1. Pathogenesis of HCC

The cells of origin of HCC are not clear. Experimental evidence supports the implication of transformed mature hepatocytes as the cell of origin, but also the possibility that the source is liver stem cells ^[1][2][24,25].

HCC is the result of either mutations, such as those in the TERT promoter or p53 suppressor gene ^[3][4][26,27], or epigenetic modifications. Some of them are directly involved or activate important signaling pathways leading to HCC ^[5][28]. Genes coding for several signaling pathways, such as Wnt/β-catenin, oxidative stress, AKT/mTOR and MAP kinase, are often mutated in HCC and are used for the molecular classification of HCC ^[6][29]. All these molecular abnormalities are triggered by external factors, such as alcohol consumption, viral infection and abnormal nutrition. In general, three mechanisms are implicated in the initiation and progress of HCC, namely, persistent liver inflammation, endoplasmic reticulum (ER) stress and abnormalities of cell signaling pathways ^[7][5]. Inflammation is an important pathogenetic factor irrespective of the etiology of the liver disease that leads to HCC ^[8][30]. Inflammation starts when hepatocytes undergoing programmed or accidental death liberate factors, such as HMGB1 and HDGF, to initiate an inflammatory response ^[9][31]. Different inflammasomes, particularly the nucleotide-binding oligomerization domain, leucine-rich repeat and pyrin domain containing 3 (NLRP3), are activated and lead to the release of the pro-inflammatory cytokine IL-1β. NLRP3 activation is mostly triggered by the production of ATP from the mitochondria of the damaged cells ^[10][32] and lysosomal disruption ^[11][33]. NLRP3 inflammasome activation in hepatocytes is a two-step process. Priming is the first step when damage-associated molecular patterns (DAMPs) from damaged cells and pathogen-associated molecular patterns (PAMPs) stimulate TLR receptors, followed by translocation of NF-kB to the nucleus and increase in pro-IL-1β and pro-IL-18 expression. The second step is triggered when extracellular ATP or active lysosomal enzymes finally lead to the activation of caspase-1. Cleavage by caspase-1 turns pro-cytokines into mature IL-1β and IL-18, while the cleavage of gasdermin D leads to a programmed cell death called pyroptosis. In this canonical activation of pyroptosis, fragments of gasdermin D form pores in the plasma membrane, killing the cell and releasing IL-1β and IL-18, which aggravates inflammation ^[12][13][34,35].

In the non-canonical pyroptosis, lipopolysaccharides (LPSs), from Gram-negative bacteria, turn the pro-caspases 4, 5 and 11 into active enzymes that cleave gasdermin D. The pores in the plasma membrane are formed, but without maturation, and release IL-1β and IL-18. This is not always the case, as caspase-11 may activate the NLPR3-dependent caspase-1 inflammasome and indirectly stimulate the release of intracellular cytokines ^[14][15][36,37].

Interestingly, inflammasome-mediated pyroptosis also occurs in non-parenchymal liver cells, implicating the gut microbiota. DAMPS and gut-derived PAMPs activate Kupffer cells that produce IL-1β and TNF. NLRP3 activation in hepatic stellate cells (HSCs) promotes the production of the profibrogenic cytokine and induces the expression of the profibrogenic molecule TGF-β. These events in concert lead to liver inflammation and fibrosis, being the link between liver damage and hepatocellular carcinoma ^[16][17][38,39].

It was suggested that cathepsin B from lysosomes is the triggering stimulus of the NLRP3 inflammasome, which is mostly mediated by the release of Cathepsin B ^[18][19][20][40,41,42]. However, mice macrophages deficient in cathepsin B showed comparable NLRP3 inflammasome activation with wild-type animals ^[21][43], suggesting that other products contribute to NLRP3 inflammasome activation. Cathepsins B, L, C, S and X are probable candidates in NLRP3 inflammasome activation by silica particles ^[22][44]. In an adenovirus infection, cathepsin B release is also a mediator of inflammation, but reactive oxygen species (ROS) inhibition reduces IL-1β secretion, indicating that ROS production might be the mechanism of the induction of inflammasome activation by cathepsin B ^[23][45].

ER stress is caused by the accumulation of unfolded or misfolded proteins in the ER lumen. Pathogens, mutations and an increased metabolic rate led to an increase in the protein secretory load of the hepatocyte. The accurate monitoring of protein folding is not maintained in ER, inducing the unfolding protein response (UPR) either to normalize protein synthesis or to induce cell death in severe ER stress ^[24][25][46,47]. ER stress in murine hepatocytes activates inflammatory pathways, such as NF-kB and TNF, leading to HCC induction ^[26][48]. Chronic ER stress and increased UPR activity have been implicated in the development of HCC and are present in HCC tumors irrespective of grade or stage ^{[27][28][29][30]}[49,50,51,52].

The abnormal activation of the oncogenic phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) signaling is associated with HCC. This is not unexpected, as this pathway is involved in various cellular functions, such as cellular proliferation, differentiation, apoptosis and metabolism ^[31][53]. The AKT/mTOR pathway is known to interfere with aerobic glycolysis regulating three limiting enzymes in the glycolytic pathway (hexokinase 2, phosphofructokinase 1, and pyruvate kinases type M2), a fact that is crucial for HCC progression ^{[32][33][34][35]}[54,55,56,57]. The mTOR pathway is analyzed later, as it is implicated in the interplay between autophagy and apoptosis.

The deregulation of Wnt/β-catenin signaling is critical in human HCC ^[36][37][58,59]. A total of 35% of human HCC tumors had a gain-of-function mutation of CTNNB1 encoding β-catenin and loss-of-function mutation of AXIN1 ^[38][39][60,61]. Resistance to sorafenib and regorafenib treatment was attributed to activated Wnt/β-catenin in HCC patients ^[40][41][62,63]. There is evidence that a second signal is required because Wnt/β-catenin alone is not sufficient to induce hepatocarcinogenesis. Oncogenic mutations of β-catenin cooperate with other oncogenes, such as c-Met ^[42][43][44][64,65,66] and K-RasV12 ^[45][67]. After activation, β-catenin induces several downstream targets that are implicated in HCC induction. c-MYC is one of the best-studied down-stream effectors of β-catenin ^[46][47][68,69].

miRNAs are regulators of several tumor-related genes in carcinogenesis, acting either as oncogenes or tumor suppressor genes. miRNAs are classified according to their implication in the main molecular pathways leading to HCC tumorigenesis ^[48][70]. miR-30a, miR-365, miR-526a, miR-377, miR-199a-5p and miR-330 all were implicated in apoptosis regulation and were either upregulated or downregulated in HCC ^[49][71]. Other mRNAs are involved in the repression of the PI3K/AKT/mTOR pathway or the Wnt/β-catenin pathway ^[31][53]. Tumor suppressor miRNAs are associated with either HCC initiation and progression ^[50][72] or with metastasis and recurrence ^[51][73]. Similarly, some pro-oncogenic miRNAs are associated with HCC initiation and progression ^[52][74], or are involved in HCC recurrence and metastasis ^[53][75].

Exosomes transfer proteins, DNAs and RNAs, such as miRNAs, long non-coding RNAs (lncRNAs) and messenger RNAs (mRNAs), between HCC and normal cells. Exosomes initiate either local or systemic reactions, participating to the initiation and progression of HCC. Exosomes are used as biomarkers and therapeutic tools in HCC ^[54][55][76,77].

Ferroptosis is an iron-dependent process of regulated cell death, with accumulation of lipid peroxides that causes damage to liver cells and is associated with the development of HCC. It is analyzed later in the paper.

Microbial products in the bowel are deeply involved in HCC pathogenesis. Bacterial products from the gut microbiota can directly or indirectly damage DNA through the pro-duction of ROS ^[56][57][78,79]. The altered composition of the gut microbiota may be one of the mechanisms underlying the action of aflatoxin and other mycotoxins as powerful inducers of HCC ^[58][80]. In addition to gut microbial products, several studies have incriminated specific gut microbiota in association with HCC. Enterococcus faecalis is increased in HCV-induced HCC development ^[59][81]. A similar study showed a significant increase in Enterococcus species in patients with viral and alcoholic cirrhosis leading to HCC ^[60][82]. Microbiotas also interfere with bile acid metabolism, contributing to HCC development. Bile acid metabolites, produced by the gut microbiota, can cause inflammation, ROS overproduction and a reduction in apoptosis in the liver, finally leading to the development of HCC. Moreover, they can modulate the function of liver immune cells, affecting HCC progression. They can also indirectly contribute to the activation of the TLR4 receptor in hepatocytes and Kupffer cells. Bile acids increase gut permeability, acting on the tight junctions, and allow for an increased transportation of LPS to the liver, thus promoting angiogenesis and the downregulation of tumor suppressor miRNAs ^[61][83].

Ca²⁺ is present in various cell compartments transported among them by transporters and exchangers, collectively known as the Ca²⁺ transportome. The impairment of the Ca²⁺ transportome contributes to HCC initiation, the formation of metastatic cells and reduction in cell death ^[62][84].

These two critical parameters in the initiation and progress of HCC are analyzed separately. Two very informative reviews on the pathogenesis of HCC have been recently published ^[7][16][5,38].

Detailed pathophysiological factors implicated in HCC pathogenesis, including gene mutation and epigenetic changes, have been recently reviewed ^[63][85].