Lysosomal quality control: molecular mechanisms and therapeutic implications

Abstract

Lysosomes are essential catabolic organelles with an acidic lumen and possessing dozens of hydrolytic enzymes. The detrimental consequences of lysosomal leakage have been well-known since lysosomes were discovered in 1950s. However, detailed knowledge of lysosomal quality control mechanisms has only emerged relatively recently. It is now clear that lysosomal leakage triggers multiple lysosomal quality control pathways that replace, remove, or directly repair damaged lysosomes. Here, we review how lysosomal damage is sensed and resolved in mammalian cells, with a focus on the molecular mechanisms underlying different lysosomal quality control pathways. We also discuss the clinical implications and therapeutic potential of these pathways.

Lysosomal quality control pathways

Lysosomes are essential catabolic organelles in animal cells. Originally known as the terminal compartments for endocytic and autophagic degradation, lysosomes are now also appreciated for their roles in nutrient sensing, growth signaling, and cellular stress handling, as well as innate and adaptive immunity [1]. These diverse functions are all dependent on the acidic luminal pH and dozens of hydrolytic enzymes in the lysosomal lumen. Thus, maintaining lysosomal integrity is critical for lysosomal activity.

When originally discovered by Christian de Duve, lysosomes were proposed as “suicide bags” of the cell due to their hydrolytic contents and potentially damaging impact on cells if lysosomal integrity was compromised [2, 3]. Since then, significant progress has been made in lysosomal leakage-induced cell death in various diseases, which typically involves lysosomal membrane permeabilization (LMP) (see Glossary) and subsequent cathepsin release [4]. LMP can be induced by many pathological conditions such as microbial infections, crystallization of uric acid or cholesterol, phagocytosis of particles, lipid oxidation, or lysosomal storage conditions [5, 6]. LMP-related cell death is associated with many human diseases, such as atherosclerosis, Alzheimer’s disease, Parkinson’s disease, stroke, and lysosomal storage diseases [4].

Despite the well-known detrimental consequences of lysosomal destabilization, little was known about lysosomal quality control mechanisms until recent years. From 2009 to 2022, multiple pathways have been discovered that replace, remove, or repair permeabilized lysosomes (Table 1, Key Table). These include lysosomal biogenesis and exocytosis by the transcription factor EB (TFEB)/TFE3 pathway [7, 8], lysosomal membrane turnover by lysophagy [9–11] or microautophagy [12–14], and direct lysosomal membrane repair by the endosomal sorting complex required for transport (ESCRT) [15–17] and the phosphoinositide-initiated membrane tethering and lipid transport (PITT) pathway [18]. In this review, we discuss how lysosomal damage or LMP is sensed and resolved in mammalian cells, with a focus on the molecular mechanisms of different lysosomal quality control pathways. We also discuss the clinical implications and therapeutic potential of these pathways.

Table 1.

Lysosomal quality control pathways

Strategy	Pathway	Mechanism of action	Key molecules		Macro-autophagy ?	ATG8 lipidation ?	Year of initial discovery	Ref
Biogenesis; Exocytosis	TFEB/TFE3	Transcriptional upregulation of lysosomal biogenesis and exocytosis	Ca2+, V-ATPase, ATG5/12/16, ATG8 (GABARAPs), mTORC1/RagC/RagD/folliculin, calcineurin/PP2A, TFEB/TFE3, TRPML1		No	Yes, in many cases	2009	[7, 8, 20–23, 25, 27, 28, 32, 34, 37–39, 45, 144]
Degradation	Lysophagy	Autophagic degradation of damaged lysosomes	Canonical	Ubiquitin ligases, ubiquitin, p62, TAX1BP1, ATGs	Yes	Yes	2013	[9, 10, 51, 137, 142]
			Noncanonical	V-ATPase, ATG5/12/16, ATG8 (LC3s/GABARAPs)	No	Yes	2019	[40, 41]
	Microautophagy	Selective degradation of lysosomal membrane proteins	Canonical	Ubiquitin ligases, ubiquitin, ESCRT	No	No	2015	[12, 13]
			Noncanonical	ATG5/12/16, ATG8	No	Yes	2020	[59]
Direct repair	ESCRT	Membrane repair by inward membrane budding	Ca2+, LRRK2, ESCRT		No	No	2018	[16, 17, 72]
	PITT	Membrane repair by direct lipid transfer	Ca2+, PI4K2A, PI4P, ORP9, ORP10, ORP11, OSBP, phosphatidylserine, cholesterol, ATG2A/B		No	No	2022	[18, 82]

TFEB/TFE3-dependent lysosomal biogenesis and exocytosis

In response to lysosomal destabilization, one general approach to maintaining lysosomal quality and activity is generating new lysosomes. Promoter analysis of lysosomal genes revealed a palindromic 10-base pair motif “GTCACGTGAC” as the coordinated lysosomal expression and regulation (CLEAR) element which mediates transcriptional upregulation of lysosomal biogenesis [7]. The CLEAR element is similar to sequences bound to the microphthalmia/transcription factor E (MiT/TFE) family of transcription factors (TFs) ─ MITF, TFEB, TFE3, and TFEC [19]. Among them, TFEB and TFE3 are the major transcription factors binding to CLEAR elements upon lysosomal stress to upregulate lysosomal biogenesis, as well as macroautophagy [7, 8, 20].

A major mechanism that regulates TFEB/TFE3 activity is their phosphorylation status which controls their subcellular localization (Fig. 1). Mammalian target of rapamycin complex 1 (mTORC1) is a master regulator of TFEB/TFE3 phosphorylation [8, 20–23]. In resting conditions, TFEB/TFE3 are phosphorylated at multiple serine/threonine residues. Phosphorylation of TFEB at S211 and TFE3 at S321 by mTORC1 promotes TFEB/TFE3 interaction with 14-3-3, which locks them in the cytosol [8, 20]. In response to multiple lysosomal stressors, TFEB/TFE3 are dephosphorylated and released from 14-3-3, undergoing nuclear translocation and transcriptional activation [7, 8, 20–22].

Figure 1. TFEB/TFE3-mediated lysosomal biogenesis and exocytosis.

TFEB and TFE3 are important transcription factors activated in response to lysosomal stress to upregulate lysosomal biogenesis and autophagy. TFEB/TFE3 are normally phosphorylated by mTORC1, leading to its sequestration by 14-3-3 protein in the cytosol. Lysosomal stress or nutrient starvation suppresses mTORC1 activity, triggering TFEB/TFE3 dephosphorylation and nuclear translocation. An emerging mechanism is the sensing of stress-induced lysosomal Ca²⁺ signaling by the proton pump V-ATPase which then recruits the ATG12/ATG5/ATG16 complex, leading to direct conjugation of ATG8 proteins on lysosomal membranes. Lysosomal conjugation of GABARAPs but not LC3 is essential for TFEB activation by recruiting and inactivating the folliculin (FLCN)/ folliculin interacting protein (FNIP) complex, which is normally required for RagC/D activation and TFEB/TFE3 phosphorylation. Lysosomal Ca²⁺ release also appears to activate TFEB/TFE3 through mechanisms independent of mTORC1 inhibition. This involves the Ca²⁺-dependent protein phosphatase calcineurin and likely also other phosphatases such as protein phosphatase 2 (PP2A) that dephosphorylate TFEB/TFE3. In the nucleus, TFEB/TFE3 bind to coordinated lysosomal expression and regulation (CLEAR) motifs to promote transcriptional upregulation of approximately 400 lysosomal and autophagy genes, leading to lysosomal biogenesis, autophagosome formation, and lysosomal exocytosis, the latter promoted by the upregulation of TRPML1, a lysosomal Ca²⁺ channel. Figure was created using  BioRender.

Loss of mTORC1 signaling due to lysosomal stress is sufficient to activate TFEB/TFE3 (Fig. 1). As mTORC1 activation depends on nutrient signaling on intact lysosomes, lysosomal stressors such as proton pump inhibition, lysosomal membrane damage or nutrient starvation are all known to inactivate mTORC1, resulting in TFEB/TFE3 activation [21, 24–26]. Upon mTORC1 inhibition by different pharmacological agents, TFEB is quickly dephosphorylated, whereupon it translocates to the nucleus [20]. Lysosomal mTORC1 activation requires nutrient signaling to the RAG GTPases and growth factor signaling to Rheb GTPase which is inhibited by GTPase-activating protein (GAP) tuberous sclerosis complex (TSC1/TSC2) when growth factors are absent [24]. However, Rheb is not required for mTORC1-mediated TFEB/TFE3 activation [27]. Indeed, in TSC deficient cells, TFEB is dephosphorylated even though mTORC1 is hyperactivated by Rheb [22]. The activation of TFEB/TFE3 requires the GDP-bound, active form of RagC/D, which directly recruits TFEB for its phosphorylation by mTORC1 [27, 28]. Folliculin, a specific GAP for RagC/D [29], is essential for TFEB recruitment to mTORC1 [27]. Loss of RagC/D or folliculin causes constitutive activation of TFEB/TFE3, whereas overexpression of constitutively active or dominant negative RagC/D locks TFEB/TFE3 in the cytosol or nucleus, respectively [20, 21, 27].

Lysosomal stress-stimulated TFEB/TFE3 activation may also involve active dephosphorylation by Ca²⁺-dependent protein phosphatases (Fig. 1). Lysosomal stress [30, 31] triggers lysosomal Ca²⁺ release [32] by transient receptor potential cation channel mucolipin subfamily member 1 (TRPML1), a major Ca²⁺ channel on lysosomes in most cell types [33]. Consistent with these observations, starvation-induced TFEB dephosphorylation is partially dependent on calcineurin, a Ca²⁺-activated phosphatase [32]. Additionally, TRPML1 agonists are sufficient to trigger TFEB dephosphorylation and nuclear translocation [32]. The underlying mechanism likely involves phosphatases such as calcineurin [31, 32] and protein phosphatase 2 (PP2A) [34, 35] and is dependent on direct lipidation of autophagy protein 8 (ATG8 lipidation, also known as ATG8ylation [36]) onto the lysosomal membrane, which can be activated by TRPML1-mediated lysosomal Ca²⁺ release [37–39].

Lysosomal ATG8 lipidation is emerging as an essential mechanism for TFEB activation in response to multiple lysosomal stressors (Fig. 1). Upon LMP or TRPML1 activation, the lysosomal proton pump V-ATPase recruits the ATG16L1/ATG12/ATG5 complex through direct ATG16L1 binding (see below), leading to ATG8 lipidation onto lysosomal membrane [39–41]. This is independent of macroautophagy or lysophagy and appears essential for TFEB activation by lysosomal Ca²⁺ release [39–41]. Indeed, loss of the ATG8 conjugation system completely blocks TRPML1-mediated TFEB activation [37–39]. Among the mammalian ATG8 homologues, GABARAPs but not LC3s are essential mediators of TFEB activation upon lysosomal damage or TRPML1 activation [37, 39]. Membrane-conjugated GABARAPs are proposed to facilitate TFEB dephosphorylation by interacting with calcineurin [37] and/or by sequestering the folliculin complex [39], the latter being otherwise important for mTORC1-mediated TFEB phosphorylation (see above) [27–29, 42]. Although TRPML1 appears to be involved in lysosomal stress or damage-induced Ca²⁺ release and subsequent TFEB activation, structural and functional studies have shown that TRPML1 is indeed inhibited at neutral pH [33, 43, 44]. It is thus likely that additional Ca²⁺ channels are responsible for TFEB/TFE3 activation when lysosomes are alkalinized.

Besides lysosomal biogenesis, TFEB/TFE3 also transcriptionally upregulates lysosomal exocytosis (Fig. 1). Following peripheral trafficking, lysosomes can fuse with the plasma membrane to release their contents into extracellular environment [8, 45]. Lysosomal exocytosis is promoted by lysosomal Ca²⁺ signaling and TRPML1 is transcriptionally upregulated by TFEB [45]. TFEB activation likely alleviates lysosomal storage diseases by promoting both increased lysosomal degradation and augmented lysosomal exocytosis [8, 45].

In summary, upon lysosomal destabilization, the TFEB/TFE3 transcription factors are activated by dephosphorylation and nuclear translocation to upregulate lysosomal biogenesis and exocytosis. Compromised mTORC1 signaling reduces TFEB/TFE3 phosphorylation, whereas stress-related lysosomal Ca²⁺ signaling triggers ATG8 lipidation onto lysosomal membrane as a potential platform for TFEB/TFE3 dephosphorylation. Of note, besides de novo lysosomal biogenesis, new lysosomes may be regenerated from damaged membranes, a process depending on ATG8 lipidation [46] and resembling autophagic lysosomal reformation where new lysosomes are reformed from autolysosome-derived membranes [47, 48].

Lysosomal turnover by lysophagy and microautophagy

Although newly synthesized lysosomes can compensate for lost degradative activity, old, injured lysosomes still need to be removed to avoid further damage from leaked cathepsins. Autophagy can often degrade damaged subcellular organelles such as mitochondria, the ER, and peroxisomes. Early studies in immunology connected autophagy machineries with phagolysosomes that are damaged by internalized pathogens [11, 49]. The concept of lysophagy as an indirect lysosomal repair mechanism was introduced by two independent groups in 2013, whereby damaged lysosomes can be selectively captured by autophagosomes for degradation by other, intact lysosomes (Fig. 2) [9, 10]. In support of this concept, autophagy-related proteins including the initiating factor ULK1, ATG14L, WIPI1, ATG5, LC3, and ATG9A are all detected on damaged lysosomes [9, 10]. The current model of lysophagy is based on poly-ubiquitination of damaged lysosomes with complex and likely redundant regulatory events, followed by the recruitment of autophagy adaptors that connect ubiquitinated lysosomes with the growing phagophore (Box 1). A commonly used LMP marker is galectin 3, a member of the galectin family that translocates from the cytosol to the lysosomal lumen where it binds to β-galactosides on glycosylated proteins [50]. Electron microscopy revealed that, upon lysosomal damage, Galectin 3- and LC3-positive compartments are swollen lysosomes sequestered by double membranes, typical structure for autophagosomes [10]. Finally, in support of a role for lysophagy in degrading galectin3-positive lysosomes, disruption of core ATG8 conjugation machineries such as ATG4B, ATG5, or ATG7 delayed the degradation of galectin3 puncta [10, 51].

Figure 2. Lysosomal turnover by lysophagy and microautophagy.

Damaged lysosomes can be completely or partially degraded depending on the extent of lysosomal damage and the degradation pathway activated. (1) In the model of lysophagy, large membrane ruptures on lysosomes cause polyubiquitination of lysosomal membrane proteins, which in turn recruits autophagy adaptors such as TAX1BP1 and p62 to initiate selective capturing of heavily damaged lysosomes for autophagic degradation. Galectin 3 (Gal3), a marker for lysosomal membrane damage, translocates from the cytosol to the lysosomal lumen, where it binds to β-galactosides on glycosylated lysosomal proteins. E3 ubiquitin ligases TRIM16 and CUL4A and an E2 ubiquitin ligase UBE2QL1 are reported to regulate polyubiquitin chain formation, which is further tuned by the ubiquitin-directed AAA-ATPase VCP/p97. TRIM16 and Gal3 might also promote the ATG12/ATG5/ATG16 complex for LC3 lipidation onto phagophores. (2) Lysosomal damage also triggers noncanonical lysophagy which depends on V-ATPase-mediated recruitment of the ATG12/ATG5/ATG16 complex for direct ATG8 lipidation onto the lysosomal membrane. This is likely followed by subsequent fusion of ATG8-labeled vesicles with other intact lysosomes. (3) Selective degradation of lysosomal membrane proteins by microautophagy happens in response to specific cellular stress, which involves substrate ubiquitination and recognition by the ESCRT complex for intraluminal sorting and degradation. (4) A noncanonical microautophagy pathway degrades selective lysosomal membrane proteins including TRPML1 in response to glucose starvation or lysosomal osmotic stress. This pathway requires the ATG8 conjugation system but is independent of macroautophagy. It is still unclear how ATG8-conjugation selectively delivers TRPML1 to intraluminal vesicles for degradation. Figure was created using  BioRender.

Box 1. Lysosomal ubiquitination and lysophagy.

Multiple studies have reported lysosomal ubiquitination as an initiating factor of lysophagy. Ubiquitin and the ubiquitin-binding autophagy adaptor SQSTM1/p62 are recruited to damaged lysosomes [9, 10, 136], suggesting selective autophagic capturing of damaged lysosomes by adaptor recognition of ubiquitination. A recent study compared the impact of multiple autophagy adaptors on the turnover of lysosomal galectin 3 and found a critical role for tax1 binding protein 1 (TAX1BP1), but not other related adaptors including optineurin, calcium-binding and coiled-coil domain-containing protein 2 (CALCOCO2), or p62 [137]. The E2 ubiquitin ligase UBE2QL1 is recruited to damaged lysosomes and is important for lysosomal ubiquitination, as well as the turnover of lysosomal galectin 3 puncta [51]. Ubiquitin, galectin 3, and tripartite motif containing 16 (TRIM16), a galectin 3-interacting E3 ubiquitin ligase, likely coordinate in the lysosomal recruitment of autophagy machineries including the ATG12/ATG5/ATG16L1 complex for ATG8 lipidation [138, 139]. Additionally, galectin 3, TRIM16, and ATG16L1 have been reported to promote autophagy-related secretion of α-synuclein upon lysosomal damage [140]. It is yet to be examined how these proteins coordinate lysophagy and secretion pathways.

The ubiquitin-directed AAA-ATPase VCP/p97, which is essential for ER-associated degradation (ERAD), is also found on damaged lysosomes [141]. VCP/p97 was proposed to work with other factors in reducing lysosomal K48 ubiquitin chains for more efficient lysophagy [141]. In contrast, another study identified CUL4A as an important E3 ubiquitin ligase for lysophagy, and CUL4A appears to promote K48-linked, but not K63-linked, poly-ubiquitination of its lysosomal substrate LAMLP2 upon lysosomal damage [142]. The actin stabilizer calponin-2 (CNN2) is reported as a regulator of lysophagy [143]. CNN2 is recruited to damaged lysosomes, dynamically ubiquitinated and removed by VCP/p97, which ensures efficient clearance of lysosomal galectin 3 puncta [143]. Thus, it appears there are complex, dynamic, and redundant ubiquitination events during lysophagy.

It is important to note that almost all lysophagy studies used the turnover of galectin 3 puncta as a readout of lysophagic activity. However, because galectin 3 puncta are in the lysosomal lumen, the degradation of lysosomal galectin 3 can also be achieved by lysosomal membrane repair [15–18], noncanonical lysophagy [40], or ATG8 lipidation-mediated lysosomal microautophagy [59], in addition to macroautophagy. Therefore, galectin 3 puncta turnover may not necessarily exclusively indicate lysophagy activity. It would be imperative to identify substrates that are specifically degraded by lysophagy but not other lysosomal degradation pathways.

While the field of lysosomal repair has focused on lysophagy for a while, a recent study identified a noncanonical lysophagy pathway that relies on direct ATG8 lipidation on damaged lysosomes (Fig. 2) [40]. V-ATPase is essential for this pathway and it directly recruits ATG16L1 upon lysosomal damage by internalized bacteria or other materials. This leads to ATG8 lipidation onto endolysosomes [40] and presumably subsequent degradation of their contents upon fusion with other intact lysosomes. This mechanism resembles direct ATG8 lipidation onto single membranes in other situations, such as phagocytosis [41, 49, 52] and lysosomal osmotic stress [53, 54]. ATG8 lipidation in all these situations is dependent on the ATG12/ATG5/ATG16L1 complex but not the ULK/FIP200/ATG13 autophagosome initiating complex, the latter being essential for macroautophagy or lysophagy.

Besides the degradation of damaged lysosomes as a whole, selective turnover of lysosomal membrane proteins is mediated by microautophagy for cellular homeostasis (Fig. 2). A highly conserved mechanism is mediated by substrate ubiquitination and ESCRT-mediated intraluminal sorting [13, 14]. This was initially discovered in yeast [12–14, 55, 56] and now also observed in human cells [57]. For example, a yeast lysosomal/vacuolar lysine transporter Ypq1 (PQLC2 in humans) is selectively degraded upon lysine starvation, which is controlled by the recognition of inactive transporter conformation by Ssh4, a lysosomal transmembrane E3 ligase [58]. ESCRT-mediated microautophagy can degrade lysosomal transmembrane proteins either constitutively or in response to cellular stress stimuli such as nutrient starvation or ion imbalance. It is likely that similar mechanisms apply to the turnover of damaged lysosomal membrane proteins.

An additional type of noncanonical lysosomal microautophagy has been described which functions to degrade selective lysosomal membranes (Fig. 2) [59]. This pathway is activated in response to glucose starvation or lysosomal stressors that trigger ATG8 lipidation onto the lysosomal membrane [59]. It requires ATG5 for ATG8 lipidation onto lysosomal membranes and subsequent intraluminal membrane budding, but is independent of the ULK1/ATG13 autophagosome initiation complex or macroautophagy [59]. It is yet to be determined whether ubiquitination or ESCRT is responsible for the intraluminal sorting of ATG8-lipidated membranes and whether this inward membrane budding process contributes to lysosomal repair in response to LMP.

In summary, lysophagy degrades the whole damaged lysosome; noncanonical lysophagy degrades lysosomal contents including membrane-damaging protein aggregates and pathogens; and microautophagy removes selective lysosomal membrane proteins. It is remarkable that lysosomal ATG8 lipidation is essential for multiple lysosomal quality control pathways including TFEB/TFE3 activation, noncanonical lysophagy, and ATG8-dependent lysosomal microautophagy. ATG8 lipidation is likely a key factor coordinating all these pathways upon lysosomal damage: (1) lipidated ATG8s (GABARAPs) immediately trigger TFEB/TFE3 activation through inhibition of the folliculin complex (Fig. 1); (2) lipidated ATG8s might work with the ESCRT complexes for more efficient membrane remodeling and repair (Fig. 2 and 3); (3) if all membrane damages are repaired by the ESCRT and PITT pathways (see below), then lysosomal lipidated ATG8s may be removed and recycled by ATG4 or otherwise further initiate noncanonical lysophagy to degrade potential membrane-damaging source in the lysosomal lumen; (4) finally, when all these efforts fail, lysophagy would remove over damaged lysosomes, with new lysosomes being generated by the TFEB/TFE3 pathway.

Figure 3. ESCRT-mediated rapid repair of small pores on lysosomal membrane.

The endosomal sorting complex required for transport (ESCRT) machineries are rapidly recruited to damaged lysosomes for the repair of small lysosomal membrane pores. The recruitment of ESCRT depends on Ca²⁺ release upon lysosomal damage, which directly recruits programmed cell death protein 6 (PDCD6)/apoptosis-linked gene 2 (ALG-2), a Ca²⁺ effector. PDCD6 further recruits ESCRT-III subunits through ALIX or alternatively through TSG101 and ESCRT-II. The assembly of ESCRT-III subunits into filament spirals are expected to remodel the lysosomal membrane and push damaged membranes into the lysosomal lumen, with the AAA+ ATPase VPS4 finishing the repair through membrane scission. ESCRT assembly on damaged lysosomes is likely regulated by leucine-rich repeat kinase 2 (LRRK2) and its kinase substrate RAB8A GTPase. Figure was created using  BioRender.

ESCRT-mediated lysosomal repair by membrane remodeling

While removing and replacing damaged lysosomes are essential for maintaining lysosomal activity, these pathways might not be sufficiently rapid to block acute lysosomal leakage, a major cause of lysosomal cell death. It is thus not surprising that mammalian cells have evolved additional direct lysosomal repair strategies. The first reported, direct lysosomal repair mechanism is mediated by the ESCRT complex [15–17, 60]. ESCRT has well-established functions in the formation of intraluminal vesicles at endolysosomes [61, 62]. It requires the sequential and coordinated assembly of ESCRT-0, -I, -II, and -III complexes, leading to membrane budding away from the cytosol and subsequent membrane scission [61, 62]. Similar membrane remodeling activity has been observed in other cellular processes such as the closure of the nuclear envelope or autophagosome, as well as the repair of damaged plasma membrane or nuclear envelope [63–67].

LMP-induced lysosomal ESCRT recruitment is rapid and Ca²⁺-dependent (Fig. 3). Multiple ESCRT subunits are immediately recruited to lysosomes damaged by various materials or pathogens [15–17, 60, 68]. The recruitment happens within minutes of lysosomal damage, clearly before galectin 3 puncta are formed [16, 17], consistent with ESCRT being involved in the rapid repair of smaller membrane pores. LMP-related lysosomal Ca²⁺ leakage likely initiates ESCRT assembly onto damaged lysosomes by recruiting programmed cell death protein 6 (PDCD6)/apoptosis-linked gene 2 (ALG-2) [17]. PDCD6 is a Ca²⁺-binding protein with five EF-hand motifs and is known to bind two ESCRT subunits ALIX and TSG101 in the context of plasma membrane repair [69]. In vitro reconstitution showed that Ca²⁺ binding to PDCD6 is sufficient to drive PDCD6 binding to membranes [70]. Ca²⁺-bound PDCD6 recruits ALIX to the membrane, releasing an auto-inhibitory domain of ALIX to allow for further recruitment of downstream ESCRT-III subunits, including CHMP2A, CHMP3, and CHMP4B, as well as the AAA+ ATPase VPS4 [70]. Alternatively, Ca²⁺-activated PDCD6 can recruit the ESCRT-I subunit TSG101, which further recruits ESCRT-III through the ESCRT-II complex [70]. Once ESCRT-III subunits are recruited, their filament spirals are expected to remodel the membrane and push damaged membranes into the lysosomal lumen, with VPS4 finishing the repair through membrane scission [61, 62]. Consistent with redundant roles for ALIX and TSG101, double knockdown of both proteins, but not the single knockdown of either, leads to an almost complete block of ESCRT-III recruitment to damaged lysosomes, causing a delay in rapid lysosomal repair [17].

The recruitment of ESCRT to damaged lysosomes appears to be regulated by leucine-rich repeat kinase 2 (LRRK2), a risk gene for both Parkinson’s disease (PD) and Crohn’s disease [71]. Specifically, LRRK2 is activated by endolysosomal damage, and it in turn recruits and phosphorylates the Ras-related protein 8A (RAB8A) GTPase [72]. Both LRRK2 and RAB8A have been found to be essential for LMP-induced lysosomal recruitment of CHMP4B, an ESCRT-III subunit [72]. Although it is not determined which step of ESCRT recruitment requires LRRK2 and RAB8A, these observations suggest that ESCRT assembly can be regulated by kinase signaling and RAB GTPases which likely remodel local membranes at ER-lysosome contacts [73] for more efficient ESCRT recruitment. Further investigation of such regulatory mechanisms might provide insights into the pathogenic mechanisms of LRRK2 mutations in human disease.

The PITT pathway for rapid lysosomal repair through direct lipid transfer

The robust lysosomal recruitment of ESCRT subunits upon LMP strongly suggests that ESCRT is likely essential for rapid lysosomal repair. However, as observed in multiple studies, loss of ESCRT only causes a partial delay in rapid lysosomal repair, indicating the existence of additional repair mechanisms. The search for such mechanisms led to the discovery of the PITT pathway for rapid lysosomal repair (Fig. 4) [18]. The PITT pathway is initiated by the rapid production of a lipid messenger, phosphatidylinositol 4-phosphate (PI4P), on damaged lysosomes [18]. PI4P in turn drives extensive ER-lysosome membrane tethering as a platform for direct lysosomal repair through multiple lipid transfer events [18].

Figure 4. The PITT pathway for rapid lysosomal repair.

The phosphoinositide-initiated membrane tethering and lipid transport (PITT) pathway rapidly repairs lysosomes by directly transferring lipids to damaged membranes. Multiple pathophysiologically relevant lysosomal damaging conditions trigger the rapid recruitment of PI4K2A to lysosomes, leading to robust generation of PI4P, a lipid messenger. PI4P further recruits multiple oxysterol-binding protein (OSBP)-related protein (ORP) family proteins including ORP9, ORP10, ORP11, and OSBP. The ORP proteins directly tether damaged lysosomes to the ER network through simultaneously binding to lysosomal PI4P and VAPA/B on the ER through different motifs. Lysosomal PI4K2A and the ER-anchored PI4P-phosphatase Sac1 maintain a steep PI4P gradient at ER-lysosome contacts that drives OSBP/ORP-mediated lipid exchanges. OSBP homodimers mediate the counter transport of PI4P and cholesterol (Chol) between the ER and damaged lysosomes, leading to increased lysosomal cholesterol contents which promotes lysosomal membrane stability. In parallel, heterodimers of ORP9 with either ORP10 or ORP11 mediate PI4P/phosphatidylserine (PS) exchange. The resulting lysosomal PS accumulation activates autophagy protein 2 (ATG2)-dependent lipid delivery to repair lysosomal pores. PS likely activates ATG2-mediated lipid delivery to both damaged lysosomes and phagophores, immature membrane structures in macroautophagy. However, lysosomal repair and macroautophagy might differ in the requirement of lipid scramblases like ATG9 on the acceptor membrane. While redistribution of newly delivered lipids to the inner and outer leaflets on phagophores appears essential in macroautophagy, such lipid scrambling activity is likely not necessary in rapid lysosomal repair as the inner and outer leaflets of lysosomal membranes are interconnected at damage sites, allowing automatic distribution of lipids to both leaflets. Figure was created using  BioRender.

The PITT pathway was identified through an unbiased analysis of lysosomal surface proteomes before and after lysosomal membrane damage (Fig. 4). This screen identified a list of PI4P-related proteins as specifically enriched on damaged lysosomes [18], including phosphatidylinositol-4 kinase type 2α (PI4K2A), an enzyme that generates PI4P [74], and four oxysterol-binding protein (OSBP)-related protein (ORP) family members ─ ORP9, ORP10, ORP11 and OSBP, all well-established PI4P-binding proteins [75–80]. Within a few minutes of LMP, PI4K2A is strongly recruited to lysosomes, accompanied by robust lysosomal PI4P production and subsequent recruitment of the ORP family proteins, all independently of the ESCRT subunits [18].

Once the ORP family proteins are recruited, they mediate extensive membrane tethering between damaged lysosomes and the ER network (Fig. 4) [18]. An ORP protein typically contains an N-terminal pleckstrin homology (PH) domain, a C-terminal OSBP-related domain (ORD), and a domain containing two phenylalanine residues (FF) in an acidic tract, termed the “FFAT” motif [81]. All four ORP proteins are recruited to damaged lysosomes through their PH domain which binds to PI4P [18, 80], but only ORP9 and OSBP have strong FFAT motifs to directly bind vesicle-associated membrane protein-associated protein A (VAPA) and VAPB on the ER [81]. Consistent with these structural insights, OSBP forms homodimers and can mediate ER-lysosomal tethering on its own, whereas ORP10 and ORP11 form heterodimers with ORP9 to form stable and functional tethers [18].

The ORP proteins are also lipid transfer proteins at membrane contacts and they usually exchange PI4P for either cholesterol or phosphatidylserine (PS) using their C-terminal ORD domains [81]. Upon lysosomal damage, ORP9, ORP10, and ORP11 specifically mediate PI4P/PS counter transfer at ER-lysosome contacts, and OSBP drives PI4P/cholesterol exchanging (Fig. 4) [18]. Similarly to the ER-Golgi contacts [80], a sharp PI4P gradient at the ER-damaged lysosome contacts is likely maintained by lysosomal PI4K2A and the ER-anchored PI4P phosphatase Sac1. Thus, PI4P stimulates ORP-mediated transfer of PS and cholesterol to damaged lysosomes [18]. Of note, another study employing an unbiased lipidomics approach also identified LMP-induced lysosomal accumulation of PS and cholesterol, which was found to be downstream of PI4K2A [82]. Remarkably, in ORP-QKO (quadruple knockout of ORP9, ORP10, ORP11, and OSBP) cells, reconstitution of ER-to-lysosomal transfer of either PS or cholesterol can rescue rapid lysosomal repair [18]. Thus, PS and cholesterol act in parallel to promote rapid lysosomal repair. It is not clear whether cholesterol further recruits proteins or vesicles for lysosomal repair, but increasing cholesterol levels is well known to enhance membrane stability [83]. Indeed, artificially loading lysosomes with higher levels of cholesterol can increase lysosomal membrane stability and cellular resistance to various LMP inducers [82, 84, 85].

With the help of lysosomal PS, the lipid transfer protein ATG2 directly repairs lysosomal membrane pores through lipid transfer (Fig. 4) [18]. Mammalian ATG2 folds into a 20 nm long rod-like structure, with an elongated groove-shaped hydrophobic cavity in the rod, while amphipathic helices attach to the C-terminal end of the rod [86, 87]. ATG2 was established as a lipid transfer protein and is believed to deliver lipids through its hydrophobic central groove from the ER membrane to growing phagophores in autophagy [88–91], although in vitro evidence for large-scale lipid transfer by ATG2 is still missing. ATG2 is dynamically recruited to damaged lysosomes, and cells lacking ATG2A and ATG2B (ATG2A/B-DKO) were markedly defective in rapid lysosomal repair [18]. The role for ATG2 in lysosomal repair is completely dependent on its lipid transfer activity through its hydrophobic central groove and requires its C-terminal amphipathic helices [18]. It appears that lysosomal PS improves the membrane embedding of the amphipathic helices of ATG2, which allows for efficient lysosomal lipid unloading to directly repair membrane pores [18]. PS probably also stimulates ATG2-mediated lipid delivery to phagophore tips during autophagosome formation, given the close proximity of PS synthase 1 to the autophagy-initiating FIP200 puncta [92] and the abnormal accumulation of PS on open phagophores in cells with autophagy defects at the phagophore stage, including ATG2-deficient cells [93, 94].

Despite similar lipid transfer mechanisms for ATG2 in macroautophagy and PITT-mediated rapid lysosomal repair, the two pathways are independent of each other. Either pathway can be selectively shut down by genetic approaches without affecting the other, including selective disruption of ATG2 functions in lysosomal repair without affecting macroautophagy [18]. The two pathways might also differ in their requirements for lipid scrambling at lipid unloading sites. The lipid scramblase ATG9A appears to work with ATG2 at phagophore tips for efficient lipid re-distribution between the two leaflets of phagophore membranes [95–98]. However, lipid scrambling is theoretically not necessary for rapid lysosomal repair since the inner and outer leaflets of the lysosomal membrane are connected at membrane damage sites, which would allow for automatic lipid re-distribution (Fig. 4).

Lysosomal lipid remodeling is emerging as an essential mechanism for rapid lysosomal quality control. The PITT pathway provides several examples of marked lipid remodeling on damaged lysosomes. These include a dramatic increase of lysosomal PI4P, PS, and cholesterol, as well as more non-selective lipid transportation by ATG2 [18]. The annexin family of Ca²⁺-binding proteins, particularly annexins A1 and A2, likely also play a role in lysosomal repair [99, 100]. As annexin A2 is known to bind PS in the presence of cholesterol [101], it may potentially improve PITT-mediated lysosomal repair by promoting PS clustering on damaged lysosomes [102]. Of note, two studies reported substantial changes of another group of lipids on damaged lysosomes ─ sphingolipids, which also appears to help mediate lysosomal repair [103, 104]. Sphingolipids are typically kept in the luminal leaflet of the lysosomal membrane. However, robust sphingomyelin exposure, mediated by Ca²⁺-activated lipid scrambling, was observed on the cytosolic leaflet of damaged lysosomal membranes [103]. The exposed sphingolipids are subsequently converted into ceramides [104] which are expected to facilitate intraluminal membrane budding. It seems likely that these events coordinate with ESCRT- and/or ATG8-mediated intraluminal membrane budding.

Lysosomal quality control in aging, diseases, and therapeutics

The integrity and activity of lysosomes are essential for human health and longevity. As such, lysosomal destabilization is a common feature of aging and age-related pathology such as neurodegeneration and cardiovascular diseases [105]. Genetic mutations impairing lysosomal function are also known to cause many diseases including early-onset lysosomal storage diseases and late-onset neurodegenerative diseases [106, 107]. In contrast, long-lived animal models typically maintain higher lysosomal activity as they age compared with their control animals [108]. Multiple lifespan-extending interventions are associated with enhanced lysosomal quality and activity [108]. Caloric restriction, fasting, and rapamycin all activate TFEB/TFE3-mediated transcriptional upregulation of lysosomal biogenesis and autophagosome formation [20–23, 25]. Exercise, which provides multiple health benefits, also activates TFEB nuclear translocation [32]. The widely prescribed antidiabetic medicine metformin, an agent linked to augmenting lifespan, has been found to achieve its health benefits by mildly inhibiting the lysosomal proton pump V-ATPase and thus activating adenosine monophosphate-activated protein kinase (AMPK) [109]. AMPK is known to activate TFEB/TFE3 independent of mTORC1 [110, 111], although it is yet to be determined if TFEB/TFE3 contribute to metformin’s benefits. Thus, mechanistic investigation of lysosomal quality control pathways may pave the way for novel therapeutics to slow aging or combat age-related diseases.

Genetic loss of lysosomal quality control is associated with pre-mature aging and neurodegeneration. For example, the deletion of PI4K2A, the first key enzyme of the PITT rapid lysosomal repair pathway [18] that is highly expressed in the brain, causes severe, pre-mature aging phenotypes in human patients and neurodegeneration in mice [112–114]. In cell-based models, deletion of ESCRT subunits or PI4K2A substantially increases tau fibril spreading, consistent with more endolysosomal escape of internalized tau seeds, a key step in the progression of Alzheimer’s disease [18, 115]. Similarly, TFEB is also highly expressed in the brain, and TFEB dysfunction is linked to neurodegenerative diseases [116].

Stimulating lysosomal quality control has been shown to alleviate lysosomal-related diseases. Among the known lysosomal quality control pathways, activation or overexpression of the TFEB/TFE3 pathway has been widely tested in cell and animal models of multiple diseases, such as lysosomal storage diseases, Huntington disease, multiple system atrophy, Parkinson’s and Alzheimer’s disease [7, 8, 117–121]. TFEB/TFE3 activation or overexpression showed protective activities by upregulating lysosomal exocytosis and more efficient clearance of intracellular cargos including disease-causing protein aggregates or pathogens [7, 8, 117–123]. Small molecule activators of TFEB/TFE3 have been developed with promising impacts on the upregulation of lysosomal biogenesis and autophagy in vitro and in vivo [124–127]. Pharmacological activation of other lysosomal quality control pathways has not yet been reported, but potential PITT activators that stimulate the lysosomal recruitment and/or activation of PI4K2A might also improve lysosomal integrity and activity. Additionally, relieving endogenous suppressors of lysosomal quality control pathways might also provide benefits. For example, removing the inhibition of ESCRT assembly in LRRK2 mutated patients will likely rescue ESCRT-mediated lysosomal repair.

Lysosomal integrity and activity are also important for innate immunity. Endocytosed disease-causing pathogens often escape from endolysosomes and compromise host lysosomal quality control pathways. For example, a bacterial effector SopF specifically inhibits noncanonical lysophagy by ADP-ribosylation of the ATP6V0C subunit of V-ATPase, which blocks ATG16L1 recruitment and subsequent ATG8 lipidation on lysosomes [40]. Through inhibiting noncanonical lysophagy, SopF promotes intracellular bacterial growth without affecting macroautophagy or lysophagy [40]. In malaria, the replication vacuole of the parasite Plasmodium is decorated by lipidated ATG8, but the parasite expresses a transmembrane protein upregulated in infective sporozoites 3 (uis3) on the replication vacuole to avoid host defense through lysosomal quality control pathways [128, 129]. Parasites lacking uis3 can still infect but cannot grow in host cells [130]. Interestingly, GABARAPs and the ATG8 conjugation machineries, but not LC3 proteins, are essential for host elimination of uis(−) parasites [131]. The underlying mechanism might involve noncanonical lysophagy (Fig. 2) and/or GABARAPs-mediated TFEB/TFE3 activation and thus upregulation of lysosomal and autophagic activity (Fig. 1).

While lysosomal quality control is desired for normal cells, it can be hijacked by cancer cells as a survival mechanism. In pancreatic cancer, myoferlin, a ferlin family member with established roles in plasma membrane repair, is hijacked by cancer cells and re-targeted to lysosomes, providing an extra layer of protection against lysosomal destabilization [132]. Of note, no ferlin family proteins are found on damaged lysosomes in other conditions. In many other cancers, PI4K2A, the master enzyme of the PITT pathway, is overexpressed and appears to form abnormal lysosomal protection complexes [133–135]. Thus, understanding the unique mechanism by which myoferlin and PI4K2A protect lysosomes in cancer cells might reveal unique targets for cancer treatments.

Concluding remarks

While lysosomes have been historically viewed as essential elements in overall cellular quality control, the quality control of the lysosome itself has been largely ignored for decades. Fortunately, multiple distinct pathways have now been identified whereupon damaged lysosomes are quickly identified by the cell. These include rapid lysosomal repair mediated by the ESCRT and PITT pathways, selective turnover of unrepairable lysosomal membranes by lysophagy and microautophagy, and compensatory biogenesis of new lysosomes driven by TFEB/TFE3. It is tempting to speculate that these pathways are activated depending on the type, strength, and duration of the lysosomal damage. For instance, ESCRT might be sufficient for the repair of mild membrane permeabilization, PITT for mild to moderate injuries, and lysophagy for severe lysosomal ruptures.

Considerable efforts have been devoted to mechanistic studies of each pathway through genetic, cell biological, and biochemical approaches. However, many questions remain to be answered (See Outstanding Questions). Among these questions are how these various pathways are regulated on damaged lysosomes and how they cooperate to achieve optimal recovery of lysosomal quality and activity. Lysosomal Ca²⁺ signaling appears to be a universal mechanism that senses diverse lysosomal perturbations. Downstream of Ca²⁺ signaling, lysosomal lipid remodeling and ATG8 lipidation likely coordinate all lysosomal quality control pathways. PITT-mediated lysosomal lipid changes are expected to impact the efficiency of membrane remodeling by the ESCRT complexes. A thorough understanding of all lipid remodeling events in lysosomal quality control will likely unravel lipid-protein interactions as a strategy for PITT communications with other lysosomal quality control pathways. It is of particular interest that ATG8 is directly lipidated onto the lysosomal membrane in response to most lysosomal stressors [39, 40, 53, 54]. Lysosomal ATG8 lipidation likely stimulates TFEB/TFE3 activation; contributes to lysosomal repair and microautophagy through intraluminal membrane budding; and communicates with lysosomal repair pathways to initiate noncanonical lysophagy. A clearer model for lysophagy is desired, which will likely depend on the identification of lysophagy-specific substrates for the development of more accurate lysophagy assays. Also intriguing is the molecular mechanism underlying cellular selection of macroautophagy versus microautophagy during the clearance of damaged lysosomal membranes.

Outstanding questions.

Is lysosomal Ca²⁺ release a general mechanism for lysosomal quality control? How does Ca²⁺ activate different lysosomal quality control pathways? For example, how does V-ATPase sense lysosomal Ca²⁺ release for ATG16L1 recruitment? How does PI4K2A sense lysosomal Ca²⁺ release for PITT activation?

How do cells coordinate different lysosomal quality control pathways? Are there differences between pathogen-mediated lysophagy and sterile lysophagy? Are there lysophagy specific substrates not degraded by other lysosomal quality control pathways? What is the mechanism for the intraluminal vesicle formation driven by ATG8 lipidation?

Are there additional Ca²⁺-activated phosphatases that activate TFEB/TFE3? How is TFEB/TFE3 activated in TSC-deficient cells?

How is ESCRT assembly regulated on damaged lysosomes? Can we boost such pathways for lysosomal protection?

What determines the directional lipid delivery by ATG2 in the PITT pathway for rapid lysosomal repair? Does ATG2 need a transmembrane protein on damaged lysosomes for lipid unloading? Are there additional lipid effectors downstream of the PITT pathway? What are the roles for other lipids such as ceramide in lysosomal repair?

Can we pharmacologically manipulate lysosomal quality control pathways for therapeutic benefit in human diseases? Can we improve the health of senescent cells by stimulating lysosomal quality control pathways?

Often deregulated in aging and diseases, lysosomal quality control pathways provide promising therapeutic targets for lysosomal-related diseases, aging, infectious diseases, and cancer. As the first identified lysosomal quality control pathway, TFEB/TFE3 have been extensively tested in vitro and in vivo with promising applications to be soon tested in clinical settings. The more lately discovered ESCRT and PITT rapid lysosomal repair pathways both appear to be connected to human diseases, but more investigation is required regarding their regulation and activation in disease models. Despite an initial lack of attention to lysosomal quality control, significant progress has been made in the past 10 years, with at least four distinct pathways discovered. It is likely, the next decade will bring even more important insights into this emerging area.

Highlights.

• Damaged lysosomes are replaced, removed, and/or repaired by lysosomal quality control pathways.

• Lysosomal stress stimulates TFEB/TFE3-mediated transcriptional upregulation of lysosomal biogenesis and exocytosis.

• Lysophagy degrades damaged lysosomes as a whole, whereas microautophagy degrades selective lysosomal membrane proteins.

• The ESCRT machineries repair small lysosomal membrane pores by direct membrane sealing.

• The PITT pathway repairs damaged lysosomes through direct lipid transfer at ER-lysosome membrane contact sites.

• Often deregulated in aging and diseases, lysosomal quality control pathways provide promising therapeutic targets for lysosomal related diseases, aging, infectious diseases, and cancer.

Glossary

ATG8 lipidation

conjugation of ATG8, a ubiquitin-like protein, onto PE or PS in cellular membranes. Mammalian ATG8 has six homologues GABARAP, GABARAPL1, GABARAPL2, MAP1LC3A (LC3A), LC3B, and LC3C. During macroautophagy, ATG8 is lipidated onto phagophore membranes, which is required for autophagosome formation. ATG8 is also lipidated onto single membrane compartments including stressed endolysosomes, where it regulates TFEB activation, microautophagy, and noncanonical lysophagy

ESCRT

endosomal sorting complex required for transport, including ESCRT-0, -I, -II, -III that are sequentially recruited to membranes to mediate vesicle budding away from the cytosol. ESCRT can act on multiple subcellular membranes such as endosomes, lysosomes, the plasma membrane, the nuclear envelope, immature autophagosomes, and so on

LMP

lysosomal membrane permeabilization, a common lysosomal stress in aging and diseases. LMP causes lysosomal leakage and is a danger signal that activates multiple lysosomal quality control pathways. If not immediately fixed, LMP can activate lysosomal cell death pathways

Lysophagy

macroautophagy-mediated selective capturing and degradation of damaged lysosomes decorated by poly-ubiquitination. Autophagy adaptors that recognize both ubiquitin and LC3 are needed to selectively capture damaged, ubiquitinated lysosomes. Autophagosomes containing damaged lysosomes fuse with healthy lysosomes to completely degrade damaged lysosomes. Note that in noncanonical lysophagy, LC3 is directly lipidated onto lysosomal membranes, and such LC3-labeled lysosomes subsequently fuse with other healthy lysosomes for content degradation. Noncanonical lysophagy might depend on lysosomal repair to ensure that fused lysosomes are intact

Microautophagy

ubiquitination-dependent, ESCRT-mediated lysosomal intraluminal sorting and degradation of selective lysosomal membrane proteins. In noncanonical microautophagy, lysosomal intraluminal vesicle formation and degradation are driven by lysosomal ATG8 lipidation

PI4P

phosphatidylinositol 4-phosphate, a signaling lipid in eukaryotic cells. PI4P is primarily generated by four different PI4-kinases in mammalian cells. In resting conditions, most PI4P is found at the trans-Golgi network and the plasma membrane, with much lower levels in endolysosomes. Lysosomal damage triggers rapid and robust PI4P generation on lysosomes by PI4-kinase type IIα (PI4K2A), the master enzyme for the PITT lysosomal repair pathway

PITT

The phosphoinositide-initiated membrane tethering and lipid transport pathway for rapid lysosomal repair. This pathway starts with LMP-induced, PI4K2A-mediated generation of PI4P, a lipid messenger on damaged lysosomes, which in turn drives extensive endoplasmic reticulum-lysosome membrane contacts as a platform for rapid lysosomal repair by multiple lipid transfer events. OSBP and ORP9/10/11 transfer cholesterol and phosphatidylserine, respectively, to damaged lysosomes at the expense of lysosomal phosphatidylinositol 4-phosphate, whereas ATG2 mediates larger-scale nonselective lipid delivery to repair lysosomal membrane pores