Curcumin’s poor absorption in vivo, rapid metabolism, low bioavailability, and short half-life in the gastrointestinal tract limit its efficacy (Anand et al., 2007). The main modes of curcumin administration include oral administration and injectable administration, and the bioavailability of the two differs considerably. The study showed that the serum level of curcumin was 0.36 ± 0.05 μg/mL in rats after intravenous injection of curcumin 10 mg/kg, while the serum level of the drug was only 0.06 ± 0.01 μg/mL after oral administration of 50 times the dose of curcumin (Yang et al., 2007). Zhang et al. studied the pharmacokinetics and absolute bioavailability of curcumin in rats after intragastrical administration, intraperitoneal administration and sublingual. The experimental results showed that the metabolic process of curcumin in rats in the three administration routes was consistent with the two-compartment model, with t1/2 of 159.28, 90.79 and 11.96 min, respectively. The area under the drug-time curve (AUC0-∞) was 86.36, 73.39 and 104.62 mg min·L⁻¹, respectively. The absolute bioavailability of intraperitoneal administration was 35. 07%, while that of intragastrical administration was only 4.13% (Zhang et al., 2011). The low oral bioavailability of curcumin may be due to a first-pass effect and a certain degree of intestinal metabolism, especially the high binding rate from glucuronidation and sulfation into its water-soluble metabolites and excretion in the urine, so that the drug concentration in the plasma is low (Sharma et al., 2007). Most of the research on curcumin in recent years has been looking to overcome the problem of low bioavailability of curcumin. Researchers have established nanoparticles, liposomes, micelle, phospholipid complexes and other methods to prolong the residence time of curcumin in vivo, increase the biological permeability of curcumin, slow down the metabolic process of curcumin, and finally improve the bioavailability of curcumin (Abd El-Hack et al., 2021). For example, curcumin loaded nanofibers can improve sustained release behavior and bioavailability and inhibite MCF-7, HepG2 and L929 cell lines. The active targeting of liposomes can improve the solubility, stability and drug loading efficiency of curcumin and the toxicity to tumor cells (Zhang et al., 2019).

After entering the body, curcumin can be distributed in plasma, liver, kidney and brain tissue, indicating that curcumin can cross the blood-brain barrier (Shehzad et al., 2010). The metabolic sites of curcumin are mainly in the intestine and liver. After oral administration, curcumin is first absorbed and metabolized from the intestine, then enters the systemic blood circulation, and finally is taken to the liver for extensive metabolism, and curcumin undergoes a complex transformation during absorption through the intestine (Wang and Qiu, 2013). The metabolism of curcumin in vivo may include two stages: reduction, dehydroxylation reaction and glucuronidation, sulfation. The metabolites of curcumin are related to the mode of administration, and when administered orally, curcumin is metabolized primarily to glucuronates and glucuronate/sulfate conjugates (Asai and Miyazawa, 2000). However, when administered intravenously or intraperitoneally, curcumin is metabolized to tetrahydrocurcumin, hexahydrocurcumin, and octahydrocurcumin (Ravindranath and Chandrasekhara, 1981). Some researchers also believe that curcumin-glucuronic acid, dihydrocurcumin-glucuronic acid, tetrahydrocurcumin-glucuronic acid and tetrahydrocurcumin-glucuronic acid are the main metabolites of curcumin in vivo (Pan et al., 1999).

There are no studies on the excretion kinetics of curcumin, and similar to most drugs, curcumin and its metabolites are mainly excreted through the kidneys and bile, while quercetin can modulate the pharmacokinetics of curcumin in vivo and promote the excretion of curcumin via bile acids (Kim et al., 2012).

Previous studies have reported that curcumin has no obvious acute toxicity, subchronic toxic damage, and no potential mutagenic or teratogenic effects in animal toxicity experiments, and few adverse reactions have been reported, indicating that curcumin has high safety (Zhou et al., 2019). Nelson et al. found that a single oral high dose of curcumin 12 g/d still did not produce any side effects (Nelson et al., 2017). According to the reports of the Joint FAO/WHO Expert Committee on Food Additives and European Food Safety Authority the Acceptable Daily Intake values for curcumin are 0–3 mg/kg (Kocaadam and Şanlier, 2017). In a phase I clinical trial, Sharma et al. gave oral curcumin capsules (0.45–3.6 g/d) to patients with advanced colon cancer. Two patients had diarrhea after 1 and 4 months of treatment, respectively, and another patient had nausea, and the symptoms spontaneously relieved after continued treatment. In addition, elevated serum alkaline phosphatase levels were found in 4 patients and elevated serum lactate dehydrogenase levels in 3 patients on blood tests (Sharma et al., 2004). In a dose-escalation study of curcumin conducted by Lao et al., two patients developed diarrhea at 1,000 mg and 12,000 mg, one patient developed headache at 4,000 mg, two patients developed yellow stools and rash at 8,000 mg, respectively, and two patients developed yellow stools and headache at 10,000 mg, respectively, all with toxicity grade 1 (National Cancer Institute, Common Toxicity Criteria v.2.0) (Lao et al., 2006).

The lysosome biogenesis involves two pathways: biosynthetic pathway and endocytic pathway. The biosynthetic pathway is mainly regulated by the microphthalmia/transcription factor E (MiT/TFE) protein family, in which transcription factor EB (TFEB) is being studied mostly. Generally, TFEB can be dephosphorylated and activated by the phosphatases calcineurin (Medina et al., 2015) and protein phosphatase 2 (Medina et al., 2011) or phosphorylated and thus inhibited by the atypical serine/threonine kinase mechanistic target of rapamycin complex 1 (mTORC1) (Martina et al., 2012) and serine/threonine kinases such as mitogen-activated protein kinase 1 (Settembre et al., 2011), protein kinase B (Palmieri et al., 2017), glycogen synthase kinase 3β (Li et al., 2016), and mitogen-activated protein kinase 3 (Hsu et al., 2018). The activated TFEB can transcribe the coordinate lysosomal expression and regulation gene, which then encodes proteins related to lysosomal structure and function, mainly including multiple soluble lysosomal hydrolases (also referred to as acid hydrolases) and integral lysosomal membrane proteins (LMPs). These proteins are first synthesized in the rough endoplasmic reticulum (ER) and then modified by N-linked glycosylation. Next, they are transferred to the Golgi, modified by N-acetylglucosamine phosphotransferase and phosphoglucosidase, and captured by sorting receptors. The common sorting receptors mainly include mannose-6-phosphate receptors, sortilin, and lysosomal integral membrane protein 2, and different sorting receptors are involved in the transportation of different types of soluble lysosomal hydrolases and LMPs (Hunziker and Geuze, 1996). Lastly, clathrin/AP-coated vesicles are formed after budding on the Golgi apparatus and transported out. Meanwhile, early endosomes are derived from cell membranes via endocytosis and their interiors can be gradually acidified through V-ATPase, simultaneously fused with the same type of endosomes to increase their volume, and gradually transformed into late endosomes (LEs). The internal environment of LEs is further acidified to release RAB5 and capture RAB7 (Rink et al., 2005; Saftig and Klumperman, 2009). Subsequently, clathrin/AP-coated vesicles are mutually integrated with LEs, receptors are recycled, and finally mature primary lysosomes are formed after the dephosphorylation of lysosomal enzymes. Given these findings, lysosome formation is a finely-tuned dynamic biological process.

During the investigation of programmed cell death, numerous investigators have gradually identified the roles of lysosomes in cell death. As a digestive center, lysosomes can participate in the regulation of cell autophagy, apoptosis, pyroptosis, and ferroptosis through many pathways.

The metabolic functions of lysosomes are mainly realized through hydrolytic enzymes such as proteases, lipases, nucleases, etc. (Perera and Zoncu, 2016). These enzymes hydrolyze macromolecules and transport nutrients to the cells through specialized channels in the lysosomal membrane (Sagné et al., 2001) Lysosomes act as the cellular energy sensor and regulation center and its specific localization in cells is also regulated: there are mainly two localization statuses, juxta-nuclear position, and peripheral position. Lysosomal trafficking is accomplished via distinct transport complexes in two directions: RAB-7/RAB interacting lysosomal protein is responsible for the trafficking in a juxta-nuclear direction; while ADP ribosylation factor like GTPase 8B/SKIP confers translocation toward the cell peripheral position. Adenosine 5‘-monophosphate-activated protein kinase (AMPK) and mTORC1 are two key kinase complexes that sense cellular nutrient levels, regulate metabolic balance, and modulate cell growth, both of which are opposite and complementary (González et al., 2020). Activation of mTORC1 requires the co-stimulation of nutrients, energy, and growth factors, all of which are indispensable. S-adenosylmethionine sensor upstream of mTORC1 is a sensor for methionine, Sestrin2 for leucine, and cytosolic arginine sensor for MTORC subunit for arginine (Rebsamen et al., 2015; Wang et al., 2015; Lawrence and Zoncu, 2019). When cells sense abundant nutrients in these sensors, growth factors induce the phosphorylation of protein kinase B (AKT) and negatively regulate TSC complex subunit 2, activating the Rheb GTPase (Gwinn et al., 2008; Merchut-Maya and Maya-Mendoza, 2021). Amino acids can activate Rag GTPases by altering their binding to guanine nucleotides (Lin and Hardie, 2018). Upon activation of Rag GTPases, RagA/B binds to GTP, and RagC/D binds to GDP (Cao et al., 2021). Activated Rag GTPases can specifically bind to Raptor, a component in mTORC1, recruit mTORC1 to lysosomes, and activate mTORC1, thereby promoting cell metabolism (Saxton and Sabatini, 2017). Interestingly, mTORC1 stimulates the anabolism of the cells with high nutrition, whereas AMPK induces the catabolism of the cells with low nutrition (Jia and Bonifacino, 2019). Activation of AMPK involves a canonical pathway and a non-canonical pathway. In the canonical pathway, the recruitment of AXIN and liver kinase B1 (LKB1) to lysosomes leads to the dissociation of mTORC1 from the membrane surface, and LKB1 phosphorylates AMPK, thereby inhibiting catabolism and cell growth, which can be facilitated by adenosine monophosphate (Sarbassov et al., 2005; Weber et al., 2020). In the non-canonical pathway, AMPK activation also requires the involvement of the framework protein AXIN. Glucose insufficiency and a decline in the levels of its metabolite fructose 1,6-bisphosphate (FBP) can be sensed by V-ATPase-bound aldolases at lysosomes. Once FBP is insufficient to bind to the aldolase, the aldolase can inhibit transient receptor potential cation channel subfamily V (TRPV) calcium channels on the ER, which converts the low glucose signal into a low calcium signal; meanwhile, TRPV can simultaneously bind to the V-ATPase on adjacent lysosomes, resulting in a conformational change that the translocation of architectural proteins AXIN and LKB1 to the lysosomal surface activates AMPK on the lysosomal membrane (González et al., 2020). Additionally, lysosomes regulate the metabolism of metal ions such as iron, calcium, and zinc to prevent their accumulation in the cytoplasm and protect against cell damage.

Based on the catabolic and degradative capacity of lysosomes, they play an important role in pathogen recognition and related signal transmission, antigen processing and presentation to T lymphocytes. Moreover, it can combine with dendritic cells, which are the bridge between innate immunity and adaptive immunity (Steinman, 2007). In innate immunity, after phagocytotic cells engulf bacteria, they form phagosomes, which combine with primary lysosomes to form secondary lysosomes and degrade the bacteria internalized by phagocytosis by hydrolytic enzymes (Novak et al., 2010). Toll-like receptors (TLRs) that recognize pathogen‐derived molecular products not only exist on the surface of immune cells, but also on the lysosomal membrane (Nakagawa et al., 2004). By restricting TLR signaling to endolysosomes, pathogens are broken down and nucleic acids are released as self-DNA and RNA are disrupted in the external environment. TLRs bind to NA ligands to endolysosome for wound processing, and because of the presence of the molecular chaperone Unc-93 homolog B1, pathogens are broken down and nucleic acids are released as self-DNA and RNA are disrupted in the external environment (Kim et al., 2008). It has also been further demonstrated that several TLRs need to undergo protein hydrolysis (Sepulveda et al., 2009). The activation of TLR9 and TLR7 signaling pathways in dendritic nuclear epithelial cells requires AEP (a lysosomal cysteine endopeptidase) hydrolysis (Ewald et al., 2011). And in macrophages, histone proteases are similarly involved in TLRs activation (Ewald et al., 2011; Jakoš et al., 2019). Alternatively, there is a negative regulation of lysosomal proteases in host defense in the innate immune response. Attenuation of the protease would enhance major histocompatibility complex (MHC) -restricted antigen presentation because it would disrupt T lymphocyte epitopes (Watts, 2022). In adaptive immunity, TLRs on the lysosomal membrane recognize pathogens, stimulate proinflammatory signals on the one hand, and hydrolyze pathogens into antigenic peptides on the other hand. It promotes the activation of MHC II molecules to cluster of differentiation (CD) 4⁺ T cells and B lymphocytes (Balka and De Nardo, 2019; Lind et al., 2022). Antigen-loaded MHC II molecules can enter the endosomal/lysosomal pathway from the cytoplasm via phagocytosis or receptor-mediated endocytosis. MHC II can become more stable after the removal of nine amino acid residues from MHC II interencapsulated bodies (MIICs) by lysosomal protein hydrolysis. MHC II extends MIICs with the aid of Ia-associated invariant chain and directs MHC II transport to MIICs cytosolic, thereby preventing premature peptide binding to MHC II molecules (Maric et al., 2001; Münz, 2016; Keller and Lünemann, 2017). With the assistance of human leukocyte antigen -DM molecular chaperones, lysosomes can reduce MIICs and class2-associated invariant chain peptide (CLIP) fragments, and catalyze the dissociation of CLIP from the binding site of MHC II molecules (Dengjel et al., 2005; Mellins and Stern, 2014). In addition, Lysosomes can degrade immune checkpoints such as programmed cell death-ligand 1 (PD-L1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and CD70 and other immune checkpoints (Wang et al., 2019; Wang et al., 2020).

Lysosomes can control tumor proliferation by regulating cell senescence induced by growth factors and relevant oncogenes (Tang et al., 2020). Growth factor signal is initiated at the plasma membrane by receptor tyrosine kinases (RTKs), such as epidermal growth factor (EGF) receptor, which is internalized into cellular structures upon EGF stimulation to promote cell growth. Lysosomes are capable of regulating growth factor signaling through the endocytosis of RTKs (Goh and Sorkin, 2013). Insulin like growth factor 2 receptor (IGF2R) can bind to a regulatory polypeptide IGF2 and be degraded in lysosomes, which inhibits IGF2-mediated tyrosine kinase activation, thus exerting a growth inhibitory effect, whereas M6P/IGF2R is responsible for the trafficking of hydrolases. Studies have documented that M6P/IGF2R is aberrantly expressed in liver cancer (Puxbaum et al., 2012), breast cancer (Iwamoto and Barber, 2007), ovarian cancer (Huang et al., 2006), head and neck squamous cell carcinoma (Jamieson et al., 2003), and prostate cancer (Hu et al., 2006). Aberrant expression of M6P/IGF2R can lead to the mislocalization of hydrolytic enzymes, resulting in the failure to activate the anti-proliferative cytokine transforming growth factor (Gemma et al., 2000). In addition, tumor cells can proliferate indefinitely and evade death. In this regard, oncogene-induced senescence in vivo and in cells is an important tumor-suppressive mechanism, where senescence-associated heterochromatin is translocated from the nucleus to the cytoplasm, and tumor cells can degrade the heterochromatin by increasing the number of lysosomes, thus delaying cell senescence (Aird and Zhang, 2015), arresting the cell growth, and inhibiting proliferation.

Infinite tumor cell proliferation requires substantial energy support, while lysosome-mediated regulation can drive tumor cell metabolism. As mentioned above, on the one hand, lysosomes can boost the synthesis of nutrients such as glucose, proteins, and amino acids through the mTORC1 pathway. Loss of GTPase activating protein for RAGA (GATOR1) has been identified in multiple tumors (Bar-Peled et al., 2013). GATOR1 is an upstream key regulatory complex capable of regulating Rag dimer. In breast cancer cells, reduced levels of GATOR1 affect the localization of mTORC1 in lysosomes and attenuate the decrease of mTORC1, thereby facilitating nutrient synthesis and tumor cell proliferation (Chen J. et al., 2018). On the other hand, KRAS-mutant tumors are featured with the ability to grow in a nutrient-deficient and anoxic microenvironment and maintain energy balance (Guo et al., 2011). Increased autophagy widely exists in these tumors and is critical for tumor growth. Lysosomes can degrade intracellular and extracellular substances for cells by autophagy. Macroautophagy can degrade serum albumin and macromolecules into amino acids in pancreatic and lung adenocarcinomas to maintain energy homeostasis in cancer cells (Commisso et al., 2013).

Curcumin can effectively inhibit the proliferation of tumors. Ribosomal DNA (rDNA) can be transcribed to ribosomal RNA (rRNA) to drive cell growth and cell proliferation. mTORC1 is an upstream regulator of rDNA transcription. Curcumin inactivates mTORC1 via repressing mTOR lysosomal localization, which affects the assembly of the RNA polymerase I transcription initiation complex at rDNA promoters and inhibits the rDNA promoter activity, resulting in decreased rRNA synthesis and, ultimately, suppressed protein synthesis as well as tumor cell growth and proliferation (Xu et al., 2020). Due to the excellent anti-tumor ability of curcumin, an increasing number of curcumin derivatives or analogs have been unveiled. As a derivative of curcumin, hydrazinobenzoylcurcumin can promote the formation of autophagolysosomes, induce autophagy, and inhibit tumor cell proliferation in lung cancer A549 cells in a dose- and time-dependent manner (Zhou et al., 2014). The single use of protein phosphatase 242 (PP242), an inhibitor of mTORC1 and mTORC2, cannot induce cell death in human renal cancer cells. However, combined use of PP242 and curcumin can downregulate the protein levels of Rictor and Akt while inducing Bax activation and reducing myeloid cell leukemia-1 and B-cell lymphoma-2 expression, thus accelerating renal cancer cell death (Seo et al., 2018). Oyanguren et al. found that, on the one hand, curcumin induced ER stress and thus caused uncontrollable unfolded protein response and calcium release, which in turn impaired the stability of the mitochondrial compartment and induced apoptosis; additionally, curcumin could mediate the activation of caspase 8 via CTS and calpains, which in turn induced LMP, leading to disruption of mitochondrial homeostasis; both two pathways mutually acted to boost tumor cell death (Sala de Oyanguren et al., 2020). Wang et al. (Wang et al., 2016) experimentally demonstrated that curcumin increased the levels of HSP70 and LAMP1 during the autophagy of HCT116 colon cancer cells, which induced autophagy through lysosomal activation, thus leading to cell death. Khaket et al. (Khaket et al., 2020) found that silenced CTSC strengthened the anti-tumor potential of curcumin, and curcumin treatment in colon cancer cells induced ER stress and dysregulation of autophagy accompanied by LMP and ROS generation. This LMP triggered the cytosolic CTSB-mediated BID-dependent mitochondrial membrane permeabilization and thus evoked caspase-dependent apoptosis. Chen et al. (Chen QY. et al., 2012) showed experimentally that curcumin caused LMP and cytosolic relocalization of CTSB and CTSD, which caused lung cancer A549 cell apoptosis. Yuan et al. (Yuan et al., 2020) found that curcumin TPP-PEG-PCL nanomicelles could target the mitochondria of breast cancer cells and escape from the capture of lysosomes, which enhanced its pro-apoptotic effect on tumor cells.

As a master regulator of autophagy-lysosome biogenesis, TFEB has been shown to be activated in a variety of tumors to exert tumor-promoting effects (Calcagnì et al., 2016). Besides, MiT/TFE transcription factors can stimulate the transcriptional activation of the RagD GTPase to regulate mTORC1, thereby accelerating tumor growth (Di Malta et al., 2017). Curcumin can regulate the functions of lysosomes to exert anti-tumor activity by directly binding to TFEB and increasing the nuclear translocation and transcriptional activity of TFEB (Zhang J. et al., 2016). Another study suggested that a synthesized curcumin derivative termed C1 could serve as a novel mTOR-independent activator of TFEB. C1 can specifically bind to TFEB and promote its nuclear translocation, thus increasing the protein levels of LAMP1, LC3B-II, and CTSD in the human neuroblastoma cell line SH-SY5Y; also, C1 can enhance autophagy in N2a mouse neuroblastoma cells and Hela human cervical carcinoma cells through the key autophagy genes, ATG5 and Beclin 1 (Song et al., 2016).

Lysosomes also play a significant role in tumor invasion and metastasis. During tumor invasion, tumor cells can induce the formation of adhesion molecules, and lysosomes can target these specific adhesions, which links the growing tumors to their surrounding environment (Hämälistö and Jäättelä, 2016). The degradation of the extracellular matrix (ECM) is essential for further metastasis. On the one hand, various CTS secreted by lysosomes can degrade ECM, which can not only affect the invasion and metastasis of tumors but also indicate a poor prognosis. Lysosome-secreted CTS can be mainly categorized into cysteine peptidases (CTSB, C, F, et al.), serine peptidases (CTSA and G), and aspartic peptidases (CTSD and E). CTS, which are highly expressed in the majority of tumors, can degrade ECM components such as type IV collagen, laminin, and fibronectin, thereby expediting tumor metastasis (Rudzińska et al., 2019). Interestingly, however, CTSX is highly expressed in the early or pre-malignant stage but lowly expressed in the late stage of several tumors compared with normal tissues, and its underlying mechanism needs to be further investigated (Fisher et al., 2015). On the other hand, ECM components can also be endocytosed by specific cell surface receptors and subsequently degraded through the lysosomal pathway. The lysosomally degraded proteins can supply essential nutrients and energy for tumor invasion. Epithelial-mesenchymal transition (EMT) is a crucial process in tumor invasion and metastasis. Cytokines produced during the EMT process can promote lysosomal degradation of E-cadherin and increase cell motility, consequently promoting tumor cell invasion and metastasis (Wu and Hirsch, 2009). Several suppressors of tumor metastasis such as NM23-H1 can also trigger downregulation mechanisms via lysosomal degradation (Fiore et al., 2014). Several regulators have also been shown to affect tumor migration by modulating lysosomal function. EGF mediates anterograde lysosomal movement by activating the p38 mitogen-activated protein kinase pathway, which affects protease secretion and causes tumor cell invasion and infiltration through the basement membrane (Dykes et al., 2017). Activation of TFE3 can promote lysosomal pathway-mediated autophagy in oral squamous cell carcinoma and thyroid carcinoma, which in turn promotes tumor metastasis (Lu et al., 2021; Fan et al., 2022).

In addition, angiogenesis is a key process in tumor growth, invasion, and metastasis. Multiple studies have confirmed that the lysosome-secreted CTSB, D, S, and L also play a vital role in angiogenesis. Upregulation of CTSB expression can increase the activity of vascular endothelial growth factor C and matrix metallopeptidase (MMP) 9, thereby promoting angiogenesis (Yanamandra et al., 2004). CTSD can promote breast cancer angiogenesis by releasing ECM-bound bFGF (Berchem et al., 2002). Active CTSS regulates tumor angiogenesis by controlling the production of type IV collagen-originated anti-angiogenic peptides and laminin 5-derived pro-angiogenic gamma2 fragments (Wang et al., 2006). CTSL acts on the local microvasculature of high-grade serous carcinomas in an autocrine fashion to trigger the transcription and release of carbohydrate-binding protein galectin-1 (Gal 1), hence eliciting tumor angiogenesis (Pranjol et al., 2019). Moreover, RAB7, an LMP, can increase the expression of IL-6 and monocyte chemoattractant protein 1 (MCP-1) through the mTOR pathway, promoting endothelial cell-stimulated melanoma invasion and migration (Zhao et al., 2017).

Curcumin mediates lysosomes to inhibit tumor invasion and migration, which effectively controls tumor progression. Mutated in multiple advanced cancers 1 (MMAC1) is a common tumor-suppressive gene, and curcumin analog C1 can reverse the migration of cholangiocarcinoma owing to MMAC1 inhibition (Jiang et al., 2022). Lysosomes and mitochondria serve as the action sites of curcumin-loaded layered double hydroxide nanoparticles (Cur/LDH NPs); autophagy can enhance the potential of Cur/LDH NPs to inhibit the migration and invasion of A172 glioblastoma cells, the mechanism of which is also confirmed to be correlated with the downregulation of the PI3K/AKT/mTOR signaling pathway (Zhang H. et al., 2016). Curcumin can not only inhibit tumor proliferation by targeting CTS but also affects tumor invasion and migration. Yao et al. (Fei et al., 2016) revealed that knockdown of CTSL significantly promoted curcumin-induced cytotoxic responses, apoptosis, and cell cycle arrest and also suppressed glioblastoma cell migration and invasion, suggesting that CTSL might be a novel target to enhance the anticancer efficacy of curcumin. Another study demonstrated that the derivative of (E)-3-((E)-4-chlorobenzylidene)-5-((5-methoxy-1H-indol-3-yl)methylene)-1-methylpiperidin-4-one (CB-2) was an autophagy inhibitor and blocked the fusion of autophagosomes with lysosomes. CB-2 can increase LC1B-II and SQSTM549 levels associated with autophagosome accumulation in A549 cells, significantly blocking autophagosome and lysosome fusion and thus inhibiting autophagy (Liu Z. et al., 2021). Another derivative of curcumin, (3E,5E)-1-methyl-3-(4-hydroxybenzylidene)-5-(3-indolymethylene)-piperidine-4-one (CUR5g), can also block the recruitment of STX17, a soluble N-ethylmaleimide-sensitive factor attachment protein receptor protein, to autophagosomes through a UVRAG-dependent mechanism, which blocks autophagosome-lysosome fusion; the single use of CUR5g can significantly inhibit the migratory ability and colony formation of A549 cells (Chen et al., 2022).

Chemoresistance is one of the main factors affecting the therapeutic efficacy of tumors, and increasing evidence has shown that lysosomes share a close association with tumor chemoresistance. Hydrophobic weak base chemotherapeutic drugs such as daunorubicin, doxorubicin, lapatinib, vincristine, and nintedanib can be internalized and sequestered by lysosomes through non-enzymatic and non-transporter-mediated mechanisms, thereby preventing them from reaching intracellular target sites and causing tumor MDR (Hurwitz et al., 1997; Groth-Pedersen et al., 2007; Herlevsen et al., 2007; Kazmi et al., 2013; Englinger et al., 2017; Zhitomirsky and Assaraf, 2017). Meanwhile, drugs in the cytoplasm can also be overexpressed inside lysosomes via ABC transporters (ATP-binding cassette transporters) such as P-glycoprotein (P-gp) and ABC transporter A3, thereby prompting lysosomal sequestration of chemotherapeutic drugs (Chapuy et al., 2009; Sinha et al., 2018). The differential pH value between lysosomes and cytoplasm is the driving force for this process (Forgac, 2007). Lysosomes cannot sequester drugs forever, and lysosomal exocytosis can also result in tumor drug resistance. Lysosomal exocytosis is a Ca²⁺-dependent process that can be positively regulated by TFEB (Rodríguez et al., 1997; Medina et al., 2011). Lysosomal exocytosis may reduce drug toxicity and availability at the action site, resulting in MDR (Halaby, 2019). Moreover, overexpression of CTS released by lysosomal exocytosis is associated with drug inactivation, and knockdown of CTSL in SKOV3 ovarian cancer cells can increase the pro-apoptotic effect of paclitaxel (Sui et al., 2016).

Importantly, several chemotherapeutic agents or tumors themselves can also affect the biogenesis of lysosomes and alter the number and function of lysosomes, ultimately enhancing tumor drug resistance. Zhitomirsky et al. experimentally demonstrated significant increases in the size and number of cellular lysosomes 72 h after mitoxantrone treatment in MCF-7 breast cancer cells (Zhitomirsky and Assaraf, 2015). For tumors themselves, on one hand, tumors regulate the size, maturation, and activity of lysosomes via PI3K (Mousavi et al., 2003). On the other hand, to avoid LMP, tumor cells can overexpress cytosolic protease inhibitors and translocate HSP70 from the cytoplasm to the lysosomal lumen, or promote the binding of HSP70 to an endolysosomal anionic phospholipid bis(monoacylglycero)phosphate and acid sphingomyelinase activity, thereby maintaining lysosomal stability and protecting themselves (Silverman et al., 1998; Nylandsted et al., 2004; Kirkegaard et al., 2010).

The combination of curcumin with anti-tumor drugs can effectively improve the efficacy of anti-tumor drugs and overcome drug resistance. The combination of curcumin derivative CUR5g with cisplatin can significantly enhance the sensitivity of human non-small cell lung cancer cells (A549) to cisplatin by suppressing autophagic flux in tumors through blocking autophagosome-lysosome fusion (Chen et al., 2022). Imatinib mesylate is the first-choice drug for the treatment of chronic myeloid leukemia, yet its clinical application is limited by its therapeutic selectivity and drug resistance. The simultaneous delivery of both imatinib and curcumin via nanomaterials induces the release of CTSB from lysosomes, elicits LMP, and promotes cell death, therefore enhancing tumor cell sensitivity to imatinib (Acharya and Sahoo, 2016). Huang et al. constructed a targeting controlled release core-shell nanocarriers based on demethoxycurcumin, which acted on lysosomes and endosomes to form nanoprecipitates and enhanced cytotoxicity, exerting a strong inhibitory effect on anoxic tumor cells with MDR, especially non-small cell lung cancer (Huang et al., 2015). Curcumin can induce excessive mitophagy in thyroid cancer cells, increase lysosomal degradation of mitochondria to trigger autophagic cell death, and promote the redifferentiation of thyroid cancer cells to improve cancer cell uptake of radioiodine-131, finally achieving radiosensitization. Furthermore, ER stress induces high expression of IL-6 and IL-8 to mediate radioiodine resistance in thyroid cancer cells; curcumin as well as the molecularly targeted drug sorafenib can inhibit ER stress to reduce the resistance of thyroid cancer cells to radioiodine (Zhang, 2017). Cai et al. showed that MTH-3, a derivative of curcumin, induced autophagy via TFEB to decrease the viability and trigger intrinsic apoptosis in cisplatin-resistant human oral squamous cell carcinoma cells (CAL27) (Tsai et al., 2022).

TLRs, an important class of protein molecules involved in nonspecific immunity, mainly exist in the membranes and lysosomal membranes of innate immune cells. Specific TLR3, TLR7, and TLR9 are expressed on lysosomes (Balka and De Nardo, 2019). Autophagy induced by TLR4 or TLR3 activation augments the production of various cytokines such as IL6, CCL2/MCP-1, CCL20/MIP-3α, and VEGFA by promoting TNF receptor-associated factor 6 ubiquitination, which ultimately drives lung cancer progression (Zhan et al., 2014). Currently, researchers have developed a variety of TLRs agonists such as poly (I: C) and lipopolysaccharide, agonists of TLR3 and TLR4, both of which can activate M1 macrophages to produce the anti-tumor effector molecule NO, thus inhibiting tumor growth (Müller et al., 2018). However, CpG ODN, a TLR9 agonist, can enhance the antigen-presenting activity of tumor-associated macrophages (TAMs) and improve the sensitivity to adaptive immune responses (Muraoka et al., 2019). In addition to the impact of TLRs on lysosomes in immune regulation, macrophages exhibit promoting impacts on both phagocytosis and polarization. Upon the uptake of particles by macrophages, massive Ca²⁺ is released from the LMP, transient receptor potential mucolipin 1, which facilitates phagosome formation and increases phagocytic efficiency (Samie et al., 2013). Meanwhile, Ca²⁺ released through the lysosomal Ca²⁺ channel mucolipin-1 (Mcoln1) can activate p38 and NF-κB, realizing M1 polarization of macrophages and ultimately contributing to tumor inhibition (Chen D. et al., 2018). At present, immune checkpoint therapy, as a type of cancer immunotherapy, has attracted increasing attention. It has been revealed that lysosomes can act as major blockers of immune checkpoint molecules because lysosomes can momentarily store immune checkpoint proteins such as CTLA-4, PD-L1, etc (Wang et al., 2020). Among them, CTLA-4 is a lymphocyte surface receptor expressed by activated T cells, and lysosomes can degrade CTLA-4 via endocytosis and transfer it to the cell surface via the T cell receptor (Chuang et al., 1997; Iida et al., 2000). Tubeimoside-1, a PD-L1 inhibitor, selectively binds to the mTOR kinase target and inhibits the activation of mTORC1, which results in TFEB nuclear translocation and induction of lysosomal biogenesis as well as reduces the abundance of PD-L1 to enhance the cytotoxicity of T cells against tuomr cells, thereby playing an anti-tumor role (Liu X. et al., 2021). Although the correlation of PD-L1 with lysosomes is less well-studied and clearly defined than that of CTLA-4, it is not difficult to discern a close relationship between PD-L1 and lysosomes.

As an immune modulator, curcumin activates the in vitro apoptotic activity of tumor cells but has no impact on normal cells within the immune system (Mohammadi et al., 2019). In terms of macrophage polarization regulation, experimental results from Mukherjee et al. showed that either intranasal delivery or intraperitoneal infusion of curcumin relieved glioblastoma and re-polarized TAMs toward a tumoricidal M1-like state (Mukherjee et al., 2016). Shiri et al. found that dendrosomal curcumin significantly reduced the size of metastatic breast cancer and promoted the polarization of TAM to M1 phenotype as well as decreased the levels of M1 phenotype-related signal transducer and activator of transcription (STAT) 4 and IL-12 (Shiri et al., 2015). Moreover, curcumin can regulate tumor immune checkpoints. The results of Liu et al. (Liu L. et al., 2021) demonstrated that curcumin could decrease the expression of immune checkpoint ligands, such as PD-L1, PD-L2, and Galectin-9, in head and neck squamous cell carcinoma and affect tumor invasion by mediating EMT. Also, it was found that curcumin effectively restored the ability of CD8 cytotoxic T cells to lyse tumor cells, boosted T cell proliferation, and increased the production of tumor-infiltrating lymphocytes and effector cytokines, while decreasing the expression of PD-1, TIM-3, inhibitory immune checkpoint receptors, and their ligands in the tumor microenvironment, suggesting re-activation of exhausted CD8⁺ T cell responses. Additionally, this study demonstrated that curcumin reduced CD4⁺ CD25⁺ FoxP3⁺ Treg cells and inhibited the expression of PD-1 and TIM-3. MaruYama et al. also revealed that the curcumin analogue GO-Y030 contributed to inhibiting the anti-PD-1 immune checkpoint and reducing regulatory T cell populations in tumor-infiltrating lymphocytes (MaruYama et al., 2021). Hayakawa et al. unveiled (Hayakawa et al., 2020) that curcumin restored the activity of dendritic cells targeting activated STAT3 and stimulated T cell activity in tumor cells and immune cells to strengthen tumor antigen-specific T cell responses, and synergistic treatment with curcumin and anti-PD-1/PD-L1 improved anti-tumor activity. Similarly, Guo et al. also unraveled (Guo et al., 2021) that the combination of curcumin and anti-PD-1/PD-L1 promoted the immune activity in liver cancer and exhibited superior anti-tumor efficacy. In addition to the PD-1/PD-L1 immune checkpoint, curcumin could downregulate the tumor immunotherapeutic molecule CTLA-4 at both protein and mRNA levels (Zhao et al., 2012).