Role of Endolysosome Function in Iron Metabolism and Brain Carcinogenesis

Abstract

Iron, the most abundant metal in human brain, is an essential microelement that regulates numerous cellular mechanisms. Some key physiological roles of iron include oxidative phosphorylation and ATP production, embryonic neuronal development, formation of iron-sulfur clusters, and the regulation of enzymes involved in DNA synthesis and repair. Because of its physiological and pathological importance, iron homeostasis must be tightly regulated by balancing its uptake, transport, and storage. Endosomes and lysosomes (endolysosomes) are acidic organelles known to contain readily releasable stores of various cations including iron and other metals. Increased levels of ferrous (Fe²⁺) iron can generate reactive oxygen species (ROS) via Fenton chemistry reactions and these increases can damage mitochondria and genomic DNA as well as promote carcinogenesis. Accumulation of iron in the brain has been linked with aging, diet, disease, and cerebral hemorrhage. Further, deregulation of brain iron metabolism has been implicated in carcinogenesis and may be a contributing factor to the increased incidence of brain tumors around the world. Here, we provide insight into mechanisms by which iron accumulation in endolysosomes is altered by pH and lysosome membrane permeabilization. Such events generate excess ROS resulting in mitochondrial DNA damage, fission, and dysfunction, as well as DNA oxidative damage in the nucleus; all of which promote carcinogenesis. A better understanding of the roles that endolysosome iron plays in carcinogenesis may help better inform the development of strategic therapeutic options for cancer treatment and prevention.

Introduction

Iron is the most abundant transition metal on Earth and the most abundant metal in human brain [1, 2]. Iron is essential for life, is an indispensable nutrient, and helps regulate numerous biological processes and mechanisms. These crucial roles of iron include DNA and RNA synthesis, embryonic neuronal development [3], oxidative phosphorylation and ATP synthesis [4], heme synthesis, oxygen transport, myelin synthesis [5], neurotransmitter synthesis [6, 7], immune metabolism, and the activity of a multitude of enzymes [8, 9]. Iron cycles between ferric (Fe³⁺) and ferrous (Fe²⁺) forms and affects redox reactions through the formation of complexes with oxygen. In addition, iron forms the iron-sulfur clusters in mitochondria and plays key roles in the electron transport chain [10]. Because of iron’s importance in regulating cellular functions, organisms have developed highly regulated homeostatic systems that balance the uptake, transport, and storage of this element [11].

Excess iron or a disrupted balance of iron homeostasis easily becomes detrimental to cells and is a source of oxidative damage through the formation of free radicals known as reactive oxygen species (ROS). Because iron is a metal capable of transitioning between Fe³⁺ and Fe²⁺ as well as transporting electrons, excess free (labile) iron can produce harmful levels of ROS via Fenton chemistry reactions:

$${\text{Fe}}^{2+}+{\text{H}}_2{\text{O}}_2\to {\text{Fe}}^{3+}+\text{OH}-+\text{OH}\cdot$$

$${\text{Fe}}^{2+}+{\text{O}}_2\to {\text{Fe}}^{3+}+{\text{O}}_{\text{2}}{\cdot }^{-}$$

Increased levels of iron can cause cellular damage and increase the risk of cancer, and dysregulated iron has been widely implicated in carcinogenesis [12, 13]. Numerous studies have examined iron regulation pathways and analyzed links between iron levels and enhanced tumor growth [14–16]. For example, high cellular concentrations of iron result in the conversion of hydrogen peroxide and superoxide into the highly reactive hydroxyl radical, which is a specific oxidant responsible for inducing DNA strand breaks and damage [17–19]. Furthermore, hydroxyl radical-induced DNA damage activates the K-ras and C-raf oncogenes via point mutations and deletions, respectively, which suggest that oncogene activation is a potential mechanism of oxidant-induced carcinogenesis [20]. In addition, increased oxygen levels make iron-sulfur clusters susceptible to oxidation and Fenton chemistry reactions that damage DNA [21, 22], and increased iron-sulfur clusters that accumulated the iron-storage complex hemosiderin were detected in macrophage deposits of brain metastatic tumors [23]. These specific mechanisms involving iron, DNA damage, and oxidant-induced carcinogenesis are supported by clinical findings associated with tumor development and progression.

Lymphomas and gliomas are the most common primary brain tumors in adults [24], and the global occurrence of all brain tumors is 10.8 per 100,000 individuals per year [25]. Glioblastoma multiforme (GBM) is the most aggressive of brain tumors, and despite recent advances, survival remains dismal; 2-year survival is 18%, 3-year survival is 11%, and 5-year survival is 4% in patients diagnosed with GBM during or after 2005 [26]. This demonstrates the need to develop novel therapeutic approaches, and studies investigating subcellular mechanisms associated with carcinogenesis in brain have the potential to inform these approaches. Studies have compared the overall survival data from patients with brain tumors to altered gene expression data necessary for iron uptake into the cell, and the results showed that most of the proteins involved in iron uptake were expressed at high levels and were predictive of poor survival [27–29]. These data suggest that iron plays important physiological roles in brain development and helps maintain healthy cellular functions, and excess iron may play a role as a brain carcinogen.

Iron Uptake into the Body

Iron is absorbed (90%) mostly by duodenum enterocytes [30] and comes from two types of dietary iron in its ferric form; heme (mainly from hemoglobin and myoglobin in red meat) and non-heme (from vegetables) [31, 32]. The divalent metal transporter 1 (DMT1) transports iron across enterocyte membranes and out of the endolysosomes into the cytosol as Fe²⁺ [33]. Ferroportin (FPN) is the primary exporter of intracellular iron that transfers duodenal iron into systemic circulation [34]. Released Fe²⁺ is then oxidized to Fe³⁺ by hephaestin or ceruloplasmin in plasma, and it is eventually bound to transferrin (Tf) in blood [35]. Within the body, it is important to note that iron is primarily complexed with the oxygen-carrying heme protein hemoglobin; 75% of the body’s iron is contained in erythrocytes and this is known as heme iron. Non-heme iron forms complexes with Tf, the iron storage protein ferritin [36] as well as amino acids, albumin, and ATP [37, 38]. The average iron concentration in serum ranges from 9 to 30 μM and it is mostly bound to Tf; including hemoglobin the iron concentration of blood is about 8 mM [39, 40].

Transferrin receptors (TfR) are ubiquitously expressed on plasma membranes of most cells and when combined with holo-Tf (Tf containing diferric iron) a complex is formed and then endocytosed [41, 42]. Fe³⁺ dissociates from Tf in the acidic environment of endolysosomes; apo-Tf (Tf with no iron) still bound to TfR can then be exocytosed back into the systemic circulation [27, 43]. Virtually no iron is excreted; almost all iron is reabsorbed by the kidneys and exported back into circulation [44]. Situations where iron is lost include blood donations, and excessive or abnormal bleeding.

Once iron is endocytosed, the six-transmembrane epithelial antigen of prostate 3 (STEAP3) ferrireductase protein reduces Fe³⁺ to Fe²⁺ within endolysosomes. Fe²⁺ can be released into the cytosol via DMT1 [42, 45, 46] where it can be stored as Fe³⁺ bound to ferritin, can be taken up into mitochondria, or it can be exported from cells by FPN [47, 48]. However, relatively little is known about the concentration of the labile iron pool in endolysosomes and its key roles in physiological and pathophysiological processes.

Iron Uptake into the Brain

Brain iron is essential for many neurological processes [49] and iron levels are mainly regulated by the blood brain barrier (BBB) [50]. General brain iron transport/homeostasis is illustrated in Figure 1. Iron transport into brain uses holo-Tf complexes and non-transferrin-bound iron (NTBI) that enters into endothelial cells [51, 52]. Iron crosses the BBB first by being taken up by the microvascular endothelial cells at the luminal membrane and second by being released into brain interstitial spaces at the abluminal membrane [53]. The holo-Tf-TfR complex is the main pathway for iron transport across the endothelial luminal membrane [54–57], and iron, likely in the form of Fe²⁺, crosses the abluminal membrane, enters into brain parenchyma, gets oxidized to Fe³⁺, and binds Tf [58–61]. The endocytosis of holo-Tf bound to TfR occurs similarly in other cell types in the human body [55]. However, iron transport into brain can bypass the BBB [62] and iron can cross the BBB faster than Tf [63]. The physiological relevance of these phenomena remains unclear [64, 65].

Figure 1. Iron transport and homeostasis in the brain.

Iron homeostasis in the brain is largely controlled by the iron transport through the blood brain barrier, comprised of blood capillary endothelial cells (BCEC). Iron transport is facilitated by transferrin (Tf) and the Transferrin receptor (TfR). Proteins playing an important role in transport and homeostasis include ferroportin (Fp) and ceruloplasmin (Cp), as well as the steap3, DMT1, and ZiP14 genes. Created with BioRender.com.

Neurons, astrocytes, and microglia can all uptake iron by means of the TfR-mediated uptake of holo-Tf [66–71]. In contrast, oligodendrocytes appear to uptake iron through processes involving H-ferritin binding to H-ferritin receptors and endocytosis; there is an absence of TfR and DMT1 in these cells [72–75]. Neurons, astrocytes, and microglia can uptake NTBI by means of a Tf-independent pathway; a process not apparent in oligodendrocytes [58, 76–80]. Because the affinity of transferrin for iron is pH-dependent, acidic environments in brain such as occurs in stroke and cancer [81, 82] results in decreased ability of transferrin to bind iron [83], dissociation of iron from Tf [84], and the uptake of NTBI by neurons and other cells [85].

Homeostatic regulation of iron levels also depends on iron efflux out of cell by iron exporters and ferroxidase. The iron exporter, FPN is expressed in neurons [86, 87], oligodendrocytes [87, 88], astrocytes [87, 89], and microglia [70, 89]. Ceruloplasmin, a ferroxidase, is found in neurons [90, 91] and astrocytes [92, 93], while the ferroxidase hephaestin is found in neurons [94, 95], oligodendrocytes [88, 96], and microglia [88].

Regulation of Iron Transport and Storage in the Brain

Iron levels in brain are second only to liver and vary between different brain regions [97, 98]. The majority of brain iron (33 to 75%) resides within glial cells [99] and is bound to ferritin in its soluble form [100]. Ferritin can bind 2,000 to 4,500 iron molecules [101] and recycle iron for heme synthesis as well as sequester labile iron to protect cells from oxidative damage. Ferritins are found in greater quantity in white matter than in gray matter [102]. Hemosiderin, which is a degradation product of ferritin, stores iron in an insoluble state [103]; iron can be mobilized from hemosiderin [104], however the physiological significance of this pool of iron remains unclear [98]. Despite the presence of iron binding proteins and export systems, incubation of cells with Fe³⁺ results in increases in the intracellular labile iron pool and oxidative damage [105], [106].

While most cells can regulate levels of iron by cell division, neurons are post-mitotic cells and must rely on iron homeostatic mechanisms to avoid iron-induced oxidative stress [105], including downregulating the expression of TfR [107, 108]. Neurons can also control levels of iron at the molecular level via post-transcriptional modification by iron regulatory proteins (IRP1 and IRP2). IRP1 and IRP2 are cytosolic proteins that bind to the iron-responsive elements (IRE) within the untranslated regions of the iron regulatory proteins ferritin, TfR and DMT1 [105].

Iron Content of Different Brain Regions

Iron levels in brain vary greatly between brain regions. The highest iron levels are found in the globus pallidus (GP) followed by the red nucleus, substantia nigra (SN), putamen, and dentate nucleus. The lowest levels of iron are found in regions of the cerebral cortex [97]. GP and SN have the highest levels of stainable iron (3.3–3.8 mM) [97, 98]. The forebrain, midbrain, and cerebellar structures had moderate staining levels of iron, while brain-stem and spinal cord had the lowest intensity of staining [97]. These early findings have been largely confirmed in later studies; high iron concentrations have been noted in GP, dentate gyrus, interpeduncular nucleus, thalamus, ventral palladium, nucleus basalis, and red nucleus [109]. Non-heme iron concentrations in the human parietal cortex were 0.65 mM [97]. Regions of brain associated with motor functions tend to have a higher content of iron [109] and this is likely due to high rates of oxidative metabolism [7], which is important in maintaining ionic membrane gradients, axonal transport, and synaptic transmissions [110].

As humans age, the brain tends to accumulate iron [48], and some regions accumulate more than others. The brain regions that accumulate the most iron with aging are SN, putamen, GP, caudate nucleus, and cortex; other regions such as the locus coeruleus continue to have low iron concentrations throughout life [111]. A number of explanations for this observation include increased BBB permeability, inflammation, redistribution of iron within the brain, and changes in iron homoeostasis [111]. In young, healthy brains, oligodendrocytes within the brain contain the highest iron concentration, while astrocytes have a low cellular iron content [112]. However, with aging, both microglia and astrocytes accumulate iron [113], which could be due to inflammatory processes as both cell types respond to inflammatory stimuli. This is not true for oligodendrocytes, which have a consistent iron content over time despite their higher concentrations of cellular iron to begin with [114].

Conditions Associated with Increased Iron in Brain

Iron accumulates in brain under normal and pathological conditions. Over and above age-induced increases in brain iron, iron accumulation also increases in individuals with certain pathologic conditions [115]. Subsequent to intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), or hypoxia-ischemia, iron is released from Tf or ferritin and accumulates [116]. In addition, following ICH and SAH neurons are exposed to blood components [117] including hemoglobin, which is highly neurotoxic [54, 117–119] and heme oxygenases (HO-1, HO-2) that cleave heme within hemoglobin producing free Fe²⁺ that induces free radicals and lipid peroxidation [54, 120]. Such insult-induced increases in reactive iron can persist for several days [121] and the observed neurological deficits appear to be due to iron accumulation because the iron chelator deferoxamine (DFO) protects against the brain damage [122].

Our diets also have a key role in iron accumulation, both systemically and in the brain. Heme iron in meat, poultry, and fish is highly absorbed; non-heme iron from plants is less absorbed [123]. Although heme iron is considered an essential nutrient [124], the high heme iron content in red meat has been associated with diseases including brain cancer [125, 126]. Thus, different lifestyles and diseases play key roles in the accumulation of iron in brain may contribute to dysregulation of iron homeostasis (Figure 2).

Figure 2: Mechanisms of iron accumulation in the brain and links to carcinogenesis.

Multiple conditions contribute to iron accumulation in the brain including advanced age, various human diseases, dietary choices, cerebrovascular bleeding, inflammation, pathogens, and increased blood brain barrier (BBB) permeability. Iron is normally stored, alongside other metals, in endolysosomes. Deregulated iron release into the cytosol appears capable of contributing to various pathogenesis including carcinogenesis. X-ray and γ-radiation can induce endolysosome membrane permeability (LMP; left side of figure) leading to increased mitochondrial iron and downstream increases in reactive oxygen species (ROS), mitochondrial division (fission), and the accumulation of mtDNA mutations. Drug-induced de-acidification of endolysosomes (right side) can be caused by a wide range of carcinogenic drugs and compounds including tamoxifen, nicotine, ethyl alcohol, and chloroquine. Iron released into the cytosol can also lead to an accumulation of iron in the nucleus where it produces hydroxyl radicals that cause oxidative damage to DNA and strand breaks leading to genomic DNA mutations. The endocytosed iron chelator deferoxamine (DFO) can decrease the downstream and deleterious effects of iron released from endolysosomes.

Heavy Metal and Trace Metal Elements in Carcinogenesis

Heavy metals can have biologic as well as toxic effects depending on their concentrations [127]. Small amounts of metals are necessary for physiological growth, development, and maintenance of cells and organisms [128]. Trace metals can function as micronutrients when they are present in healthy tissues and have known biological functions. Such essential microelements include aluminum, bromine, cobalt, copper, chromium, fluorine, germanium, iron, iodine, lithium, magnesium, manganese, molybdenum, selenium, strontium, zinc, nickel, tin, vanadium, and silicon. Non-essential microelements such as antimony, arsenic, barium, bismuth, cadmium, gold, lead, mercury, rubidium, silver, titanium, and zirconium, are present in animal tissues, but have either unknown or unclear biological functions and/or known neurotoxic impacts [127, 129–134].

Trace metal elements continue to be implicated in cancer progression; their concentrations differ between cancerous and noncancerous tissues, between cancer types, and between different malignancy grades [135, 136]. Numerous studies have focused on trace element levels in cancerous and noncancerous tissue from different human organs including brain [135, 137]. The International Agency for Research on Cancer (IARC) developed a four group classification system to evaluate the carcinogenicity of agents to humans based on scientific evidence from human and animal studies (International Agency for Research on Cancer; IARC) with Group 1 representing those agents labeled as “carcinogenic to humans” and defined as having “sufficient evidence of carcinogenicity”. Group 2 represents likely carcinogenic agents, divided into two subgroups (probable and possible). Groups 3 and 4 represent agents “not able to be classified as carcinogenic” and “probably not carcinogenic”, respectively. Here we will consider how iron works in conjunction with two non-essential heavy metal elements arsenic and cadmium as well as two essential elements chromium and nickel; all four of these heavy metals (arsenic, cadmium, chromium, and nickel) are classified into Group 1 by IARC.

Heavy metals localize to and accumulate in endolysosomes of brain cells (Figure 2) [138, 139]. This organellar store of heavy metals might affect the structure and function of other organelles including mitochondria because cadmium and arsenic are strong uncouplers of oxidative phosphorylation [140, 141], and the nucleus, because nickel can inhibit RNA polymerase activity [142] and cadmium severely inhibits RNA synthesis [143]. Thus, disrupted functions in the nucleus and in mitochondria might originate from heavy metals found in endolysosomes. In contrast to other cancers [135, 137] relatively little is known about trace metals in brain tumors [144], and what information is available is generally highly specific to the tumor type and tumor grade. Astrocytomas (grade I-III) have significantly higher concentrations of arsenic than healthy brain tissue [137]. In contrast, nickel concentrations do not significantly change between glioblastoma (grade IV) and healthy brains [145, 146]. Iron and chromium are significantly increased in astrocytoma (III) and glioblastoma (IV) tumors compared to normal adjacent tissue [147, 148], but the higher the malignancy grade between astrocytoma (I-III) and GBM (IV), the lower are the iron and cadmium concentrations [149, 150].

Distribution of trace metals within individual tumors has also been examined, although studies of tumor tissue distribution was frequently limited by lack of direct within patient tumor-to-normal comparisons, relying instead on comparisons of tumor samples to brain samples from other age-matched individuals [145, 149]; Nevertheless, changes in distribution of trace elements have been observed, with lead, zinc, calcium, and uranium showing homogeneous tumor distribution, while copper and iron are heterogeneously distributed [151, 152].

Mechanisms of Iron and Other Metal Carcinogenesis in Brain Malignancy

Over the last three decades, biochemical studies have contributed to our understanding of metal-induced carcinogenesis. Heavy metals including, but not limited to, arsenic, chromium, cadmium and nickel affect cellular organelles and enzymes involved in metabolism, detoxification, and damage repair [139, 153]. Heavy metal-induced increases in oxidative stress alter mitochondrial permeability and swelling, mutates mtDNA, reduces ATP synthesis, inhibits respiratory chain complexes, and increases apoptotic cell death [127, 154–160]. These metals also affect tumor suppressor genes, DNA repair processes, and enzymatic activities concerned with metabolism and oxidative damage [161]; all can contribute to carcinogenesis [140, 153, 162].

Mutations of genomic DNA capable of activating oncogenes or inactivating tumor suppressor genes are key features of tumor initiation, which is further facilitated by transcription factor activation, amplification of oncogenes, and recombination [163–166]; metals affect the four stages of carcinogenesis - initiation, promotion, progression, and metastasis. Although superoxide and hydroxyl radical formation can be caused by iron, copper, chromium and cobalt, iron appears to be of greatest importance. Free pools of copper are too low to produce free radicals via Fenton chemistry reactions. Cadmium, nickel and mercury cannot directly catalyze Fenton-type reactions and their main role in toxicity is depleting glutathione and binding to sulfhydryl groups of proteins [167–169]. In contrast, iron, chromium and cobalt can undergo redox-cycling reactions leading to oxidative stress [167], but iron must be present for arsenic to be oxidized [167, 170]. Thus, with regard to oxidative stress, iron is critical for excess ROS production and redox cycling. Indeed, the vulnerability of astrocytes to oxidative stress, loss of mitochondrial membrane potential and ATP depletion have been shown to be closely correlated to the iron content [171].

Occupational exposures to different heavy metals can lead to increases in brain cancer [172, 173]. One long term study of iron, lead, cadmium, and chromium that spanned several decades found increased risks of brain tumors to be highest for people with an occupation that exposed them to iron; the standardized incidence ratio was 2.15, which means there was a 115% higher incidence of observed cases than would otherwise be expected [174]. In the same study, cancer incidence was increased by 88% in people working with chromium, 33% in people working with lead, and 47% in people working with cadmium; no differences were found for people working with arsenic and there was a 51% increase for nickel [174]. This study linking iron and brain malignancy is especially strong because the cohort study included 413,877 people, over 183 different job titles, 25 different agents including asbestos and engine exhaust, and the robust finding that the highest brain tumor incidence was in people that had occupational exposures to iron [174].

Toxicity of Iron Accumulation in the Brain

The accumulation of iron in brain can induce neuronal damage. Even after being bound to ferritin, iron in the form of Fe²⁺ can be released from ferritin; this release is potentiated when the extracellular fluid is acidic [40] and in the presence of superoxide radicals [175] or ascorbate [176]. Iron toxicity occurs when the size of the labile iron pool exceeds the ability of ferritin to sequester iron [177] and is mediated by the increased production of ROS. When Fe³⁺ is reduced by ascorbate or other reductants, Fe²⁺ is regenerated and re-enters the oxidation-reduction cycle; a process known as redox cycling. Thus, there is this continuous production of ROS that is sustained and can continuously damage the cell; the quantity of the highly reactive hydroxyl radical is directly proportional to the concentration of Fe²⁺ [105]; the oxidative capacity of Fe²⁺ is 5 times greater than that of Fe³⁺ [178]. Thus, free radical production may overwhelm the cell’s antioxidant mechanisms, and because brain has low antioxidant defenses it is especially vulnerable to metal toxicity [42, 179, 180].

Numerous studies have emphasized the connection between dysregulation of iron metabolism including uptake, storage, and export, and human diseases including cancer [14, 16, 181, 182]. In cancer cells, however, disruption of iron metabolism occurs at the different steps of tumor initiation, progression and metastasis. Iron is very much involved in transformation of malignant glioblastoma cells. The ZIP14 (a metal-ion transporter) and the TfR iron uptake genes are both upregulated, STEAP3 is upregulated leading to an increase in Fe²⁺, and Dcytb, which is a ferric reductase enzyme that catalyzes the reduction of Fe³⁺ to Fe²⁺ required for dietary iron absorption in the duodenum of mammals is upregulated. The upregulated ZIP14 expression in GBM is associated with a significant shortening of patient survival rates [29].

Changes in iron metabolism are now considered by many to be metabolic and immunologic hallmarks of cancer, and further studies are required to determine how alterations of this element might affect tumorigenesis and progression. Clearly, cancer cells require large quantities of iron, and exhibit a strong dependence on iron, a situation that has been termed the “iron addiction” of cancer [183]. Tumor iron demands exist especially for proliferation [184, 185]. Specifically, the increased expression of the TfR has been shown to affect cancer cell proliferation, migration, invasion, and metastasis [186]. These findings suggest that TfR may be a target for cancer therapy. Furthermore, iron reductase and members of STEAP family are important for iron uptake and reduction in endolysosomes [187] and numerous studies have demonstrated altered expressions of STEAP 1–4 in different types of tumor including gliomas [188]. STEAP3 was highly expressed in malignant gliomas, especially in the mesenchymal glioma molecular subtype cell as well as isocitrate dehydrogenase 1/2 (IDH1/2) wild-type gliomas. Expression levels of STEAP3 in gliomas correlated inversely with patient survival [188].

Iron is also essential in forming lipid peroxyl radicals which can become cytotoxic aldehydes that can react with DNA and have damaging effects [189]. Moreover, several iron-dependent enzymes can indirectly produce large amount of ROS including superoxide radicals and peroxides. For example, cytochrome P450 enzymes, nitric oxide synthases, NADPH oxidases, and lipoxygenases are some of the most important enzymes involved in ROS generation [190, 191]. In a normal state, the antioxidant enzymes including glutathione peroxidase, superoxide dismutase and catalases intervene in detoxifying ROS [192], but become overwhelmed in a malignant state.

Astrocytes are more resilient to iron toxicity than neurons

Excess iron and hemoglobin are more toxic to neurons than they are to glial cells [193]. Astrocytes are the most abundant cell type in brain and help maintain an extracellular environment that protects neurons [194, 195]. Among numerous roles, astrocytes are more resistant than neural cells to ROS and iron-accumulated toxicity in brain [193, 196]. Astrocytes have high levels of antioxidants such as metallothionein, reduced glutathione, and manganese superoxide dismutase that all protect against metal-induced toxicity [197–199].

Endolysosome Iron, Mitochondrial Dysfunction, DNA Damage, and Carcinogenesis

Endolysosomes and the labile iron pool

Endolysosomes are acidic organelles containing more than 60 hydrolases, contain regulated stores of hydrogen (H⁺) ions, and have a vacuolar ATPase (v-ATPase) enzyme, which is the main proton pump responsible for maintaining the acidic lumen. v-ATPase has been identified as a novel therapeutic target for glioblastoma [200, 201]. Endolysosomes contain readily releasable stores of cations, including Fe²⁺. Although much is known about how Fe²⁺ is formed in endolysosomes and how that Fe²⁺ is delivered to the cytosolic pool for storage and use, less is known about the labile pool of iron [202, 203].

The cellular labile iron pool largely resides within the cytosol; however, it may exist in additional subcellular spaces and organelles. In rat liver endothelial cells, chelatable iron concentrations were 7.3 μM in the cytosol, 9.2 μM in mitochondria, 11.8 μM in the nucleus, and 16 μM in endolysosomes [204]. The heterogeneity among the labile iron pools and findings of the highest concentrations of labile iron in endolysosomes has implications for oxidative stress and the actions of iron chelators [204]. Further evidence of the role of labile iron in endolysosomes comes from mineral iron drugs used to treat iron deficiency, such as iron-dextran. Iron-dextran complexes have been shown to induce cancer tumor formation [205]. Because dextran is endocytosed, the administered iron-dextran accumulated in endolysosomes. Following endocytosis, the iron dissociates from the dextran, and release of iron from endolysosomes into the cytosol stimulates the formation of ROS [206, 207]. Of note, iron-dextran is classified as a Group 2B “possible carcinogenic to humans” on the IARC monograph.

Deferoxamine is specific to endolysosomes

Deferoxamine (DFO), specifically chelates the endolysosome labile iron pool because it cannot diffuse across the plasma membrane. DFO instead enters cells via fluid phase-endocytosis, and accumulates and remains in endolysosomes [203, 208–212]. DFO is a strong hydrophilic iron chelator that firmly binds all six coordinates of iron and prevents Fe²⁺/Fe³⁺ redox-cycling [203]. Furthermore, DFO is capable of crossing the BBB and has been shown to inhibit glioblastoma proliferation [213, 214]. Because DFO can chelate other metals such as aluminum and chromium [215, 216], DFO may find use against metal toxicity.

Radiation induced lysosomal membrane permeability

Radiation is used clinically for diagnostic and therapeutic reasons [217]. However, repeated exposure to radiation can cause increased lysosomal membrane permeabilization (LMP), increased release of endolysosome iron into the cytosol, and cell death (Figure 2). Drugs that induce LMP can cause similar effects [218, 219]. Radiation-induced apoptosis was largely prevented when endolysosome iron was chelated by DFO [220]. Thus, certain treatments and drugs can disrupt endolysosomes, promote excess intracellular labile iron, and increase levels of ROS and oxidative stress. Such findings led the IARC to classify X-ray radiation and γ-radiation into Group 1 “carcinogenic to humans” for the brain [221].

Drug-induced de-acidification of endolysosomes release cations including Fe²⁺

Numerous drugs including chloroquine (CQ), fluoxetine, imipramine, dimebon, tamoxifen, rotenone, chlorpromazine, amitriptyline, and verapamil can de-acidify the lumen of endolysosomes (Figure 2) [222, 223]. Lysosomotropic weak bases, which de-acidify endolysosomes, release endolysosome calcium into the cytosol [224]. This is consistent with findings that collapse of endolysosome transmembrane pH gradients with monensin causes an increase in cytosolic calcium levels [224]. In addition, free labile iron in endolysosomes is released into the cytosol upon bafilomycin A1 treatment, a known v-ATPase proton pump inhibitor that de-acidifies endolysosomes (Figure 2) [201, 206] and bafilomycin A1-induced iron release from endolysosomes was decreased by DFO [206]. Thus, regulating the pH of endolysosomes may result in new therapeutic strategies [200].

Ethanol (EtOH) is a major constituent of alcohol beverages and induces brain damage by triggering inflammatory processes in glial cells through activation of Toll-like receptor 4 (TLR4) signaling [225]. The IARC has alcoholic beverages classified as Group 1 carcinogens for certain cancers, and TLR4 participates in ethanol-induced de-acidification of endolysosomes [226]. Additionally, smoking tobacco is associated with endolysosome stress, mitochondrial stress, and swelling of neural stem cells [227]. The increase in endolysosome pH by nicotine, likely due to it being a weak base and activating two-pore channels [228], can diffuse across the cell membrane and accumulate in endolysosomes by proton trapping [229, 230]. As noted earlier, the de-acidification of endolysosomes results in labile iron release into the cytosol and increased ROS production [206].

Endolysosome iron causes mitochondrial fragmentation and dysfunction

Mitochondria are dynamically regulated by fission (fragmentation) and fusion events [231]. Molecular mechanisms that control mitochondrial fission and fusion help preserve mitochondrial integrity through mitophagy and disposal of dysfunctional mitochondria. Mitochondrial fragmentation, also known as the thread-grain transition, is induced by oxidative stress, occurs during apoptosis, and results in the presence of numerous smaller mitochondria [232]; [233]. Mitochondrial fragmentation and mitophagy function to maintain healthy mitochondria; mitophagy removes mitochondria that have mitochondrial DNA (mtDNA) mutations and this helps reduce mtDNA heterogeneity and mitochondrial heteroplasmy [234].

Iron accumulation promotes increases in ROS levels, mitochondrial fragmentation, and cellular apoptosis, a chain of events that was blocked by DFO [235]. Similar results were demonstrated in hippocampal neuronal cells; FAC-induced iron accumulation triggered mitochondrial fragmentation and apoptosis that was prevented by DFO [236]. Many drugs can de-acidify endolysosomes, induce mitochondrial fragmentation, and cause loss of mitochondrial membrane potential including CQ, artesunate, diazepam, oligomycin, antimycin A, and rotenone. These effects that were blocked by DFO [223, 237–242]. Thus, endolysosome iron release appears to be an important event upstream of increased mitochondrial ROS, fragmentation, dysfunction, and apoptosis.

Effects of mitochondrial fragmentation on disease

Mitochondrial fragmentation causes mitochondria to divide into smaller, more numerous vesicles and this phenomenon has been implicated in oxidative stress-related apoptosis and various pathological phenotypes including initiation of brain tumors (Figure 1) [242–245]. Glioblastomas have an increased frequency of mitochondrial fission [246–248] and suggest that mitochondrial fragmentation may be more pronounced in stem cells [245]. DRP1, a regulator of mitochondrial fission, helps maintain the tumorigenic potential of brain tumor initiating cells (BTIC) and reductions in DRP1 causes loss of the BTIC phenotype and tumor formation [245]. In contrast, mitochondrial division inhibitor 1 (Mdivi-1) inhibits mitochondrial division and decreases BTIC tumor formation [245]; Mdivi-1 acidified endolysosomes and genetic inhibition of DRP1 with si-DRP1 de-acidified endolysosomes [249]. Although total levels of DRP1 are similar in normal and neoplastic brain tissues, phosphorylation of DRP1 (DRP1^S616) is increased in glioblastomas and there is a strong inverse correlation between phosphorylation of DRP1 (S616) and poor glioblastoma patient survival [245].

Mitochondrial ROS promotes mtDNA mutations and carcinogenesis

mtDNA mutations in cancer occur in genes of complexes I, III, IV, and V and affect oxidative phosphorylation and the regulation of ROS [250]. Although still actively investigated, somatic and germline mtDNA mutations have been reported in numerous different cancer types including astrocytic tumors [251–254]. The importance of mtDNA mutations in neoplastic transformation [251, 255] is supported by findings that exchanging cancer cell mtDNA with other sources of mtDNA changes cancer cell phenotypes including production of ROS (Figure 2) [256–258]. Elevated levels of mitochondrial ROS are toxic, are strong mitogens, and damage mitochondria and mtDNA [250].

Endolysosome iron mediates oxidative-stress-induced DNA damage

DNA mutations known to induce malignancies can be a consequence of cellular oxidative stress and the generation of highly reactive and short-lived hydroxyl radicals in close proximity to DNA [259–263]. Similar to the labile iron pool found in the cytosol, a chelatable and labile iron pool exists in mammalian nuclei [204], and transition metals and hydrogen peroxide can induce hydroxylation and other DNA modifications [261, 262, 264–266]. The specific mechanisms by which iron enters the nucleus remain unclear. Labile iron bound to low molecular weight ligands may freely diffuse through nuclear pores and/or get actively (ATP-dependent) transported across nuclear membranes; nuclear uptake of iron can be blocked by DFO [267]. These findings suggest that endolysosomes are the source of nuclear free labile iron. DFO is able to protect against oxidative disruption induced by hydrogen peroxide and prevents strand breaks in nuclear DNA induced by hydrogen peroxide [268]. Thus, iron from endolysosomes can be translocated to the nucleus after endolysosome membrane rupture [268]; a form of lysosome membrane permeabilization (LMP) (Figure 2).

Oxidative DNA damage leads to DNA mutations and cells harboring DNA mutations that escape programmed cell death are capable of initiating tumors. Oxidative DNA damage of guanine, due to its low oxidation potential, leads to the formation of 8-oxo-7,8-dihydroguanine (8-oxo-dG) and incorporation into DNA [269]. In addition, oxidative stress and DNA damage via 8-oxo-dG directly inhibits telomerase transcriptional activity of TERT [269]. There is a direct correlation between increased stores of iron in the body with increased risks of cancer [270, 271], the mutagenicity of iron is mediated by ROS [272–274].

Human glioblastoma arises from cells in the subventricular zone (SVZ) that contain key driver mutations (p53, PTEN,, EGFR, and TERT) [275]; astrocyte-like neural stem cells (NSC) migrating from the SVZ led to the development of malignant gliomas in distant brain regions of the cortex. The loss of TERT, via genetic mutation or downstream of DNA damage via 8-oxo-dG, may be an early genetic event by which NSCs in the SVZ acquire more driver mutations that eventually develop into glioblastoma [275]. However, it remains uncertain whether the presence of 8-oxo-dG in DNA is sufficient to cause tumor formation in part because other pathological conditions associated with elevated oxidative DNA damage have no association with increased carcinogenesis.

Conclusions

Numerous risk factors including stress, tobacco, environmental pollutants, radiation, viral infection, diet, and infection are linked with chronic diseases such as cancer through the production of ROS [276], and increased levels of ferrous (Fe²⁺) iron can generate reactive oxygen species (ROS). ROS activates nuclear factor kappa-light-chain-enhancer of activated B cells [NF-κB], hypoxia-inducible factor-1α (HIF-1α), and signal transducer and activator of transcription 3 (STAT3) resulting in increased expression of proteins that increase inflammation, cellular transformation, cell survival, tumor cell proliferation and invasion, angiogenesis, and metastasis. ROS is also known to damage cellular proteins, lipids and DNA, and contribute to carcinogenesis [277, 278]. Chronic inflammation, which is one of the major mediators of tumorigenesis, is regulated by ROS and an association between ROS, DNA mutations, and malignant transformation has been clearly shown from studies of gain-of-function mutations in oncogenes and loss-of-function mutations in tumor suppressor genes [167, 279–281].

As discussed above, many factors contribute to elevated iron and other heavy metal levels in brain and endolysosomes contain a relatively large concentration of free labile iron. Once released from endolysosomes, iron can affect many different organelles through Fenton chemistry reactions and the production of ROS. Dispersion of the endolysosome labile iron can be caused by drug-induced de-acidification of endolysosomes, radiation or an overabundance of iron accumulation. Fe²⁺ released from endolysosomes can damage mitochondria and cause oxidative damage to nuclear DNA. However, despite links between dysregulated iron metabolism and carcinogenesis, relatively little has been done to directly test whether modulation of iron may be a valuable treatment strategy. Iron chelators initially used to treat acute iron overload have received some attention due to high levels of iron in cancer cells [282], and preliminary studies with these agents such as DFO have shown that they may have a role in inhibiting tumor growth [282, 283]. Although clinical trials of DFO in cancer therapy are limited, some positive results have been published. DFO was found to have a combined 88% full or partial response of tumor decreases in neuroblastoma in a clinical trial with 7% of cases progressing, However, significant side effects were reported that would need to be addressed before advancing these therapies further and argue for a more targeted approach. Another study found chelating iron with DFO enhanced fluorescence contrast to provide a more effective fluorescence-guided resection of brain tumors [284, 285]. Overall, while many key questions remain, we believe that there is sufficient evidence to suggest that regulation of iron homeostasis of endolysosomes plays a key role in carcinogenesis in the brain, and this may warrant additional consideration of therapeutic strategies which modulate iron for cancer treatment and prevention.