An update on the bridging factors connecting autophagy and Nrf2 antioxidant pathway

Macroautophagy/autophagy is a lysosome-dependent catabolic pathway for the degradation of intracellular proteins and organelles. Autophagy dysfunction is related to many diseases, including lysosomal storage diseases, cancer, neurodegenerative diseases, cardiomyopathy, and chronic metabolic diseases, in which increased reactive oxygen species (ROS) levels are also observed. ROS can randomly oxidize proteins, lipids, and DNA, causing oxidative stress and damage. Cells have developed various antioxidant pathways to reduce excessive ROS and maintain redox homeostasis. Treatment targeting only one aspect of diseases with autophagy dysfunction and oxidative stress shows very limited effects. Herein, identifying the bridging factors that can regulate both autophagy and antioxidant pathways is beneficial for dual-target therapies. This review intends to provide insights into the current identified bridging factors that connect autophagy and Nrf2 antioxidant pathway, as well as their tight interconnection with each other. These factors could be potential dual-purpose targets for the treatment of diseases implicated in both autophagy dysfunction and oxidative stress.


Introduction
Autophagy is a metabolic process that takes place in eukaryotic cells, which use lysosomes to degrade proteins and damaged organelles for self-eating and self-renewal (Wong et al., 2020). Autophagy dysfunction contributes to a variety of pathologies, including lysosomal storage diseases, neurodegenerative diseases, chronic metabolic diseases, and cancer (Klionsky et al., 2021). Of note, these diseases are usually accompanied by elevated levels of reactive oxygen species (ROS) (Giordano et al., 2014;Ryter et al., 2019;Wu W. et al., 2021).
ROS are highly reactive and diffusive molecules, and endogenous ROS mainly come from cellular respiration in mitochondria Sies et al., 2022). In addition, various organelles (endoplasmic reticulum and peroxisome) and enzymes (NADPH oxidase and p450 cytochrome) also produce ROS during metabolic processes Sies et al., 2022). ROS are regarded as a "double-edged sword." Physiologically low levels of ROS serve as a signal, promoting cell viability and immune function and maintaining cell homeostasis (Sies and Jones, 2020;Sies et al., 2022). Conversely, excessive ROS oxidize and damage cellular components, leading to cellular death (Davalli et al., 2016), which is implicated in many diseases (Luo et al., 2020;Yang and Lian, 2020). To maintain ROS at low levels, cells are controlled by multiple antioxidant systems, including various non-enzymatic and enzymatic antioxidant pathways (Alkadi, 2020;Forman and Zhang, 2021), as well as several antioxidant transcription factors. Furthermore, these antioxidants have different subcellular localizations, therefore eliminating ROS in different compartments (Brieger et al., 2012).
Interestingly, ROS also interplay with autophagy (Gao et al., 2020). Autophagy can participate in the redox balance by clearing oxidative stress-damaged molecules and organelles (Yun et al., 2020). The Nrf2-Keap1-ARE pathway plays a central role in adaptive cellular redox response (Shaw and Chattopadhyay, 2020;Esteras and Abramov, 2022). Studies have reported that several proteins serve as bridges between the Nrf2 pathway and autophagy Zhang et al., 2021). In this review, we focus on the bridging factors that establish networks between autophagy and the Nrf2 antioxidant pathway.

FIGURE 1
Balance between ROS and antioxidant responses. Under normal conditions, cells maintain redox homeostasis by balancing ROS levels and antioxidant activity. Cellular ROS including oxygen anions, superoxide (O 2 1.2 Nrf2-Keap1-ARE signaling pathway Nuclear factor erythroid 2-related factor 2 (Nrf2) is a basic leucine zipper transcription factor belonging to the Cap 'n' Collar family. Under quiescent conditions, cytosolic Nrf2 directly binds to Kelch-like ECH-associated protein 1 (Keap1), a substrate adapter protein for the Cul3-E3-ligase complex, resulting in rapid degradation of Nrf2 (Ulasov et al., 2022). Keap1 is also a cysteine (Cys)-rich sensor of redox damage (Yamamoto et al., 2018). However, with exposure to stimuli such as oxidative/xenobiotic stress, Nrf2 escapes from Keap1-mediated repression and translocates to the nucleus, where it dimerizes with small musculoaponeurotic fibrosarcoma proteins (sMaf). The activated Nrf2-sMaf heterodimer binds to antioxidant response element (ARE) sequences in the promoter of a tandem of antioxidant genes, including heme oxygenase-1 (HMOX1), aldo-keto reductases, glutathione, and glutathione-S-transferase (Hayes and Dinkova-Kostova, 2014;Tonelli et al., 2018), to trigger their expression. Numerous studies have revealed the complexity and diversity of the Nrf2-Keap1 pathway in regulating biological processes, including cell proliferation, differentiation, antiinflammation, and cytoprotection (Murakami and Motohashi, 2015;Yamamoto et al., 2018). Nrf2 has also been demonstrated to regulate mitochondrial biogenesis (Esteras and Abramov, 2022). Two key regulators of mitochondrial biogenesis, namely, proliferator-activated receptor gamma coactivator 1-α and nuclear respiratory factor 1, which are responsible for mitochondrial DNA transcription, are under the control of Nrf2 (Piantadosi et al., 2008;Baldelli et al., 2013;Merry and Ristow, 2016).
It is known that ROS also interact with autophagy. ROS could promote the formation of autophagy. Autophagy, in turn, could reduce oxidative damage by engulfing and degrading oxidized substances.

ROS and autophagy interplay
Autophagy is a catabolic process that is responsible for the degradation of intracellular proteins and damaged organelles in response to endogenous and exogenous stresses, including oxidative stress, endoplasmic reticulum stress, hypoxic stress, and nutrient and growth factor starvation (Galati et al., 2019). Promoting autophagy has been reported to have beneficial effects on (1) ROS directly oxidize the cysteine residues of α and β subunits of AMPK, and activated AMPK inhibits the activity of the mTORC1 complex or phosphorylates the ULK1 complex, subsequently promoting autophagy (top). (2) Oxidation of Atg4 by ROS results in inactivation of LC3-PE deconjugation activity and accumulation of autophagic LC3-II isoforms, supporting autophagosome formation (middle). (3) The oxidation of catalytic thiols on Atg3 and Atg7 by ROS prevents the lipidation of LC3, thereby inhibiting autophagy (bottom). Autophagy, in turn, contributes to reduction of ROS and damaged organelles through pathways such as mitophagy. Damaged mitochondria are recruited into the autophagosome by binding with LC3. The mature autophagosome fuses with the lysosome to form the autolysosome, and the damaged mitochondria are subsequently degraded.
Frontiers in Cell and Developmental Biology frontiersin.org longevity, anti-infection, and disease prevention, including diseases such as myocarditis, tumors, and neurodegenerative diseases (Petrache et al., 2008). Autophagy dysfunction has also been identified in many diseases, including Niemann-Pick type C (NPC) disease, atherosclerosis, and non-alcoholic steatohepatitis (Fischer et al., 2009). It is broadly recognized that excessive ROS can regulate autophagy (activation or inhibition) in various conditions (Gao et al., 2020;Ornatowski et al., 2020;Zhou et al., 2022). Autophagy, in turn, can also reduce oxidative damage by degrading oxidized substances  ( Figure 2). Thus, ROS and autophagy interact closely with each other. The process of autophagy is generally divided into the occurrence of phagophores, the formation of autophagosomes, the fusion of autophagosomes and lysosomes, and degradation. A variety of autophagy-related proteins (Atg) constantly regulate the whole process of autophagy (Levine and Kroemer, 2019).
Previous studies have identified that H 2 O 2 and O 2 − are major inducers of autophagy (Filomeni et al., 2015). H 2 O 2 is relatively stable, long-lived, and highly selective to cysteine (Bienert and Chaumont, 2014). Instead, O 2 − is unstable and easily converted to H 2 O 2 spontaneously or catalyzed by enzymes (Fujii et al., 2022). In the initiation stage of autophagy, ROS (H 2 O 2 and O 2 − ) promote autophagy mainly via regulating the mechanistic target of rapamycin (mTOR). mTOR is a receptor of amino acids and ATP that plays a gating role in autophagy. ROS can activate AMP-activated protein kinase (AMPK) or inhibit protein kinase B (Zhao et al., 2017), which represses mTORC1 complex activity and enhances autophagy Guo et al., 2022). In the autophagosome formation stage, Atg4, an essential protease in autophagy, has been identified as a direct target of H 2 O 2 (Scherz-Shouval et al., 2007). H 2 O 2 directly oxidizes the Cys81 of Atg4 (Scherz-Shouval et al., 2007), thus leading to the inactivation of Atg4 and promoting the lipidation of light chain 3 (LC3, also the mammalian homolog of Atg8), thereby supporting the formation of autophagosomes (Pérez-Pérez et al., 2021). Interestingly, ROSinduced autophagy is primarily caused by O 2 − during prolonged starvation (Chen et al., 2009). In contrast, H 2 O 2 is produced immediately after starvation and regulates autophagy (Zhou et al., 2022). In addition, the crosstalk between autophagy and ROS also varies in different cell types. For example, in renal tubular cells, increased ROS can activate autophagy, leading to mitochondrial destruction and kidney damage (Duan et al., 2018). However, in tumor cells, ROS-induced autophagy activation can promote cell growth and cancer progression (Taucher et al., 2022).

Autophagy regulates ROS
Mitochondria are the primary sites of ROS production (approximately 90% of cellular ROS) (Balaban et al., 2005). Thus, mitochondria also contain a large number of antioxidants (proteins and enzymes), including glutathione, glutathione reductase, peroxiredoxin, and superoxide dismutase . Mitochondrial antioxidants reduce oxidative damage to mitochondria and maintain their homeostasis (Liang et al., 2012;Ristić et al., 2015;Kang et al., 2020). In pathological conditions, excessive ROS leads to oxidative stress, causing mitochondrial dysfunction and cellular damage (Tirichen et al., 2021). The process of specific and selective degradation of damaged mitochondria is called mitophagy. Mitophagy generally consists of two steps: the initiation of general autophagy and the priming of damaged mitochondria for autophagic recognition (Figure 2). Recently, studies have shown that ROS, as cellular signaling molecules, could trigger autophagosome formation and autophagic degradation (Filomeni et al., 2015;Roca-Agujetas et al., 2019).

Pexophagy
Peroxisomes are small and highly dynamic organelles that are ubiquitous in eukaryotic cells and play a key role in cell metabolism and the regulation of redox homeostasis (Germain and Kim, 2020). Pexophagy, a process of selective autophagy for damaged peroxisomes, contributes to maintaining the intracellular redox balance (Germain and Kim, 2020). Under ROS stimulation, the peroxisome localization protein PEX5 (peroxisomal biogenesis factor 5) is phosphorylated at Ser141 by ataxia telangiectasia mutated kinase and subsequently ubiquitinated by E3 ligase . SQSTM1 then recognizes the ubiquitinated PEX5 complex and triggers pexophagy . In addition, PEX2 is identified as an E3 ubiquitin ligase during amino Frontiers in Cell and Developmental Biology frontiersin.org acid starvation (Sargent et al., 2016). PEX2 can ubiquitinate PEX5, thereby activating NBR1-dependent pexophagy (Sargent et al., 2016).

Bridging factors connecting autophagy and Nrf2-Keap1 pathway
Both autophagy dysfunction and oxidative stress have been implicated in many diseases. Treatment targeting one aspect shows limited effects. For example, N-acetylcysteine (NAC), a potent antioxidant, has been examined in clinical trials of NPC disease, which shows increased oxidative stress and dysfunctional autophagy (Fu et al., 2013;Wheeler and Sillence, 2020). Results show that NAC significantly reduces oxidative stress in NPC disease but has no effect on autophagic dysfunction-related cholesterol accumulation, thus failing in clinical trials (Fu et al., 2013). Of note, identifying the bridges that connect autophagy and antioxidant pathways could be extremely beneficial to novel dual-target therapies.

SQSTM1
SQSTM1 (sequestosome1, also called p62) is an autophagy adapter protein that participates in the degradation of protein aggregates and cytoplasmic bodies (Lin et al., 2013). Numerous studies have described the effects of SQSTM1 on a variety of diseases, such as Parkinson's disease (PD) (Martins-Marques et al., 2015;Chu, 2019;Kumar et al., 2022). Under normal conditions, SQSTM1 recruits α-synuclein to form aggregates and eliminates them through autophagy. In contrast, in PD patients, dysfunctional SQSTM1 fails to recruit impaired α-synuclein, which accumulates in the central nervous system and aggravates PD progression . In addition, SQSTM1 can recognize ubiquitinated Parkin, a key protein of mitophagy, and disrupted SQSTM1 mediates mitophagy deficiency (Song et al., 2016), which is an important cause of PD. SQSTM1 recruits ubiquitinated proteins and organelles to the autophagosome through its simultaneous interaction with multiple proteins, including protein kinase C and mTOR (Katsuragi et al., 2015).
In autophagy-deficient cells, SQSTM1-induced Nrf2 activation is critical in tumor progression (Komatsu et al., 2010). SQSTM1 can directly bind to Keap1, disrupting Keap1regulated Nrf2 degradation and promoting aberrant Nrf2activation (Komatsu et al., 2010). The DPSTGE motif in the Keap1 interaction region (KIR) domain of SQSTM1 resembles the ETGE motif in Nrf2 (Komatsu et al., 2010) and illuminates the competitive binding of Keap1 between SQSTM1 and Nrf2. However, the binding ability of SQSTM1-KIR to Keap1 is much weaker than that of Nrf2-ETGE. Nevertheless, under exposure to ROS or other electrophilic stimuli, Ser351/349 within the DPSTGE motif in the KIR region will be phosphorylated, and consequently, the binding affinity of SQSTM1 to Keap1 can be drastically increased (Ichimura et al., 2013;Sánchez-Martín and Komatsu, 2018). SQSTM1 can recruit Keap1 and LC3 to form the LC3-SQSTM1-Keap1 complex, which is then degraded by selective autophagy (Liao et al., 2019) (Figure 3). Nevertheless, the interaction between LC3, SQSTM1, and Keap1 still remains controversial, as the binding competition to SQSTM1 between Keap1 and LC3 can be attributed to the spatial adjacency of the LIR domain and KIR domain in SQSTM1 (Jain et al., 2010). Moreover, depleting SQSTM1 with siRNA leads to an almost twofold increase in the half-life of Keap1 (Copple et al., 2010), indicating that SQSTM1 could influence the basal level of Keap1 and regulate its degradation via clearance of the

FIGURE 3
Bridging factors that connect autophagy and Nrf2 antioxidant pathway. A working model illustrates the mechanisms of bridging factors (SQSTM1, TFEB, Sestrin2, TRIM16, Ca 2+ , and miRNAs) connecting autophagy (left) and the main antioxidant Nrf2-Keap1-ARE pathway (right) and the feedback loops between these factors.
Frontiers in Cell and Developmental Biology frontiersin.org SQSTM1-Keap1 complex by the ubiquitin-proteasome system, which is triggered by binding of the ubiquitin-associated domain (UBA) of SQSTM1. Upon oxidative stress, the escape of Nrf2 from Keap1 also contributes to the formation of LC3-SQSTM1-Keap1 aggregates . In addition, Nrf2 regulates SQSTM1 gene expression (Ho and Gorski, 2019). Overall, an SQSTM1-Keap1-Nrf2 positivefeedback loop  represents the mutual regulatory relationship between the Nrf2 antioxidant system and autophagy (Figure 3).
Overexpression of TFEB leads to degradation of bulk autophagic substrates such as damaged mitochondria (Nezich et al., 2015), longlived protein aggregates (Martini-Stoica et al., 2016), and lipid droplets (Settembre et al., 2013), indicating the crucial role of TFEB in organelle-specific autophagy, such as mitophagy and lipophagy (Settembre et al., 2013;Nezich et al., 2015). Importantly, TFEB has no effect on the basal transcriptional levels of downstream target genes but rather regulates their transcriptional levels in response to stimuli. Under normal conditions, TFEB is phosphorylated by mTORC1 and exists mainly in an inactive state in the cytosol (Martina and Puertollano, 2013). Upon stimuli such as starvation or oxidative stress, TFEB is dephosphorylated, subsequently translocates to the nucleus, and promotes the activation of its target genes (Sancak et al., 2010;Medina et al., 2015).
Interestingly, Bo Pan et al., 2020 have delineated a relationship between SQSTM1 and TFEB. They proposed that the protein aggregates formed by SQSTM1 during systemic proteasome inhibition sequester mTOR and prevent its TFEB phosphorylation. Furthermore, SQSTM1 is capable of interacting with Raptor, a part of mTORC1, and colocalizing with Rag GTPases at the lysosomal membrane, hence being responsible for the activation of mTORC1 (Duran et al., 2011). SQSTM1 can also recruit the tumor necrosis factor receptor-associated factor 6 in an amino acid-dependent manner and activate mTOR as well (Linares et al., 2013). Notably, Nrf2/NFE2L2 is identified as a target gene of TFEB; thus, TFEB regulates redox homeostasis via targeting Nrf2 . Therefore, a network forms that connects Nrf2, SQSTM1, TFEB, and mTOR ( Figure 3).

Other non-canonical bridging regulatory factors
Other non-canonical autophagy regulatory proteins like Sestrin2 (Bae et al., 2013) and tripartite motif-containing protein 16 (TRIM16) (Kimura et al., 2015) also participate in the regulation of Nrf2 and mTOR via direct or indirect interactions, providing novel targets for pondering the connection between autophagy and the Nrf2 antioxidant pathway.

Sestrin2
Sestrin2, a highly conserved protein belonging to the Sestrin family, is involved in multiple biological functions, including regulating redox reactions, metabolic homeostasis, and the aging process Gong et al., 2021). Sestrin2 has been proposed to be the therapeutic target for various diseases, such as cardiac diseases, respiratory diseases, and non-alcoholic fatty liver disease (Ren D. et al., 2020;Sun et al., 2020;Wu D. et al., 2021). Sestrin2 has been reported to protect insulin-resistant cells by activating the AMPK pathway or upregulating the mTOR pathway (Gong et al., 2021), thus promoting a beneficial effect on diabetes (Tian et al., 2022).
Studies have shown that Sestrin2 overexpression can induce autophagic degradation of Keap1, leading to the upregulation of Nrf2 activity (Bae et al., 2013). In addition, this process requires the phosphorylation of the UBA domain of SQSTM1 at Ser409 (Polito et al., 2014;Rhee and Bae, 2015), resulting in the degradation of LC3-SQSTM1-Keap1 aggregates by selective autophagy.

TRIM16
TRIM16 has been identified as a regulator of oxidative-stressresponsive proteins (Ren X. et al., 2020) and is associated with tumor suppression , cell motility (Huo et al., 2015;Li et al., 2016), apoptosis , and autophagy (Chauhan et al., 2016;New and Thomas, 2019). TRIM16 deficiency significantly exacerbates cardiomyocyte hypertrophy, while overexpression of TRIM16 inhibits cardiac hypertrophy, suggesting that a TRIM16 inhibitor could be a novel inhibitor of pathological heart hypertrophy and heart failure (Liu J. et al., 2022). Jena et al. (2018) have unveiled the latent significance of TRIM16 in protecting cells from oxidative stress-induced cytotoxicity by interacting with Nrf2, Keap1, and SQSTM1. TRIM16, as an E3 ligase, promotes the SQSTM1-induced autophagic degradation of Keap1 and regulates Nrf2 activity (Jena et al., 2018).
A recent study has revealed that TRPV1-evoked Ca 2+ influx promotes CaMKII phosphorylation and subsequently promotes Nrf2 nuclear translocation (Lv et al., 2021). In addition, the Ser558 site in the Neh1 domain of Nrf2 has been identified as a direct phosphorylation site of AMPK (Joo et al., 2016), which results in Nrf2 activation via the Ca 2+ -CaMKKβ-AMPK pathway. Moreover, Ca 2+ -dependent protein kinase C can phosphorylate the Ser40 site in the Neh2 domain of Nrf2 (Niture et al., 2014), thereby activating Nrf2.

miRNAs
At post-transcriptional level, microRNAs (miRNAs) have been reported to impact both the regulation of autophagy and Nrf2 antioxidant signaling. miR-144 is the first miRNA identified as an Nrf2 negative regulator (Sangokoya et al., 2010), and miR-144 can also promote autophagy (Chen et al., 2015). Conversely, Nrf2 can upregulate miR-129-3p, which inhibits mTOR and leads to the initiation of autophagy (Sun et al., 2019). In addition, a study investigated miR-93 and Nrf2 and demonstrated that miR-93 can regulate the mRNA and protein levels of Nrf2 (Singh et al., 2013), while it can target Atg16L to affect autophagy in breast cancer cells (Lu et al., 2014).

Conclusion
In summary, examining the crosstalk between Nrf2 antioxidant signaling and autophagy provides insights into how they are interconnected and the proteins that mediate their communication. In the aspect of autophagy and the Nrf2-Keap1 antioxidant pathway, SQSTM1, TFEB, Sestrin2, and TRIM16 are all involved in different conditions and act as the scaffolds connecting the two pathways (Table 1). Moreover, studies also reveal that intricate feedback loops form between these bridging proteins. These factors are potential therapeutic targets for diseases with both autophagy dysfunction and oxidative stress. However, since these regulatory proteins seem intricately entangled, potential side effects in practical scenarios should also be taken into consideration. Nevertheless, further studies on understanding the complex crosstalk between autophagy and antioxidant pathways are yet to be conducted.

Author contributions
Conceptualization, DL; data curation, DL, BN, and SH; writing-original draft preparation, BN, SH, and CM; writing-review and editing, DL, BN, SH, and WZ; supervision, DL; project, DL. All authors contributed to the article and approved the submitted version.

Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.