Role of lysosomes in physiological activities, diseases, and therapy

Lysosomes are composed of acid lumen and a layer of lysosomal membrane formed by phospholipid bilayer. The acid lumen contains a host of hydrolytic enzymes, including nucleases, proteases, phosphatases, lipases, sulfatases, and others. These enzymes are synthesized in endoplasmic reticulum, and modified in Golgi, where the proteins are added mannose-6-phosphate residue, a tag targeting lysosomes [1]. On lysosomal membranes, the overarching protein maintaining acid luminal pH is vacuolar H+-ATPase (v-ATPase) that hydrolyses adenosine triphosphate (ATP) to transport protons into lysosomes [2]. Other key lysosomal membrane proteins include several ion channels, through which ions move inward/outward [3]; lysosome-associated membrane proteins (LAMPs), including LAMP1-5, among which LAMP1 and LAMP2 are well documented to be involved in phagocytosis, chaperone-mediated autophagy (CMA), macroautophagy, lipid transport, and aging [4]; soluble N-ethylmaleimide-sensitive factor activating protein receptors (SNAREs) and RAB GTPases that are involved in trafficking and fusion [5]; toll-like receptors (TLRs) that sense pathogen-associated molecular patterns (PAMPs) and initiate inflammatory response [6]; and mammalian target of rapamycin (mTOR) that co-ordinates several homeostatic signaling pathways [7]. Notably, a series of terms have been proposed to classify lysosomes, such as terminal lysosomes, endolysosomes, protolysosomes, phagolysosomes, and autolysosomes indicating heterogeneity of lysosomes. In a word, the complex lysosomal luminal and membrane proteins indicate lysosomes to be pleiotropic organelles.

Lysosomes are formed through fusion of endosomes and vesicles budded from trans-Golgi network. The main regulator of lysosomal biogenesis is microphthalmia/transcription factor E (MiT/TFE) protein family including transcription factor EB (TFEB), transcription factor EC (TFEC), transcription factor binding to IGHM enhancer 3 (TFE3), and melanocyte inducing transcription factor (MITF) [8], among which TFEB is the best studied. TFEB can be phosphorylated on several sites by a number of kinases and phosphatases, such as mechanistic target of rapamycin complex 1 (mTORC1) on Ser211 [9], Ser142 [10], and Ser122 [11], glycogen synthase kinase 3β (GSK3β) on Ser134 and Ser138 [12], extracellular signal-regulated kinase (ERK) on Ser142 [13], mitogen-activated protein kinase kinase kinase kinase 3 (MAP4K3) on Ser3 [14], and protein kinase B (PKB) on Ser467 [15] repressing nuclear translocation of TFEB. Dephosphorylation of TFEB by calcineurin [16] and protein phosphatase 2 (PP2A) [17] lets TFEB enter nuclear to regulate transcription. For example, if nutrient is in abundance, TFEB is phosphorylated by mTORC1, which facilitates 14-3-3-dependent retention of TFEB in cytoplasm and prevents transcription function of TFEB [9]. On the contrary, in the context of starvation or mTOR inhibitor treatment, inactivated mTOR leads to accumulation of dephosphorylated TFEB resulting in nuclear translocation of TFEB [10]. Alternatively, in some cases, lysosomes release Ca²⁺ through transient receptor potential mucolipin 1 (TRPML1) to activate calcineurin that dephosphorylates TFEB [16]. In nucleus, TFEB transcribes coordinated lysosomal expression and regulation (CLEAR) network genes that code proteins highly related to structure and function of lysosomes [18] including receptors and enzymes participating in lysosomal sorting. Both soluble and membrane lysosomal proteins should be recognized by sorting receptors. The best-known sorting receptor is the mannose-6-phosphate receptors (M6PRs) that recognize M6P-tagged proteins in the Golgi complex [19]. Newly synthesized lysosomal proteins are successively modified by oligosaccharyltransferase [20], GlcNAc-1-phosphotransferase, and uncovering enzyme [21] to be tagged with M6P residues. In Golgi complex with pH 6.7, tagged proteins bind with M6PRs, while in acidic endosomes with pH 6, the proteins are released [19]. Additionally, a subgroup of M6PRs, cation-independent MPR, is able to localize at plasma membrane, which underpins retrieving of escaped extracellular lysosomal proteins [19]. Besides M6PRs, two sorting receptors are also found, sortilin and lysosomal integral membrane protein 2 (LIMP-2). Sortilin is involved in the sorting of prosaposin [22], acid sphingomyelinase [23], cathepsin D, and cathepsin H [24]. LIMP-2 mediates the transport of β-glucocerebrosidase [25], whose dysfunction is associated with Gaucher disease (GD). For lysosomal membrane proteins, tyrosine-based or dileucine-based signals are essential for their lysosomal targeting through adaptor proteins or Golgi-localising, Gamma-adaptin ear domain homology, ARF-binding proteins (GGAs) [26, 27]. Early endosomes budding from post-Golgi complex engage multiple rounds of membrane fusion and fission, resulting in formation of late endosomes and lysosomes [28, 29]. In a word, the lysosomal biogenesis pathway is accurately controlled to keep the homeostasis of cell.

Recent discoveries have changed the former consideration of lysosomes as organelles that degrade and recycle cellular waste. It is now clear that lysosomes are key organelles in degradation, innate and adaptive immunity, and nutrient sensing [28]. Tempting researches have highlighted many signal pathways interacting with lysosomal activities manifesting the central role of lysosomes in many physiological processes [29]. In addition, progressions of several diseases are also regulated by lysosomes [30].

The development of atherosclerosis may result in coronary syndromes [47], ischaemic strokes [48], intermittent claudication [49], and aneurysms [50], owing to plaque growth and thrombus formation. Low-density lipoprotein cholesterol (LDL-C) and oxidized LDL (oxLDL) [51] engulfed by macrophage-derived foam cells accumulation in the intima of arteries is the key during atherosclerosis, which causes chronic inflammation [52]. Afterward, these cells may engage programmed cell death resulting in a necrotic core covered by a fibrous cap, whose rupturing exposes tissue factor and provokes thrombosis [53]. As the main source of foam cells, macrophages have been rigorously studied, in which lysosomes are demonstrated to be involved in almost the whole process of atherosclerosis.

During the initiation of atherosclerosis, membrane scavenger receptors such as CD36 [54, 55], scavenger receptor A (SR-A) [55], and lipoprotein receptor-1 (LOX-1) [56] are responsible for the internalization of lipids, especially oxLDL, into lysosomes. The internalized LDL is hydrolyzed by enzymes like lysosomal acid lipase to produce free cholesterol that is exported to endoplasmic reticulum for re-esterification and storage. Substantial decrease in lysosomal acid lipase activity leads to development of premature atherosclerosis in human [57]. The lipid processing course can be in trouble when macrophages are exposed to oxLDL [54]. Treatment with oxLDL or cholesterol crystal caused lysosomal dysfunction featured with elevated lysosomal pH, increased lysosomal membrane permeability, diminished degradation capacity, and morphological changes [58]. Together, dysfunction of lysosomes owing to oxLDL or decreased lipase may lead to pathological changes that cause atherosclerosis through regulating autophagy, inflammasomes, apoptosis, and lysosomal biogenesis.

Autophagy has shown its protective role in atherosclerosis development through promoting lipid droplet hydrolysis and eventual efflux of free cholesterol from foam cells [59]. Inhibition of autophagy by autophagy-related 5 (Atg5) deletion accelerated atherosclerotic plaque development [60]. Unfortunately, atherosclerotic plaque development did impair autophagy conversely in macrophages [61, 62] accounting for cholesterol crystal accumulation and dysfunctional mitochondria, which are responsible for inflammasome hyperactivation, M1-like polarization, and apoptosis [63]. The cholesterol crystal accumulation activated NOD-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome that releases interleukin-1 beta (IL-1β) [62] through lysosomal leakage containing cathepsins [54]. Impaired autophagy also caused dysfunctional mitochondria that triggered apoptosis through reactive oxygen species (ROS) and cytochrome C [63]. In addition, it has been proved in other disease models that cathepsins from lysosomes are also the inducer of apoptosis in cytoplasm [64, 65]. Apoptosis of macrophages, or foam cells, is the key step in the formation of lipid-rich necrotic core [53]. In a word, both activated inflammasome and apoptosis in macrophages are crucial for atherosclerotic plaque development.

The aforementioned deteriorations initiate several signal changes, which are largely based on activated mTOR indirectly by cholesterol via a lysosomal transmembrane protein SLC38A9 [66]. Activated mTOR may start several detrimental mechanisms. First, mTOR activation inhibits autophagy, whose suppression causes damages described in the foregoing paragraph. Second, cholesterol trafficking from lysosomes to endoplasmic reticulum is obstructed by activated mTOR [67]. Third, activated mTOR inhibits nuclear translocation of TFEB [9, 10] that reverses lysosomal dysfunction [58]. Therapeutic benefits of mTOR disruption by Cre/loxP recombination [68] or small molecular inhibitors [69] are in accordance with these mechanisms. What’s more, although detailed mechanisms are still fuzzy, the association between mTOR and macrophage polarization are proposed [70, 71]. Inhibiting mTOR elevated M2-like polarization that stabilizes plaques and hampered M1-like polarization that promotes plaque rupture [70, 72‐74]. It’s noteworthy that a more complex grouping of macrophages in plaques has been put forward based on single-cell RNA sequencing [75]. A better understanding of mTOR-related pathways is tempting for mediation of atherosclerotic plaque development.

In summary, lysosomes in macrophages are involved in initiation, development, and progression of atherosclerotic plaque, which have been exhibited in Fig. 1 and Table 1. They are the crucial nodes linking lipid degradation, autophagy, apoptosis, inflammasome, lysosomal biogenesis, and macrophage polarization. Researches on lysosomes and mTOR signal are promising for prevention and therapy of atherosclerosis, several of which have inspired lysosome-targeting clinical trials listed in Table 4.

Aggregation of abnormal proteins and organelles is common in neurodegenerative diseases. The autophagy-lysosome pathway (ALP) is essential for the survival of non-dividing neurons by contributing to the removal of abnormal large protein aggregates and organelles [76, 77]. Defective clearance by ALP and/or increased abnormal protein aggregation is common in the pathogenesis of several neurodegenerative diseases, including Alzheimer's disease (AD), Parkinson’s Disease (PD), and Huntington disease (HD) [78]. It is difficult to determine whether the defect of ALP is the cause or result of neurodegeneration. Indeed, aggregation of toxic proteins, impaired organelles, oxidative stress, and insufficient ATP can interact with defective ALP leading to a worse situation and eventual cell death [79, 80]. Several genetic defects and lysosome-targeting clinical trials have been summarized in Tables 1 and 4, respectively. Here we summarily review the role of lysosomal dysfunction in progressive neurodegenerative disorders.

Present researches shed light on the involvement of lysosomes in pathophysiology of pancreatitis [131]. Normally, pancreatic acinar cells secrete digestive enzymes, many of which are inactive zymogens packed in zymogen granules until they reach the intestines [132]. A canonical hypothesis for pancreatitis is intra-acinar trypsinogen activation [131]. Excessive secretagogue-receptor or bile salt receptor activation causes abnormal peak-plateau calcium signal, which disrupts zymogen granule exocytosis, blocks secretion, and initiates the formation of endocytic vacuoles [133, 134]. The vacuoles containing trypsin and trypsinogen fuse with lysosomes containing cathepsin B that transforms trypsinogen into trypsin [131, 135]. Alternatively, macrophages phagocytose extracellular zymogen-containing vesicles and generate active trypsin [136]. The unstable endocytic vacuoles would rupture and release trypsin and cathepsin B into cytoplasm, thus provoking apoptosis or necrosis [137]. Additionally, calcium or tumor necrosis factor-α (TNF-α) from recruited leukocytes are able to initiate necroptosis in acinar cells [138, 139].

However, debates about the intra-acinar trypsinogen activation exist, because cathepsins co-localize with digestive enzymes in physiological autophagy without manifestation of pancreatitis [140]. Lysosomal dysfunction in pancreatitis is indicated by the findings that the maturation of cathepsin B and cathepsin L are reduced and that autophagosome formation is increased while lysosomal degradation is decreased [140]. Since cathepsin B converts trypsinogen to trypsin [141] while cathepsin L digests both trypsinogen and trypsin [142], the imbalance between cathepsin B and cathepsin L, especially insufficient cathepsin L maturation, may contribute to trypsin accumulation and development of pancreatitis [140]. Besides abnormal cathepsins, disruption of autophagy-related genes, ATG5 [143], ATG7 [144], or LAMP2 [145], induces spontaneous or chronic pancreatitis implying the importance of autophagy in pancreatitis pathology.

In summary, the detailed involvement of lysosomal dysfunction and cathepsin activities in pancreatitis is still controversial. Quantification or semi-quantification of autophagy activities and cathepsin activities in patients or animal models may help clarify the mechanisms. The lysosomal dysfunction in pancreatitis has been summarized in Table 1.

Multi-facetted role of lysosomes in autoimmune disorders has been demonstrated by the finding that autophagy, cathepsin expression, and luminal pH are altered in different immune cells in patients or mouse models [30]. In lupus, activated autophagy supports survival, development, and differentiation in B cells and T cells [146, 147]. Imbalance between pro-inflammatory Th17 and immunosuppressive regulatory T cells (Tregs) is attributed to autophagy activation in these T cell subsets [148]. However, mice deficient in LC3-associated phagocytosis, a non-canonical autophagy, showed blunted clearance of dying cells and elevated inflammatory cytokines, immune complex deposition, and autoantibodies as well as more sever lupus symptom [149]. Autophagy deficiency is also indicated in Crohn’s disease [150, 151] and T cells from rheumatoid arthritis [152], while intensified autophagy in T cells is highlighted in multiple sclerosis [153]. The regulation of glucocorticoid receptors by lysosomal degradation is a contentious topic that lysosome inhibitors, chloroquine or bafilomycin A1, enhance therapeutic effects of glucocorticoid in rheumatoid arthritis [154], while autophagy activator, rapamycin, declines prednisone requirement in lupus patients [155]. The above findings imply various autophagy alterations in different cell subsets from different autoimmune disorders. Thus, interventions targeting autophagy pathway should be carefully designed and tested. The lysosomal dysfunction and lysosome-targeting clinical trials for autoimmune disorders have been summarized in Tables 1 and 4, respectively.

LSDs are a group of heterogeneous inherited metabolic disorders, most of which are attributed to mutations of particular lysosomal hydrolases. The pathology of LSDs can be explained by the accumulation of the substrates of the hydrolases or the secondary products formation instigated by the substrates [30]. An exception is mucolipidosis type IV, where mutated MCOLN1, a non-selective cation channel gene, impairs the fusion of lysosomes with both autophagosomes and late endosomes [156]. Common detrimental autophagy alterations and following changes have been revealed in LSDs. Deregulated mitophagy leads to increased damaged mitochondria engendering reactive oxygen species, ATP production, and imbalanced Ca²⁺ [157]. Microautophagy is impaired in Pompe disease [158]. Defective CMA caused by mutated LAMP2 has been claimed in Danon disease [159]. Detailed descriptions of the deficient proteins in LSDs have been listed in Table 1. The enzyme replacement therapies for LSDs are underpinned by the mechanisms that some cation-independent MPRs are localized at plasma membrane to retrieve escaped extracellular lysosomal proteins [19]. Thus, supplying the appropriate enzymes is able to reverse the pathology of LADs. Alternatively, gene therapies are newly developed to restore deficient genes in patients. Clinical trials testing enzyme replacement therapies and gene therapies of LADs have been listed in Table 4. It will be an interesting topic to compare enzyme replacement therapies with gene therapies.

A special feature for GD is its relationship with increased risk for malignancies, particularly hematological malignancies [160, 161]. The biallelic glucosylceramidase β gene mutation causes glucocerebroside accumulation in lysosomes in macrophages, the Gaucher cells (GCs) [162] that conglomerate to form Gaucheromas in liver, spleen, bone marrow, and even nervous system [160, 161]. Patients with GD are more frequent to develop multiple myeloma and several hematological malignancies [161, 163]. Additionally, elevated frequencies of malignancies in kidney, liver, prostate, testis, brain, bone, colon, and melanoma are also related with GD [161, 164]. The predisposition of malignant tumor could be explained, at least partially, by an altered microenvironment related to phenotypes of GCs [160]. Impaired autophagy in GCs led to p65-nuclear factor-kappaB (NF-κB) activation following with IL-1β secretion by inflammasome pathway [165]. However, anti-inflammatory interleukin-13 (IL-13) was also increased [166]. Several works demonstrated that the phenotype of GCs resembled alternatively activated macrophages [167, 168]. In fact, M2-like GCs were surrounded by spleen cells with strong and frequent M1-like markers such as IL-1β and monocyte chemotactic protein-1 (MCP-1) [167]. We propose that the existence of both M1-like and M2-like macrophages in GD tissues resembles a sterile chronic inflammation environment facilitating malignancies development in various mechanisms [169‐172]. In a word, abnormal lysosomal storage disorder in macrophages is closely related to macrophage activation that may create an inflammatory microenvironment suitable for tumorigenesis.

Massive knowledge about tumor cells has been collected over the past decades. Tumor cells are now considered as a group of cells that contain unstable genome with numerous mutations and are able to proliferate continuously, evade cell death, induce angiogenesis, invade, metastasize, avoid immune supervision, resistant to chemotherapy, and resistant to radiotherapy [174]. Paradigms such as epithelial-mesenchymal transition and cancer stem cells have been proposed to explain and link these malignant behaviors. Additionally, activities of lysosomes are implicated in the phenotypes of malignant cells. The overarching lysosomal activity, autophagy, has been shown to inhibit tumorigenesis, yet promote tumor progression. Autophagy prevents tumorigenesis by inhibiting the transformation of pre-malignant cells [175, 176]. Genotoxic ROS and dysfunctional mitochondria are eliminated by autophagy [177, 178]. Moreover, autophagy could process genomic instability, such as extruded cytoplasmic chromatin fragments [179], micronuclei [180], and endogenous retrotransposons [181]. However, elevation of autophagy has been shown in primary tumor cells and cell lines [182, 183]. Comparing the differences of autophagy between pre-malignant cells and cancer cells will be an interesting topic. Autophagy, as well as other lysosome-related mechanisms facilitate tumor progression by promoting malignant phenotypes mentioned above. Here we focus on the mechanisms through which lysosomes regulate proliferation, invasion, radiotherapy resistance, and chemotherapy resistance.

Massive infiltrated TAMs in tissues always indicate chemotherapy resistance, tumor angiogenesis, immune suppression, metastasis, and poor prognosis [238‐240]. Therefore, TAMs have been a hotspot for a few decades. Polarization of TAMs has been proposed to give an outline of diverse phenotype changes of TAMs, which are able to assume distinct and metastable phenotypes in response to different environmental stimuli [241, 242]. Although more complicated spectrum of macrophage polarization have been proposed [243, 244], the paradigm of M1/M2 is now wildly used. M1 macrophages, namely activated macrophages, show pro-inflammatory phenotype, which is stimulated by ligands of TLRs, interferon gamma (IFN-γ), tumor necrosis factor alpha (TNF-α), and pathogenic microorganisms [241, 245‐247]. On the contrary, transforming growth factor beta (TGF-β), interleukin-4 (IL-4), interleukin-10 (IL-10), and IL-13 induce M2 macrophages, also known as alternatively activated, macrophages that serve as anti-inflammatory factors [241, 242, 248‐251]. TAMs are broadly M2-like macrophages even though they may temporarily express M1-like phenotypes at the early stage of tumor growth [252]. Lysosomes in TAMs have been shown to participate in polarization regulation, as well as additional behaviors distinct from macrophage polarization, such as immune mediation, stromal degradation, and angiogenesis.

Besides macrophages, DCs are the other subset of antigen-presentation cells that prime CD8⁺ T cells by MHC-I or CD4⁺ T cells by MHC-II. Functions of DCs are negatively regulated by a number of factors in TME, such as IL-6, IL-10, and TGF-β [292]. Mechanistically, IL-6 reduced human leukocyte antigen (HLA)-DR surface expression levels by lysosomal protease, cyclooxygenase 2, arginase, and STAT3 activation, thus attenuating T cell stimulation [293]. Despite the mechanisms that inhibit activities of DCs, a group of LAMP3⁺ DCs in hepatocellular carcinoma have been found to possess the ability to regulate multiple T cell subsets and migrate from tumors to hepatic lymph nodes [294]. An evaluation of esophageal squamous cell carcinoma samples from 80 patients also found that infiltrating LAMP3⁺ DCs were positively correlated with intratumoral CD8⁺ T cells and favorable prognosis indicating a potential important role of LAMP3 in antigen presentation by DCs [295]. Although lysosomal proteases are involved in inhibition of HLA-DR and antigen presentation [293], conditional knockout of vacuolar protein sorting 34 (Vps34), an autophagy-related protein, in DCs caused defect in corpse-associated antigen cross-presentation to MHC-I-restricted T cells and higher incidence of lung metastases by melanoma [296]. However, classical MHC-I and MHC-II antigen presentation pathways were enhanced in Vps34 knockout DCs [296], indicating multi-facetted mediation of antigen presentation by autophagy. In summary, the role of lysosomes in DCs during infection has been extensively studied, but it remains largely fuzzy how lysosome-related mechanisms in DCs are involved in anti-tumor immunity and immune evasion in TME.

In tumor, another prevalent subset of stromal cells is CAFs that interact with tumor cells through oxidative stress and NF-κB activation to secrete plenty of cytokines (e.g., IL-6, IL-8, IL-10, and IFNγ), growth factors (e.g., basic fibroblast growth factor (FGFβ) and TGFβ2), and other factors (e.g., MMP-2, MMP-9, and fibronectin) [297, 298]. Several cytokines, such as IFNγ, IL-6, IL-8, IL-10, TGFβ, and TNFα, induce autophagy in CAFs [298]. Autophagy in CAFs degrades caveolin-1, which in turn enhances autophagy in CAFs through a feed-forward mechanism [299]. Autophagy in CAFs promotes tumor progression by providing recycled nutrients, protecting from apoptosis, and inducing genomic instability in cancer cells [299, 300]. Additionally, down-regulation of breast-cancer susceptibility gene 1 (BRCA1) [301] and activation of peroxisome proliferator-activated receptor γ (PPARγ) [302] in CAFs also increases autophagy and promotes tumor growth in vivo.

Downstream mechanisms that are mediated by CAF autophagy during tumor progression include promoting chemotherapy resistance, EMT, and cancer stem cells. Contradictory associations have been reported in CAF autophagy and chemotherapy resistance of pancreatic cancer, where inhibition of CAF autophagy augmented [303] or diminished [304] anti-proliferation effect of gemcitabine. Triple-negative breast cancer cells showed potentiated migration, invasion, proliferation, and EMT after treatment with conditional medium from CAFs with a high level of autophagy, while conditional medium from CAFs pretreated with 3-Methyladenine, an autophagy inhibitor, did not have this effect [305]. Stemness of luminal breast cancer cells was enhanced depending on TLR4 stimulated by high-mobility group box 1 (HMGB1) from autophagic CAFs indicating high LC3II/TLR4 to be poor prognosis markers for breast cancer [306].

In summary, cross-talk between CAFs and tumor cells through oxidative stress, cytokines, and nutrients depend largely on autophagy resulting in more malignant behaviors of tumor cells. Disturbing autophagy in CAFs seems to be promising to control tumor progression.

In TME, T cells are the group of cells responsible for killing tumor cells. Upon recognition of tumor antigen by T-cell receptor, cytotoxic T lymphocytes release perforin and granzyme B from secretory lysosomes to kill target cells [307]. Autophagy has been revealed to influence T cell homeostasis, differentiation, function, and aging [308]. Yet, there are only a few researches focusing on the detailed role of lysosomal activities, such as autophagy, in T cell behaviors in TME.

Elevated extracellular potassium owing to necrotic cells reduced nutrient uptake in CD8⁺ T cells thereby inducing autophagy [309]. Elevated autophagy triggered metabolic and epigenetic reprogramming resulting in suppressed CD8⁺ T cell effector yet preserved stem-like behaviors including self-renewal, expansion, and multipotency, which suppressed B16 tumor growth and improved survival of mice in general [309]. Differently, another research shows that knockout of autophagy-related genes in T cells enhanced their immunity against breast cancer characterized as promotion of effector memory phenotype and IFN-γ and TNF-α production [310]. Atg5^−/− CD8⁺ T cells presented promoted glucose metabolism, altered histone methylation, and upregulated transcription of metabolic and effector genes [310]. The contradictory results may be explained by different baselines of control groups, in other words, high extracellular potassium in melanoma may not exist in the model of breast cancer.

As for CD4⁺ T cells, autophagy selectively degrades PU.1, the main T_H9 cell transcription factor, and inhibited anti-tumor immunity [311]. On the contrary, genetic or pharmacological inhibition of autophagy restored anti-tumor effects by enhancing IL-9 production from T_H9 cells [312]. Other CD4⁺ subsets were not modulated by autophagy in TME [312].

Treg cells are an immune-suppressive subset of T cells in TME. Conditional knockout of Atg5 or Atg7 in Treg cells impaired survival and stability of Treg cells, thus suppressing colon adenocarcinoma growth, which was underpinned by upregulated mTORC1, c-Myc, and glycolytic metabolism owing to autophagy deficiency [313]. Similar mTORC1-related hyper-glycolytic metabolism has been found in Treg cells deficient in lysosomal TRAF3-interacting protein 3 (TRAF3IP3) [314]. Knockout of TRAF3IP3 in Treg cells also booted antitumor responses [314].

In general, genetic or pharmacological inhibition of autophagy in T cells is able to promote anti-tumor immunity. The link between autophagy and alteration of metabolic states in different subsets of T cells from TME needs more intensive illustrations.

Although lysosomal enzymes are usually synthesized in ER and modified in Golgi network, some extracellular enzymes can be taken up and delivered to lysosomes, which makes enzyme replacement therapy possible [315]. Decades ago, patients with GD were treated with glucocerebrosidase extracted from placentae. Afterward, recombinant enzymes produced by CHO cells were invented [316]. At present, enzyme replacement therapy for type 1 Gaucher disease has been a successful commercial therapy [315]. Additional clinical trials applying enzyme replacement therapies for other LSDs have been summarized in Table 4. More effective enzyme replacement therapies are expectable. However, there is still a big challenge for the application of enzyme replacement therapy, because we lack efficient methods for systematical delivery, especially across blood–brain barrier. Thus, both recombinant enzymes and delivery strategies are of great importance for enzyme replacement therapies.

Autophagy is the key process in the development of atherosclerosis, neurodegeneration diseases, autoimmune disorders, and tumor, especially in orchestrating macrophage phenotypes. As mentioned above, activating autophagy by Latrepirdine and CCI-779 are potential therapy methods for PD [114] and HTT [124], respectively. This complex cellular process has many feasible targets for intervention. Among the set of proteins constituting the pathway that regulates autophagy, mTOR is the central one for targeting. Table 4 shows plenty of clinical trials where mTOR inhibitors, such as temsirolimus, sirolimus, and rapamycin, are applied alone or in combination with other therapies for atherosclerosis, autoimmune disorders, and malignant tumors. Although it is well known that inhibition of mTOR would elicit autophagy, limited results are available to prove that antiatherogenic effects of mTOR inhibitors are achieved via autophagy induction [74]. Everolimus cleared macrophages in plaques through autophagy-dependent cell death, a downstream event of mTOR inhibition [317]. In fact, more researches demonstrate anti-inflammatory effect to be responsible for antiatherogenic outcome of mTOR inhibitors, such as rapamycin [318], everolimus [319], and sirolimus [320]. The anti-inflammatory effect may also be caused by balancing M1 and M2-like macrophages through mTOR intervention [70]. Moreover, two mTOR inhibitors, LY294002 and PP242, evoked transport of TFEB to nucleus [9], which would initiate lysosomal biogenesis leading to ameliorating lysosomal dysfunction [58]. Besides targeting mTOR, other agents can also modulate cell phenotypes via regulating autophagy. For example, trehalose and hydralazine might have therapeutic effects for AD through inducing autophagy (NCT04663854 and NCT04842552). By inhibiting autophagy, bafilomycin A1 [256], chloroquine [321], and 3-methyladenine [322, 323] promoted pro-inflammatory markers or suppressed anti-inflammatory markers in macrophages, which may, at least partially, underpin the huge amount of clinical trials for malignant tumors applying chloroquine or hydroxychloroquine alone or in combination with other therapies in Table 4. In contrast, anti-inflammation function of sorafenib [324], urolithin A [325], and spermine [326] are autophagy-dependent. It’s an interesting question why both autophagy stimulator, such as rapamycin, and inhibitor, such as hydroxychloroquine, are largely applied in trials for tumors. A possible explanation is the multifaceted role of mTOR pathway that not only inhibits autophagy but also regulates anabolism and proliferation [327]. An additional reason might be diverse autophagy-related mechanisms in different cell subsets at different stages. For instance, autophagy suppresses M2-like polarization of TAMs [253] but promotes survival and stability of Treg cells [313]. The positive and negative relationships between autophagy and cell death have been found in different cancer cells [232‐237]. Thus, administration of autophagy inhibitors or activators should be carefully determined and specifically delivered.

As mentioned above, cathepsins from both tumor cells and TAMs promoted malignant phenotype of tumors, including invasion, EMT, radiotherapy resistance, angiogenesis, and anti-inflammatory TME. In the laboratory, cathepsin B inhibitor, CA074Me, inhibited invasion [192] and 5FU-induced inflammasome activation that intensified growth of tumor cells [289]. Similarly, CA-074 dampened E-cadherin down-regulation induced by cathepsin B, implying its role in impeding cathepsin B-induced EMT [195]. Cathepsin S inhibitory antibody, Fsn0503, also inhibited invasion of colon adenocarcinoma cell line MC38 [286]. An unnamed cathepsin S selective inhibitor, compound 6, reduced pro-inflammatory CCL2 expression in a dose-dependent manner [291]. Z-FY-CHO, specific cathepsin L inhibitor, increased radiosensitivity in glioma cell line U251 [214, 216]. Cathepsin L/K inhibitors, KGP94 and KGP207, were able to reduce invasion of breast cancer cell line 4T1 and M2-like markers of TAMs [284]. A general cysteine cathepsin inhibitor, E64, significantly mitigated invasion of tumor cells [283]. JPM-OEt, a cell-permeable derivative of E64, diminished drug resistance [279], tumorigenesis, angiogenesis, proliferation, and invasion [197, 254]. Although a host of cathepsin inhibitors have been developed in pre-clinical researches, only a few, such as Odanacatib, have entered clinical trials. Unfortunately, the effectiveness of cathepsin inhibitors is not always satisfying. Side effects may be caused by the inhibited important physiological function of cathepsins and nonspecific inhibition [328]. Odanacatib, the only selective inhibitor that entered phase III clinical trials, discontinued due to stroke-related side effects [190]. Further cathepsin inhibitors may need specific delivery or local administration and combination with other therapies. For example, since cathepsins regulate radioresistance [214], a combination of radiotherapy with cathepsin inhibitors may potentiate therapeutic effect.

Overexpression of tumor suppressor genes or downregulation of tumor promoter genes is reasonable strategies for therapies. A large number of transfection agents have been developed nowadays, but in vivo transfection of primary cells, especially macrophages, seems to be more difficult than others. More than a decade ago, a nonviral DNA delivery system, bacterial ghost, was invented [329]. Gram-negative bacterial cells were perforated after induction by gene E from bacteriophage to create empty envelopes for DNA loading. The envelopes filled with DNA were termed bacterial ghost that was able to be engulfed by macrophages. In some ways, could DNA escape from these bacterial ghost entrapped in phagoendolysosomes. Although the details remain elusive, the genes in the plasmid DNA could be expressed in macrophages [329]. Recently, this system was applied for LAMP2A knockdown in TAMs from breast cancer, which curtailed TAM anti-inflammatory activation and tumor growth [260]. Besides bacterial ghost, a novel nanomaterial named PEG = MT/PC NPs was invented [330]. This nanomaterial with dual pH-responsiveness was able to deliver specific siRNA to knockdown vascular endothelial growth factor (VEGF) and placental growth factor (PIGF) in both TAMs and breast cancer cells. PEG = MT/PC NPs internalized in TAMs via mannose-mediated access escaped from endosome/lysosome and released siVEGF and siPIGF resulting in suppression of tumor growth and lung metastasis [330]. At present, there is a lack of clinical trials applying such delivery systems for gene transfection. The application of nucleic acid delivery systems targeting lysosomes needs to be more intensively explored.

Since TAMs promote tumor development by too many mechanisms, some researches attempted to kill TAMs instead of reprogramming them. In order to selectively delete tumor cells and TAMs, a photosensitizer, mannose-conjugated chlorin (M-chlorin) was used in photodynamic therapy [331]. M-chlorin localized mainly in lysosomes and endoplasmic reticula revealing strong cytotoxicity for both cancer cells and M2 macrophages [331]. Modified classical chemotherapy agent, sialic acid-polyethylenimine-cholesterol modified liposomal doxorubicin (DOX-SPCL), was also able to selectively bind to TAMs, concentrate in lysosomes, and exhaust TAMs [332]. Tumor-baring mice administrated with DOX-SPCL manifested exhausted TAMs and inhibited tumor growth without significant side effect [332]. Alternatively, MEN 4901/T-0128 was designed as a cytotoxic prodrug that could release cytotoxic T-2513 after digestion by lysosomal cathepsin B at pH values ranging from 3 to 5 [333]. THP-1 derived macrophages or primary macrophages from peritoneum and spleen internalized MEN 4901/T-0128 into lysosomes and released T-2513 in the culture media, which generated sufficient concentration of T-2513 to kill human carcinoma A2780 cells [266, 333]. Further improvement of these strategies might be realized if they are combined with methods that inhibit macrophage recruitment.

Several kinds of nano/micro-particles have been developed to carry tumor antigens into antigen presentation cells, such as DCs and macrophages. These novel particles excel at selective transport and enhanced immune stimulation. Porous silicon micro-particles developed by Xia et al. achieved prolonged early endosome localization, potentiated cross-presentation, and activated type I interferon responses when engulfed by DCs [334]. Cancer vaccines based on this micro-particles loaded with human epidermal growth factor receptor 2 (HER2) stimulated robust CD8⁺ T cell immunity against HER2⁺ mammary gland tumors in mice [334]. Besides protein, mRNA vaccine can also be loaded into nanoparticles. Wang et al. created lipid-coated calcium phosphate nanoparticles to co-deliver tyrosinase-related protein 2 (TRP2) mRNA and programmed cell death-ligand 1 (PD-L1) siRNA into DCs [335]. The cargoes can be released when calcium phosphate core is dissolved in endo-lysosomal compartment, which elicited strong specific T cell response and humoral immune response and downregulated PD-L1 in DCs in mouse model of B16F10 melanoma [335]. Another interesting nanoparticle, metal–organic framework, was fabricated in an attempt to achieve powerful delivery of tumor-associated antigens [336]. Loaded tumor-associated antigens could be released when the metal–organic framework is degraded in acidic lysosomes in macrophages. This novel nanoparticle exerted enhanced prominent antitumor outcome in B16-OVA melanoma model without obvious toxicity in mice [336]. In summary, these novel nano/micro-particles have achieved targeted delivery of antigens and strengthened immune stimulation, both of the two advantages are crucial for the development of tumor vaccines in the future.

In order to reverse M2-like polarization of TAMs, several compounds have been invented, and their effects are lysosome-dependent. In a mouse breast cancer model of 4T1, cationic polysaccharide spermine modified pullulan (PS) was internalized into lysosomes of macrophages and reprogramed M2 macrophages to M1, which increased CD4⁺ and CD8⁺ T cells and inhibited tumor progression [337]. Other prevalent compounds that influence phenotypes of TAMs are the family of TLRs agonists. As mentioned above, TLR3/4/7/9 are located in lysosomal membranes, whose agonists are able to elicit pro-inflammatory phenotype of TAMs. Targeting TLR7 by imiquimod (R837) which was delivered to TAMs via polymer micelles stimulated antitumor immune response [338]. In clinical trials listed in Table 4, these agonists are usually combined with chemotherapy, radiotherapy, and bio-therapy. Combination of TLR agonists with tumor antigens and anti-PD1 antibodies seems to be promising for immunotherapy.

In summary, a great deal of methods targeting lysosomes have been designed. In laboratory, these strategies have yielded promising therapeutic effects. Although a great many compounds have been developed, lysosome-targeting interventions in clinical trials are limited. Rapamycin, hydroxychloroquine, and chloroquine are the most commonly used interventions. Compelling results might be found in clinical trials applying novel lysosome-targeting interventions. Notably, the side effects and off-target effects of these compounds are the major concerns in their further application. Temporary and controllable, instead of permanent, inhibition or activation may facilitate limiting side effects, because they can minimize perturbation of important physiological function. Furthermore, specific groups, such as mannose, would help improving the specificity of these lysosome-targeting nanoparticles.