Autophagy in the pathogenesis and therapeutic potential of post-traumatic osteoarthritis

Abstract Autophagy, as a fundamental mechanism for cellular homeostasis, is generally involved in the occurrence and progression of various diseases. Osteoarthritis (OA) is the most common musculoskeletal disease that often leads to pain, disability and economic loss in patients. Post-traumatic OA (PTOA) is a subtype of OA, accounting for >12% of the overall burden of OA. PTOA is often caused by joint injuries including anterior cruciate ligament rupture, meniscus tear and intra-articular fracture. Although a variety of methods have been developed to treat acute joint injury, the current measures have limited success in effectively reducing the incidence and delaying the progression of PTOA. Therefore, the pathogenesis and intervention strategy of PTOA need further study. In the past decade, the roles and mechanisms of autophagy in PTOA have aroused great interest in the field. It was revealed that autophagy could maintain the homeostasis of chondrocytes, reduce joint inflammatory level, prevent chondrocyte death and matrix degradation, which accordingly improved joint symptoms and delayed the progression of PTOA. Moreover, many strategies that target PTOA have been revealed to promote autophagy. In this review, we summarize the roles and mechanisms of autophagy in PTOA and the current strategies for PTOA treatment that depend on autophagy regulation, which may be beneficial for PTOA patients in the future.

summarize the roles and mechanisms of autophagy in PTOA and the current strategies for PTOA treatment that depend on autophagy regulation, which may be beneficial for PTOA patients in the future.
Key words: Autophagy, Post-traumatic osteoarthritis, mTOR, Noncoding RNAs Background Osteoarthritis (OA) is a common musculoskeletal disease with pain, disability and economic loss in patients, which is related to a variety of risk factors such as age, joint injuries, obesity, genetic factors and sex [1,2]. Post-traumatic OA (PTOA) is a subtype of OA, accounting for >12% of the overall burden of OA. PTOA is often caused by joint injuries including ligament rupture and intra-articular fracture. It was shown that people who had a knee injury were more prone to develop knee OA compared with those who did not have a knee injury. Moreover, people who suffered ligamentous and meniscal knee injuries had a significantly increased risk of Knee osteoarthritis (KOA) compared with controls. In addition, more than half of patients with fractures of the distal tibial articular surface will develop OA after the trauma [3][4][5]. Up to now, the pathogenesis of PTOA has still not been fully understood and the current clinical treatments for PTOA have been unsatisfactory [6][7][8][9]. Therefore, it is urgent to strengthen the related research to analyze and clarify PTOA pathogenesis so as to provide new strategies for PTOA therapy [10]. The pathological features of PTOA involve the whole joint tissues, including cartilage degradation, subchondral bone remodeling, osteophyte formation and synovial inflammation. Among them, articular cartilage degeneration is the core pathogenesis of PTOA [9,11,12].
Autophagy is a highly conserved metabolic degradation process by which cells recycle substrates to maintain homeostasis [13,14]. Autophagy includes macroautophagy (hereafter referred to as autophagy), microautophagy and molecular chaperone-mediated autophagy as different modes of degradation [15]. After the initiation of autophagy, autophagosomes gradually formed and wrapped the damaged organelles, macromolecular proteins and other substances, then they fused with lysosomes and finally degraded their contents by lysosomal acid hydrolase [16]. The general process of microautophagy is the non-specific or specific encapsulation of substrates by cells through the tonoplast or lysosomal membrane, followed by their degradation in lysosomes [17,18]. Chaperone-mediated autophagy is a process in which HSP70 recognizes specific substrates, then binds to LAMP-2 on the lysosomal membrane and finally transfers the substrates to lysosomes for degradation [19,20]. Autophagy is involved in many physiological and pathological processes including immunomodulation, inflammatory reaction, aging, metabolic diseases and tumors [21][22][23]. In the past decade, an increasing number of studies have shown that autophagy plays a crucial role in maintaining chondrocyte homeostasis and is a novel potential target for PTOA treatment [24]. This article summarizes the roles and mechanisms of autophagy in PTOA and the current strategies for PTOA treatment that depend of autophagy regulation.

Review
Characteristics of PTOA PTOA is commonly initiated after acute joint damage including anterior cruciate ligament (ACL) rupture, meniscus tear, shoulder dislocation, patellar dislocation, ankle instability or articular surface injuries [10,25,26]. People with experience of joint trauma have an obviously increased risk of PTOA. It was reported that nearly half of the people who suffered from significant joint injury would progress to OA at a later stage [3,25,26]. Because of the different causes of the disease, patients with PTOA are relatively younger than those with aging-related OA. Similar to aging-related OA, patients with PTOA have clinical symptoms of joint pain and motion limitation, which may present as progressive aggravation and even ultimately lead to disability. Although a variety of methods, including surgery, have been developed to treat acute joint injury, the current measures have limited effect on reducing the incidence of PTOA and improving its clinical symptoms in long-term follow-up studies.
In terms of pathological changes, PTOA is characterised by cartilage damage, subchondral bone remodeling, synovitis, meniscus injury etc., which are similar to those in agingrelated OA. As to the mechanism, the abnormality of the biomechanics and the inflammatory response are mainly involved in the pathological process of PTOA. The trauma can result in joint instability, which gradually aggravates the degeneration and injury of cartilage as well as subchondral bone remodeling by influencing the shear and contact stresses on the articular surface [27]. Excessive mechanical loading can promote cartilage degeneration and aggravate OA progression via the gremlin-1-NF-κB pathway [28]. Local delivery of an NF-κB inhibitor following joint injury could reduce chondrocyte death and influence pain-related sensitivity in a non-invasive loading model of PTOA [29]. Apart from the mechanical action, multiple biological responses, especially inflammation, greatly contribute to the pathological destruction of the joint at different stages of PTOA [30][31][32]. Once the joint is injured, an acute inflammatory reaction appears immediately in the joint cavity, which is characterized by an increased number of immune cells, production of proinflammatory factors and activation of the pro-inflammatory signaling pathway. After the acute phase, the inflammatory reaction gradually decreases, but low-level inflammation still exists in the joint cavity of most patients and is closely related to the progression of PTOA. The immune cells in synovial tissue can release pro-inflammatory factors such as interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), which inhibit the repair process of the traumatic joint and further aggravate cartilage damage [33]. Marks et al. selected patients who had ACL transection (ACLT) injury for >6 months and investigated the change in the inflammatory cytokine profile of synovial fluid. Their data revealed that the levels of IL-1β and TNF-α are higher in patients with ACL ruptures than in controls and that this change is associated with chondral damage [34]. These pro-inflammatory cytokines can activate the classical NF-κB pathway in different types of cells in the joint cavity such as chondrocytes and synovial cells [35][36][37]. The activated NF-κB resulted in an increase of inflammatory factors (IL-8, CCL5, IL-1β, IL-6, TNF-α), catabolic factors [matrix metalloproteinase 1 (MMP-1), MMP-13, a disintegrin and metalloprotease with thrombospondin motifs 4 (ADAMTS4), ADAMTS5] and angiogenic factors [vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF)], which further promotes an inflammatory response and cartilage destruction [35,38]. Targeting the NF-κB signaling pathway could be considered as a potential strategy for PTOA therapy [37,39,40].
In addition, metabolic disorders can also affect the inflammatory reaction in the articular cavity, such as obesity-associated inflammation, which further influences the progression of OA induced by aging, trauma etc. [41,42]. The up-regulation of pro-inflammatory cytokines (IL-1β, IL-6, TNF-α) as well as altered levels of adipokines (leptin and adiponectin) are characteristics of obesity-associated inflammation, which are also increased in the process of PTOA [43]. In general, different types of fatty acids have different roles in inflammatory regulation. The saturated fatty acids including arachidonic acid and its derivatives (prostaglandins and leukotrienes) present a pro-inflammatory effect, while the unsaturated fatty acids such as docosahexaenoic acid and eicosapentaenoic acid have an anti-inflammatory function [41]. The levels of fatty acids and their derivatives are often changed in serum and synovial fluid of OA patients compared to controls, which may contribute to systemic or local oxidative damage and inflammatory responses, and ultimately influence the pathological injury of the joint during the OA process. Recently, Kimmerling et al. revealed that conversion of omega-6 into omega-3 polyunsaturated fatty acids can reduce the severity of PTOA induced by destabilisation of medial meniscus (DMM), indicating that this strategy may be beneficial for PTOA in obese patients following injury [43]. In brief, the clinical symptoms and pathological changes of PTOA patients are similar to those of other types of OA, but the risk factors and pathological mechanisms have some characteristics of their own.
Animal models of PTOA A number of different animal species, e.g. mouse, rat, hamster, pig, cat, rabbit and dog, have been used for OA basic research, among which the mouse and rat are the most common [10,[44][45][46]. The existing models for OA research can be divided into four categories, including surgical model, chemical model, genetic model and naturally occurring model (Table 1). These OA models have their own advantages and disadvantages and are selected for experiments according to the different research purposes [10,45,46]. The surgically induced OA model mainly damages the stability of joints by injuring ligaments, meniscus and other structures, thus inducing the occurrence and progress of OA. This model can simulate the pathological process of PTOA in the clinic and is considered as the best model for studying PTOA [44,45]. In addition, the OA phenotype can be rapidly induced in this model compared with the naturally occurring model. Moreover, this OA model is relatively stable with a high repetition rate. Therefore, the surgically induced OA model, especially the DMM-induced model, has been widely used for studying other types of OA in addition to PTOA. To some extent, the surgical model is used as a general model for OA research at present [44,47]. In this review we mainly included the relevant literature using the surgical model, such as DMM, ACLT, etc.
Potential role of autophagy in the pathogenesis of PTOA Autophagy is precisely orchestrated by different autophagyrelated genes (ATGs) and mainly includes six key steps ( Figure 1). Firstly, the activation of autophagy depends on inhibition of the activity of mammalian serine/threonine kinase target of rapamycin complex 1 (mTORC1) by AMPK under starvation. Inhibition of mTORC1 results in the dephosphorylation of AGT13 and unc-51-like autophagy activating kinase 1/2 (ULK1/2), and further activating the ULK1/2 complex that consists of AGT13, ULK1/2 and RB1 inducible coiled-coil 1 [48][49][50]. Subsequently, the activation of the ULK1/2 complex makes the cells progress to the stage of vesicle nucleation of autophagy, beginning with the activation of the class III phosphatidylinositol 3-kinase complex that is composed of Beclin1 (BECN1), phosphoinositide-3-kinase regulatory subunit 4 (PIK3R4), phosphatidylinositol 3-kinase catalytic subunit type 3 (PIK3C3) and ATG14/ATG14L proteins [49,51]. Then, cells come to the stage of vesicle expansion of autophagy, where two ubiquitin-like ligation systems are recruited to conjugate the membrane. One is the ATG5-ATG12 complex formed under the regulation of E1-like enzymes ATG7 and E2-like enzymes ATG10, while the other is the ATG4, ATG7 and ATG3 complex, which transforms light chain 3 (LC3) precursor into  [55]. In addition, Zhang et al. found that the expression of autophagy-associated genes is upregulated in animal OA models at the OA early stage compared with that at the late stage, indicating that autophagy is gradually inhibited with the progression of OA [56]. Similarly, the ratio of LC3-II to its free form LC3-I was significantly decreased with the progression of PTOA in an experimental rabbit OA model [57]. Wu et al. found that ULK1, one of the autophagy-related genes, was significantly down-regulated in the cartilage of OA patients compared with healthy people [58]. In primary chondrocytes derived from OA patients, ATG5 expression was significantly reduced [59]. In addition, the levels of autophagy-related genes ATG3, ATG4a, ATG5, ATG7 and ATG12 gradually decreased with the progression of OA disease in a rat OA model caused by weight-bearing [56,60,61]. In a rabbit cartilage injury model, the levels of ATG3 and ATG7 were negatively correlated with a cartilage injury [62]. In a monosodium iodoacetate (MIA)-induced cartilage injury model, ATG5 and ATG7 expression were also revealed to be significantly down-regulated [63]. Similarly, in an IL-18 and IL-1-induced chondrocyte injury model, the ATG5 and ATG7 expression level of chondrocytes was also significantly reduced [64,65]. Furthermore, Bouderlique et al. employed mice with ATG5-specific knockout in chondrocytes to construct an aging OA model and observed that the ATG5 deficiency led to an increase of cell apoptosis, which in turn aggravated the symptoms of aging OA mice. However, loss of ATG5 in chondrocytes did not regulate PTOA progression in a mouse DMM-induced PTOA model [66]. In addition, a few studies suggested that ATG5 and ATG7 negatively regulated cell death of chondrocyte [67,68]. Furthermore, autophagy can inhibit the expression of chondrocyte catabolism genes and promote the expression of cartilage anabolism genes [69]. Activation of autophagy down-regulated inflammatory catabolic genes such as MMP-3 and -9, ADAMTS5 and CCL-1, -2 and -5 via inhibiting the NF-kB signaling pathway in chondrocytes, which may attenuate the destruction of cartilage [70]. Recently, Zhu et al. investigated the effects of meniscal autophagy on the pathogenesis of PTOA. They found that meniscus injury happened prior to the degeneration of articular cartilage in rats after ACLT. In addition, the secretion from meniscus cells can decrease the levels of MMP13 and ADAMTS5 in IL-1β-treated chondrocytes, indicating that activation of autophagy in meniscal cells may be a potential strategy to delay PTOA progression [71]. Collectively, the key genes regulating autophagy initiation and elongation in chondrocytes are partially inhibited during spontaneous OA or PTOA processes, which negatively influences the homeostasis of chondrocytes and results in cartilage damage via regulation of cell death, cartilage matrix synthesis and inflammatory signaling. However, as the mechanisms of autophagy inhibition and the potential effects of autophagy on cartilage in spontaneous OA and PTOA may be different, further studies on this issue are required in the future.
Regulation mechanisms of autophagy in PTOA mTOR-dependent regulation of autophagy in PTOA mTOR is a class of macromolecules that mainly participates in the formation of two different complexes, mTORC1 and mTORC2. Of the two, mTORC1 is a well-known target of autophagy inhibition [72][73][74]. Moreover, upstream signaling pathways such as PI3K/AKT, AMPK and MAPK are regulators of mTOR. The PI3K/AKT pathway in articular cartilage tissue from OA patients was significantly inactivated compared with healthy people [75]. Consistently, the PI3K/AKT pathway was also suppressed in animal PTOA models or injury chondrocytes in vitro, which was related to the decrease in ECM synthesis. However, inhibition of mTORC1 by PI3K/AKT can also enhance autophagic flux, which could be beneficial for cartilage homeostasis, suggesting that the PI3K/AKT/mTOR signaling pathway was double-edged for cartilage degeneration and OA progression [76]. Furthermore, Zhang et al. employed mice with mTOR specifically knocked out in chondrocytes and uncovered that mTOR deficiency enhanced autophagy activity by mediating ULK1/AMPK activation, thereby improving articular cartilage degeneration, apoptosis and synovial fibrosis in a DMM-induced PTOA model [77]. Carames et al. revealed that the mTOR inhibitor rapamycin can induce autophagy, decrease ADAMTS5 and IL-1β levels of chondrocytes and reduced the severity of experimental OA in DMM mice [78]. Besides, Ribeiro et al. demonstrated that rapamycin treatment activated autophagy to inhibit chondrocytes' MMP-13 and IL-12 expression, maintain chondrocyte homeostasis and reduce the damage of cartilage in a DMM-combined diabetic mouse model [79]. Furthermore, activation of autophagy by rapamycin can protect chondrocytes from IL-18-induced apoptosis and improve OA symptoms in a rat PTOA model [64]. Cheng et al. used another mTOR inhibitor (Torin 1) to activate autophagy and observed a powerful protective effect of cartilage degeneration in mice after DMM operation [80]. In addition, Alvarez Garcia et al. found that the expression level of DNA damage response 1 (REDD1), as an endogenous inhibitor of mTOR, was decreased during OA, while overexpression of REDD1 activated autophagy by inhibiting mTOR signaling [81]. Furthermore, REDD1 deficiency aggravated OA symptoms in articular cartilage, meniscus, subchondral bone and synovium of a mouse PTOA model [82]. Additionally, Xue et al. found that enhancing autophagy activity by inhibition of PI3K, AKT and mTOR signaling pathways significantly reduced the inflammatory response of chondrocytes in vitro [65]. Duan et al. demonstrated that the adiponectin receptor agonist AdipoRon that targets the AMPK-mTOR signaling pathway in chondrocytes can activate autophagy and significantly reduce chondrocyte calcification [83]. Discoidin domain receptor 1 (DDR1) is a kind of transmembrane receptor that can regulate cell proliferation, differentiation and death. In an ACLT surgeryinduced mouse PTOA model, DDR1 inhibitor (7rh) reduced cartilage degradation and reduced chondrocyte apoptosis by decreasing mTOR expression and promoting autophagy [84]. 15-Lipoxygenase-1 can upregulate the expression of TGF-beta1 of osteoblasts by regulating the autophagy level in human OA, which was involved in inhibition of AMPK and the subsequent activation of mTORC1 [85]. The above studies suggest that mTOR and its related pathways are closely involved in autophagy activation of OA-related cartilage, which further affects the progression of PTOA and is a promising target for the prevention and treatment of PTOA ( Figure 2).
Noncoding RNAs related to autophagy in PTOA Noncoding RNAs are a large class of RNAs without proteincoding ability, including microRNAs (miRNAs), long noncoding RNAs (lncRNAs) and circular RNAs (circRNAs), which function in cell growth, differentiation, senescence and apoptosis. Recent studies have shown that non-coding RNAs actively participate in the onset and progression of OA by regulating chondrocyte autophagy [86][87][88] (Table 2). miRNAs miRNAs are a class of endogenous small RNAs of ∼20-24 nucleotides in length [89]. In animal OA models, Yu et al. described that the miR-206 level was upregulated in a rat OA model. Moreover, miR-206 overexpression reduced autophagy activity and delayed the cell cycle by targeting Insulin-like growth factor-1 (IGF-1), which further activated the PI3K/AKT/mTOR signaling pathway [90]. Sun  was down-regulated in the serum and joint tissues of OA patients. In addition, it was suggested that SNHG7 participated in the regulation of chondrocyte autophagy and apoptosis by adsorbing miR-34a-5p [101]. In a ACLT-induced PTOA model, the expression of lncRNA GAS5 was notably Roles and mechanisms of mTORC1 in PTOA. mTORC1 is inhibited by rapamycin or Torin 1 to activate autophagy, and a protective effect in chondrocytes or cartilage was shown. REDD1, an endogenous inhibitor of mTOR, is thought to be a protector in OA progression by promoting the activation of autophagy. AdipoRon, an adiponectin receptor agonist, targeted the AMPK-mTOR signaling pathway in chondrocytes and was able to activate autophagy and reduce chondrocyte calcification. 7RH, the inhibitor of DDR1 that is located the upstream of mTOR signaling pathway, reduced cartilage degradation and chondrocytes apoptosis by promoting autophagy. 15-Lipoxygenase-1 inhibited AMPK with subsequent activation of mTORC1 and then up-regulated the expression of TGF-beta1 in osteoblasts by regulating the autophagy level. Sirt3 sirtuin 3, APPL1 adaptor protein, phosphotyrosine interacting with PH domain and leucine zipper 1, LKB1 liver kinase B1 homolog, AMP adenosine monophosphate, ATP adenosine triphosphate, UHRF1 ubiquitin-like with PHD and ring finger domains 1, PI3K phosphatidylinositol 3-kinase, AKT serine/threonine kinase 1, mLST8 mammalian lethal with SEC13 protein 8, PRAS40 40 kDa proline-rich AKT substrate, REDD1 protein regulated in development and DNA damage response 1, DDR1 epithelial discoidin domain-containing receptor, PTOA post-traumatic osteoarthritis increased in cartilage tissue. Moreover, overexpression of GAS5 in chondrocytes significantly activated the mTOR signaling pathway, which in turn inhibited autophagy and promoted cell apoptosis [102]. The above research shows that lncRNAs, which were traditionally considered as 'junk RNA', can also participate in the maintenance of chondrocyte homeostasis by regulating its target miRNA. Up to now, the research on lncRNAs in PTOA has been relatively limited, and the specific involvement of lncRNAs in cartilage degeneration through regulating autophagy of different cells remains to be futher studied.
circRNAs circRNAs are a class of non-coding RNA molecules without a 5 end cap and a 3 end poly(A) tail and form a circular structure with covalent bonds. circRNAs are broadly present in organisms and are involved in the transcriptional and post-transcriptional regulation of gene expression [103,104]. Recently, researchers have shown that circRNAs are also involved in OA progression [105]. Sui et al. found that hsa_circ_0037658 was significantly up-regulated in cartilage tissue from OA patients. Besides, autophagy could be activated by silencing hsa_circ_0037658 in IL-1βtreated CHON-001 cells [106]. Zeng et al. discovered that  [114]. In addition, peroxisome proliferatoractivated receptor α (PPARα), as a transcription factor, protected articular chondrocytes from cartilage degradation in LPS-induced chondrocytes as well as in DMM-induced PTOA mice. Zhou et al. found that the PPAR α agonist (Wy14643) activated autophagy and increased chondrocyte proteoglycan synthesis, which finally produced chondroprotective effects in vitro and in vivo [115]. As a major participant in regulating gene expression in cells, the changes of transcription factors may be closely related to PTOA diseases ( Table 3). The expression of key autophagy genes driven by transcription factors is very important for the maintenance of chondrocyte homeostasis and microenvironment balance, which could be a potential target for PTOA therapy.
Acetylation is related to autophagy in PTOA Acetylation, a process that transfers the acetyl group of acetyl-CoA to protein amino acid residues by acetylase [116], regulates various signaling pathways in cells, including autophagy [117]. Sirtuin (SIRT) is a class of highly conserved deacetylases from bacteria to humans. The human SIRT family consists of seven members, SIRT1-SIRT7 [118]. Wu [121]. In addition, HDAC6, as a deacetylase targeting histones, was revealed to be upregulated in a mouse PTOA model, and inhibition of HDAC6 significantly activated cartilage autophagy and reduced cartilage damage [122]. Moreover, Sacitharan et al. found that spermidine maintained the homeostasis of chondrocytes through up-regulating acetyltransferase EP300 [123]. Briefly, acetylation can affect the degeneration and destruction of cartilage by regulating the function of autophagy-related proteins (Figure 3). Enzymes that can regulate the acetylation of autophagic proteins, including acetylases and deacetylases, could be the candidate targets of anti-PTOA strategies in the future. Lin et [128]. Knockout of alpha7-nAChRs in primary chondrocytes decreased LC3 levels under normal conditions and made the cells more sensitive to MIA-induced apoptosis, which could be related to MIA-induced pain behavior and cartilage degradation [129]. The mRNA level of Lysyl oxidase like 3 (LOXL3) increased in OA patients and OA rats, which was positively correlated with the leptin concentration in joint synovial fluid. Leptin significantly promoted cell proliferation and autophagy in primary chondrocytes. In

ATF4
• Anterior part of medial meniscus excision: induced PTOA in mice by the excision of anterior part of medial meniscus. • Tunicamycin-treated chondrocytes.
• Intra-articular injection of WY14643 attenuated articular cartilage degeneration, promoted proteoglycan synthesis and enhanced autophagy in vivo. • Activation of PPARα by WY14643 increased proteoglycan synthesis via up-regulation of autophagy in LPS-treated chondrocytes. [115] FoxO (FoxO1/3/4) • AcanCreERT-TKO mice developed mild cartilage lesions (2 months) and progressed to full-thickness cartilage defects (5 months) after tamoxifen administration. • The severity of OA induced by DMM and treadmill running was obviously higher in AcanCreERT-TKO mice. • Overexpression of FoxO1 up-regulated autophagic genes such as LC3b, Sesn3 and Prkaa2, and partially reversed the inflammatory and cartilage catabolic genes in IL-1β-treated chondrocytes. [113] ATF4 activating transcription factor 4, PTOA post-traumatic Osteoarthritis, OA-Exo exosomes from osteoarthritis mice, TFEB transcription factor EB, DMM destabilization of the medial meniscus, TBHP tert-butyl hydroperoxide, PPARα peroxisome proliferators-activated receptors alpha, LPS lipopolysaccharide, FoxO forkhead box O, IL-1β interleukin 1 beta addition, overexpression of LOXL3 inhibited autophagy of chondrocytes mainly through activating mTORC1 [130,131]. Wang et al. showed that upregulation of p63, as a member of the p53 family, inhibits chondrocyte autophagy in OA chondrocytes, leading to OA progression [132]. In conclusion, as an important mechanism for maintaining cell homeostasis, the regulation of autophagy level is involved with a variety of signaling pathways from gene expression to protein modification. An in-depth study of the molecular mechanisms of autophagic changes in chondrocytes will provide strong support for autophagy-based treatment in clinical practice.

Autophagy-related PTOA therapeutic strategies
An increasing number of studies have demonstrated that promoting autophagy is an important mechanism for different PTOA treatment strategies, including drugs, physical therapy and biological therapy ( Figure 4). These strategies can delay the progression of OA by enhancing the autophagy level of chondrocytes, preventing chondrocyte death, attenuating matrix degradation and reducing the inflammatory response.
Chemically synthesized drugs that target autophagy in PTOA Metformin, the first-line agent in type 2 diabetes treatment, was reported to increase LC3II, p62 and Lysosomal associated membrane protein 1 (LAMP1) expression levels, which further enhanced the level of autophagy. Na et al.
reported that oral metformin can protect cartilage tissue and delay OA progression in an MIA-induced rat model [133]. d-Mannose is a monosaccharide with an immunity regulation function and anti-osteoporosis effect through the AMPK pathway that can enhance autophagy activity. Lin   [135]. CDDO-Im is a novel synthetic triterpene and autophagy booster.
In a DMM mouse model, Dong et al. found that CDDO-Im decreased the release of inflammatory mediators and alleviated knee cartilage erosion. Consistently, CDDO-Im dose-dependently enhanced autophagy in vitro, highlighting its cartilage protection and anti-OA activity [136]. Vitamin D, as an anti-osteoporosis drug, plays a vital role in bone health. Kong et al. reported that vitamin D could activate autophagy through the AMPK-mTOR signaling pathway and then inhibit the inflammatory response. In addition, OA symptoms were relieved after vitamin D treatment in a mouse model of PTOA [137]. Hydroxytyrosol has an antiinflammation function and has been used for treatment of inflammatory diseases. Zhi et al. found that hydroxytyrosol can inhibit the inflammatory response of chondrocytes through SIRT6-mediated autophagy [138]. Diazoxide, as one of the potential drugs to prevent OA, can significantly improve the severity of experimental PTOA, which was related to the restoration of impaired autophagy [139]. An increasing number of studies has found that different compounds can cause changes in autophagy levels. However, the mechanisms of these compounds on cells may be complex and autophagy regulation may be one of them. We need to study more compounds with the capacity of relatively specific autophagy activation so as to accurately regulate the autophagic flux in OA and reduce the potential side effects of the compounds.

Natural extracts that target autophagy in PTOA
The therapeutic effects and detailed mechanisms of natural products have been widely studied in the past decades. Previous studies have suggested that different types of natural products can reduce joint inflammation, prevent cartilage degradation and delay PTOA progression in vitro and in vivo ( Table 4).
Phenols that target autophagy in PTOA As a polyphenolic drug, punicalagin possesses various pharmacological functions, one of which is autophagy activation and restoration. A previous study revealed that punicalagin can activate autophagy and restore autophagic flux in TBHP-treated chondrocytes, while it ameliorated the degeneration of articular cartilage in mice following DMM [140]. Gallocatechin 3-gallate, an active component of green tea, could activate autophagy and regulate chondrocyte apoptosis by reducing the expression level of mTOR and enhance the expression of microtubule-associated protein light chain 3, BECN1 and p62. In an ACLT-induced OA model, gallocatechin 3-gallate treatment reduced cartilage degradation [141]. Mangiferin is a natural polyphenol. As a potential therapeutic drug for OA, mangiferin could enhance autophagy by activating the AMPK signaling pathway. In a TBHP-induced chondrocyte injury model and mouse PTOA model, Li et al. treated mice with mangiferin and reduced chondrocyte apoptosis and cellmatrix degradation [142].
Flavonoids that target autophagy in OA Quercetin is widely present in plants and is able to promote autophagy, mediated by the TSC2-RHBE-mTOR signaling pathway. Lv et al. reported that quercetin reduced chondrocyte apoptosis and relieved knee joint injury in a DMM-induced PTOA model [143]. Rhoifolin, a flavonoid extracted from the beech tree, demonstrated autophagy regulation functions through the P38/JNK and PI3K/AKT/mTOR signaling pathways. Yan et al. found that rhoifolin could reduce cartilage degradation and delay cartilage damage in a mouse PTOA model [144]. As the main component of epimediums, icariin has various pharmacological effects, including inhibiting the PI3K/AKT/mTOR signaling pathway, up-regulating autophagy genes and reducing chondrocyte apoptosis. In a rat model of OA, icariin treatment significantly alleviated the pathological damage of cartilage [145]. Glabridin is a natural antioxidant small molecule with a strong scavenging effect on free radicals, possessing an autophagy activation function. Glabridin can up-regulate the expression of ECMrelated genes including Collagen II, aggrecan, SRY-box 9 and proteoglycan 4, accompanied by an increase of autophagy level in vitro. Moreover, intra-articular injection of glabridin protected chondrocytes against apoptosis and alleviated OA progression in ACLT rats, which was partially related to mTOR-mediated autophagy [146].
Terpenoids that target autophagy in PTOA Terpenoids are the most diverse group of secondary metabolites in quantity and structure that are derived from natural sources and have various bioactivities. However, how terpenoids function in OA is still not clear. Alantolactone (ALT), a sesquiterpene lactone compound of terpenoids, is purified from Inula helenium L. Previous studies revealed that the ALT can inhibit the phosphorylation of Signal transducer and activator of transcription 3 (STAT3) selectively and restrain its transportation from cytoplasm to the nucleus. STAT3 is activated consistently in IL-1β-induced chondrocytes, while ALT treatment attenuated the inflammatory response and cartilage degeneration by abolishing STAT3 sensitization [147]. Tetrahydrohyperforin, a tetrahydro derivative of hyperforin and a member of the terpenoid group, is one of the main active components of Hypericum perforatum L, demonstrated to be able to activate ATG5-depended autophagy. Zhang et al. discovered that tetrahydrohyperforin could improve the pathological severity of a joint injury in a rat model that was chemically induced by intra-articular injection of collagenase solution [148]. The role of autophagy regulation in chondrocytes by tetrahydrohyperforin should be further investigated in PTOA models.

Other natural extracts that target autophagy in PTOA
Oxoglaucine is the extractor from Magnoliaceae and possesses an autophagy activation function, manifested as accelerating autophagic influx and vesicle formation. In chondrocytes from OA patients and animal OA models, oxoglaucine was found to activate autophagy, delay cartilage matrix degradation and prevent cartilage damage through the TRPV5/calmodulin/CAMK-II pathway [149]. As a natural extract, shikimic acid (SA) shows strong antiinflammatory properties and can promote autophagy by up-regulating the expression of ATG7, BECN1 and LC3, which was related to the inhibition of MAPK and NF-κB pathways. Moreover, the progression of OA was significantly alleviated by SA in a trauma-induced PTOA rat model [150].  [144] (continued)  Trehalose was first extracted from Ergot bacteria and is a novel mTOR-independent autophagy inducer. Tang et al. found that trehalose ameliorated oxidative stress-mediated mitochondrial damage and ER during the progression of OA after trehalose treatment in a TBHP-induced chondrocyte injury model and a DMM-induced mouse PTOA model, and eventually relieved joint symptoms [151]. β-Ecdysterone, an active ingredient isolated from Achyranthes with broad pharmacological effects, can activate chondrocyte autophagy through regulating the PI3K/AKT/mTOR signaling pathway. In an MIA-induced rat model, the inflammation of chondrocytes was significantly reduced after β-ecdysterone treatment [152]. In future studies, the therapeutic effect of β-ecdysterone on PTOA needs to be evaluated in surgically induced animal models. From the current research, the natural products regulating autophagy of chondrocytes mostly belong to phenols and flavonoids. The key autophagy genes and signaling pathways regulated by these natural compounds are not exactly the same, suggesting the complexity of their mechanisms of action on chondrocytes. More research needs to be done to explore the pharmacodynamic structures of the active components of these compounds so as to help optimize the structure of existing natural products and make them more specifically regulate autophagy levels in chondrocytes. So far, many natural products have been used in the clinical research of PTOA treatment. With the in-depth research in future, the natural products that target autophagy may be greatly improved in anti-PTOA therapy.
Physical therapy that targets autophagy in PTOA Mechanical loading can affect the pathological progression of early and late PTOA. In a mouse OA model, Zheng et al. found that mechanical loading promoted the phosphorylation of eukaryotic translation initiation factor 2a to regulate the expression of LC3II/I and p62, and ultimately increased the resistance of chondrocytes to injury [153]. Reasonable exercise can enhance the stability of the knee joint to prevent or delay the occurrence and progression of OA. Li et al. found that moderate-intensity exercise promoted autophagy activation through modulating the AMPK/mTOR signaling pathway, resulting in the inhibition of chondrocyte apoptosis and the alleviation of cartilage degeneration in vitro [154]. Ozone can inhibit the inflammation of OA and regulate cartilage metabolic balance. In a PTOA model, Xu et al. found that ozone could up-regulate the expression level of the autophagy-related protein LC3II, then reduce the inflammation and inhibit MMP-13, which finally reduced cartilage degeneration and pain symptoms [155]. Zhang et al. found that 30 min of treadmill exercise promoted autophagy in articular cartilage, decreased inflammation and increased the expression of type II collagen, resulting in the protection of chondrocyte degeneration in a rat model of OA induced by MIA [156]. In general, the activation of autophagy may be an important mechanism of some physical therapies for PTOA patients. We need to further optimize the parameters of physical therapy on the basis of its action mechanism so as to produce better effects in PTOA treatment.
Biological therapy that targets autophagy in OA Exosomes are nanoscale extracellular vesicles secreted by cells that can carry biologically active substances such as lipids, proteins and complex RNAs. In a DMM-induced mouse PTOA model, chondrocyte exosomal vesicles inhibited macrophage ATG4B expression via miR449a-5p, then activated inflammation and promoted IL-1β maturation by inhibiting endotoxin-induced autophagy, leading to an increase in synovial inflammation, indicating that targeting exosomal vesicles from inflammatory chondrocytes is a potential strategy for PTOA treatment [157]. In an IL-1βinduced chondrocytes injury model, Wen et al. found that human mesenchymal stem cell-derived exosomes (MSC-Exo) upregulated KLF3-AS1 expression, then activated the PI3K/Akt/mTOR signaling pathway and finally inhibited autophagy, mediated by lncRNA (KLF3-AS1). Then, MSC-Exo were found to reduce chondrocyte apoptosis and improve OA symptoms, which is expected to be a potential target for OA therapy [158]. As a novel scavenger of ROS and reactive nitrogen species, dopamine melanin (DM) nanoparticles possess low cytotoxicity and a powerful ability to sequester ROS and reactive nitrogen species. In an animal model of OA, Zhong et al. treated mice with intra-articular injection of DM and found that it could alleviate cartilage degradation and reduce inflammatory cytokine release and proteoglycan loss [159]. IL-1β plays a critical role in the progression of OA and is a vital factor for cartilage damage. In a IL-1β-induced chondrocyte injury model and surgically induced mouse PTOA model, Wang et al. found that blocking IL-1Ra can facilitate autophagy recovery and delay ECM degradation [160]. MSCs are a kind of early undifferentiated cell with self-renewal, self-replication, unlimited proliferation and multi-directional differentiation potential. MSCs can secrete cytokines, reduce inflammation, reduce tissue cell apoptosis and promote endogenous stem progenitor cell proliferation in sexual tissues and organs. In ACLT and DMM surgically induced rat PTOA models, Chen et al. performed phosphate-buffered saline or MSCs-conditioned medium treatment and uncovered that MSC-conditioned medium could enhance autophagy, inhibit chondrocyte apoptosis, protect subchondral bone microstructure and balance the ratio of MMP-13 to TIMP-1 in cartilage, so as to relieve PTOA symptoms [161]. In addition, Zhou et al. implanted adipose-derived MSCs into the right knee joint of a surgically induced rat PTOA model and found that they reduced the secretion of pro-inflammatory factors and decreased apoptosis by activating autophagy [162]. Biological therapy, especially related to stem cells and their derivatives, has attracted much attention in PTOA clinical therapy. Activating the autophagy level of chondrocytes by biological products to prevent cartilage degeneration is a potentially feasible strategy for OA treatment, which deserves further exploration in the future.

Perspective
PTOA is a whole joint disease, involving cartilage, subchondral bone, synovium and other joint tissues. Existing studies have mainly focused on the roles of autophagy in the maintenance of cartilage homeostasis. In recent years, some studies also suggest that autophagy can affect the progression of PTOA by affecting synovial inflammatory cells and subchondral osteoblasts, indicating that autophagy participates in the progression of PTOA by targeting different joint tissues [71,163]. Therefore, more studies are needed to reveal the roles and detailed mechanisms of autophagy during PTOA, so as to more effectively improve the effect of anti-PTOA treatment based on autophagy. With the development of the life sciences and technology, it is believed that we will have a more comprehensive and in-depth understanding of the roles and mechanisms of autophagy in the process of PTOA.
As an important mechanism to maintain cell homeostasis, autophagy in PTOA has attracted increasing attention. Previous studies have mainly focused on the role of macroautophagy in the maintenance of cartilage homeostasis, but other types of autophagy need more attention. Recent studies have shown that mitophagy, a novel manifestation of autophagy, is similarly closely associated with PTOA progression [164]. Mitophagy is generally thought to be a dominant mechanism of mitochondrial quality control. The intact mitochondrial structure is a prerequisite for the survival of chondrocytes [165]. Mitochondrial dysfunction causes metabolic disorders and inflammatory responses in chondrocytes, which in turn contributes to PTOA progression. Zhang et al. showed that mitophagy and mitochondrial dynamics are involved in chondrocyte biological stress regulation [166]. Moreover, in an AGE-induced chondrocyte injury model, Tang et al. found that bone marrow MSC-Exos inhibited chondrocyte apoptosis and cartilage matrix degradation by drp1-mediated mitophagy [167]. Apart from mitophagy, the roles and mechanisms of other types of autophagy in PTOA should also be given more attention in the future.
Although many strategies have been developed for autophagy evaluation, most of the existing methods are static evaluations and can hardly achieve accurate dynamic monitoring [168]. mRFP-GFP-LC3 is an optional method for the dynamic detection of autophagy, but there is still the problem of inaccurate fluorescent quantification, which needs further improvement and optimization. In addition, the current drugs that activate or inhibit autophagy lack specificity, which may increase the risk of side effects during treatment [169,170]. Therefore, we need more in-depth research on mechanisms of autophagy regulation in order to develop more specific autophagy targeting agents and ultimately provide more effective methods for PTOA therapy in clinic.

Conclusions
In conclusion, the decrease of autophagy is very common during the PTOA process, and its change is closely related to the occurrence and development of PTOA. As an important mechanism for the maintenance of chondrocyte homeostasis, autophagy can effectively prevent cartilage degeneration by regulating cell death, anabolism and catabolism, hypertrophy and inflammation. The regulatory mechanisms of autophagy during PTOA are highly associated with gene transcription, mRNA stability, protein modification, etc. Some potential therapeutic strategies display their anti-PTOA effects partially through enhancing autophagy. Rational enhancement of autophagy will be a promising mode to prevent the occurrence and development of PTOA in the future. Therefore, we need to further study the mechanisms of autophagy and identify specific targets in PTOA so as to provide effective treatments for clinical PTOA patients.