p62/SQSTM1 in liver diseases: the usual suspect with multifarious identities

p62/Sequestosome‐1 (SQSTM1) is a selective autophagy receptor that recruits and delivers intracellular substrates for bulk clearance through the autophagy lysosomal pathway. Interestingly, p62 also serves as a signaling scaffold to participate in the regulation of multiple physiological processes, including oxidative stress response, metabolism, inflammation, and programmed cell death. Perturbation of p62 activity has been frequently found to be associated with the pathogenesis of many liver diseases. p62 has been identified as a critical component of protein aggregates in the forms of Mallory–Denk bodies (MDBs) or intracellular hyaline bodies (IHBs), which are known to be frequently detected in biopsy samples from alcoholic steatohepatitis (ASH), non‐alcoholic steatohepatitis (NASH), and hepatocellular carcinoma (HCC) patients. Importantly, abundance of these p62 inclusion bodies is increasingly recognized as a biomarker for NASH and HCC. Although the level of p62 bodies seems to predict the progression and prognosis of these liver diseases, understanding of the underlying mechanisms by which p62 regulates and contributes to the development and progression of these diseases remains incomplete. In this review, we will focus on the function and regulation of p62, and its pathophysiological roles in the liver, by critically reviewing the findings from preclinical models that recapitulate the pathogenesis and manifestation of these liver diseases in humans. In addition, we will also explore the suitability of p62 as a predictive biomarker and a potential therapeutic target for the treatment of liver diseases, including NASH and HCC, as well as recent development of small‐molecule compounds for targeting the p62 signaling axis.

Introduction p62/Sequestosome-1 (SQSTM1) is a well-known selective autophagy receptor which functions to recruit and deliver intracellular substrates for bulk clearance through the autophagy-lysosome pathway [1,2]. Mounting evidence, however, suggests that p62 also serves as a signaling scaffold to participate in the regulation of multiple physiological processes, including oxidative stress response, metabolism, inflammation, and programmed cell death [3][4][5]. Not surprisingly, mutations in p62/SQSTM1 or dysregulated levels of p62 are frequently found to be associated with pathogenesis of many diseases, as summarized in Table 1. In this review, we will be focusing on the function and regulation of p62 and its possible roles in contributing to regulating liver physiology and diseases, with the aim to explore the suitability of p62 as a predictive biomarker and potential therapeutic target for the treatment of certain liver diseases.

Structure and functions of p62
p62, which is encoded by the SQSTM1 gene, is a 62-kDa protein consisting of distinct domains: Phox and Bem1p (PB1) domain, zinc finger (ZZ) motif, LIM protein-binding (LB) domain, tumor necrosis factor receptor-associated factor 6 (TRAF6)-binding (TB) domain, LC3-interacting region (LIR), Kelch-like ECH-associated protein 1 (Keap1)-interacting region (KIR), and ubiquitin-associated (UBA) domain ( Fig. 1). As an autophagy receptor, p62 is associated with the poly-ubiquitylated protein substrates via its Cterminal UBA domain to assemble them into aggregates that can subsequently be incorporated into the autophagosome [1,2,6]. Aggregation of p62 involves oligomerization through trans interaction of its Nterminal PB1 domain via hydrogen bond formation between the lysine 7 (K7) residue and the aspartic acid 69 (D69) residue among the p62 molecules [1,7,8]. Oligomerization of p62 promotes and stabilizes its interaction with ubiquitylated substrates and LC3 [9]. The binding between p62 and LC3 is mediated by a single LIR motif at the C-terminal region of p62 [6]. All the key domains of p62 known to date that confer interaction with various effectors are summarized in Table 2.
To date, there are two isoforms of p62 reported in humans: p62-H1 is the largest isoform consisting of 440 amino acid residues, while p62-H2 is a truncated form partly devoid of the PB1 domain [10,11]. These isoforms were recently shown to exhibit different properties in aggregation in a cell line-specific manner, with p62-H2 forming larger aggregates generally [11]. The PB1 domain is deemed to be critical for aggregation of p62. The observation that the p62-H2 isoform lacking an intact PB1 domain could still form aggregates in a specific cell type may suggest that its aggregation can also be influenced by other factors that exist in a context or cell type-dependent manner. More details on how p62 aggregation can be regulated are provided in the section on how p62 is regulated.

p62 as a selective autophagy receptor
The physiological role of p62 in selective autophagy was first discovered in the context of facilitating clearance of protein aggregates, a process known as aggrephagy [1,2,12]. Apart from aggrephagy, p62 is also involved in facilitating autophagic clearance of damaged mitochondria (mitophagy), lipid droplets (lipophagy), and pathogens (xenophagy) based on in vitro and in vivo studies using mouse models [13][14][15][16]. The specificity of p62 in recognizing these substrates is dictated primarily by their K63-linked poly-ubiquitin chains [17,18]. Recently, specific protein sequences in the substrates such as the type 1 and type 2N-terminal degrons (N-degron) were reported to render them to be specifically recruited by p62 for autophagic degradation [19][20][21][22][23]. p62 also binds directly to autophagylinked FYVE protein (ALFY), an adaptor protein that   High level of p62 observed in NASH and NAFLD patients is associated with disease progression to fibrosis. [156,157] Alcoholic steatohepatitis (ASH) Accumulation of p62 observed in ASH patients is associated with disease progression to fibrosis. [157] Primary biliary cirrhosis (PBC) Aggregation of p62 was observed in the inflamed damaged small bile ducts of PBC patients. [157,158] Copper storage and intoxication diseases, e.g., Wilson disease (WD) and idiopathic copper toxicosis (ICT) High level of p62 observed in WD and ICT patients and is associated with the formation of intracellular hyaline bodies (IHBs). [97,157,159] Cancer Hepatocellular carcinoma (HCC) Accumulation of p62 aggregates is frequently observed in HCC patients and correlates with poor prognosis. [29,100,157,160] Ovarian cancer Elevated levels of p62 are observed in ovarian cancer tissues and associated with resistance to cisplatin in ovarian cancer patients. [161,162] Pancreatic ductal adenocarcinoma (PDAC) Accumulation of p62 correlates with the maintenance of malignant status in PDAC patients. [30] Breast cancer p62 is highly expressed in cancerous breast tissues and correlates with poor prognosis. [163,164] Head and neck squamous cell carcinoma (HNSCC) Progressive accumulation of p62 is observed in HNSCC patients and associates with drug resistance and poor prognosis. [165,166] Neurodegenerative disease Amyotrophic lateral sclerosis (ALS) / Frontotemporal lobular degeneration (FTLD) Mutant form(s) of p62/SQSTM1 (e.g., P348L) associated with increased aggregation of p62, TDP-43, and SOD1.
[175] Skeletal and muscular disorders Paget's disease of bone (PDB) Mutation of p62/SQSTM1 resulting in mutant p62 in the PDB domain (P392L) is associated with the overactive state of osteoclasts in PDB. [176] Cardiovascular disease Genetic cardiomyopathies Accumulation of p62 protein aggregates associated with desmin and phospholamban cardiomyopathy. [177] is critical in mediating selective autophagy [24,25]. Binding of p62 with ALFY facilitates formation and selective autophagic degradation of p62-enriched protein aggregates. The role of p62 in selective autophagy is also facilitated by wild-type Huntingtin, a ubiquitously expressed protein that serves as a scaffold protein interacting with unc-51-like autophagy activating kinase (ULK1) and p62 to couple the initiation of autophagy to promote subsequent incorporation of the cargo into the autophagosome [15]. Interestingly, p62 was recently shown to recruit FIP200, a subunit of the ULK1 complex, to initiate autophagosome formation via the residues 326-380 of p62, which are referred to as the FIP200-interacting region (FIR) [26]. Binding of LC3 to p62 via its LIR, which is located within the FIR, was found to displace FIP200 in an in vitro system [26], suggesting that the FIP200-containing autophagy initiation complex might be released from binding to p62 when it is being incorporated into the autophagosome.
p62 as an activator of Nrf2-mediated antioxidant signaling and metabolic reprogramming In the aggregated state, p62 recruits and sequesters Keap1, an adaptor of the cullin-3 (CUL3) E3 ubiquitin ligase complex, that regulates protein stability of nuclear factor erythroid 2-related factor 2 (Nrf2), a Sequestered by p62 bodies, resulting in stabilization of Nrf2 and activation of its antioxidant response and metabolic reprogramming.
KIR [27][28][29] Raptor Mediates activation of mTORC1 downstream targets in response to amino acid stimulation in the presence of p62.
LB [31] TRAF6 Catalyzes K63-linked poly-ubiquitylation of mTOR to enhance activation of mTORC1 signaling in the presence of p62.
TB [32] Caspase-8 p62 promotes aggregation of poly-ubiquitylated pro-caspase-8, allowing it to undergo self-cleavage and achieve full activation to drive cells for commitment to apoptosis.
ZZ [34,42] transcription factor that can activate genes involved in cellular defense against oxidative stress [27]. Sequestration of Keap1 by p62 releases and uncouples Nrf2 from CUL3-Keap1-mediated ubiquitin-proteasome system (UPS) regulation. This leads to stabilization of Nrf2, which then translocates to the nucleus to activate transcription of genes encoding antioxidant proteins and detoxifying enzymes, including heme oxygenase 1 (HMOX1) and NAD(P)H quinone dehydrogenase 1 (NQO1), as well as enzymes in the glutathione synthesis and glucose and glutamine metabolism to induce metabolic reprogramming [27][28][29]. While this can protect cells against oxidative and metabolic stresses, hyperactivation of Nrf2 is thought to confer fitness advantage to cancer-initiating cells to promote tumorigenesis in the liver and pancreas in the mouse models [5,30].
p62 as a critical mediator of nutrient sensing by activating the mTORC1 signaling Under condition of high amino acid abundance, p62 binds to the raptor and mechanistic target of rapamycin (mTOR), which are integral components of the mTOR complex 1 (mTORC1) that regulates metabolism and protein synthesis to drive cell growth and proliferation [31]. In the in vitro cell model, p62 was found to be necessary for mediating amino acid-induced activation of downstream targets of mTORC1, including the ribosomal protein S6 kinase (S6K1) and the eukaryotic translation initiation factor 4E-binding protein 1 (4EBP1), thereby acting as a mediator of nutrient sensing. In response to high abundance of amino acids, p62 also recruits TRAF6 to the mTORC1 complex where it catalyzes K63-linked poly-ubiquitylation of mTOR to further enhance activation of mTORC1 signaling [32].
p62 as a mediator of canonical NF-kB signaling Nuclear factor kappa-B (NF-kB) signaling complex, which consists of a group of signaling proteins, including IKKa/b (the catalytic subunits), IKKc (the regulatory subunit, also known as NF-jB essential modifier, NEMO), p65, and p50 transcription factors, is a master regulator of inflammation and immune response and a critical determinant of cell survival. Under the basal state, the transcription factors p65 and p50 are kept inactive by tight association with inhibitory jB (IjB) in the cytoplasm, which, however, will be released on phosphorylation by IKKa/b in response to inflammatory signals, leading to degradation of IjB through the UPS, resulting in release of p65/p50 to mediate the activation of NF-kB signaling [33]. p62 regulates NF-kB through multiple mechanisms. p62 acts as a scaffold to link the atypical protein kinase C (aPKC) to the death domain serine/threonine kinase receptor-interacting protein (RIP). RIP then binds to TNFR1-associated death domain protein (TRADD) and mediates TNF-a-induced activation of NF-kB signaling by activating transforming growth factor betaactivated kinase 1 (TAK1) which phosphorylates and activates IKKb [34]. The aPKC also connects IKKb with p62, thereby establishing a signaling cascade of interactions involving TRADD/RIP/p62/aPKC/IKKb to induce NF-kB signaling [34]. In addition, p62 also mediates activation of NF-kB signaling through facilitating poly-ubiquitylation of TRAF6, which is critical for the assembly of signaling proteins in the NF-kB pathway such as TAK1 [35,36]. In the mouse model of lung cancer, p62 is required for Ras-mediated activation of NF-kB signaling to promote cell survival during oncogenic transformation by inducing transcription of genes encoding antioxidant proteins such as superoxide dismutase (SOD) to reduce reactive oxygen species (ROS) and counteract oxidative stress [36]. Consequently, deficiency of p62 leads to elevation of ROS, which induces cell death which help reduce the cancerinitiating cells, thereby impairing Ras-induced oncogenic transformation and lung tumorigenesis.
p62 as a facilitator of caspase-8-mediated apoptosis and a regulator of necroptosis signaling Caspase-8 is an initiator caspase that mediates cleavage of executioner caspase-3 and À7 in extrinsic apoptosis signaling (in "type I" cells), as well as Bid into the activated form of Bid, tBid (in "type II" cells) to trigger Bax-/Bak-dependent mitochondrial outer membrane permeabilization (MOMP) in intrinsic apoptosis signaling [37]. Interestingly, initiation of caspase-8 activation requires it to first undergo CUL3-mediated polyubiquitylation [38]. p62 recognizes and promotes aggregation of poly-ubiquitylated pro-caspase-8, allowing it to undergo self-cleavage to achieve full activation. Furthermore, p62, together with the ATG8 conjugation system made up of the ATG5/7/16L complex, was shown to promote formation of intracellular death-inducing signaling complex (iDISC) on autophagosomal membranes [39][40][41]. The autophagosomal membranes are thought to provide a platform to bring the polyubiquitylated pro-caspase-8 oligomers in close proximity, thereby facilitating their maturation into the active form [39]. In the mouse model, genetic ablation of MAP3K7 which encodes for TAK1, facilitates p62 recruitment of receptor-interacting protein kinase 1 (RIPK1) via its ZZ domain to mediate assembly of necrosomes to the autophagosomal membranes [42]. Ths primes cells for necroptosis instead of apoptosis upon stimulation by TNF-related apoptosis-inducing ligand (TRAIL). Interestingly, this necroptosis signaling mechanism can be re-routed to apoptosis in the absence of p62.
p62 as a negative regulator of inflammation NOD-, LRR-and Pyrin domain-containing Protein 3 (NLRP3) and Absent In Melanoma 2 (AIM2) are intrinsic cellular sensors that detect a wide range of pathogen and damage-associated cues and signals [43]. Formation of the NLRP3 or AIM2 inflammasome promotes caspase-1-dependent processing of the proinflammatory cytokines (IL-1b and IL-18) and gasdermin D to induce inflammation and pyroptotic cell death, respectively. These inflammasomes undergo ubiquitylation and recruit p62 which facilitates their clearance through the autophagy-lysosome pathway [44]. p62, therefore, acts as a negative regulator of the NLRP3 or AIM2 inflammasome. Furthermore, upon stimulation of macrophages by NLRP3 agonists, p62 is upregulated by NF-kB signaling to mediate mitophagy to remove damaged mitochondria to stop them from releasing damage-associated molecular patterns (DAMPs) which can activate the NLRP3 inflammasome, thereby preventing excessive activation of the NLRP3 inflammasome in macrophages [45].

How p62 is regulated?
Multiple mechanisms, including transcriptional and post-transcriptional control, post-translational modifications, non-covalent interactions, and autophagic degradation, have all been reported to modulate p62 signaling activities. As the ability of p62 to form aggregates is key to recruit its binding partners, such as Keap1 and ubiquitylated substrates, to activate the downstream signaling events, factors that regulate its aggregation can, therefore, serve as an efficient way for facilitating rapid activation or inactivation. While degradation through the autophagy-lysosome pathway is generally regarded as a primary mechanism by which p62 and its associated protein complexes are being inactivated, recent evidence suggests p62 activity can also be suppressed in an autophagy-independent manner through direct disruption of its aggregation.

Post-translational modifications (PTMs)
At the basal state, when p62 is not involved in selective autophagy, p62 is being kept inactive through homodimerization of its UBA domain, thereby preventing it from binding to ubiquitylated cargo [4,54,55]. Upon induction of autophagy signaling, p62 is phosphorylated at its serine 407 (S407) by ULK1, resulting in disruption of the dimerization of the UBA domain and allowing it to bind to ubiquitylated substrates to subsequently deliver them for autophagic degradation [56]. The S407 phosphorylation of p62 is usually ensued by phosphorylation at S403 catalyzed by either ULK1 and casein kinase 2 (CK2) or TANKbinding kinase 1 (TBK1), which can further enhance recruitment of ubiquitylated cargo to p62 [57][58][59]. p62 can also be phosphorylated at its S349 residue in the KIR domain in a manner dependent on mTORC1, leading to enhanced ability to sequester Keap1 to activate the Nrf2-mediated antioxidant response [14]. Moreover, phosphorylation of p62 at its threonine 269 (T269) and serine 272 (S272) residues by a complex cascade consisting of the mitogen-activated protein kinase 3/6 (MEK3/6), p38delta, and the PB1containing kinase MEKK3 promotes its interaction with TRAF6, which can ubiquitylate mTOR, leading to activation of mTORC1 signaling that drives cell proliferation [60]. Recently, oxidative stress was shown to promote phosphorylation of p62 at serine 28 (S28) by fructokinase A (KHK-A) kinase, which is expressed abundantly in hepatocellular carcinoma (HCC) cells [61]. Phosphorylation at S28 inhibits ubiquitylation of p62 at its K7 residue [61], which is previously known to be catalyzed by the Tripartite Motif Containing Protein 21 (TRIM21) E3 ubiquitin ligase [62]. As the K7 residue is critical for establishing hydrogen bond, with D69 in the PB1 domain to mediate oligomerization of p62 which is critical for initiating aggregation [7,8], ubiquitylation of K7 by TRIM21 can, therefore, inhibit p62 aggregation and attenuate its function in activating the Nrf2 signaling [62]. Conversely, inhibition of K7 ubiquitylation through phosphorylation of p62 at S28 would result in enhanced aggregation of p62 and its sequestration of Keap1, thereby promoting Nrf2 activation. Furthermore, Keap1, which serves as an adaptor for the CUL-3 E3 ubiquitin ligase complex, promotes ubiquitylation of p62 at the lysine 420 (K420) residue within the UBA domain [63]. Overexpression of Keap1/Cullin3 elevates formation of inclusion bodies enriched in p62 and ubiquitylated proteins and promotes recruitment of LC3 to p62, resulting in enhanced autophagic degradation of p62 inclusion bodies and alleviating proteotoxicity. p62 can also be ubiquitylated in a manner dependent on UBE2D2 or UBE2D3 E2 conjugating enzyme in a condition known as ubiquitin (Ub)+ stress, where excessive free Ub and ubiquitylated substrates are present due to proteasomal inhibition or other stresses such as heat shock [64]. This E2-mediated ubiquitylation of p62, which occurs at multiple lysine residues including the K420, can disrupt dimerization of the UBA domain of p62, allowing it to recruit a ubiquitylated substrate for autophagic clearance. p62 has been reported to be ubiquitylated at lysine 91 (K91) and lysine 189 (K189) by the RNF166, an E3 ubiquitin ligase that is critical for mediating xenophagy [16]. However, it remains unclear how ubiquitylation of these residues alters p62 ability in forming aggregates or recruiting ubiquitylated cargo. Interestingly, p62 is also subjected to other types of post-translational modifications, such as oxidation and acetylation. Oxidation of p62 in the region between the PB1 and the adjacent ZZ domains promotes disulfide bond formation between p62 monomers, facilitating their dimerization and aggregation [65]. During nutrient starvation, p62 is acetylated primarily at K420 and lysine 435 (K435) by the histone acetyltransferase TIP60, which can be reversed by HDAC6 deacetylase [66]. Acetylation of p62 at these sites is necessary for aggregation and delivery of ubiquitylated substrates for autophagic clearance. In HCC cells, p62 was reported to be acetylated at K295 by the acetyltransferase GCN5, which is antagonized by Sirt1 deacetylase [67]. The K295 acetylation of p62 reduces its protein stability through enhanced interaction with Keap1 that can ubiquitylate p62 and target it for proteasomal degradation. Sirt1 can, therefore, uncouple p62 from Keap1-mediated UPS regulation.

Interference of self-oligomerization of p62 inhibits its activity
While numerous positive factors that can promote signaling activity and function of p62 have been reported, such as PTMs and LLPS described above, little is known on the negative regulation of p62. It is generally thought that p62 signaling is attenuated through degradation of p62 bodies and protein complexes via the autophagy-lysosome pathway. Remarkably, vault RNAs (vtRNAs), which are small non-coding RNAs transcribed by RNA polymerase III found in most eukaryotes, were recently found to bind to p62 via its ZZ domain to interfere with its self-oligomerization and interaction with autophagy effector proteins including LC3. During nutrient starvation, vtRNA binding from p62 is reduced, thereby releasing p62 to facilitate autophagy [77]. More recently, Modulator-of-apoptosis1 (MOAP-1) was recently reported to exhibit disaggregase-like activity for dissociating p62 bodies which results in the suppression of Nrf2 signaling [73].
Originally identified as a low-abundance Bax-binding protein that is tightly regulated by the UPS, MOAP-1 facilitates apoptosis signaling through promoting Baxdependent apoptotic function via activating the binding function of the mitochondria receptor, MTCH2, for tBid [78][79][80][81][82][83]. Interestingly, upon induction of cellular stresses that stimulate formation of p62 bodies in cells, MOAP-1 is recruited to the p62 bodies to reduce their levels independent of the autophagy pathway [73]. Mechanistically, MOAP-1 disassembles p62 bodies by binding to the N-terminal region of p62 containing the PB1 and ZZ domains, thereby interfering with its ability to stay in an oligomeric state, thereby resulting in destabilizing of the p62 bodies. Point mutants of MOAP-1 with exquisite function in regulating apoptosis or p62 bodies were identified, suggesting that functions of MOAP-1 in regulating apoptosis and p62 bodies might be separable [73]. Moreover, MOAP-1 was recently found to associate with LC3 and promote closure of the phagophore, thereby facilitating autophagy in the context of starvation mediated by nutrient withdrawal under EBSS condition [84]. Although disassembly of p62 bodies by MOAP-1 does not appear to depend on autophagy [73], it remains possible that association of MOAP-1 with LC3 may subsequently aid clearance of the remnants from the disassembled p62 bodies.
The multi-faceted roles of p62 in liver diseases Viral hepatitis B or C infection, obesity, and type 2 diabetes are among the common risk factors for promoting development of certain liver diseases, including non-alcoholic fatty liver disease (NAFLD) and hepatocellular carcinoma (HCC). HCC is the primary form of liver cancer, which is the fourth most common cause of cancer death worldwide and with a 5-year survival rate of about 18% [85,86]. Through comprehensive and integrative next-generation sequencing effort, more than 28 000 genetic mutations have been identified in the HCC patients, with the most frequently mutated loci identified in the TERT promoter and coding regions of TP53, CTNNB1, ALB, KEAP1, and NFE2L2 [87,88]. Despite the advancement in genetic profiling, treatment of HCC remains highly unfavorable due to limited therapeutic options and suitable biomarkers to aid prognosis and measure responses to the treatment. NAFLD affects 25% of the global population and is predicted to be the most common risk factor for HCC by 2030 [89]. NAFLD is characterized by a set of progressive liver abnormalities starting from steatosis, which can further advance to non-alcoholic steatohepatitis (NASH), cirrhosis, liver failure, and even HCC [90][91][92][93][94][95]. There is currently no effective treatment for NAFLD and NASH, which warrants urgent identification of new therapeutic targets for drug development.
Prior to the seminal discovery of its role in autophagy [1,2], p62 was known as a critical component of protein aggregates in the forms of Mallory-Denk bodies or intracellular hyaline granules which were frequently detected in the livers of patients suffering from HCC, NASH, alcoholic steatohepatitis (ASH), and several types of chronic liver diseases [96,97]. Despite emerging evidence of p62 as a hallmark of these liver diseases, the role of p62 in contributing to their pathogenesis was poorly understood, and the complexity is further compounded by the cell type-and contextspecific functions of p62 in the liver. In this section, we will review the multifarious pathophysiological roles of p62 in the liver by paying special attention to findings from the clinical and preclinical models, especially the mouse models that recapitulate the disease progression and manifestation in humans.
p62 promotes HCC initiation by activating Nrf2, mTORC1, and c-MYC pro-oncogenic pathways in hepatocytes HCC development involves a complex multi-step process which begins with hepatocytes acquiring driver mutations to disrupt cell cycle control [98]. With heightened propensity to proliferate at the expense of the metabolic equilibrium, these cancer-initiating cells are primed for cell death due to oxidative and metabolic stresses [5]. However, mutations in the NFE2L2 and KEAP1 that lead to activation of the Nrf2mediated antioxidant signaling and metabolic reprogramming can potentially promote survival of the HCC-initiating cells to actively facilitate cancer development. Apart from these genetic alterations, HCCinitiating cells may acquire alternative fitness mechanism via accumulation of p62 aggregates that can sequester Keap1, leading to stabilization of Nrf2 and further upregulation of Nrf2 signaling. In support of this notion, increased levels of p62 inclusion bodies are frequently detected in the livers from HCC patients and correlated with the tumor grades, recurrence, and disease-free survival [99][100][101]. In HCC biopsies, heightened phosphorylation of p62 at S28 and S349, which can promote aggregation and ability of p62 to sequester Keap1, respectively, are often observed in concomitance with hyperactivation of Nrf2 signaling [14,29,61]. As a proof of concept to substantiate the critical role of p62 in driving HCC development, liver-specific KO of p62 was demonstrated to attenuate tumorigenesis in several mouse models of HCC [100,102]. Furthermore, ectopic expression of wild-type or UBA-deleted (DUBA) mutant of p62, but not its Keap1-binding defective (DKIR) mutant, is sufficient for driving HCC formation in wild-type mice [100]. Thus, p62 function in activating Nrf2 signaling through sequestering Keap1, but not mediating autophagy, appears to be critical for driving HCC development. However, as constitutive activation of Nrf2 in the liver does not induce spontaneous tumors in mice per se [103], it is conceivable that p62 promotes HCC development through additional mechanisms. Analysis of TCGA and Oncomine HCC databases reveal intimate correlation between p62 and mTORC1-and c-MYCresponsive genes [100]. In the liver-specific tuberous sclerosis 1 (Tsc1), KO mice that exhibit constitutive activation of the mTORC1 signaling and develop spontaneous HCC and ablation of p62 reduces mTORC1 activity and inhibits HCC development [100,104]. Furthermore, mTORC1-dependent phosphorylation of p62 can enhance its function in activating the Nrf2 signaling [14], suggesting that a positive feedback loop exists between p62 and mTORC1 to synergistically promote hyperactivation of Nrf2 (Fig. 2). Apart from Nrf2, p62 also promotes the activity of the c-MYC oncogene during HCC development [100]. c-MYC, which is essential for tumor growth by promoting cell proliferation and metabolic reprogramming, is often upregulated in HCC [105][106][107]. The oncogenic function of c-MYC appears to be highly dependent on the mTORC1 signaling [108]. The mTORC1-c-Myc signaling axis drives hepatocarcinogenesis at least, in part, by upregulating the solute carrier (SLC) transporters, including SLC1A5 and SLC7A6, to enhance uptake of essential amino acids such as glutamine.
Defective autophagy drives liver tumorigenesis in mouse models and has been associated with increased HCC risk in the patients [102,109,110]. While p62 and Nrf2 have been demonstrated to be required for mediating hepatocarcinogenesis in the liver-specific autophagy-deficient mice, several other proteins, such as  Fig. 2. p62 mediates activation of Nrf2, mTORC1, and c-MYC pathways in the liver to enhance antioxidant defense, metabolic reprogramming, cell growth, and proliferation to promote initiation and development of hepatocellular carcinoma (HCC). p62 recruits and sequesters Keap1, thereby releasing Nrf2 from CUL3-Keap1-mediated ubiquitin-proteasome system (UPS) regulation. This leads to stabilization of Nrf2, which activates the antioxidant and metabolic reprogramming to confer fitness advantage to HCC-initiating cells. p62 also binds to the mechanistic target of rapamycin (mTOR) and mediates its activation under nutrient-rich condition. mTORC1-dependent phosphorylation of p62 enhances its function in activating Nrf2 signaling, creating a positive feedback loop between p62 and mTORC1 to hyperactivate Nrf2. Furthermore, during HCC development, p62 also promotes the activity of the c-MYC oncogene, which is essential for tumor growth by promoting cell proliferation and metabolic reprogramming. The oncogenic function of c-MYC appears to be highly dependent on the mTORC1 signaling. Inset: elevated level of p62 aggregates, which display enhanced capability in sequestering Keap1 to activate the Nrf2 pathway, is a hallmark of liver cancer in patients. Tumor biopsy from liver cancer patients was immunostained with anti-p62 (in red) antibody and visualized on a TissueFlex imaging system. Nuclei were counterstained with DAPI (in blue). Scale bar, 20 lm. p62-containing inclusion bodies were indicated by white arrows. Diagram in the left panel is adapted from [100].
yes-associated protein (YAP), transcriptional coactivator with PDZ-binding motif (TAZ), and phosphatase and tensin homolog (PTEN), were recently reported to play a critical role as well [111,112] YAP, an effector of the Hippo signaling pathway that promotes cell growth and proliferation, was identified to be regulated by autophagy [111]. Deletion of YAP in the liver-specific Atg7 KO mice inhibits liver tumorigenesis. In the Atg78/ YAP DKO livers, however, the p62-Nrf2 signaling axis is maintained, supporting a parallel role of YAP in driving liver tumorigenesis induced by autophagy deficiency. Furthermore, accumulation of YAP and TAZ in autophagy-deficient hepatocytes induces the hepatocyte to dedifferentiate into biliary-like liver progenitor cells (LPCs). These LPCs then promote the development of HCC in autophagy-deficient hepatocytes. Loss of PTEN, a tumor suppressor gene, accelerates the development of HCC in autophagy-deficient livers with the hepatomegaly phenotype [112].

p62 inhibits fibrosis via suppressing activation of hepatic stellate cells
Liver fibrosis occurs as a result of chronic liver injury and is characterized by extensive deposit of extracellular matrix (ECM) proteins, especially collagens, by the hepatic stellate cells (HSCs) that reside in the stroma of the liver to replace the damaged tissue [113,114]. Excessive activation of the HSCs can exacerbate fibrosis and promote HCC through secretion of proinflammatory cytokines [115,116]. Interestingly, in human HCC, while p62 expression is elevated in the hepatocytes, it is downregulated in the HSCs [117]. Whole body or HSCspecific ablation of p62 was found to potentiate activation of HSCs, leading to increased liver fibrosis and inflammation, thereby accelerating HCC progression in the diethylnitrosamine (DEN) coupled to a high-fat diet (HFD) mouse model [117]. Mechanistically, p62 directly binds to the vitamin D receptor (VDR) and retinoid acid receptor-alpha (RXRa) to facilitate their heterodimerization, which is crucial for the activation of the VDR: RXR target genes that can suppress HSC activation. Thus, paradoxical to the tumor-promoting role of p62 in the liver parenchyma, p62 appears to exert tumorsuppressive function in the stroma of the liver by suppressing activation of HSCs (Fig. 3).
The complex role of p62 in the development of NAFLD NAFLD includes a spectrum of liver abnormalities that are initiated primarily by hepatic fat accumulation, a condition known as steatosis or fatty liver. Steatosis is triggered by excessive accumulation of triglyceride (TG) in the hepatocytes as a result of imbalance between lipid acquisition (e.g. increased uptake and de novo lipogenesis) and clearance (e.g. defective lipolysis and lipophagy) [118]. Although TG accumulation was initially thought to induce oxidative stress and lipid peroxidation leading to lipotoxicity that can, in turn, trigger inflammation and fibrosis, recent evidence from animal models challenges this notion by demonstrating that TG accumulation might, in fact, be a protective mechanism to counteract lipotoxicity [119][120][121]. Other non-triglyceride fatty acid metabolites, such as saturated free fatty acids (FFAs) and ceramides, also have a role in inducing lipotoxicity by activating the c-Jun N-terminal kinase (JNK) and mitochondrial apoptosis signaling pathways to trigger the endoplasmic reticulum (ER) and oxidative stresses leading to hepatocyte death and inflammation RXR p62 VDR Activation of hepatic stellate cells Fig. 3. p62 inhibits fibrosis via suppressing activation of hepatic stellate cells (HSCs). Whole body or HSC-specific ablation of p62 potentiates activation of HSCs, leading to increased liver fibrosis, inflammation, and accelerated HCC progression. In HSCs, p62 binds to the vitamin D receptor (VDR) and retinoid acid receptoralpha (RXRa) to facilitate their heterodimerization, which is crucial for the activation of their target genes to result in suppression of HSC activation to inhibit fibrosis. Diagram adapted from [117]. [121,122]. However, as TG accumulation always appears as a benign bystander during lipotoxicity induced by dysregulated lipid metabolism [122], formation of TGenriched lipid droplets, a phenomenon referred to as steatosis, remains a key marker in the study of NAFLD. More biomarkers that can distinguish benign steatosis from NASH are needed for better stratification and understanding of the different stages of NAFLD.
Upregulation of p62 is frequently detected in the livers of patients and mouse models of NASH [100,123]. In these livers, due to the chronic state of oxidative and ER stresses induced by excessive lipid levels, p62 is upregulated as part of the stress response mechanism [100,124]. While genetic ablation of p62 in the liver reduces fibrosis, oxidative stress, and progression to HCC in the MUP-uPA mice fed with a HFD which exhibit NASH prior to HCC development, p62 deficiency does not prevent steatosis in this context [100]. In the NASH-HCC model using streptozocin (STZ) to induce diabetic background in the mice before being subjected to HFD, liver-specific KO of p62 inhibits liver tumorigenesis, but not steatosis or fibrosis [100]. In another study, however, liver-specific p62 KO was found to elevate lipotoxicity in mice fed for a prolonged period on a HFD and subjected to acute treatment regime involving overnight fasting followed by refeeding with a high-carbohydrate and fat-free diet [124]. Mechanistically, p62 promotes phosphorylation of ULK1 by facilitating its interaction with AMPactivated protein kinase (AMPK) and ULK1, thereby activating autophagy and Keap1 degradation that would lead to activation of Nrf2. Induction of Nrf2 signaling is known to alleviate NAFLD and NASH through multiple mechanisms, including antioxidant response and suppression of metabolic enzymes involved in fatty acid synthesis [5,125]. Unlike liverspecific p62 KO mice, whole body p62 knockout (KO) mice were found to develop steatosis spontaneously even under normal chow diet from 5-month-old onwards [126][127][128]. These p62-deficient mice were also found to develop fibrosis. Interestingly, the p62 and Nrf2 double KO mice develop fibrosis at a much higher degree than the p62 KO mice [126]. In line with this, p62 and Nrf2 were reported to confer synergistic effect on protecting Hepa1c1c7 murine hepatoma cells against lipotoxicity triggered by a high exogenous level of palmitic acid, a common FFA [129]. Therefore, depending on the contexts, p62 and Nrf2 appear to exert synergistic effect on protecting hepatocytes from lipotoxicity and development of NAFLD.
While it is intuitive to conceive that in certain contexts, p62 would exert a beneficiary effect on alleviating steatosis and lipotoxicity, p62-containing inclusion bodies are, however, thought to contribute to the pathological progression of NAFLD to NASH and HCC (Fig. 4). The abundance of Mallory-Denk bodies, which are enriched in p62, has been proposed to be a marker that differentiates benign steatosis from NASH in the clinic [130][131][132]. The level of p62 inclusion bodies is also associated with the M1 polarization of macrophages (which are proinflammatory) in the NAFLD patients [133]. In both in vitro and in vivo experimental models, formation of these p62 bodies can be potently induced in the HepG2 hepatoma cells by treatment with palmitic acid, or in livers from mice fed on a HFD for 3 months [134]. Interestingly, formation of p62 bodies under these conditions requires its phosphorylation at S403 by Tank1 binding kinase 1 (TBK1), which is activated by the cyclic GMP-AMP synthase (cGAS) and the stimulator of interferon genes (STING) pathway. Genetic ablation or inhibition of TBK1 by the small-molecule compound BX795 suppresses the formation of p62 bodies in HepG2 and livers of mice fed on a HFD. While BX795-mediated inhibition of TBK1 or liver-specific TBK1 KO does not seem to alleviate steatosis and liver injury, it dramatically reduced fibrosis and oxidative stress in the livers from HFD-fed mice [134]. The mRNA and protein levels of p62 was shown to be positively regulated by an immune surveillance protein DDX58/Rig-1 (DExD/H box helicase 58) [135]. Overexpression of DDX58 stimulates autophagy and prevents the accumulation of p62 bodies induced by palmitic acids, thereby reducing lipotoxicity and inflammation induced by excessive levels of FFAs.
Roles of p62 in chronic liver diseases via control of obesity, insulin, and leptin signaling NAFLD is commonly diagnosed in 80% of obese people and~50-60% of type 2 diabetes mellitus (T2DM) patients [136,137]. Obesity can be a consequence of over-nutrition and/or genetic predisposition that leads to defects in the control of whole body and cellular metabolism [138]. In the patients suffering from NAFLD, obesity, and T2DM, perturbation in the p62 expression has been reported in the liver, adipose tissues, beta-pancreatic cells, and other organs involved in controlling systemic metabolism, as reviewed previously [139]. p62 is normally expressed in human betapancreatic cells, but was found to be expressed only focally, very weakly, or even absent in pancreatic pathologies, including insulinomas, glucagonomas, non-functioning pancreatic neuroendocrine tumors, or carcinomas [140]. Interestingly, whole body p62 KO mice were found to develop mature-onset obesity and insulin resistance under normal chow diet [127,141]. In these p62-deficient mice, hyperactivation of the ERK pathway was observed in white adipose tissue (WAT), a specialized lipid storage tissue. Enhanced differentiation of p62-deficient white adipocyte was also detected in comparison to the wild-type counterpart in vitro, lending credence to the idea that an inhibitory role of p62 in adipogenesis is through suppression of ERK activity [127]. Moreover, through systematic analysis of organspecific p62 KO mice, p62 was also found to regulate energy metabolism via facilitating p38 MAPKmediated mitochondrial function and thermogenesis in brown adipose tissue (BAT), which serves a specialized function to convert energy into heat [128]. Like total p62 KO mice, adipocyte-specific p62 KO mice develop spontaneous steatosis under chow diet [128].
Insulin resistance, which is a hallmark for T2DM and can be a contributing factor for the development of NAFLD, can be caused by multiple factors, including defects in the insulin signal transduction pathway, chronic excess of metabolic fuels such as glucose and lipids, and hyperactivation of inflammatory and ER stress pathways [142]. Upon stimulation by insulin, p62 interacts avidly with insulin receptor substrate-1 (IRS-1), which is an adaptor protein of the insulin receptor that mediates insulin signaling in the target cells [143]. Depletion of p62 reduces activation of the insulin-mediated downstream signaling events, including activation of Akt, GLUT4 translocation, and glucose uptake. In vivo evidence to support a central role of p62 in regulating insulin signaling using liverspecific p62 KO mice is currently lacking. In a study using liver-specific knockdown of growth factor receptor binding protein 14 (Grb14) that leads to enhanced insulin signaling, p62 was reported to mediate activation of Nrf2 which, in turn, represses the lipogenic nuclear liver X receptor (LXR)-mediated de novo lipogenesis [144]. Young pre-obese p62 KO mice exhibit leptin resistance and hyperphagia, characterized by increased appetite and eating behavior [145]. p62 is expressed at an abundant level in the hypothalamic region of the brain which is the site of leptin action. In the p62-deficient hypothalamic neurons, phosphorylation and nuclear translocation of STAT3 appear to be impaired, thereby impairing leptin signaling [145]. Altogether, apart from the roles of p62 in facilitating , is a hallmark of NASH. Liver sections of C57BL/6 mouse subjected to the STAM model of NASH as previously described [154] were subjected to staining with H&E (upper panel) and immunohistochemistry using anti-p62 antibody (lower panel) and were visualized on a TissueFlex imaging system. Nuclei were counterstained with DAPI (in blue). Scale bar, 100 lm. Ballooned hepatocytes with Mallory-Denk bodies (red arrow) and infiltration of immune cells (black arrow), which are key features of NASH, were detected in the liver of mice subjected to the NASH model of STAM. p62 inclusion bodies were indicated by white arrows.
storage and expenditure of energy in WAT and BAT, respectively, p62 can also play a role in regulating obesity and insulin resistance via direct control of insulin and leptin signaling pathways.
Is p62 a viable therapeutic target for liver diseases?
Based on the findings from numerous clinical and preclinical studies, p62 appears to play multi-faceted roles in the development of chronic liver diseases, such as NAFLD and HCC, with conflicting conclusions in distinct contexts. While the conundrum could stem from the tissue and cell type-specific functions of p62, it may also be due to the differences in the genetic background of the mouse strains, treatment regime, and other experimental conditions used in the studies that can influence the outcome. For NAFLD, further investigation is still required to fully comprehend the roles of p62 during its pathogenesis and progression to NASH. It is possible that the integrative effect of p62 on development of NAFLD is dictated or largely influenced by the states of p62, whether it exists in a soluble or insoluble, aggregated form, its post-translational modifications such as phosphorylation and ubiquitylation, or its interaction with effector proteins involved in autophagy such as ULK1, FIP200, Huntingtin, or mediators, such as Keap1, TRAF6 and other yet-to-be uncovered players, that may impact on Nrf2, NF-kB, TGF-b and other critical signaling pathways. It is noteworthy to point out that the abundance of p62 inclusion bodies, i.e., in the form of Mallory-Denk bodies, has been put forward as a definitive marker to differentiate simple steatosis from NASH [130][131][132]. However, little is known on the p62 bodies in terms of their exact composition and biophysical states during development of NASH and HCC. These p62 inclusion bodies may represent biomolecular condensates that can exist either in the liquid, solid, or gel form. By determining the precise phase and composition constituting the p62 condensates present in NASH and HCC, it may yield further insights into the nature and dynamics of these aggregates and their relationship in many signaling pathways. Furthermore, as condensates have also been suggested to influence the drug's efficacy [146], it would be interesting to find out whether these p62 bodies could have an effect on the efficacy of the drugs used for the treatment of liver diseases such as HCC.

Is p62 druggable?
Structures of p62, including its PB1, ZZ, and UBA domains; binding interfaces with Keap1, ubiquitin, and arginylated proteins; as well as the oligomeric and filamentous forms that make up the p62 bodies, have been solved [8,27,54,[147][148][149]. These structural details would facilitate drug discovery endeavor to target the p62-associated signaling axis for the treatment of the liver diseases with distinct aberrant p62 activity. In an effort to identify smallmolecule compounds that can inhibit the interaction between S349-phosphorylated p62 and Keap1, N-[2acetonyl-4-(4-ethoxybenzenesulfonylamino) naphthalene-1-yl]-4-ethoxybenzenesulfonamide, which is named as K67, was identified from a fluorescence polarization-based high-throughput screen [29]. Another compound, named compound 16 (Cpd16), which was previously identified as an Nrf2 activator, has an analogous structure to K67 [29,150]. Both K67 and Cpd16 were found to exhibit comparable inhibitory effects on the p62/Keap1 interaction and Nrf2 signaling [29]. In the Huh1 liver cancer cells, which have a high basal level of S349-phosphorylated p62, treatment with K67 or Cpd16 reduces cell proliferation and viability, as well as elevating their sensitivity to sorafenib, which is the standard of care for advanced, unresectable HCC [29,151]. These inhibitory effects, however, were not observed in the liver cancer cells, Huh7, which exhibit a low level of S349 phosphorylation, suggesting the specificity of K67 and Cpd16 in suppressing liver cancer cell proliferation and resistance to anticancer drug by inhibiting the p62-Keap1-Nrf2 signaling axis. Using a virtual screening and molecular docking approach followed by validation with pull-down assays, two small-molecule compounds, termed XIE62-1004 and XIE2008, were identified to exhibit high affinity in binding to the ZZ domain of p62 [21]. These compounds induce aggregation of p62 and promote autophagosome biogenesis and degradation of p62-enriched mutant Huntingtin aggregates via the autophagy-lysosome pathway. Another compound, XRK3F2, was identified that can bind to the ZZ domain of p62 through a similar virtual screening approach [152]. Unlike XIE62-1004 and XIE2008, XRK3F2 inhibits p62 function as a signaling hub for NF-jB, p38MAPK, and JNK pathways, thereby suppressing osteoclastogenesis and multiple myeloma cell growth that are key contributors of multiple myeloma bone disease [152]. Although these ZZbinding compounds have been shown to be effective in activating or inhibiting p62 functions in mediating autophagy and other signaling mechanisms, they remain to be evaluated in other contexts, including liver cancer. Future work in identifying compounds that bind to specific p62 domains and alter distinct functions may hold promise for facilitating development of treatment for specific type of liver diseases, such as HCC and NAFLD.

Conclusion
Emerging evidence from preclinical studies revealed multi-faceted roles of p62 in regulating the development of chronic liver diseases, including NAFLD and HCC. In patients, the abundance of p62-enriched Mallory-Denk bodies correlates with onset of NASH. Furthermore, accumulation of p62 aggregates was frequently observed in HCC patients which also correlates with poor prognosis. These findings lend support to the idea that p62 bodies can potentially be used as a biomarker to grade severity of NAFLD and HCC. Whether p62 can be a therapeutic target for treating chronic liver diseases will require further insights from clinical and preclinical research using disease relevant models. It is envisaged that deeper understanding of the precise function of p62 in physiological and pathological contexts in distinct liver diseases will be required for paving the way forward to devise effective therapeutic strategies in targeting p62 for therapeutic purpose. It is encouraging to know that smallmolecule compounds that bind to specific domains of p62 are able to effect desirable function of p62 which certainly will facilitate development of potential therapeutic approaches for treating certain chronic liver diseases in future.

H&E staining and Immunohistochemistry
Formaldehyde fixed and paraffin embedded (FFPE) tissue section of human HCC biopsy was described previously [153]. FFPE tissue section of mouse NASH was prepared from C57BL/6 mouse subjected to the STAM paradigm as described [154]. The experimental procedure on mouse was approved by and conducted in accordance with the regulations and guidelines of the Institutional Animal Care and Use Committee of the National University of Singapore. The human and mouse FFPE sections were subjected to H&E staining and/or immunohistochemistry staining as previously described [73,155]. p62 was detected using mouse monoclonal antibody against p62/SQSTM1 (Abcam, ab56416). Images were acquired with a TissueFlex imaging system. Single focal plane images were shown.