalis was majorly wrapped in single-membrane vesicles, where it survived intracellularly (El-Awady et al., 2015). The finding indicated that Mfa-1 P. gingivalis engaged with DC-SIGN promoted the evasion of xenophagy. However, the activation of TLR2 cannot escape the autophagic attack. In addition, transcription factor forkhead box O1 (FOXO1) was phosphorylated and inactivated by Akt in DCs infected with Mfa-1 positive strains Pg381, causing FOXO1 to relocate to the cytoplasm (Meghil et al., 2019). FOXO1 deletion could inhibit the migration of DCs to lymph nodes and reduce the capacity of DCs to promote the formation of plasma cells and decrease the production of bacteria-specific antibodies (Xiao et al., 2015). Thus, the inactivation of FOXO1 promotes the survival of intracellular P. gingivalis.

The invasion of oral P. gingivalis into blood circulation is related to several systemic inflammatory diseases. For example, P. gingivalis-induced periodontitis could increase the risk of atherosclerosis, and P. gingivalis could also be detected in human atherosclerotic plaques (Kozarov et al., 2005). Research found that P. gingivalis could induce human umbilical vein endothelial cells (HUVECs) ER stress-mediated apoptosis and autophagy. The autophagic response protects HUVEC from apoptosis and silencing of LC3 with siRNA significantly increased P. gingivalis-induced apoptosis (Hirasawa and Kurita-Ochiai, 2018).

Infection with P. gingivalis can also accelerate the development and progression of non-alcoholic fatty liver disease (NAFLD) (Yoneda et al., 2012), a disease characterized by intracellular accumulation of lipid droplets in hepatocytes. Zaitsu et al. (Zaitsu et al., 2016). found that internalized P. gingivalis was localized in autophagosomes or lysosomes rather than lipid droplets in HepG2 hepatocytes, indicating that P. gingivalis could not use the lipid droplets for nutritional purposes. However, lipid droplets could affect the formation of autolysosomes for degradation of P. gingivalis and increase the existence of P. gingivalis in the cells at an early phase of infection. Thus, delayed elimination of P. gingivalis prolonged the inflammation and cellular damage, which raise the risk for the development of NAFLD (Zaitsu et al., 2016).

P. gingivalis can escape from autophagy to avoid degradation by hindering the combination of autophagosomes with lysosomes. In this situation, P. gingivalis could colonize and proliferate in a vacuole encapsulated by a monolayer membrane after its invasion. The autophagosome with a monolayer membrane is modified by P. gingivalis and cannot fuse with lysosomes to form autolysosomes, exerting its degradation effect. Instead of the lysosomal marker protein LAMP-1, only rough ER protein attaches to the surface around the vacuole (Dorn et al., 2001). Lee et al. (Lee et al., 2018). found P. gingivalis adhered to the surface of GECs and was trafficked into the ER-rich autophagosomes through lipid rafts. After initially localized in early endosomes in GECs, some P. gingival entered the late endosomes where they were then sorted to lysosomes for degradation, some were delivered to Rab11 and RalA positive recycling endosomes for exocytosis while others escaped from endosomes to autophagosomes (Takeuchi et al., 2011; Amano et al., 2014). Hiding in autophagosomes, P gingivalis could utilize the substances such as protein for its survival. Inhibition of autophagy by 3-MA or Atg5 siRNA significantly decreased the survival of internalized P. gingivalis in HGECs (Belanger et al., 2006; Lee et al., 2018). P. gingivalis could colonize and replicate in GECs by inhibiting pro-apoptotic Bad through Akt to facilitate its long-term survival (Yao et al., 2010).

Epidemiological studies have also linked PD to age-related macular degeneration (AMD) (Winning and Linden, 2017). Recent researches reported that P. gingivalis invasion and autophagy evasion might contribute to the pathogenesis of retinal degenerative diseases (Arjunan et al., 2020). This study showed that P. gingivalis could invade the human-retinal pigment epithelial (ARPE) cells and was able to escape from the autophagic vesicles and traffic into single membrane structures, freely occupying the cytoplasm of ARPE cells (Arjunan et al., 2020). The underlying mechanism may be associated with MREG-dependent LC3-associated phagocytosis (LAP). Lysosome maturation required the binding of LC3 to melanoregulin (MREG), and the MREG-mediated LC3-dependent degradation pathway was an essential process for P. gingivalis clearance (Martinez et al., 2011; Frost et al., 2015). Frost et al. (Frost et al., 2015) observed the co‐localization of LC3 and MREG in GECs of patients with severe periodontitis, but not in mild periodontitis or healthy people. The result showed that the expression of LC3, Beclin-1, and MREG in periodontitis was less than that in healthy people. It is speculated that P. gingivalis may affect the expression of MREG and LC3, and inhibit the normal autophagic process in HGECs of periodontitis patients.

As mentioned above, P. gingivalis expressed two forms of fimbriae and could interact with different PRRs on the host cell, affecting the fate of engulfed bacteria. Both long and short fimbriae were critical for the initiation of inflammatory responses and alveolar bone loss in periodontitis (Umemoto and Hamada, 2003). However, the interaction between Mfa-1 positive strains Pg381 with DC-SIGN could promote Akt phosphorylation and decrease the expression of LC3-II, Rab5, and LAMP1, thus inducing lower intracellular killing of P. gingivalis in DCs (El-Awady et al., 2015). However, the activation of TLR2 by FimA positive strains cannot escape from autophagic attacks. A similar phenomenon was found in Pg381 strain-infected human coronary artery endothelial (HCAE) cells. The survival rate of intracellular P. gingivalis decreased when HCAE cells were pretreated with 3-MA or wortmannin (Dorn et al., 2001). However, the pathogenic mechanisms varied in different P. gingivalis strains (Rodrigues et al., 2012; Blasi et al., 2016). (Rodrigues et al., 2012). compared different P. gingivalis strains (PgW83, PgA7436, Pg381, and Pg33277) and their abilities of adherence, internalization, and survival rate in HCAE cells. Both PgW83 and Pg381 were trafficked through the autophagic pathway, but only Pg381 could survive, depending on the autophagic pathway. Furthermore, Pg381 strain infected-HACE cells resulted in the highest production of pro-inflammatory cytokines IL-6, IL-8, and TNF-a.

NLRP3 inflammasome is a multiprotein complex that contains adapter protein apoptosis-associated speck-like protein (ASC) and procaspase-1. Its biochemical function is to activate caspase-1, which leads to the maturation of IL-1β and IL-18 (Shao et al., 2015; He et al., 2016a). Recent studies found a serine/threonine kinase named NEK7, which is bound to the leucine-rich repeat domain of NLRP3. NEK7 is also an essential mediator of NLRP3 activation (He et al., 2016b; Shi et al., 2016).

The NLRP3 inflammasome pathway activation involves two steps. In the initiating step, TLRs sense the invading P. gingivalis outside the cell and in intracellular endosomes and lysosomes (Medzhitov, 2001; Zhou et al., 2005; Zhang J. et al., 2018), leading to activation of NF-κB-mediated signaling, which in turn induces the transcription of NLRP3, pro-IL-1β (Bauernfeind et al., 2009; Hornung and Latz, 2010). The second step of inflammasome activation is the assembly of NLRP3, ASC, and pro-caspase-1 into a complex (Petrilli et al., 2007; Shao et al., 2015; Jo et al., 2016). P. gingivalis LPS also engages in the non-canonical inflammasome pathway (Ding et al., 2020; De Andrade et al., 2021), in which caspase-4 (known as caspase-11) is activated, promoting pyroptosis through cleavage of the pore-forming protein gasdermin D, resulting in maturation and release of IL-1β (Kayagaki et al., 2011; Hagar et al., 2013; Kayagaki et al., 2013). Different from the canonical NLRP3 inflammasome pathway, TLR4 senses the extracellular P. gingivalis LPS while caspase-4/11 senses the cytosolic LPS (Kayagaki et al., 2013; Downs et al., 2020; Xie et al., 2020)

The molecular mechanisms involved in NLRP3 inflammasome activation include mitochondrial ROS generation (Sorbara and Girardin, 2011; Heid et al., 2013), lysosomal rupture (Heid et al., 2013), pore formation, and potassium (K+) efflux (Petrilli et al., 2007). The P2X7 receptor is an ion channel activated by extracellular adenosine triphosphate (eATP) that induces IL-1β secretion through NLRP3 inflammasome activation (Park et al., 2014; Di Virgilio et al., 2017; Almeida-da-Silva et al., 2019).

Recent research reported that autophagy could suppress the activation of NLRP3 inflammasome. Enhancing autophagy with rapamycin reduced the expression of ASC as well as caspase-1 p20, and decreased the secretion of IL-1β, which in turn attenuated myocardial ischemia-reperfusion injury in diabetic rats (Zhang et al., 2020). Besides, Atg16L1-deficient macrophages produce high amounts of cleaved caspase-1 and IL-1β following stimulation of LPS (Saitoh et al., 2008). However, the mechanism of how autophagy suppressed the activation and assembly of inflammasomes upon P. gingivalis infection is intricate. It is supposed that this may be linked to clearing pathogens, degrading inflammasome components, and interrupting the molecular mechanism (Cao et al., 2019).

Xenophagy effectively limits intracellular bacterial survival and replication. Cytoplasmic bacteria, such as Salmonella. Typhimurium (S. Typhimurium), can be enveloped by ubiquitin proteins and recognized by autophagy receptors, such as p62, NDP52, OPTN, and TAX1BP1. With these receptors’ help, cytoplasmic bacteria are targeted to autophagosome (Thurston et al., 2009; Wild and Farhan, 2011; Tumbarello et al., 2015; Lee et al., 2022). In addition to ubiquitin, galectin-8, a cytosolic lectin, detects S. Typhimurium invasion by binding host glycans after being exposed to damaged Salmonella-containing vacuole (SCV). Galectin-8 can recruit NDP52 transiently and then NDP52 are recruited in a ubiquitin-dependent manner to activate xenophagy (Thurston et al., 2012). Although xenophagy machinery of targeting bacteria for degradation is well established, the ubiquitination mechanism and how autophagy receptor senses bacterial, especially P. gingivalis, are not fully clear. Only p62 and NDP52 receptors are demonstrated to bind to cytosolic P. gingivalis so far (Lee et al., 2018), and whether OPTN and TAXIBP1 or other ubiquitin adaptor proteins could target P. gingivalis to autophagosome is not clear. However, infecting P. gingivalis can be wrapped in ER-rich-double-membrane autophagosomal-vacuoles for survival. Inhibiting autophagy with 3-MA or ATG5 siRNA significantly reduced the viability of intracellular P. gingivalis. Besides, p62 and NDP52 receptors only bind to cytosolic free P. gingivalis while the vacuolar-P. gingivalis, which reveals a novel mechanism of P. gingivalis survival from xenophagy (Lee et al., 2018).

The activation of NLRP3 inflammasome and its components could be inhibited by autophagy. During the priming step of NLRP3 inflammasome activation, autophagy can inhibit NF-κB activation and the expression of pro-IL-1β (Wu et al., 2020), but the mechanism is not clear. During the process of inflammasomes’ assembly, the ubiquitination or phosphorylation modification of components can trigger selective autophagy. Spalinger et al. (Spalinger et al., 2017) found that only phosphorylated NLRP3 is found in autophagosomes and NLRP3 lacking phosphorylation sites cannot be recruited to phagophores upon inflammasome activation. Phosphorylated NLRP3 binds to p62 and is recruited into phagophores for degradation, which is dependent on NLRP3-PYCARD interaction (Spalinger et al., 2017). NLRP3 does not interact with SQSTM1in PYCARD-deficient cells, indicating that PYCARD may be the tag recognized by p62 for selective autophagy. The specific interaction and mechanism are worth further exploration. In addition to NLRP3, ASC under ubiquitination can recruit p62. With its assistance, ASC is delivered to autophagosomes for degradation in THP-1 treated with LPS and ATP (Shi et al., 2012). Besides to ubiquitination or phosphorylation modification-mediated selective autophagy, TRIMs family protein-mediated a highly exact process termed precision autophagy also plays an important role in the inactivation of NLRP3 inflammasome (Kimura et al., 2016). TRIM family proteins functioned as receptor-regulator proteins, both recognized targets and mediated autophagy. Different from selective autophagy, TRIMs could recognize targets through a direct protein-protein binding way without ubiquitin or galectin intermediates. And TRIMs function as platforms for the assembly of the core regulators of autophagy, such as the ULK1 complex or Atg16L1 complex (Kimura et al., 2016). For example, TRIM20 directly recognizes NLRP3 inflammasome and its component procaspase-1 and presents them for autophagic degradation (Kimura et al., 2015). The receptor and regulatory features enable TRIMs to mediate the inflammasome degradation via a highly exact process termed precision autophagy.

Autophagy also interacts with the molecular mechanisms involved in NLRP3 inflammasome activation, such as ROS clearance and P2X7 receptor. It has been demonstrated that P. gingivalis LPS-induced cytosolic ROS plays an important role in the pathogenesis of periodontitis (Chapple and Matthews, 2007; Golz et al., 2014). Mitophagy could preserve mitochondrial integrity and eliminate ROS through the degrading of damaged mitochondria, thereby reducing NLRP3 inflammasome activation (Lin et al., 2019). P2X7 receptor expression is enhanced in the P. gingivalis infection model, which is essential for NLRP3 activation promotes. The activation of the P2X7 receptor could promote P. gingivalis elimination in macrophages (Almeida-da-Silva et al., 2019). The possible mechanism may be the activation of P2RX7 can activate the AMPK/ULK1 pathway to promote autophagy flux for degradation of intracellular pathogens (Dong et al., 2021).