Mitochondrial cholesterol: Metabolism and impact on redox biology and disease

Abstract

Cholesterol is a crucial component of membrane bilayers by regulating their structural and functional properties. Cholesterol traffics to different cellular compartments including mitochondria, whose cholesterol content is low compared to other cell membranes. Despite the limited availability of cholesterol in the inner mitochondrial membrane (IMM), the metabolism of cholesterol in the IMM plays important physiological roles, acting as the precursor for the synthesis of steroid hormones and neurosteroids in steroidogenic tissues and specific neurons, respectively, or the synthesis of bile acids through an alternative pathway in the liver. Accumulation of cholesterol in mitochondria above physiological levels has a negative impact on mitochondrial function through several mechanisms, including the limitation of crucial antioxidant defenses, such as the glutathione redox cycle, increased generation of reactive oxygen species and consequent oxidative modification of cardiolipin, and defective assembly of respiratory supercomplexes. These adverse consequences of increased mitochondrial cholesterol trafficking trigger the onset of oxidative stress and cell death, and, ultimately, contribute to the development of diverse diseases, including metabolic liver diseases (i.e. fatty liver disease and liver cancer), as well as lysosomal disorders (i.e. Niemann-Pick type C disease) and neurodegenerative diseases (i.e. Alzheimer's disease). In this review, we summarize the metabolism and regulation of mitochondrial cholesterol and its potential impact on liver and neurodegenerative diseases.

Graphical abstract

Highlights

• •Cholesterol availability in MIM is the rate-limiting step in steroids and bile acids generation.
• •Cholesterol is delivered to MIM via a lipid transfer multiprotein complex of which STARD1 is crucial.
• •Mitochondrial cholesterol overload impacts mitochondrial function and redox homeostasis.
• •High mitochondrial cholesterol loading is associated with various pathological conditions.

1. Introduction

Cholesterol is an integral component of biological membranes that plays a crucial role in the integrity and regulation of signaling pathways [1]. The plasma membrane is essential for cell physiology, with cholesterol being an important determinant of membrane organization as the content of this lipid governs membrane fluidity. Furthermore, the plasma membrane of cells contains organized microdomains enriched in cholesterol termed lipid rafts that serve as platforms for the assembly of signaling molecules, allowing a closer interaction with receptors and promoting favorable contacts necessary for the transduction of signals [2,3]. In the brain, cholesterol is an indispensable component of myelin membranes that surround the axonal bodies of neurons and provide neuroprotection [4]. At the functional level, cellular cholesterol homeostasis is maintained through several mechanisms, including cholesterol esterification, leading to generation of cholesterol ester, which besides triacylglycerol, it is a crucial component of lipid droplets (LDs). Indeed, LD-associated cholesterol ester hydrolysis regulates cholesterol efflux out of cells, and in the case of macrophages this event emerges as an antiatherogenic mechanism by promoting cholesterol efflux [5,6]. Besides the generation of cholesterol ester, cholesterol acts as a precursor for the biosynthesis of metabolites like bile acids (BAs), steroid hormones, and vitamin D, which in turn have important biological roles as signal transducers and lipid solubilizers [7].

Cholesterol is obtained mostly from the diet, but can be also synthesized by almost all cells in the endoplasmic reticulum (ER). However, its distribution is not equal in all cellular lipid bilayers, being more abundant in the plasma membrane. Therefore, to ensure adequate levels of cholesterol within membranes and to maintain its homeostasis, cholesterol is delivered to other subcellular membranes, including mitochondria, through various mechanisms comprising vesicle-mediated transport as well as protein-mediated non-vesicular transport [8] (Fig. 1). Fluctuations in lipid content can largely affect cellular functions and essential physiological roles in different organelles, and emerging evidence recognizes alterations in the content and metabolism of mitochondrial cholesterol as a key factor in several disorders, including steatohepatitis, carcinogenesis and neurodegeneration. The present review summarizes current knowledge on the biology of mitochondrial cholesterol, its trafficking and functions, focusing in its role in cell metabolism, redox biology and pathophysiology.

Fig. 1.

Intracellular cholesterol distribution and functions. Cells can obtain cholesterol from the diet via endocytosis of lipoproteins, or it can be synthesized de novo by the mevalonate pathway in the endoplasmic reticulum (ER). Depending on the cellular demands, the newly internalized/synthesized cholesterol can be stored as lipid droplets (made up of esterified cholesterol and triacylglycerols) acting as energy reservoirs (1), or it can be metabolized to meet cellular demands. Once synthesized in the ER, cholesterol is delivered to other subcellular membranes through vesicular or protein-mediated transfer by specific carriers. One of the main functions of cholesterol is to be a structural component of biological membranes (2). Cholesterol content controls membrane rigidity and this feature is crucial for cellular function as it determines transmembrane import of molecules or cell death processes. This lipid is highly enriched in membrane domains termed “lipid-rafts”, organized membrane microdomains that provide a platform for the assembly of signaling molecules, allowing a closer interaction with receptors. In addition, it is an important component of the myelin sheaths covering the axonal bodies of neurons. The surrounding of nerve cells by myelin sheaths is essential for the correct transmission of electrical impulses. Besides, cholesterol exerts important metabolic functions as it is the precursor for the synthesis of steroid hormones and neurosteroids and bile acids (3). To do so, cholesterol has to be transported to mitochondria, where it is metabolized by i) CYP11A1 yielding pregnenolone which is a central precursor of steroids, and ii) CYP27A1, the enzyme that converts the mitochondrial cholesterol in 27-hydroxycholesterol, which is a precursor of the acidic bile acid synthesis pathway. In the brain, cholesterol is metabolized by CYP46A1 in the ER to 24-hydroxycholesterol, which crosses the blood-brain barrier (BBB) and is transported to the liver for its further catabolism. HDL: high-density lipoprotein.

2. Mitochondria as source of ROS

Oxidative stress refers to the dysregulation between the production of free radical oxygen species (ROS) and the antioxidant defense strategy inside the cell, wherein excess of ROS cannot be overcome by antioxidant defense mechanisms. ROS are a family of free radicals that includes a number of molecular species resulting from oxygen metabolism, such as superoxide anion (O₂^•−), hydrogen peroxide (H₂O₂) or hydroxyl radical (•OH). Although H₂O₂ is not strictly a free radical, it is a potent oxidant which can serve under certain conditions as a precursor of •OH (e.g. Fenton reaction). These toxic radicals or oxidants interact with cellular components such as DNA, proteins, lipids, and other molecules, which are modified and may cause cell death. Among the cellular sources of ROS, the mitochondrial electron transport chain (ETC) is one of the predominant sources [9]. During the transfer of electrons to molecular oxygen (O₂) for the generation of ATP in the ETC, O₂^•− and H₂O₂ are formed as a collateral consequence of the oxidative phosphorylation (OXPHOS), particularly O₂^•−, while H₂O₂ is generated either as a product of superoxide dismutase (SOD2) or by spontaneous dismutation [9]. There are two major respiratory chain regions where ROS are produced, one being complex I (CI, NADH coenzyme Q reductase) and the other complex III (CIII, ubiquinol cytochrome c reductase) [10,11]. Electrons are transferred along the ETC until they reach CIII and CIV. Once they arrive at CIV, electrons are trapped by O₂ to form water. Yet, electrons have a tendency to leak spontaneously at CI and CIII, resulting in the formation of O₂^•− [12,13].

Likewise, ROS are also generated by the action of several mitochondrial flavoenzymes. Especially relevant in the brain are monoamine oxidases, enzymes located in the outer mitochondrial membrane (OMM) that catalyze the oxidation of monoamine neurotransmitters employing one oxygen molecule to cleave their amine group accompanied by the release of H₂O₂ [14]. Another contributor to increase the mitochondrial pool of ROS is the flavin-containing enzyme dihydrolipoyl dehydrogenase within the ketoglutarate dehydrogenase complex (KGDHC), a rate-limiting enzyme of the tricarboxylic acid cycle [15,16]. Under conditions of elevated NAD⁺/NADH ratio in the mitochondrial matrix, KGDHC catalyzes the decarboxylation of α-KG (also known as 2-oxoglutarate) to succinyl-CoA using NAD⁺ as electron acceptor, resulting in the generation of O₂^•−. Therefore, at the same time that its activity is governed by redox status through the availability of NAD⁺, KGDHC exerts pro-oxidant effects as it stimulates ROS generation. Moreover, α-KG is the precursor of the excitotoxic neurotransmitter glutamate, meaning that changes in KGDHC may alter correct neural activity. Indeed, deficiency of KGDHC is associated with a number of neurological disorders, including Alzheimer's disease (AD) and its activity is diminished by oxidative stress [17].

Besides mitochondria, ROS are also originated in other subcellular sites such as i) the ER during protein folding [18] or by microsomal monooxygenases [19], ii) peroxisomes as byproducts in the β-oxidation of fatty acids [20], iii) the plasma membrane where ROS are generated by NADPH oxidases, lipoxygenases and cyclooxygenases [21] or iv) at the extracellular space via xanthine oxidase [22]. Of note, ROS can easily cross membranes through specific channels, so while the contribution of each organelle might be different, the transport of ROS from one cellular compartment to another may lead to the accumulation of these signaling molecules in the proximity of any of these organelles [23]. Besides the classic notion of oxidative stress as the imbalance between ROS and antioxidant defenses, it has been shown that a strict equilibrium between antioxidant defense is also crucial to avoid the accumulation of oxidants and free radicals, as illustrated in the case of liver disease characterized by impaired glutathione (GSH) redox cycle in the mitochondria relative to the action of SOD2 [24].

3. Mitochondrial cholesterol transport and regulation

Cholesterol is synthesized de novo by the mevalonate pathway in the ER, where acetyl-CoA is transformed into cholesterol through different steps that require oxygen consumption [25]. The transcription factor SREBP-2 regulates cholesterol synthesis by inducing key enzymes from the mevalonate pathway, such as hydroxymethylglutaryl CoA (HMGCoA) reductase (HMGCR), the target of statins that inhibit cholesterol synthesis. Once it is synthesized, cholesterol prevents its overproduction by suppressing the processing of SREBP-2 and promoting HMGCR degradation. Briefly, to assure cholesterol homeostasis, GP78, a membrane-anchored ubiquitin ligase, associates with Insig-1, which binds to the sterol-sensing domain of HMGCR, and couples sterol-regulated ubiquitination to degradation of HMGCR. However, cholesterol is not only derived from de novo synthesis. Cholesterol can be taken up from the diet as a component of lipoproteins that are distributed through different organs, mainly the liver, and enter cells by endocytosis and subsequent delivery to endolysosomes (Fig. 1). Further, excess cholesterol from peripheral tissues can be transported to the liver by a process known as “reverse transport” for its excretion [7]. In this process, cholesterol is released from cells associated to apolipoproteins, in particular apolipoprotein A-I, which upon further lipidation is converted to mature high-density lipoproteins particles. Besides vesicle-mediated interorganellar transport, there is another non-vesicular protein-mediated transfer mechanism, in which other proteins work as transporters of cholesterol determined by the presence of lipid-binding pocket domains.

Mitochondrial cholesterol represents about 2–4% of the total cholesterol pool in the cells in comparison to the plasma membrane, which accounts for 25–30%, and is essential for the synthesis of oxysterols, steroids and hepatic BAs [26]. The transport of cholesterol from the OMM to the inner mitochondrial membrane (IMM) occurs via non-vesicular transport by the steroidogenic acute regulatory protein 1 (STARD1, also known as StAR). The START family is a large family of proteins that transport cholesterol, oxysterols, phospholipids, sphingolipids and fatty acids, and includes the first discovered member STARD1, the founding member of the family, and other members such as STARD3 (also called MLN64) and STARD4 subfamily (STARD4, STARD5 and STARD6) [26]. Once cholesterol is delivered to the IMM, it is metabolized by different enzymes depending on the tissue. For instance, in neurons and steroidogenic tissues, mitochondrial cholesterol is converted by the cytochrome P450 side-chain cleavage enzyme (P450scc; CYP11A1) into pregnenolone, the precursor of steroid hormones. The loss of function of STARD1 causes the development of lipoid congenital adrenal hyperplasia (lipoid CAH) [27], indicating its relevance in cholesterol transport to the mitochondria for steroidogenesis. Mice with global STAR deletion fail to grow normally and die by lethal CAH within 7–10 days after birth. These findings unravelled a critical role of this protein in steroidogenesis and hence in the trafficking of cholesterol to the IMM for processing as other members of the STAR family cannot compensate for the STARD1 deficiency [28].

Human STARD1 is synthesized as a cytosolic 37 kDa protein with a sterol-binding domain that contains a signal sequence in its N-terminal that directs the protein to the mitochondria, where it is processed to its 30 kDa mature form. In addition to STARD1, there are some other critical proteins that work together in the delivery of cholesterol to the mitochondria. This protein complex, known as the transducome, is composed of OMM proteins including 18 kDa translocator protein (TSPO), the voltage-dependent anion channel (VDAC), acyl-coenzyme A binding domain containing 3 (ACBD3) and the protein kinase A regulatory subunit 1α (PKA-RIα), and IMM protein ATPase family AAA domain-containing protein 3A (ATAD3A) [29]. The formation of this multiprotein complex prevents unwanted crosstalk of cholesterol with other pathways, including interference with general mitochondrial function and metabolism, and optimizes substrate availability. Upon hormonal stimulation, the transducome is assembled and promotes the pooling of cholesterol at the OMM. VDAC anchors STARD1 and TSPO together and assisted by PKA-RIα initiates the transport of cholesterol into the IMM. Moreover, ATAD3A forms a bridge connecting OMM and IMM, allowing the segregation of cholesterol into the mitochondrial matrix [29]. However, recent data question the role of TSPO in mitochondrial cholesterol regulation as some models with TSPO deletion exhibited no obvious change in steroidogenesis, suggesting a regulatory role of the protein or possible functional redundancy [30,31]. Furthermore, a translocator role for TSPO has been proposed, involving the transfer of cholesterol to a protein partner rather than a role in intermembrane trafficking [32].

Finally, another existing lipid transfer protein family is the OSBP/ORPs (oxysterol binding protein and OSBP-related proteins) family of lipid transfer proteins, which transport cholesterol, oxysterols and phospholipids among different organelles, including ER, Golgi, lysosomes, mitochondria and plasma membrane, through membrane contact sites [33]. Of note, this lipid transfer protein family has been involved in pathophysiological processes such as malignant tumor, trafficking of the amyloid precursor protein and obesity [[34], [35], [36]].

Thus, while current evidence indicates that cholesterol import to mitochondria requires STARD1, VDAC, TSPO and OSBP/ORPs might also play a role in controlling cholesterol import to the mitochondria, although their specific contribution is not completely understood [37]. Further research is required to uncover the molecular mechanisms of the delivery of cholesterol to the IMM for metabolism.

4. Impact of cholesterol accumulation on mitochondrial redox biology

The mitochondrial pool of cholesterol is highly regulated and its availability in the IMM acts as the rate-limiting step in BAs and steroids synthesis. Giving the crucial role in determining the structural and physical properties of bilayers, the accumulation of cholesterol in mitochondria exceeding its biotransformation into pregnenolone or 27-hydroxycholesterol has an important impact on the fluidity of mitochondrial membranes decreasing the transition of liquid-ordered to liquid-disordered phases of membranes, with a wide-range impact on mitochondrial function and redox biology. Unlike the OMM, the IMM is fairly impermeable to solutes due to the lack of porins, and thus requires specific carriers for transport into the matrix. Therefore, although cholesterol enrichment in mitochondria occurs in both the OMM and IMM, the impact on the latter is of relevance for mitochondrial solute transport and homeostasis. For instance, genetic or diet-induced mitochondrial cholesterol enrichment decreases fluidity of the IMM [[38], [39], [40]], which in turn, affects the function of specific membrane carriers of the mitochondrial solute carrier family (SLC25), which are critical for the maintenance of mitochondrial performance as they are involved in the transport of fatty acids, carboxylic acids, amino acids, cofactors, inorganic ions and nucleotides across the IMM [41]. Among the isoforms of SLC25, SLC25A10 (dicarboxylate carrier, DIC) and SLC25A11 (2-oxoglutarate carrier, OGC) members have been involved in the transport of GSH from the cytoplasm into mitochondria in kidney, liver, brain, heart and colonic epithelial cells [[42], [43], [44], [45]]. While the SLC25A10 transports dicarboxylates (e.g. malonate) across the IMM in exchange for phosphate, sulfate and thiosulfate, the SLC25A11 exchanges matrix 2-oxoglutarate for cytosolic GSH. It was shown in isolated kidney mitochondria that about 80% of the GSH is transported to the IMM by both carriers [46]. In addition, functional expression studies in X. laevis indicated that oocytes microinjected with the SLC25A11/FLAG-tagged cRNA construct showed increased GSH transport in isolated mitochondria that were sensitive to competition with phenylsuccinate, a relative specific inhibitor of the 2-oxoglutarate carrier [47]. However, in contrast to the evidence reported from rabbit kidney [46], the functional expression in oocytes microinjected with the SLC25A10 cRNA from HepG2 cells or rat liver did not result in significant GSH transport activity [47]. Thus, this approach provides further evidence that SLC25A11 acts as a mitochondrial GSH transporter in the liver. Recent studies reported that inhibition of SLC25A10 and SLC25A11 aggravated ferroptosis in cardiomyocytes with increased mitochondrial ROS, membrane depolarization and GSH depletion [44]. In adipocytes, mitochondrial SLC25A10 deletion causes unrestrained lipolysis and promotes liver lipotoxicity, and its overexpression prevents liver lipotoxicity, thus pointing to the DIC-succinate axis as a potential therapeutic target in non-alcoholic fatty liver disease (NAFLD) [48]. Quite intriguingly, recent findings showed that the expression of SLC25A11 in the cytoplasmic membrane of Lactococcus lactis (L. lactis) and subsequent fusion of L. lactis membrane vesicles with liposomes failed to show significant GSH transport activity [49], which contrast with the functional expression of mitochondria-targeted SLC25A11 from HepG2 cells in X. oocytes and the reconstitution of partially purified SLC25A11 in proteoliposomes from kidney mitoplasts [46,47]. Whether the findings of Booty et al. [49] reflect specific biochemical and biophysical properties of cytoplasmic membrane vesicles from L. Lactis in which the SLC25A11 was reconstituted or if the putative function of SLC25A11 as a mGSH transporter in physiological settings requires other partners remains to be further investigated. In addition, a novel mitochondrial carrier SLC25A39 has emerged as a putative GSH transporter [50]. However, whether SLC25A39 is sensitive to the disruption of membrane fluidity imposed by the accumulation of cholesterol or not remains to be established, and this feature constitutes an essential mark for the mitochondrial GSH transport activity. As GSH is a cofactor for the GSH redox cycle, the depletion of mGSH by increased cholesterol accumulation determines the tone of oxidative stress and the detoxification of H₂O₂ as well as other lipid peroxides, which can compromise mitochondrial performance and cell fate [51,52]. In line with this, mice deficient in caveolin-1 protein, a membrane protein associated with endocytosis, extracellular matrix organization and cholesterol distribution, showed enhanced sensitivity towards steatohepatitis and neurodegeneration due to increased influx and accumulation of cholesterol in mitochondrial membranes, resulting in disrupted membrane permeability, OXPHOS capacity and GSH mitochondrial import, which are reversed upon GSH ethyl ester (GSH-EE) addition, a permeable precursor of GSH [38].

One of the predicted consequences of increased oxidants and ROS generation reflecting lower GSH defense by cholesterol loading in mitochondria is the impact on the lipid composition targeting sensitive lipid species. This is of most relevance to cardiolipin, an anionic phospholipid found exclusively in the IMM, which is essential in the performance of mitochondrial energy generation in the respiratory chain and in the maintenance of IMM structure [52,53]. Cardiolipins are characterized by the presence of four unsaturated fatty acyl chains, which determine their sensitivity to ROS attack. In healthy cells, cardiolipin is present in the IMM and close to the protein complexes of the OXPHOS, and supports their activity [[54], [55], [56], [57]] and their organization in supercomplexes [58,59]. Oxidative alterations of cardiolopin such as peroxidation affect CI, CIII and CIV activities [[60], [61], [62]] and are associated with mitochondrial dysfunction in several pathologies, including diabetes, NAFLD, aging, Parkinson's disease and heart failure [53,63,64]. Cardiolipin also plays a crucial role in apoptosis by regulating the release of cytochrome c and binding to Bcl-2 family protein Bid to induce Bax and Bak oligomerization [65]. Of relevance to metabolic liver diseases, such as NAFLD and alcoholic-related liver disease (ALD), hepatocytes with mGSH depletion due to mitochondrial cholesterol loading become sensitive to TNF-induced apoptosis involving a mechanism whereby acid sphyngomyelinase mediated ROS generation induces cardiolipin peroxidation and promotes Bax oligomerization and OMM permeabilization (OMMP) [66]. Data with large unilamellar vesicles indicated that the reconstitution of the bilayer with peroxidized cardiolipin potentiated oligomerized Bax-mediated OMM-like liposome permeabilization by restructuring the lipid bilayer, pointing that the proapoptotic role of mitochondrial cholesterol loading is determined by mGSH depletion and subsequent cardiolipin peroxidation. However, interestingly, in established hepatocellular carcinoma (HCC) it has been observed that mitochondria from HCC cells exhibited increased mGSH levels despite cholesterol accumulation and decreased membrane fluidity, due to a compensatory increased expression of SLC25A11 mediated by hypoxia inducible factor-1 stabilization [45]. Thus, in the progression of non-alcoholic steatohepatitis (NASH) towards HCC development, mitochondrial cholesterol accumulation switches from a proapoptotic role in the early phase of NASH, while acting as an antiapoptotic factor in HCC, in which the maintenance of mGSH levels preserves cardiolipin oxidant status, and the accumulation of mitochondrial cholesterol-mediated loss of membrane fluidity impairs OMMP and apoptosis. Overall, the final action of mitochondrial cholesterol as a proapoptotic or antiapoptotic factor appears to be dependent on the level of mGSH and status of cardiolipin peroxidation, suggesting that the modulation of the rheostat mGSH/cardiolipin peroxidation status may be a potential target in the progression of metabolic liver diseases towards HCC development.

Besides the impact of mitochondrial cholesterol on cell fate, recent findings have disclosed a critical role of cholesterol in the determination of the performance of OXPHOS [40]. Nutritional-induced increased liver cholesterol in mice by feeding a cholesterol-enriched diet stimulated the trafficking of cholesterol to mitochondria due to induction of STARD1 and decreased CI and CII-driven state 3 respiration and mitochondrial membrane potential. Decreased respiratory and uncoupling control ratio from CI was also observed after in situ enrichment of mouse liver mitochondria with cholesterol, indicating a direct effect of cholesterol in the function of OXPHOS. Moreover, in vivo cholesterol loading did not affect the levels of respiratory complexes CI, CIV and CV but decreased the level of CIII₂ and the assembly of respiratory supercomplexes I₁+III₂+IV and I₁+III₂. The disrupting effect of cholesterol in OXPHOS was reversed upon mGSH replenishment with GSH-EE, as treatment of mice with GSH-EE restored the levels of respiratory CIII₂ [40]. Thus, these findings underscore the critical role of cholesterol in the regulation of mitochondrial respiration through changes in mitochondrial membrane physical properties, consistent with previous findings in which reconstitution of CI and CIII in proteoliposomes with different lipid composition revealed that the ratio of phospholipid to protein determines the assembly of supercomplex I₁-III₂ by the impact on membrane viscosity [67]. Thus, cholesterol trafficking to the IMM has a dual facet, with a bioactive profile determined by its role as precursor for steroids or BAs, or as an structural regulator of IMM physical and functional properties. The switch between these seemingly differential roles could be determined by the extent of cholesterol trafficking to the IMM for its biotransformation by the steroidogenic CYP11A1 or hepatic CYP27A1, as cholesterol availability in IMM is the rate-limiting step and governed by STARD1.

5. Mitochondrial cholesterol accumulation in disease

5.1. Liver diseases

5.1.1. Non-alcoholic fatty liver disease (NAFLD)

NAFLD is a spectrum of liver alterations that begins with fat accumulation in the liver (in more than 5% of hepatocytes) in the absence of alcohol consumption, viral infection or drugs that can promote steatosis (abnormal retention of fat). NAFLD is a multifactorial disease linked with numerous risk factors including genetic, epigenetic and environmental factors, and embraces a wide array of hepatic modifications ranging from simple steatosis (non-alcoholic fatty liver, NAFL), which can further evolve to NASH, liver fibrosis, and eventually, cirrhosis and HCC [68]. NAFLD is identified as the most predominant chronic liver disease globally with an occurrence of 25% among the normal population. Obesity markedly amplifies the chances of developing NAFLD and particularly its advanced stages of NASH and HCC, due to the onset of insulin resistance (IR), diabetes mellitus type 2, or metabolic syndrome, which impacts the dysregulation in the lipid metabolism characteristic of NAFLD. Indeed, the prevalence of NAFLD in obese and type 2 diabetic patients is greater than 50% [69]. Originally, a “two-hit” hypothesis described the pathogenesis of NAFLD. First, IR promotes liver steatosis by the accumulation of free fatty acids (FFAs). This excess of fat sensitizes the liver to “second-hits” such as oxidative stress, largely due to mitochondrial dysfunction or the action of inflammatory cytokines underlying the characteristic inflammatory state, which coexists with lipotoxicity and fibrosis that synergize to promote NASH [70,71]. This initial concept was replaced by the “multiple hit” theory [72], which contemplates many factors acting in parallel and synergistically. Currently, a novel notion has been suggested to define the heterogeneous population of individuals with this disease termed metabolic-dysfunction-associated fatty liver disease (MAFLD), as the preceding nomenclature NAFLD overrates alcohol consumption and underestimates metabolic risk factors [73]. This new classification covers patients with hepatic steatosis, together with overweight/obesity, type 2 diabetes, or signs of metabolic dysregulation [74].

As mentioned above, steatosis is the first event in the development of NAFLD (or MAFLD) [75]. Several lipid species are stored in hepatocytes in the evolution of MAFLD to NASH, including triacylglycerols, diacylglycerol, FFAs, ceramides and cholesterol. Therefore, unbalanced lipid metabolism is the most direct promotor of MAFLD. In this regard, augmented hepatic cholesterol level, involving its specific trafficking and accumulation into the mitochondria, has emerged as a critical contributor to NASH progression [76,77]. Accordingly, nutritional and genetic models of hepatic steatosis driving mitochondrial cholesterol accumulation exhibited compromised membrane fluidity and mGSH levels and increased sensitivity to inflammatory cytokines [78,79]. STARD1 overexpression implies an elevated mitochondrial cholesterol content [80], correlating with the mGSH depletion observed in human NASH patients [81]. Hence, these data demonstrate that the preservation of the appropriate cholesterol homeostasis is vital to ensure an efficient supply of mitochondrial antioxidant defenses, in particular mGSH. In line with this, it has been shown that the level of mGSH is a critical determinant of the therapeutic relevance of O₂.- scavengers to avoid the generation of H₂O₂ and subsequent •OH, which can contribute to the perpetuation of mitochondrial dysfunction in NASH [24].

Moreover, a transgenic mouse model of intestine-specific overactivation of SREBP-2 fed a high fat-high cholesterol diet developed hepatic steatosis and displayed elevated sensitivity to liver damage, inflammatory response and fibrosis [82]. These observations provide a new scenario that highlights the importance of cholesterol homeostasis not only in the liver but also in other organs, like the intestine, in the development of liver injury, and underscores an inter-organ communication in which cholesterol accumulation emerges as a crucial player.

In addition, crosstalk between mitochondria and inflammasome has been described in NASH and ALD [25,83]. Inflammasomes are cytosolic multiprotein oligomers of the innate immune system responsible for the activation of inflammatory responses that are activated by diverse PAMPs (pathogen-associated molecular patterns) and DAMPs (danger-associated molecular patterns), such as lipopolysaccharide (LPS) and cholesterol crystals [84]. Ethanol intake and FFA/cholesterol-enriched diets can modify intestinal permeability and PAMPs and DAMPs such as FFA or microbial products might penetrate through the disturbed tight junctions and promote NLRP3 inflammasome activation in liver cells. Besides damaged hepatocytes, further findings revealed that inflammasome-mediated cholesterol crystallization-mediated activation of NLRP3 in Kupffer cells (KCs) is also an important contributor during the progression of hepatic inflammation in NASH [85]. Of note, the administration of benzyl isothiocyanate, diminished LPS- and cholesterol crystal-induced NLRP3 inflammasome activation in KCs and prevented NASH development [86]. Therefore, cholesterol metabolism is a critical determinant of the progression of NASH to liver fibrosis as it regulates the inflammatory response in the liver through the activation of the NLRP3 inflammasome. Recent evidence has revealed that cholesterol-rich diets activate the processing of gasdermin, underlying the onset of pyroptosis as an important initial player in NASH [87]. Besides KCs, the accumulation of cholesterol in hepatic stellate cells (HSCs) has been identified as an important player in liver fibrosis and NASH development [88,89]. Since STARD1 is induced not only in hepatocytes but also in HSCs during NAFLD [90,91], it is likely that the accumulation of free cholesterol in mitochondria of HSCs may be an important mediator of HSC activation and liver fibrogenesis, which deserves further investigation in the future.

5.1.2. Hepatocellular carcinoma (HCC)

HCC is the most common type of liver cancer and the second driving cause of cancer-associated death worldwide due to late diagnosis and poor therapeutic outcomes [92,93]. HCC is the end stage of chronic liver disease driven by different etiologies, including viral hepatitis (HBV and HCV), ALD or NASH [94,95]. A rise in the incidence of NASH-related HCC is estimated to occur due to its correlation with obesity and type 2 diabetes and the effective treatment of HBV and HCV [96]. Along these lines, increased hepatic cholesterol has not only been identified in the onset of NASH [76,77] but has also been established as a tumor promoter in the NASH-to-HCC progression [[97], [98], [99]]. In this regard, using a model of relevance to NASH which exhibit chronic ER stress (MUP-uPA mice), a previously unrecognized role of STARD1, and consequently mitochondrial cholesterol, in NASH-driven HCC pathogenesis has been defined, wherein overexpression of STARD1 promotes the synthesis of primary BAs through the acidic mitochondrial pathway [100]. BAs are critical molecules involved in lipid digestion and signaling pathways. The predominant synthesis of BAs in the liver occurs through the classic pathway, which is regulated by the metabolism of cholesterol through 7-α-hydroxylase (CYP7A1), the rate-limiting enzyme in BAs synthesis expressed in hepatocytes, which forms two primary BAs, cholic acid (CA) and chenodeoxycholic acid (CDCA) in human liver [101]. Complementing this pathway, CDCA can be also synthesized in the mitochondria via an alternative pathway in which cholesterol is delivered to the IMM for metabolism by 27-hydroxylase (CYP27A1) [101]. The product of this reaction generates 27-hydroxycholesterol, which then feeds the synthesis of CDCA following the action of 25-hydroxycholesterol-7-α-hydroxylase (CYP7B1) (Fig. 2). The rate-limiting step in this alternative pathway of BAs synthesis is the delivery of cholesterol to the IMM rather than the activity of CYP27A1 [102]. An additional difference between the classic and the alternative pathways is that the expression of CYP7A1 is feedback regulated by BAs via enterohepatic circulation, while the expression of CYP27A1 and CYP7B1 are refractory to this inhibition. In this regard, feeding mice a high-cholesterol diet supplemented with sodium cholate results in increased liver cholesterol content and decreased Cyp7a1 and Cyp8b1 mRNA levels, but not Cyp27a1 or Cyp7b1 mRNA suggesting continued availability of cholesterol to feed the alternative pathway of BAs [40]. Thus, the increased expression of STARD1 emerges as a crucial player in determining the levels of BAs in NASH-driven HCC, and the mitochondrial-derived BAs induce the expression of transcription factors involved in self-renewal, stemness and inflammation, characteristic of HCC [100].

Fig. 2.

Bile acids biosynthesis. Two major bile acid biosynthetic pathways are depicted. In the classic pathway, cholesterol is converted to 7α-hydroxycholesterol by the rate-limiting enzyme CYP7A1. 7α-hydroxycholesterol is immediately metabolized to 7α-hydroxy-4-cholesten-3-one (C4), which is converted to 7α,12α-dihydroxy-4-cholesten-3-one by sterol 12α-hydroxylase (CYP8B1) leading to the synthesis of CA. Without the action of CYP8B1, C4 is converted to CDCA. In the alternative pathway, cholesterol is initially transformed to 27-hydroxycholesterol by CYP27A1. Oxysterol 7α-hydroxylase (CYP7B1) catalyzes the hydroxylation of 27-hydroxycholesterol to 3β,7α-dihydroxy-5-cholestenoic acid, which is ultimately converted to CDCA. In the intestine, bacterial dehydroxylases convert CA to DCA and CDCA to LCA and other secondary bile acids, including MCA, THCA, TMDCA, THDCA, and TUDCA. CA: cholic acid; CDCA: chenodeoxycholic acid; DCA: deoxycholic acid; LCA: lithocholic acid; MCA: muricholic acid; THCA: taurohyocholic acid; TMDCA: tauromurideoxycholic acid; THDCA: taurohyodeoxycholic acid; and TUDCA: tauroursodeoxycholic acid.

In addition to the chronic ER stress model, STARD1 has also been reported to mediate NASH-driven hepatic carcinogenesis and HCC development in which mice were pretreated with DEN, a toxic nitrosamine that produces liver carcinogenesis, and then fed high fat-high cholesterol diet, as the signs of NASH and the incidence of tumor multiplicity and size were markedly reduced by STARD1 deletion in hepatocytes [100]. The presence of cholesterol in the diet potentiates NASH-driven HCC compared to mice fed a high-fat diet, indicating that cholesterol per se rather than steatosis promotes the progression of NASH toward HCC [76,100]. Of note, atorvastatin improved high fat-high cholesterol diet-induced HCC development [103], and it has been shown that atorvastatin increased doxorubicin-mediated HCC growth arrest and cell death in an in vivo model of HCC [104], giving further support to the employment of strategies aimed at attenuating liver cholesterol content as coadjuvants during chemotherapy for the treatment of NASH-driven HCC. In addition, mitochondrial cholesterol loading shields mitochondrial membrane, impairing Bak/Bax oligomerization in OMM and subsequent OMMP, and therefore, it represents an additional mechanism of cell death resistance of HCC cells and hence of HCC tumor growth [104]. In line with these findings, in vitro studies using HCC cells have shown that mitochondrial cholesterol accumulation contributes to sorafenib resistance of HCC [105], further demonstrating the crucial role of the mitochondrial pool of cholesterol in HCC development.

However, despite the evidence for a tumor promoter role of cholesterol in HCC, there are findings indicating that cholesterol suppresses HCC development. For instance, by promoting CD44 localization in lipid rafts, cholesterol impedes HCC invasion and metastasis [106]. Moreover, cholesterol loading inhibited HCC progression in vivo and in vitro, which was associated with a decreased SCAP (SREBF Chaperone) expression in cholesterol-loaded liver cancer cells [107]. Unfortunately, these studies did not address the specific contribution of mitochondrial cholesterol, thus raising the possibility that cholesterol may exert a differential role in HCC depending on whether it accumulates in the mitochondrial membrane or extramitochondrial compartments.

5.1.3. Alcoholic-related liver disease (ALD)

ALD embraces a variety of liver modifications due to the consumption of high amounts of ethanol, which begins with simple steatosis that can progress to alcoholic hepatitis and cirrhosis, and ultimately to HCC. The oxidative metabolism of ethanol activates diverse mechanisms that participate in ALD progression. However, despite significant efforts, the pathogenesis of ALD is still incompletely understood, which has limited the development of efficient therapy [108]. Alcohol metabolism in the liver arises with the alcohol dehydrogenase enzyme, which catabolizes ethanol to acetaldehyde. Acetaldehyde is in turn oxidized to acetate by the acetaldehyde dehydrogenase (ALDH). Alternatively, ethanol can be also metabolized via the microsomal system cytochrome P450, CYP2E1, which also transforms alcohol into acetaldehyde. Acetaldehyde and its derivative malonaldehyde produce modified noxious macromolecules (adducts), which can be engulfed by KCs, endothelial and HSCs, and trigger the activation of an inflammatory response that participates in ALD progression [109].

The mechanisms underlying the impact of alcohol metabolism on ALD are multifactorial and involve alterations in methionine metabolism, ER stress, disruption of lipid metabolism, genetic and environmental factors, and mitochondrial dysfunction with changes in mitochondrial dynamics, respiration, changes in membrane composition, mitochondrial DNA oxidation and ROS production [110]. Recent findings indicated that ALDH2, which plays a critical role in the detoxification of acetaldehyde produced during ethanol metabolism, regulates cholesterol synthesis [111]. The inactivation of ALDH2, responsible for alcohol flux in nearly 8% of the world population and about 40% of Asians, has been shown to increase the expression of HMGCR. Interestingly, the ALDLH2-mediated regulation of cholesterol synthesis is linked to the mitochondria-associated ER membranes (MAM), a specific domain of membranes in which ER and mitochondria contact, which plays a critical role in the transit of calcium and lipids, including cholesterol, between ER and mitochondria. Thus, these findings suggest that the impairment in ALDH2, results not only in the perpetuation of acetaldehyde levels and the adverse consequences derived from its extreme reactivity (e.g. formation of adducts with proteins among others), but also in the increased availability of cholesterol to traffic to mitochondria via MAMs. In line with this possibility, it is known that chronic alcohol feeding induces mitochondrial cholesterol accumulation and subsequent mGSH depletion due to the alterations of mitochondrial membrane fluidity caused by cholesterol loading in mitochondrial membranes [[112], [113], [114]]. Strategies aimed to reverse the decrease in membrane fluidity (e.g. S-adenosyl-l-methionine) replenished the mitochondrial transport of GSH and restored mGSH pool following alcohol intake [115,116]. The depletion of mGSH by alcohol is of relevance in determining susceptibility to inflammatory cytokines, as hepatocytes from alcohol-fed models exhibited an increased susceptibility to TNF-mediated cell death due to the priming of mitochondria to TNF-mediated OMMP by Bax through cardiolipin peroxidation [66,117]. The mechanism of alcohol-induced cholesterol accumulation in mitochondria involved acid sphingomyelinase-mediated ER stress, which led to the induction of STARD1 leading to increased trafficking of cholesterol to mitochondria [118,119].

5.1.4. Primary biliary cholangitis (PBC)

Primary biliary cholangitis (PBC) is a chronic and progressive form of cholestatic liver disease (CLD), caused by the retention of products that are normally excreted into bile, such as cholesterol and BAs, due to the dysfunction of bile ducts [120]. Although PBC is immune-mediated, as specific anti-mitochondrial antibodies are found in most patients [121], and rare, other causes of CLD due to genetic defects, drug toxicity or hepatobiliary malignancy constitute a major cause of liver transplantation. PBC is characterized by an increase of liver and serum BAs, which are believed to contribute to hepatocellular and intrahepatic bile duct injury, inflammation and fibrosis that can progress eventually to biliary cirrhosis. Besides BAs, cholesterol metabolism seems to play a key role in PBC as markers of lipid peroxidation and cholesterol self-oxidation products have been identified in serum samples from PBC patients [122,123]. Changes in serum cholesterol content are related to altered BAs production through mechanisms that involve modifications of hepatic CYP7A1 expression/activity [124,125]. As mentioned above, BAs are synthesized in hepatocytes from cholesterol in the classic pathway, and hence most BAs accumulating in PBC and other forms of CLD are believed to derive from this pathway regulated by the expression of CYP7A1. Although the alternative mitochondrial pathway of BAs generation is considered to contribute to a minor extent to the hepatocellular BAs pool, it is conceivable that induction of STARD1 may boost this event and emerges as a significant alternative pathway contributing to BA-mediated hepatocellular injury and disease progression. Whether or not STARD1 expression is induced in models and patients with CLD remains to be investigated in the future. In this regard, we have observed that mice with bile duct ligation, a well-characterized model of CLD, exhibited higher expression of STARD1 compared to sham-operated counterparts as well as increased mitochondrial cholesterol loading (Conde de la Rosa et al., unpublished observations). Future studies are underway to determine the contribution of the alternative pathway of BAs synthesis in mice following BDL.

5.2. Neurodegenerative diseases

The human brain makes up only 2.1% of body weight but it contains around 23% of the total body cholesterol (10 times higher than other tissues) reflecting the major role of this lipid in this organ. Neurosteroids are synthesized from cholesterol, meaning that the brain requires adequate levels of this lipid to guarantee normal functionality [126]. In recent years, a complex bidirectional interplay between the disruption of brain cholesterol metabolism and congnitive function has been revealed. For instance, murine models of diabetes and aging display a reduction in brain cholesterol biosynthesis associated with cognitive phenotypes [127,128]. Further, genetic inactivation of pregnenolone production by STARD1 deletion and subsequent metabolism of mitochondrial cholesterol by CYP11A1 in hypothalamic POMC neurons in mice, deteriorated recognition memory independently of metabolic disturbances [129]. Thus, despite its low physiological content, the mitochondrial import of cholesterol in the brain is a highly regulated process, as alterations in the pool of cholesterol in this organelle significantly contributes to the development of neurodegenerative diseases.

5.2.1. Alzheimer's disease (AD)

AD is a type of dementia characterized by the extracellular accumulation of amyloid β (Aβ) peptides and intraneuronal strings of hyperphosphorylated tau proteins known as neurofibrillary tangles, which develops gradually and affects memory and thinking skills. The accumulation of Aβ peptides is generally due to mutations in the amyloid β precursor protein (APP) or other AD-related genes. The hippocampus is the brain area responsible for memory, and hence, this area is particularly vulnerable in AD and shows extensive neuronal loss [130].

AD is a multifactorial disease and several players contribute to its progression, including aging and disruption of cholesterol homeostasis. Indeed, existing evidence indicate that hypercholesterolemia is a major risk factor for AD development [131]. However, in spite of the association between hypercholesterolemia and AD, cholesterol's function in AD is still not well-known and remains debatable. From one side, a body of literature supports a link between increased cholesterol levels in the brain and the progression of AD. The first evidence that cholesterol may have an impact on the progression of AD was proposed by Sparks et al., in 1994, who observed that experimental models fed with cholesterol-enriched diets showed an increased intracellular accumulation of Aβ peptides in the brain compared with the control group [132]. This was later confirmed by other experimental approaches in which diet-induced hypercholesterolemia accelerated AD pathology [[133], [134], [135]]. Further, the presence of the enzymes involved in the generation of toxic Aβ peptides in lipid rafts provides strong evidence of the association between high cholesterol levels with Aβ aggregate deposition and AD development [136,137]. Moreover, the E4 isoform of the cholesterol carrying protein ApoE has long been related with an increased risk of developing AD. Consistently, statins-based therapy is associated with a lower incidence of AD [138]. Given that the blood-brain barrier (BBB) is not permeable to plasma lipoproteins, meaning that all the cholesterol inside the central nervous system is obtained from de novo synthesis, observations relating dietary cholesterol intake to AD progression are puzzling. However, increased fluidity of the BBB has been described in AD, which could account for this association [139]. Despite the strong association between elevated cholesterol levels and increased risk of AD, there is also data in the literature that indicates the opposite [140]. This inverse relationship is of particular relevance during aging, as low levels of cholesterol in the hippocampus region of AD patients as well as cognitively-deficient in old rodents were found [140,141]. However, it is important to remark that the highest peak of cholesterol synthesis in the brain takes place during active myelination at prenatal stages, which is carried out by oligodendrocytes. After myelination has been completed, cholesterol synthesis decreases by ∼90% and continues at a low rate in adults, supported mainly by astrocytes rather than neurons [142]. Once synthetized, cholesterol cannot be cleared from the brain. Instead, cholesterol 24S-hydroxylase (CYP46A1), which is produced by neurons, converts cholesterol into 24S-hydroxycholesterol, which can cross the BBB and is transported to the liver where it is degraded [143]. As a strategic regulator of cholesterol metabolism in the brain, CYP46A1 may play a role in protecting neurons as reduced expression of this enzyme is involved in hippocampal nerve injury induction or progression of AD pathology, followed by an increased accumulation of cholesterol in the brain [144].

In addition to its extracellular deposition, there is data demonstrating Aβ accumulation in other intracellular spaces, including mitochondria or the ER. Beel et al. proposed that the APP C-terminal fragment acts as a cholesterol-sensing protein in the ER membrane [145]. Indeed, pathogenic accumulation of the 99-aa C-terminal domain of APP in MAMs induces the uptake of extracellular cholesterol as well as its trafficking from the plasma membrane to the ER in AD cell and animal models, causing alterations in lipid homeostasis and composition [146]. Although it is unclear whether APP processing and Aβ generation occur in mitochondria, it has been observed that Aβ aggregates target mitochondria to induce ROS formation [147]. Moreover, expression of aberrant APP protein was identified in mitochondria from AD patient's brains and correlated with disease severity [148]. Thus, based on this background, we hypothesized that mitochondrial cholesterol may emerge as a novel player in AD. Indeed, experiments in AD mouse models have shown that high mitochondrial cholesterol sensitizes neurons to Aβ-mediated neurotoxicity and inflammation through the depletion of mGSH levels [149]. Besides, we provided evidence for a determinant role of mitochondrial cholesterol in AD progression as the overexpression of SREBP-2 gene in APP/PS1 mouse accelerated AD phenotype [150]. The molecular events contributing to the aggravation of the disease were dependent on mGSH depletion, as in vivo treatment with the cell-permeable GSH-EE attenuated synaptic degeneration, improved cognitive function, and restored mGSH levels and mitochondrial cholesterol loading.

Additionally, administration of the ER stress inhibitor 4-phenylbutyric acid prevented mitochondrial cholesterol loading and mGSH depletion, thereby protecting APP/PS1 mice against Aβ-induced neurotoxicity, and was effective in normalizing STARD1 expression in SH-SY5Y cells after Aβ-exposure [151], suggesting that STARD1 was a target of ER stress signaling, in line with data in primary mouse hepatocytes treated with tunicamycin [119]. Thus, it is plausible that Aβ-induced ER stress may be a key mechanism in AD pathology that promotes cholesterol accumulation in the brain due to increased expression of STARD1 [152]. Remarkably, elevated STARD1 levels have been reported in the cytoplasm of hippocampal pyramidal neurons from brain samples of AD patients, suggesting a mechanistic link between STARD1 and mitochondrial cholesterol loading in human AD [153]. STARD1 is expressed in specific neuronal and astrocyte populations in the cortex, hippocampus, hypothalamus, cerebellum and cranial nerve motoneurons. However, it is worth mentioning that its expression in the nervous system is much lower in comparison to endocrine tissues [154], making the analysis of STARD1 expression in the brain more difficult. In this line, we have recently identified high expression of STARD1 and NPC1 (NPC Intracellular Cholesterol Transporter 1) proteins in hippocampus and cortex postmortem samples from patients with AD [155]. Of note, STARD1 expression discriminated control subjects from AD patients with better accuracy than the routinely employed pathological marker Aβ₄₂. Hence, STARD1 could be a potential pre-clinical marker associated with early stages of AD pathology. Future studies are required to address the causal relationship between STARD1 upregulation and mitochondrial cholesterol accumulation in AD, and will entail the generation of cell-type-specific STARD1 deletion models in the brain to examine the individual contribution to the pathology.

5.2.2. Niemann-Pick type C disease (NPC)

Niemann-Pick type C disease (NPC) is an autosomal recessive, neurodegenerative disorder characterized by the impaired intracellular transport of cholesterol and other lipids. Typical symptoms include splenomegaly, ataxia, seizures, cataplexy and dementia. While most cases of the disease (around 90%) are linked to mutations in the NPC1 gene [156], a small part corresponds to alterations in the NPC2 gene [157]. Both proteins function closely in the export of lysosomal cholesterol [158], meaning that the defects in NPC1/NPC2 result in the intracellular accumulation of free cholesterol and glycosphingolipids [159]. The accumulation of lipids then initiates a pathological cascade that includes deficient oxysterol production, peroxisomal dysfunction, aberrant sphingosine metabolism, neuroinflammation, activation of apoptosis, and impaired synthesis of neurosteroids [160]. Even if neurodegeneration is the major feature in NPC, some affected individuals display early severe hepatic failure and some patients die from liver dysfunction even before the onset of the neurodegenerative effects [161]. The pathogenic mechanism explaining how the accumulation of intracellular cholesterol leads to cell death remains still elusive.

Although increased cholesterol accumulates intracellularly in affected organs, particularly in the endolysosomal compartment as a direct consequence of NPC1/NPC2 defects, cholesterol has been shown to accumulate in mitochondria of NPC cells, which has a negative impact on mitochondrial function by causing impairment of mitochondrial antioxidant defense and alterations in energy metabolism, such as increased lactate production, often observed when mitochondrial function is compromised as is the case of NPC disease [78,162,163]. Indeed, aberrant ATPase activity in Npc1^−/− mice brain mitochondria has been causally linked to cholesterol accumulation in mitochondrial membranes as treatment with methyl-β-cyclodextrine (βCD) (an agent that extracts cholesterol from membrane bilayers) restored enzyme activity [163]. Very recently, redox imbalance and mitochondrial dysfunction associated with cholesterol accumulation have been reported in NPC patients’ fibroblasts, which were prevented by the combination of βCD with antioxidants [164]. Consistent with these findings, mitochondria from Npc1^−/− mice exhibited decreased mGSH levels that were replenished by GSH-EE, while N-acetyl-l-cysteine failed to increase mGSH due to the impaired transport imposed by the accumulation of cholesterol [42]. Importantly, restoration of mGSH by GSH-EE protected against oxidative stress and cell death, improved locomotor impairment and neurodegeneration of Npc1^−/− mice and extended the average and maximal life span of the mice [42].

While the accumulation of cholesterol in mitochondria emerges as an additional characteristic of NPC disease in addition to its accumulation in endolysosomes due to defective NPC1 function [165,166], the mechanism responsible for the increased trafficking to mitochondria has not been fully elucidated yet [78,149,163]. Interestingly, we have observed increased expression of STARD1 in the liver and brain from Npc1 null mice by an ER stress-independent pathway [39]. Whether the increased expression of STARD1 in the brain of NPC models is cell specific is not fully understood, although consistent with this possibility, Chen et al. showed decreased levels of STARD1 protein and mRNA in astrocytes from Npc1 deficient mice [167]. Although our findings dissociated the increase in STARD1 expression with the onset of ER stress, we uncovered a previously unrecognized inverse relationship between acid ceramidase (ACDase) and STARD1 in NPC disease. ACDase has been shown to repress STARD1 expression through binding to the nuclear transcription steroidogenic factor SF-1 [168]. Quite interestingly, the expression of ACDase decreased in the liver and brain of Npc1^−/− mice and in fibroblasts from patients with NPC disease. Further, ACDase overexpression in fibroblasts from NPC patients prevents STARD1 upregulation, mitochondrial cholesterol accumulation, and improves mitochondrial activity, suggesting that enhanced levels of STARD1 in NPC disease may occur through an ER stress-independent mechanism involving STARD1 derepression through ACDase downregulation.

In addition to the recognized involvement of STARD1 in NPC, another putative candidate that can contribute to mitochondrial cholesterol trafficking in NPC disease is STARD3, also known as MLN64. Balboa and colleagues demonstrated that overexpression of STARD3 in normal and NPC1 knockout hepatocytes, promotes mitochondrial cholesterol overload, resulting in mGSH exhaustion and mitochondrial impairment [169]. Furthermore, along with STARD3, NPC2 has been also shown to participate in the delivery of endosomal cholesterol to mitochondria [170]. Recent findings have demonstrated that aberrant cholesterol transport was restored after reestablishing the activity of sphingosine kinase 1 enzyme, which is reduced in NPC, highlighting a novel link between sphingolipid pathway and lysosomal cholesterol accumulation [171]. Thus, in light of these findings, it is feasible that the increase of mitochondrial cholesterol observed in NPC1 deficient cells could derive from the combination of STARD1, STARD3 and NPC2, as the expression of these proteins is altered in NPC disease [169,172]. Whether these proteins work in close association or synergism with each other in the transport of cholesterol to mitochondria remains to be established, and further understanding their specific role in promoting mitochondrial impairment in NPC disease may be of relevance for the identification of novel therapeutic opportunities.

6. Conclusions

Due to the critical role of cholesterol in cell physiology and function, alterations in cholesterol homeostasis and metabolism have been linked to a myriad of pathological conditions. Here, rather than analyzing the role of total cholesterol levels, we focus on understanding how the intracellular cholesterol pools, particularly in mitochondria, influence metabolism and redox biology and trigger the onset of prevalent liver diseases, such as NAFLD, NASH or HCC, as well as neurodegenerative diseases like AD and NPC disease. Although the pathogenesis of these disorders is different, increased accumulation of cholesterol in mitochondria, which exceeds its metabolism, emerges as a common denominator in all of them [42,100,150,173]. Upregulation of STARD1 expression has been recognized as the mechanism underlying cholesterol trafficking and overload in mitochondria. Although ER stress is known to induce the upregulation of STARD1 in AD and NASH, there is no evidence of activation of ER stress in NPC [39], meaning that the induction of STARD1 might be different in each particular pathology. In agreement with this, we found that upregulation of STARD1 following ER stress mediates acetaminophen-induced liver injury via SH3BP5 (or SAB) mitochondrial protein and phosphorylation of c-Jun N-terminal kinases (JNK) 1 and 2, JNK1 and JNK2 [173].

One important conclusion we extracted from our studies is that the accumulation of cholesterol in mitochondrial membranes exerts similar effects despite the affected organ or disease's etiology. As a summary, Fig. 3 illustrates the impact of mitochondrial cholesterol accumulation on redox biology of brain and liver tissues, and the molecular drivers contributing to the development of neurodegenerative and liver disorders. Cholesterol overload impairs the transport of GSH from the cytosol to the mitochondrial matrix, rendering brain mitochondria more susceptible to Aβ-induced ROS generation and neurotoxicity, while in the liver, in addition to the oxidative damage coming from excess ROS formation, it promotes BAs production in the mitochondrial acidic pathway and enhances hepatic tumorigenesis. Thus, based on these findings, mitochondrial cholesterol may emerge as a new concept of study and therapies targeted to boosting mitochondrial antioxidant armamentarium may be a promising approach for the treatment of diseases that share mitochondrial cholesterol imbalance as a common hallmark.

Fig. 3.

Impact of mitochondrial cholesterol overload on brain and liver pathology. From one side, increased levels of cholesterol in mitochondrial membrane impairs GSH transport from cytosol to mitochondria, as the GSH transporter 2-OGC is dependent on membrane fluidity. This leads to mitochondrial GSH depletion and consequently, elevated ROS production. High level of ROS impairs mitochondrial function and induces mitochondrial permeability transition (MPT) pore formation, followed by cellular death. In addition, the oxidation of cardiolipin by ROS facilitates the release of cytochrome c, further promoting MPT. However, in cancer cells, it has been observed that this “shielding” of the mitochondrial membrane by cholesterol provides resistance to chemotherapy and favors cell survival. From the metabolic point of view, increased levels of cholesterol in mitochondria stimulate the generation of bile acids through the alternative pathway and increase the synthesis of steroid hormones. In Alzheimer's disease (AD) the localization of amyloidogenic components at the MAMs (mitochondrial-associated-membranes) induces the uptake of extracellular cholesterol as well as its trafficking from the plasma membrane to the ER, inducing a dysregulation of cholesterol homeostasis that leads to a reorganization of ER-membrane domains. In NPC disease, the fusion of endosomes and lysosomes for debris clearance is impaired, hence, free cholesterol accumulates in mitochondria, resulting in mGSH depletion. Upregulation of STARD1 protein due to the activation of ER stress signaling pathway has been defined as the underlying mechanism for the increase in mitochondrial cholesterol trafficking in NASH, ASH, NASH-induced HCC and AD. However, no ER stress induction has been observed in NPC disease. ΔΨ: membrane potential; Aβ: amyloid beta protein; AD: Alzheimer's disease; ASH: alcoholic steatohepatitis; ATP: adenosine triphosphate; Chol: cholesterol; ER: endoplasmic reticulum; GSH: glutathione; H₂O₂: hydrogen peroxide; HCC: hepatocellular carcinoma; IMM: inner mitochondrial membrane; MAM: mitochondrial-associated-membrane; mChol: mitochondrial cholesterol; mGSH: mitochondrial glutathione; MPT: mitochondrial permeability transition; NASH: non-alcoholic steatohepatitis; NPC: Niemann-Pick type C; NO: nitrogen oxide; OMM: outer mitochondrial membrane; OXPHOS: oxidative phosphorylation system; STARD1: steroidogenic acute regulatory protein 1; TSPO: translocator protein; VDAC: voltage-dependent anion channel; 2-OGC: 2- oxoglutarate carrier.

Author contributions

L.G.B., L.C.d.l.R., S.T.N., C.G.-R. and J.C. F.-C. revised the literature and discussed reported findings and conceived the focus of the review. L.G.B. and S.T.N. conceived figures. L.G.B., L.C.d.l.R., S.T.N., C.G.-R. and J.C.F.-C. wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

We acknowledge the support from grants PID2019-111669RB-100 and PID2020-115055RB-I00 from Plan Nacional de I+D funded by the Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) and from the CIBEREHD; the center grant P50AA011999 Southern California Research Center for ALPD and Cirrhosis funded by NIAAA/NIH; as well as support from AGAUR of the Generalitat de Catalunya SGR-2017- 1112, European Cooperation in Science & Technology (COST) ACTION CA17112, Prospective European Drug- Induced Liver Injury Network, the 2018-102799-T “Enfermedades Metabólicas y Cancer” from the Red Nacional of the Spanish Health Ministry and the Project 201916/31 Contribution of mitochondrial oxysterol and bile acid metabolism to liver carcinogenesis 2019 by Fundació Marató TV3. In addition, this project has received funding from the European Horizon's research and innovation program HORIZON-HLTH-2022-STAYHLTH-02 under agreement No 101095679.

Declaration of competing interest

The authors declare no conflict of interest.

Abbreviations

Aβ

amyloid β peptide

ACDase

acid ceramidase

AD

Alzheimer's disease

ALD

alcoholic liver disease

APP

amyloid-beta precursor protein

ATP

adenosine triphosphate

ASH

alcoholic steatohepatitis

BA

bile acid

BBB

blood-brain barrier

βCD

methyl-β-cyclodextrine

CA

cholic acid

CDCA

chenodeoxycholic acid

CYP27A1

sterol 27-hydroxylase

CYP7A1

7α-hydroxylase

CYP7B1

25-hydroxycholesterol 7-α-hydroxylase

CYP8B1

microsomal sterol 12α-hydroxylase

DAMPs

danger-associated molecular patterns

DIC

dicarboxylate carrier

DNA

deoxyribonucleic acid

ER

endoplasmic reticulum

ETC

electron transport chain

FFA

free fatty acid

GSH

glutathione

GSH-EE

glutathione ethyl ester

H₂O₂

hydrogen peroxide

HCC

hepatocellular carcinoma

HMGCoA

hydroxymethylglutaryl CoA

HMGCoAR

hydroxymethylglutaryl CoA reductase

HSC

hepatic stellate cells

IMM

inner mitochondrial membrane

IR

insulin resistance

JNK

c-Jun N-terminal kinases

KCs

Kupffer cells

LD

lipid droplet

MAFLD

metabolic-dysfunction-associated fatty liver disease

mGSH

mitochondrial glutathione

NAD⁺

nicotinamide adenine dinucleotide (oxidized form)

NADH

nicotinamide adenine dinucleotide (reduced form)

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

NPC

Niemann-Pick Type C

NPC1

Niemann-Pick Type C intracellular cholesterol transporter 1

O₂•−

superoxide anion

O₂

molecular oxygen

OGC

2-oxoglutarate carrier

•OH

hydroxyl radical

OMM

outer mitochondrial membrane

OMMP

OMM permeability

OXPHOS

oxidative phosphorylation system

PAMPs

pathogen-associated molecular patterns

PBC

primary biliary cholangitis

PKA-RIα

protein kinase A regulatory subunit 1α

ROS

reactive oxygen species

SLC25

mitochondrial solute carrier family

SOD2

superoxide dismutase 2

SREBP-2

sterol regulatory element-binding protein

STARD1

steroidogenic acute regulatory protein

START

steroidogenic acute regulatory protein-related lipid transfer

TGF-β

transforming growth factor β1

TNFα

tumor necrosis factor alpha

TSPO

18 kDa translocator protein

VDAC-1

voltage-dependent anion channel

Data availability

This MS is a review