Liver cancer is the third leading cause of cancer-related deaths worldwide [1]. In 2020, there were more than 900,000 new cases and 800,000 deaths attributed to this malignancy [1]. Unsettlingly, the incidence of this cancer type is growing, and it is estimated that more than 1 million new cases will be diagnosed per year worldwide by 2025 [2]. Hepatocellular carcinoma (HCC) is the most frequent type of primary liver cancer, accounting for 75–90% of all cases ^[1][2][1,2]. The greatest risk factors for HCC are viral infections by hepatitis B, which accounts for more than half of all liver cancer cases and deaths, and hepatitis C ^[1][2][1,2]. In the West, hepatic injury leading to non-alcoholic steatohepatitis (NASH) is becoming a more prominent risk factor for HCC due to increased rates of obesity, diabetes, and metabolic syndrome [2]. Treatment options for advanced HCC (which accounts for 40% of HCCs at diagnosis) are limited to supportive care and systemic therapies including the multi-tyrosine kinase inhibitors sorafenib, lenvatinib, cabozantinib, and regorafenib ^[3][4][3,4]; immune checkpoint inhibitors; and monoclonal antibodies [2]. The recently approved combination therapy employing antibodies against programmed cell death 1 ligand 1 (atezolizumab) and vascular endothelial growth factor (bevacizumab) has led to a marked improvement in progression-free and overall survival compared with the standard of care agent, sorafenib [2]. Despite increased incidence of serious adverse events (^[5]reviewed in [5]), atezolizumab plus bevacizumab is currently the first-line treatment for HCC patients (^[6]reviewed in [6]). Although novel therapeutic approaches have been developed, palliative care applies to many HCC patients. Therefore, investigating the molecular mechanisms underpinning HCC initiation and progression may provide vital clues to aid the development of novel therapeutic strategies for the prevention and treatment of HCC. As such, new therapeutics for the treatment of liver cancer are highly desired and are a global health priority.

Glycosphingolipids (GSLs) contain lipid and sugar moieties. They are important components of the cellular membrane and act as signaling molecules in cellular processes such as apoptosis [7]. Uridine diphosphate (UDP)-glucose ceramide glucosyltransferase (UGCG) catalyzes the first glycosylation step in the synthesis of GSLs. By transferring UDP-glucose to ceramide, glucosylceramide (GlcCer, cerebroside) is produced, which is the precursor for all complex GSLs. Therefore, UGCG, which resides in the Golgi apparatus, is the key enzyme of GSL metabolism. For more detailed insight into GSL production^[8], we refer to our review from 2020 [8]. By adding a galactose molecule to GlcCer, lactosylceramide (globoside) is synthesized. More complex GSLs, known as globosides, such as globotriaosylceramide (Gb3) and gangliosides such as monosialodihexosylganglioside (GM3) are produced by adding monosaccharides to lactosylceramide (^[9]reviewed in [9]). The precursor for sulfatide (glycosphingolipid sulfate) is galactosylceramide, which is produced in the endoplasmic reticulum (ER). Sulfatides carry a sulfate ester group attached to the carbohydrate moiety.

2. UGCG/GlcCer

Previous studies have shown that UGCG expression is altered in 0.8% of HCC tumors (TCGA, Firehose Legacy) ^[10][11][10,11]. UGCG gene expression is increased in HCC tissue compared to non-cancerous tissue ^[10][11][12][10,11,34] and Li et al. showed upregulation of GlcCer (hexosylceramide) in serum samples of HCC patients ^[13][35]. Interestingly, in a review by Simon et al. it was reported that during HCC development, cellular ceramide levels decrease, and sphingosine-1-phosphate (S1P) levels increase ^[14][19]. This decrease in ceramide supports tumor growth and inhibits apoptosis, which correlates with the high proliferative capacity of HCC (^[14]reviewed in [19]). Besides S1P production, ceramide clearance is also achieved by UGCG, which again poses the question of how GSLs are involved in HCC development. Jennemann et al. showed delayed tumor growth in diethylnitrosamine (DEN)-induced liver tumors in mice, which exhibit a liver specific UGCG knockout (KO) ^[15][36]. The effect is mediated by decelerated cytokinesis, but the precise molecular mechanisms are unknown. Interestingly, the sphingomyelin concentration is increased in normal liver and tumor tissues of liver specific UGCG KO mice, which could be a cellular mechanism to avoid ceramide induced apoptosis ^[15][36]. Since a lack of GSLs per se does not prevent liver tumor development, other proteins/pathways, beside UGCG, may play a role in HCC development. Indeed, UGCG mRNA expression and GlcCer levels are increased in the livers of mice with mechanistic target of rapamycin (mTOR)-activated HCC tumors, and UGCG inhibition reduced proliferation of hepatocytes, tumor burden, and markers of liver damage ^[16][37]. These data suggest that UGCG is indeed involved in tumor development, but only when carcinogenesis was already activated. The role of UGCG in liver tumorigenesis is likely mTOR-dependent and mTORC2 might be a potential target to treat HCC. Another study showed sorafenib induced UGCG expression leading to sorafenib resistance in liver cancer cells ^[17][13]. However, following UGCG inhibition, no alterations in phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) and RAF)/mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase signaling were detected in liver cancer cells ^[18][30]. These, compared to normal liver cells, are contradictory study results. This contradiction could be related to the induction of the mentioned signaling pathways rather in the onset of carcinogenesis than to the induction of these signaling pathways in established cancer cells. Interestingly, downregulation of ORMDL sphingolipid biosynthesis regulator 3 (ORMDL3) in the human liver cancer cell line HepG2 led to increased levels of GlcCer (C16, C22, C24:0) and downregulation of ORMDL1 also led to an increase in GlcCer levels (C16, C20, C22, C24) to an even greater extent ^[19][58]. Furthermore, irritant-induced inflammation decreased ORMDL protein expression and increased GlcCer levels (C22, C24) in the livers of mice. These data indicate that ORMDLs may be involved in regulation of ceramides during interleukin-1-mediated sterile inflammation in liver cancer cells ^[19][58]. Ying et al. utilized data from the UALCAN web resource to show that the expression of glucosylceramidase beta 3 (GBA3) is significantly decreased in HCC tissues ^[20][59], leading to GlcCer accumulation. In contrast to GBA1 (lysosomal) and GBA2 (extra-lysosomal), GBA3 is localized in the cytosol and exhibits its highest activity at neutral pH (GBA3 identified by Hayashi et al. ^[21][60]). GBA3 mRNA expression is significantly lower in HCC than in non-tumor liver tissue (328 HCC samples, 151 non-tumorous liver tissues) ^[20][59].

3. Globoside Lactosylceramide

There is a limited number of published studies about lactosylceramide and how they impact HCC. One study identified lactosylceramide as a biomarker in five types of HCC cell lines ^[22][38]. Another study analyzed serum samples of HCC patients and showed that lactosylceramide is increased ^[13][35]. Souady et al. also showed enrichment of lactosylceramide in cancerous liver tissue compared to healthy tissue ^[23][39]. Further research is needed to clarify whether this GSL species might be a novel therapeutic target in HCC.

4. Globosides (DSSG, Gb3, Gb2, Globo H, Gb4, iso-Gb4)

Disialosyl galactosyl globoside (DSGG), Gb3, and Gb2 are increased in the serum of HCC patients compared to the serum of healthy individuals ^[24][40]; therefore, they may serve as prognostic markers for this disease. This is the first time DSGG has been shown to be highly expressed in HCC ^[24][40]. DSSG has been linked to metastatic potential in renal cell carcinoma ^[25][61]. Contradictory data from Souady et al. showed reduced expression of Gb3 and Gb4 in malignant liver tissue ^[23][39]. However, Globo H (Fucosyl-Gb5) might be a potential prognostic marker for HCC patients as well, since Zhu et al. showed that Globo H is only expressed in human HCC tissues, but not in peritumoral tissues ^[26][41]. Furthermore, Su et al. showed that Globo H is expressed on cancer stem cells generated from Hepa-1 cells ^[27][42], while Ariga et al. showed accumulation of iso-Gb4 in female rat hepatomas ^[28][43]. Thus, globosides may play a role in HCC development and serve as markers for malignant liver tissue.

5. Gangliosides (GM1, GM2, GM3, GD3, NeuGcGM3 Ganglioside)

Very early studies showed that GM2 levels are increased in two HCC samples compared to normal liver tissue ^[29][30][44,45] and that GM3 levels are decreased in two HCC samples compared to normal liver tissue ^[29][44]. However, GM3 is involved in cell migration, which is an important step in the process of metastasis. Li et al. showed that GM3, but not GM2, inhibited epidermal growth factor-stimulated cell migration and promoted the hepatocyte growth factor-stimulated migration in Hepa1–6 cells. These effects are mediated through the activation of the PI3K/Akt signaling pathway ^[31][32][33][46,47,48]. Su et al. showed that ganglioside synthesis is increased in the livers of mice in an animal model featuring activation and expansion of liver progenitor-like cells and liver cancer (stem) cells ^[34][49]. Together with elevated ganglioside synthesis, the expansion of mouse hepatic stem/progenitor cells was increased. Furthermore, GM1 ganglioside levels were significantly higher in the epithelial cellular adhesion molecule (EpCAM) positive cancer stem cell (CSC)-like HCC cell line JHH7. D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (D-PDMP), an inhibitor of UGCG, decreased ganglioside synthesis and suppressed cell proliferation and spheroid growth of JHH7cells; whereas apoptotic and necrotic cell death were not impacted ^[34][49]. PDMP effects were attributed to decreased expression of Aurora kinase A, Aurora kinase B, protein kinase TKK, kinetochore protein NDC80 homolog, Ki67, and CCNB1; whereas p53 was increased. Accordingly, p53 may have led to the decrease in expression of the aforementioned cell cycle/cytoskeletal-regulatory proteins. Interestingly, the inhibition of ganglioside synthesis also changed the lipid composition of liver cancer cells ^[34][49]. The researcheuthors conclude that blocking ganglioside synthesis might be a novel therapeutic option for HCC patients ^[34][49]. However, since PDMP blocks UGCG at the first key step in GSL synthesis, the study results from Su et al. do not clearly indicate that the described effect is ascribable to gangliosides or possibly GlcCers. Interestingly (and contrary to other studies), natrin (purified from snake venom and a member of the cysteine-rich secretory protein (CRISP) family) exhibits anticancer activity in HCC by inhibiting cell proliferation and inducing apoptosis. Lu et al. identified gangliosides as potential biomarkers for natrin-induced apoptosis ^[35][50]. Following natrin treatment, gangliosides increased in concentration, as well as the ratio of Bax to Bcl-2, which indicates higher susceptibility of cells to apoptosis. Since UGCG/gangliosides are increased in HCC, the question arises whether cells might reach a point of excessive ganglioside accumulation and therefore apoptosis would be induced. Additionally, more research is required about which specific ganglioside species are changed in HCC. Another important parameter to consider is the immune response in HCC development. Zhu et al. investigated immune responses based on the HCC cohort of The Cancer Genome Atlas (TCGA) database ^[36][62]. By assigning immune cell types to three different immunity groups (low, medium, and high) and following Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, the researcheuthors showed that the immunity groups differ in gene expression of ganglio series GSL producing proteins ^[36][62]. This poses the question as to whether the immune response (immune cell infiltration of the tumor) can be influenced and therefore immunotherapy enhanced. Wu et al. showed increased fucosyl GM1 levels in the serum of HCC patients, which indicates a potential prognostic marker function of this ganglioside ^[24][40]. However, during induction of rat hepatoma by DEN, GD3 increased in malignant tissues compared to normal tissue, as well as the appearance of precancerous hepatocytes ^[37][51]. Furthermore, the ganglioside GD3 induces cell death by targeting mitochondria and engagement of the apoptosome leading to inhibition of survival pathways in HCC^[38], which is reviewed here [52]. However, GD1α is increased in two HCC samples compared to normal liver tissue ^[29][44]. CD75s-(Neu5Acα6Galβ4GlcNAcβ3Galβ4Glcβ1Cer) and iso-CD75s-gangliosides (Neu5Acα3Galβ4GlcNAcβ3Galβ4Glcβ1Cer) are increased in HCC tissue, but independent of ST6GAL1 and ST3GAL6 expression ^[39][53]. NeuGcGM3 ganglioside is overexpressed in HCC as well ^[40][54]; therefore it might be a potential target for HCC therapy. Additionally, ganglioside depletion reduces integrin-mediated cell adhesiveness of rat hepatoma cells ^[41][63].

6. Lacto/Neo-Lacto Series Glycosphingolipids

Zhu et al. showed that fucosylated GSLs are overexpressed in HCC samples compared to the adjacent tissue ^[26][41]. In detail, Fuc(Hex)₃HexNAc-Cer (Fucα₂Galβ₃GlcNAcβ₃Galβ₄Glcβ₁Cer, Fucα₂Galβ₄GlcNAcβ₃Galβ₄Glcβ₁Cer (H5-2), Fucα₃(Galβ₄)GlcNAcβ₃Galβ₄Glcβ₁Cer, Fucα₄(Galβ₃)GlcNAcβ₃Galβ₄Glcβ₁Cer), Fuc₂(Hex)₃HexNAc-Cer (Fucα₂Galβ₃(Fucα₄)GlcNAcβ₃Galβ₄Glcβ₁Cer (Leb-6), Fucα₂Galβ₄(Fucα₃)GlcNAcβ₃Galβ₄Glcβ₁Cer (Ley-6)), and Fuc(Hex)₄HexNAc-Cer (Fucα₂Galβ₃GalNAcβ₃Galα₄Galβ₄Glcβ₁Cer) are expressed in HCC samples ^[26][41]. More research is needed to shed light on the role of the fucosylated lacto/neo-lacto series GSLs in HCC and in cancer in general.

7. Sulfatides (Glycosphingolipid Sulfates)

Sulfatides are highly expressed in HCC ^[42][55] and regulate integrin αV expression and cell adhesion in hepatoma cells ^[43][56]. The regulatory mechanisms are based on complexing of sulfatide with paired amphipathic helix protein (SIN3B) leading to reduced binding affinity of SIN3B to histone deacetylase 2 (HDAC2). Subsequently, HDAC2 recruitment to the integrin αV gene promoter is reduced and the promoter thereby activated. This leads to enhanced tumor metastasis ^[44][57].