The predominant mechanism to provide the need of cellular cholesterol is its de novo synthesis from acetate in the form of acetyl-CoA in an energy-consuming metabolic route called known as the mevalonate pathway. In addition to cholesterol, several other non-sterol components are generated through the mevalonate pathway. This pathway is initiated in the ER by the conversion of acetyl-CoA into 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) and then converted to mevalonate by the rate-limiting enzyme, HMG-CoA reductase (HMGCR). Afterwards, a series of enzymatic reactions convert mevalonate into farnesyl pyrophosphate (FPP) which can be diverted either to the sterol synthesis or to the non-sterol pathway. In the sterol branch, squalene synthase commits the FPP to the sterol biosynthesis [48] by converting the FPP to squalene and afterwards to lanosterol and finally cholesterol by two related pathways. The Block pathway use desmosterol and 7-hydroxycholesterol as the immediate precursors of cholesterol through their reduction by the 24-dehydrocholesterol reductase (DHCR24) and the Kandutsch-Russel pathway use the the DHCR7. UV radiation can convert the 7-dehydrocholesterol into vitamin D₃, which has an important effect on cancer biology [49]. The biotransformation of lanosterol to cholesterol involves several redox reactions with oxygen requirement. Consequently, an additional mechanism that regulates cholesterol synthesis is oxygen availability whereas hypoxia has been shown to stimulate HMGCR degradation through both accumulation of lanosterol and Insigs induction [50].

In the non-sterol pathway, FPP can generate by additional enzymatic steps important cellular components such as dolichol, ubiquinone, heme A and isoprenoids for protein prenylation. Ubiquinone and heme A are essential for the mitochondrial transport chain and dolichol is a transmembrane carrier for glycol units in the synthesis of glycated proteins and lipids [51]. Isoprenoid generation in the mevalonate pathway is an essential mechanism of posttranslational modification of certain proteins by anchoring target proteins to cell membranes. FPP is used to add a prenyl moiety to proteins of the Ras family, while geranylgeranyl pyrophosphate (GGPP) is used to prenylate those of the Rho family [52]. Prenylation of G-proteins is required for intracellular signalling between receptors and effector enzymes [53]. Prenylation of Ras protein is an important step in translocation and activation of Ras, so that it can regulate cell growth, while inhibition of Ras prenylation leads to cell growth blockade [54].

Given the importance of cholesterol for cell homeostasis and the high energy used for its biosynthesis, the de novo cholesterol synthesis is a highly-regulated process controlled by an efficient feed-back mechanism. Accumulation of sterols in the ER membrane triggers binding of the membrane domain of HMGCR to a Insig1 and Insig2, which carry a membrane-anchored ubiquitin ligase called GP78 which ubiquitinates HMGCR, marking it for proteasomal degradation [55]. Moreover, HMGCR gene expression is regulated by ER-bound membrane transcription factor sterol regulatory element protein 2 (SREBP2), whose activation depends on the sterol-sensitive SREBP cleavage activating protein (SCAP) which is the true sterol level sensor of the ER. When cholesterol levels are low in the ER, SCAP binds to SREBP2 inducing SREBP2 to leave the ER and move to the Golgi, where it is sequentially cleaved by the specific proteases S1P and S2P. This event releases the soluble N terminus domain of SREBP2, which travels to the nucleus to activate the transcription factor activity of SREBP2, triggering expression of HMGCR and several enzymes in the mevalonate pathway and cholesterol metabolism [41].

The de novo synthesized cholesterol leaves the ER rapidly is targeted to the plasma membrane, mostly by non-vesicular mechanisms that bypass ER-Golgi membrane transport thereby helping to maintain low ER sterol content [56]. Also, cholesterol equilibrates with the preexisting cholesterol pool and accesses other sites such as endosomal compartments [57]. An important buffering mechanism to maintain steady levels of sterols is its esterification. Cholesterol can be esterified with long chain fatty acyl-CoA to cholesterol esters in a reaction catalyzed acyl-CoA cholesterol acyltransferase (ACAT). Esterified cholesterol increases the hydrophobicity of cholesterol and is usually a neutral form of storage of cholesterol within lipid droplets along with triglycerides.

ACAT has been reported to be overexpressed in different types of cancers, including glioma, pancreatic and liver cancer [58, 59]; although its role in liver cancer is not well understood with reports indicating a lower expression of ACAT in tumor areas from patients with HCC [60]. These controversial results may indicate different requirements of the different types of tumors for non-esterified cholesterol for their survival.

In addition to the de novo synthesis, there is an important cholesterol input from diet to the liver, which plays a key role in its redistribution. Enterocytes and hepatocytes package cholesterol and triglycerides into lipoproteins, circumventing the physical problem of transport of hydrophobic molecules in a hydrophilic media in the circulation. Dietary cholesterol is absorbed very efficiently by gut enterocytes which package it along with triglycerides and apolipoprotein B-48 into chylomicrons. Once in circulation, an important amount of chylomicron triglycerides is hydrolyzed within few minutes while new apoproteins, such as ApoE are added, generating chylomicron remnants that are taken up by hepatocytes in the liver through the low-density lipoprotein (LDL) receptor. These particles are endocytosed by clathrin-coated vesicles and transported to the lysosomal compartment, where cholesteryl esters are hydrolyzed by acid lipase rendering unesterified cholesterol. The multivesicular late endosomes harbor two transporters the Niemann-Pick type C protein 1 and 2 (NPC1 and NPC2) which play a crucial role in cholesterol trafficking out of the endosomal system. Indeed, deficiency of either of these proteins leads to the accumulation of LDL-derived unesterified cholesterol in late endocytic organelles, which is characteristic of the lysosomal genetic disorder NPC disease [61]. Another site in the endocytic pathway for active sterol exchange is the recycling compartment. Recycling endosomes are enriched in cholesterol and sphingolipids and function as acceptors for non-vesicular sterol flux through the cytosol [62]. Once released from the endolysosomal system, cholesterol is delivered mainly to plasma membrane, lipid droplets for its esterification or may be transferred to other membranous organelles such as ER, recycling endosomes, and mitochondria.

Hepatocytes, in turn, redistribute cholesterol throughout the body by the secretion of in very low-density lipoprotein (VLDL) particles which carry lipids packaged with apolipoprotein B-100. VLDL particles are processed in the circulation by addition of several apolipoproteins and the hydrolysis of their triglycerides into LDL, the main lipoprotein that delivers cholesterol to peripheral cells through its uptake by the LDL receptor. When peripheral cells have excess cholesterol, it can be released to high density lipoproteins (HDL), through a series of transmembrane receptors of the ATP-binding cassette (ABC) transporter family such as ABCA1 by a process known as cholesterol efflux. Indeed, nascent HDL particles carrying ApoA-I and ApoA-II are released by the liver, they uptake unesterified cholesterol exerting the cholesterol efflux from peripheral tissues and this cholesterol is esterified by the HDL-associated enzyme lecithin-cholesterol acyltransferase (LCAT). These HDL particles exchange lipids and apolipoproteins with other lipoproteins in the blood and hepatocytes take up the cholesterol from mature HDL particles, returning the cholesterol to the liver in a process called reverse cholesterol transport [41]. Moreover, LDL particles and VLDL remnants can be cleared by hepatocytes returning the cholesterol and cholesterol esters to the liver for its redistribution or transformation in bile acids.

The pool of nonesterified cholesterol, both the newly synthesized in the liver and the one derived from the intestinal absorption or from circulating lipoprotein uptake, can be excreted from the hepatocyte directly to the bile by the function of the heterodimer of ABC transporters ABCG5/G8 found on the canalicular membrane of the hepatocytes. However, one essential pathway of cholesterol metabolism in hepatocytes is its conversion to bile acids. This complex process in which either CYP7A1 or CYP27A1 mediate the first limiting reaction generates an array of bile acid derivatives with diverse physical properties and physiological actions [63]. The microsomal CYP7A1 catalyzes the first step of the main classical bile acid synthesis pathway also known as the neutral pathway, performing the rate-limiting 7α-hydroxylation of cholesterol. Afterwards, the microsomal sterol 12α-hydroxylation by CYP8B1 is specifically required for cholic acid synthesis and its derivatives, after a series of enzymatic transformations. Conversely, the mitochondrial CYP27A1 starts the alternative and quantitatively minor bile acid synthesis (acidic pathway). In hepatocytes, mitochondrial cholesterol is metabolized by CYP27A1 to 27-hydroxycholesterol followed by CYP7B1, which then feeds the alternative mitochondrial pathway of bile acid synthesis leading mainly to chenodeoxycholic acid generation [64–66]. In mouse liver, but not in humans, chenodeoxycholic is metabolized to muricholic acids (α-MCA and β-MCA) [67]. This specie-specific difference has important implications given that chenodeoxycholic acid acts as a farnesoid X receptor (FXR) agonist, while muricholic acids display FXR antagonistic activity. Nevertheless, there are evidences showing that the transport of cholesterol to mitochondria, mediated by the cholesterol transporter steroidogenic acute regulatory protein (STAR) rather than CYP27A1 is the real rate-limiting step of the mitochondrial pathway of bile acid synthesis. Indeed, in primary hepatocytes or HepG2 cells, the overexpression of STAR resulted in a 5-fold increase in the rate of bile acid synthesis, while transfection with CYP27A1 upregulated bile acid synthesis only by 2-fold [68, 69]. In addition to the bile acid synthesis, STAR-mediated trafficking of cholesterol to mitochondria has important implications in its functionality and as a consequence, in MASLD pathogenesis and HCC development [40, 70], as will be discussed later. In humans, the mitochondrial acidic bile acid synthesis pathway of bile acid synthesis is considered to contribute approximately to 10% of the total bile acid pool and its physiological relevance is not well understood.

The primary cholic acid and chenodeoxycholic acids bile acids are quickly conjugated with the amino acids, taurine or glycine, by the action of amino acid N-acyltransferase (BAAT), changing their physical properties and lowering their cell toxicity. These conjugated bile acids are secreted into bile via bile salt export pump (BSEP) to the bile that along with phospholipids and cholesterol where they have critical roles in lipid digestion. In the ileum, conjugated bile acids are secreted into portal blood via organic solute transporter α/β and circulated back to hepatocytes via sodium-taurocholate co-transport peptide (NTCP). Furthermore, conjugated bile acids are deconjugated and 7α-dehydroxylated in the colon by specific bacterial flora. As a result, part of cholic acid and chenodeoxycholic acid are converted to deoxycholic acid and lithocholic acid, respectively. These secondary bile acids, which show higher toxicity, are absorbed and circulated to the liver like the conjugated and unconjugated primary bile acids. Approximately 95% bile acids are recovered and circulated back to the liver (enterohepatic cycle) and the 5% of bile acids that are lost in feces which can be compensated by de novo synthesis in the liver from cholesterol [71].

In the neutral pathway, two mechanisms have been proposed for the feed-back regulation of CYP7A1 gene transcription by primary bile acids. First, primary bile acids activate directly FXR in the liver to induce small heterodimer partner (SHP), which represses trans-activation of the CYP7A1 and CYP8B1 genes. Moreover, primary bile acids entering the enterocytes through the ASBT transporter, activate FXR in the intestine to induce fibroblast growth factor 15 (FGF15, or human orthologue FGF19), which reaches the liver activating hepatic FGF receptor 4/Klotho signaling to inhibit CYP7A1 gene transcription via extracellular regulated protein kinases 1/2 (ERK1/2)/c-Jun of the mitogen-activated protein kinase (MAPK) pathway [72]. The inhibition of ASBT reduces FGF19/15 secretion from the ileum, promoting bile acid synthesis and decreasing levels of hepatic cholesterol. However, the alternative mitochondrial pathway is not controlled by these mechanisms responding directly to bile acids or cholesterol. However, given that the limiting step of this pathway is the transport of cholesterol to mitochondria mediated by STAR, the regulation of expression and activity of this transporter is pivotal for the availability of cholesterol for the CYP27A1. However, little is known about regulation of STAR in the liver. STAR has been reported to be regulated by LXR activation [65] and induced by ER stress [73], and both features play important roles in MASLD and likely to the initiation and progression of HCC.

The accumulation of lipid species in hepatocytes is initiated in MASLD by de novo lipogenesis, whose rate is up to 3-fold higher in MASLD patients [81]. Hence, lipogenesis was speculated to contribute to MASH progression by the generation of lipotoxic free fatty acids (FFA) and the induction of liver damage; fibrosis and inflammation in the transition from bland MASLD to MASH are believed to be a consequence of lipotoxicity. However, some studies attribute the main lipotoxicity in the MASH progression to nonesterified cholesterol accumulation [82–84]. Importantly, given that de novo lipogenesis and cholesterol synthesis share regulatory pathways, increased hepatic triglyceride levels are associated with increased cholesterol synthesis in the liver [85, 86]. Therefore, the lipotoxicity attributed to triglyceride accumulation in steatosis in MASLD may be, at least in part, due to increased non-sterified cholesterol in different cell compartments. In line with this, mitochondrial cholesterol loading cause mitochondrial dysfunction in hepatocytes [15] rendering them more susceptible to a second insult exerted by TNF-alpha leading to liver damage, which is a characteristic feature of MASH. Of note, TNF-alpha is also a critical mediator of MASH and HCC progression [13].

An important part of the lipotoxicity found in MASH is manifested with the persistent hepatocellular ER stress. Compared to normal liver tissues, MASLD tissues have been detected to have higher levels of C/EBP homologous protein (CHOP), cleaved ATF6 and spliced X-box binding protein 1 (XBP1) indicating the activation of the unfolded protein response to cope with ER stress in MASLD [87]. Moreover, the presence of chronic ER stress in livers from obese individuals is not simply by protein overloading but also driven by compromised folding capacity, in which lipid metabolism may have a function given the known ability of the incorporation of palmitate and cholesterol to the ER membrane to induce ER stress [88]. The implication of ER stress in MASH and the development of HCC were demonstrated in experimental animals where by means of chemical chaperones or liver overexpression of the protein chaperone BiP/Grp78 completely abrogated the MASH signs [13]. In line with this, a SREBP2 activation mechanism was described whereby ER stress drives continued cholesterol synthesis through the activation of Casp2-mediated activation of Sp1 [75]. This mechanism overrides the inhibition of de novo cholesterol synthesis where ER is sensitive to cholesterol levels, creating a cholesterol overload as found in states of chronic liver disease preceding HCC development such as MASH [82, 85]. Moreover, cholesterol accumulation triggers ER stress by altering the critical cholesterol-to-phospholipid ratio of the ER membrane, necessary to maintain its fluidity. The resulting stiffening of the ER membrane inhibits enzyme conformational changes and impairs ER activity and eventually triggering cell apoptosis. Even minimal increases of cholesterol loading on ER are able to impair the activity of SERCA, a sarco/ER calcium ATPase Ca²⁺ pump, decreasing the intra-ER Ca²⁺ concentrations, resulting in an impairment of protein-folding capacity and ultimately triggering ER stress [88]. This axis would create a feed forward mechanism of exacerbated cholesterol synthesis and ER stress elicited by ER membrane cholesterol overload [88], both situations found in MASH and HCC.

Lastly, ER stress is also associated with long-term inflammation, as a result of an overproduction of ROS and the activation of nuclear factor kappa-light-chain-enhancer of activated B (NF-κB) and JNK pathways. This creates a reciprocal and bilateral relationship between ROS production and ER stress, which drives the progression from MASLD to MASH. Oxidative stress and inflammation feeds on persistent ER stress in liver with MASH which in turn induces a deregulated increase of cholesterol synthesis and accumulation of free cholesterol [75]. This accumulation of cholesterol and persistent ER stress in HCC precursor cells might perpetuate the ER stress and increase its transport to other organelles. It is plausible that these occurrences contribute to lipotoxicity, promoting necrotic and apoptotic cell death, tissue damage and compensatory proliferation [13], potentially leading to a heightened mutagenesis rate and thereby promoting HCC initiation.

Cancer cells are characterized by an array of metabolic transformations, which promote their unrestricted growth and ability to escape apoptosis by counteracting cell death pathways and evasion of immune surveillance [30, 89]. Cholesterol metabolism is deregulated in tumors [47] and it has been previously described that cholesterol trafficking to mitochondria in tumor cells, due to overexpression of the mitochondrial cholesterol transporter STAR, increase the resistance of HCC cells to cell death induction [90, 91], in addition to the role of STAR-mediated trafficking of cholesterol to mitochondria in MASLD pathogenesis [40, 70]. Moreover, this cholesterol transporter has been found upregulated in HCC cells [82, 85] and in patients with MASH-derived HCC [74]. Consequently, the combined dysregulation of cholesterol metabolism and its increased trafficking to mitochondria may be impacting in the initiation, survival, cell death evasion and immune surveillance of HCC cells by different mechanisms, as will be discussed below. A diagram illustrating the possible ways in which specifically the buildup of cholesterol in mitochondria could play a role in the formation of HCC is shown in Figure 1.

Another critical survival characteristic of tumors is their ability to evade immune surveillance, which is the process by which transformed cells are usually detected and eliminated by the immune system. As transformed cells usually display neoantigens in their surface, this ability allows that only cells with the capacity to escape immune surveillance are able to establish a tumor during the microevolution process towards malignancy. Definitely, the immune system in the liver plays a vital role in tumor surveillance and the failure of this system can lead to the development of liver cancer [143]. Additionally, the interaction between immune cells and cancer cells is influenced by the metabolic changes that occur in immune cells [144]. The adaptive immune system, particularly CD8+ T cells and lymphotoxin β, has been identified as a key mediator in liver injury and the development of HCC [145]. In this regard, the antitumor function of natural killer cells is known to increase with high serum cholesterol levels, leading to a reduction in the growth of liver tumors [146, 147]. While cholesterol in some immune cells activate their activity, the accumulation of cholesterol in liver tissue and/or the presence of its metabolites, such as oxysterols or bile acids, may create an immunosuppressive microenvironment in this tissue. Indeed, there is an observed immunosuppressive effect of a cholesterol-enriched tumor microenvironment, which contributes to unrestricted tumor growth [148, 149]. Consistent with this, an increase in liver cholesterol levels was found to enhance the expression of lymphocyte exhaustion genes such as Pdl1 and Ctl4 among others, in mice with DEN-HFHC-induced HCC and in the spontaneous MuP-UPA transgenic mice model of HCC. Importantly, pharmacological treatment with ezetimibe, which reduces liver cholesterol levels, repressed the expression of these genes and the incidence and growth of tumors, suggesting that this treatment could potentially restore the cytolytic activity of CD8+ cells [140] in both the induced and spontaneous model. Furthermore, in another study using the MuP-UPA transgenic spontaneous HCC model, treatment with checkpoint inhibitors was able to induce regression of HCC tumors [150]. These results, although very promising, showed a partial response as it has been found in humans in the clinical setting, indicating that additional factors are blunting the cytolitic response against tumors in HCC that cannot be completely corrected with checkpoint inhibitors. These results prompt to the use of pharmacological interventions for reducing liver cholesterol burden as a coadjuvant to immunotherapies.