Metabolic reprogramming-driven resistance to multi-kinase inhibitors in hepatocellular carcinoma: molecular mechanisms and therapeutic opportunities

Abstract

Hepatocellular carcinoma (HCC), the most common form of primary liver cancer, is frequently diagnosed at advanced stages, limiting curative options. Multi-kinase inhibitors (MKIs), such as sorafenib and lenvatinib, serve as first-line therapies for unresectable HCC. However, the widespread development of drug resistance significantly diminishes the clinical efficacy of MKIs, and current treatments lack effective strategies to enhance MKI sensitivity. Metabolic reprogramming, a hallmark of cancer cells that facilitates unchecked growth and metastasis, has emerged as a critical mechanism driving MKI resistance in HCC. This review comprehensively examines the roles of glycolysis, lipid metabolism, and amino acid metabolism in promoting MKI resistance, with a focus on key molecular regulators that could serve as potential targets to reverse resistance. Additionally, this review synthesizes preclinical and clinical evidence of therapeutic agents that synergize with MKIs by modulating metabolic pathways, and discusses the regulatory role of metabolic reprogramming in the tumor immune microenvironment (TIME) of HCC, offering innovative strategies to improve treatment outcomes for patients with HCC. These findings highlight metabolic reprogramming as a crucial target for developing novel interventions aimed at overcoming MKI resistance in clinical practice.

Introduction

Nutrient metabolism is a vital physiological process in cells. In contrast to normal cells, cancer cells exhibit uncontrolled proliferation, migration, and invasion. To meet their bioenergetic and biosynthetic demands in nutrient-deprived microenvironments, cancer cells undergo metabolic reprogramming, a process that sustains their growth and survival [1]. Metabolic reprogramming includes alterations such as the Warburg effect, lipid metabolism, and amino acid metabolism [2, 3]. These changes are marked by enhanced aerobic glycolysis and increased biosynthesis of macromolecules, such as lipids and amino acids [4–6]. Several signaling pathways, such as PI3K/AKT/mTOR, Myc, and MEK/ERK, drive metabolic reprogramming in cancer, promoting tumor growth, metastasis, and resistance to therapy [7–9]. Targeting cancer metabolism has emerged as a promising approach for both cancer treatment and overcoming drug resistance [10, 11].

Liver cancer ranks as the third leading cause of cancer-related mortality worldwide, with 757,948 deaths reported in 2022 [12]. Hepatocellular carcinoma (HCC) accounts for approximately 80% of primary liver cancer cases [13]. Due to its subtle progression, most patients with HCC are diagnosed at an advanced, unresectable stage, where only palliative pharmacological treatments are available [14]. Conventional chemotherapies, such as 5-fluorouracil (5-FU), cisplatin, and gemcitabine, show limited efficacy in advanced HCC [15]. Consequently, molecularly targeted therapies have become a pivotal strategy for treating advanced HCC [16, 17]. Multi-kinase inhibitors (MKIs), such as sorafenib, lenvatinib, and regorafenib, target multiple signaling pathways and have been approved as first-line treatments for unresectable HCC [17]. However, approximately 25% of HCC patients develop primary resistance to sorafenib [18]. Even though some patients may derive short-term benefits, most will develop acquired resistance during subsequent treatment, which significantly limits their survival time [19, 20]. In fact, despite having a higher objective response rate (ORR) than sorafenib’s 9.2%, lenvatinib only achieves an ORR of 24.1% [21]. Regorafenib, an optional treatment after sorafenib failure, also has an ORR only slightly higher than that of sorafenib [22]. The widespread development of drug resistance means that only a subset of patients benefit from MKI therapy [23, 24]. Recently, A. D. Ladd et al. reported mechanisms contributing to therapeutic resistance in HCC. These mechanisms involve alterations in signaling pathways, drug efflux, changes in the tumor microenvironment (TME), and dysregulation of apoptosis, among others [25]. However, the role of metabolic reprogramming in HCC drug resistance has not yet been elucidated in detail. Understanding the molecular mechanisms underlying metabolism-mediated MKI resistance and identifying novel agents to reverse MKI resistance and enhance MKI efficacy is crucial for improving the prognosis of patients with HCC. This review explores the role of metabolic reprogramming in HCC and how it contributes to MKI resistance, proposing the combination of targeted metabolic agents as a promising strategy to overcome resistance in HCC.

Metabolic characteristics driving MKI resistance in HCC

HCC is a rapidly growing and highly invasive malignancy [26]. The proliferation and survival of HCC cells are closely tied to energy supply under varying environmental conditions. To sustain their energy demands, HCC cells reprogram multiple metabolic pathways [27, 28]. Glycolysis facilitates the accumulation of glucose intermediates in HCC, with elevated expression of glycolysis-related enzymes observed in these cells [29]. Through glycolysis, glucose is converted to pyruvate, which is further converted to lactate by pyruvate dehydrogenase under anaerobic conditions [29]. Additionally, HCC exhibits abnormal lipid synthesis, degradation, and storage capabilities [30]. Specifically, HCC cells have a high capacity for de novo fatty acid (FA) synthesis and utilize FAs as an energy source for tumor cell proliferation [31, 32]. Lipids also function as mediators of signaling pathways and as structural components of cell membranes [33, 34]. Furthermore, the uptake and catabolism of glutamine are enhanced in HCC cells [35]. Glutamine, serving as a carbon and nitrogen donor, drives biosynthetic processes crucial for the onset and progression of HCC [36]. The following section will examine the distinct metabolic features of HCC in greater detail (Fig. 1 and Table 1).

Fig. 1.

Metabolic Reprogramming in HCC. HCC cells are characterized by aberrant glycolysis (blue arrows), lipid metabolism (yellow arrows), and amino acid metabolism (gray arrows). In HCC cells, pyruvate generated via glycolysis tends to be converted into lactate by LDHA, rather than entering the mitochondria to participate in the TCA cycle. The synthesis of FAs and cholesterol is enhanced. FAs supply ATP for HCC cells via the FAO pathway. Additionally, FAs contribute to phospholipid synthesis, which are essential components of cell membranes. Glutamine is converted into glutamate by GLS, which then participates in GSH synthesis or is further metabolized to α-KG to enter the TCA cycle. Serine is synthesized from 3-PG of the glycolytic pathway through catalysis by PHGDH, PSAT1 and PSPH. Aspartate is generated from oxaloacetate and glutamate via catalysis by GOT2, a process that occurs within the mitochondria. These metabolic pathways are interconnected and collectively mediate HCC growth, metastasis, and MKI resistance

Table 1.

Key Enzymes in metabolic dysregulation of HCC

Enzymes	Functions	Reference
Glycolysis
HK2	Catalyzes the reaction of glucose and ATP to produce glucose-6-phosphate	42
PFK1	Catalyzes the conversion of fructose-6-phosphate to fructose-1,6-bisphosphate	41
PKM2	Catalyzes the conversion of phosphoenolpyruvate to pyruvate	43,44
PDH	Catalyzes the oxidative decarboxylation of pyruvate to produce Acetyl-CoA	46
LDHA	Catalyzes the conversion of pyruvate to lactate	47
Lipid metabolism
ACLY	Catalyzes the conversion of citrate to Acetyl-CoA and oxaloacetate	74
ACC	Catalyzes the combination of Acetyl-CoA and CO₂ to produce Malonyl-CoA	75
FASN	Mediates de novo fatty acid synthesis	76
ACSL1/4	Activates long-chain fatty acids in cells to form fatty acyl-CoA	81,82
CPT1	Mediates the transport of fatty acyl-CoA from the cytoplasm to the mitochondria	83,84
SCD1	Promotes the production of MUFAs	95
SPT, SPK, SMS, CERK	Promote the synthesis of sphingolipids	104
Amino acid metabolism
GS	Promotes the synthesis of glutamine	113
GLS	Catalyzes the conversion of glutamine to glutamate	115
PHGDH	Catalyzes the serine synthesis	121
GOT2	Catalyzes the conversion of oxaloacetate and glutamate to aspartate	126

HCC Hepatocellular carcinoma, HK2 Hexokinase 2, ATP Adenosine triphosphate, PFK1 Phosphofructokinase 1, PKM2 Pyruvate kinase M2, PDH Pyruvate dehydrogenase, Acetyl-CoA Acetyl coenzyme A, LDHA Lactate dehydrogenase A, ACLY ATP citrate lyase, ACC Acetyl-CoA carboxylase, Malonyl-CoA Malonyl coenzyme A, FASN Fatty acid synthase, ACSL Acyl-CoA synthetase long-chain family, CPT1 Carnitine palmitoyltransferase 1, SCD1 Stearoyl-CoA desaturase 1, MUFAs Monounsaturated fatty acids, SPT Serine palmitoyltransferase, SPK Sphingosine kinase, SMS Sphingomyelin synthase, CERK Ceramide kinase, GS Glutamine, PHGDH Phosphoglycerate dehydrogenase, ASNS Asparagine synthetase, GOT Glutamic-oxaloacetic transaminase

Aerobic glycolysis in HCC

In normal cells, glucose is metabolized to CO₂, H₂O, and a substantial amount of ATP through oxidative phosphorylation (OXPHOS). However, even in the presence of sufficient oxygen, cancer cells preferentially utilize glycolysis, a phenomenon known as the Warburg effect [37]. In HCC, the highly expressed glucose transporters (GLUT) 1, 3, and 4 enhance glucose uptake, contributing to the elevated glycolytic rate in HCC cells [38, 39]. Several key enzymes in the glycolytic pathway, including hexokinase (HK), phosphofructokinase 1 (PFK1), and pyruvate kinase (PKM), are upregulated in HCC cells [40, 41]. Moreover, these glycolytic enzymes exhibit distinct isoform expressions during HCC progression. For instance, the high-affinity isoforms HK1 and HK2 are commonly upregulated in HCC [42], while PKM tends to splice preferentially into the PKM2 isoform, which promotes glycolysis [43, 44]. Glucose accumulates in HCC cells via GLUT1 and is subsequently converted to pyruvate by HK2, PFK, and PKM2 [45]. Pyruvate is then transported into mitochondria by pyruvate dehydrogenase (PDH) to generate acetyl-CoA and ATP [46]. However, PDH activity is inhibited in HCC, leading to a higher proportion of pyruvate being converted to lactate by lactate dehydrogenase A (LDHA), resulting in an acidic tumor microenvironment (TME) [47]. The glycolytic process in HCC is regulated by key signaling pathways, including PI3K/AKT, AMPK, HIF-1α, and c-Myc [41]. Activation of these pathways can directly upregulate GLUT1 and glycolytic enzymes or function as transcription factors to regulate glycolysis-related gene expression, thereby enhancing glycolysis [48, 49].

Recent studies have highlighted increased glycolysis as a key factor contributing to HCC resistance to MKIs. Clinical data show that elevated total lesion glycolysis (TLG) correlates negatively with PFS and OS in patients with HCC, indicating poor responses to sorafenib treatment [50]. Pan et al. developed an aerobic glycolysis index (AGI) model to predict poor prognosis and sorafenib sensitivity in HCC [51]. Glycolysis provides adenosine triphosphate (ATP) for ATP-binding cassette (ABC) transporters in cancer cells [52]. ABC transporters, a well-established class of drug efflux pumps, are upregulated in HCC cells, promoting sorafenib efflux and contributing to resistance [53, 54].

As the end product of glycolysis, lactate plays a significant role in cancer progression and drug resistance. Monocarboxylate transporters (MCTs) participate in lactate transport and enhancing lactate shuttling between cells. MCT1 and MCT4 are the most important MCT subtypes in HCC. MCT4 serves as the primary lactate efflux carrier, preventing intracellular acidification. In contrast, MCT1 is responsible for lactate uptake by neighboring cancer cells and immune cells, in addition to being involved in a minor portion of lactate efflux [55–57]. Excessive extracellular lactate creates an acidic pH environment, which may hinder the uptake of weakly alkaline chemotherapeutic agents such as vincristine and paclitaxel by cancer cells due to dissociation [58]. Chen et al. demonstrated that using calcium carbonate nanoparticle-loaded gelatin microspheres (CaNPs@Gel-MS) as an emerging embolic agent can significantly enhance the uptake and cytotoxicity of doxorubicin (DOX) during transarterial chemoembolization (TACE) by neutralizing lactate [59]. Furthermore, lactate can upregulate ABC transporters, including multidrug resistance-associated protein 1 (MRP1), thereby increasing the efflux of antitumor drugs from cancer cells [60, 61]. Excessive exposure of cancer cells to lactate can prevent DNA damage and cytotoxicity induced by cisplatin. Lactate reduces DNA damage and enhances DNA recombination and repair by stimulating the expression of genes involved in mismatch repair and nucleotide excision repair pathways, thereby conferring chemoresistance to the cells [62]. Under hypoxic conditions, HIF-1α-induced lactate alters intracellular pH, reduces the release of free Fe²⁺, and enhances the SLC7A11 system, thereby increasing cellular resistance to ferroptosis [63]. Simultaneously, lactate-mediated AMPK inactivation in HCC modulates lipid metabolism, reducing cellular susceptibility to lipid peroxidation and ferroptosis [64]. Lactate promotes lactylation and modulates intracellular signaling pathways in cancer cells [65, 66]. In a lenvatinib-resistant HCC model, glycolysis facilitates the lactylation of IGF2BP3, thereby maintaining elevated levels of PCK2 and NRF2, which strengthens the antioxidant response [67]. Histone H3 lysine 18 lactylation (H3K18la) is frequently reported to be modified in response to lactate upregulation, thereby participating in tumor progression [65, 68]. In HCC, lactate-mediated H3K18la suppresses ferroptosis in HCC cells by enhancing the transcription of NFS1 cysteine desulfurase (NFS1), a key gene in the iron-sulfur cluster biosynthesis pathway [69]. Lactate secreted by cancer cells also promotes angiogenesis in HCC tumor tissues. It activates the G protein-coupled receptor GPR81 on the surface of both cancer and immune cells, acting as a signaling molecule that promotes angiogenesis through autocrine or paracrine mechanisms, thereby enhancing tumor cell survival by increasing local oxygen and nutrient supply [70]. Additionally, as an important signaling molecule secreted by HCC cells, lactate mediates the tumor immune microenvironment (TIME) in HCC, which will be discussed in greater detail in subsequent sections. Notably, MKI treatment may itself contribute to the onset of glycolysis. Sorafenib suppresses OXPHOS in HCC, partially reducing ATP supply [71]. However, this inhibition leads to a compensatory upregulation of aerobic glycolysis, enabling cell survival and contributing to sorafenib resistance [72].

Aerobic lipid metabolism in HCC

In addition to the aberrant activation of glycolysis, abnormal lipid metabolism plays a pivotal role in the malignant progression and drug resistance of HCC [30]. Lipid molecules in HCC encompass FAs, cholesterol, triglycerides, sphingolipids, and phospholipids [30]. FA metabolism is central to lipid metabolism in HCC, with cancer cells increasing intracellular FA accumulation through enhanced external uptake or de novo synthesis [73]. Recent studies have shown upregulation of key enzymes in the FA synthesis pathway, including ATP citrate lyase (ACLY), acetyl-CoA carboxylase (ACC), FA synthase (FASN), and stearoyl-CoA desaturase 1 (SCD1), promoting FA synthesis in HCC. Acetyl-CoA in the mitochondria combines with oxaloacetate to form citrate, which is transported to the cytoplasm and cleaved by ACLY to generate acetyl-CoA [74]. ACC then catalyzes the conversion of acetyl-CoA to malonyl-CoA, a critical step in de novo FA synthesis [75]. Malonyl-CoA and acetyl-CoA are condensed by FASN to produce FA products, primarily palmitic acid (PA) [76]. The de novo synthesis of FAs, mediated by FASN, is regulated by the transcription factor sterol regulatory element-binding protein 1 (SREBP-1) [77]. SREBP-1 also promotes lipid synthesis in HCC cells and is a key factor driving the progression from fatty liver disease to malignant HCC [78]. Furthermore, SREBP-1 upregulates glycolysis in HCC cells by enhancing the transcription of hexokinase domain-containing 1 (HKDC1) [79].

FA β-oxidation (FAO) is another critical pathway by which FAs support HCC cell functions [80]. Under energy-deficient conditions, FAs are converted into acyl-CoAs in the cytoplasm through the action of long-chain acyl-CoA synthetase (ACSL) 1/4 [81, 82]. These acyl-CoAs are then transported into the mitochondria by carnitine palmitoyltransferase (CPT) 1 and 2 located on the mitochondrial membrane [83, 84]. In the mitochondria, they undergo oxidative breakdown into acetyl-CoA and ATP [85]. Upregulation of FAO has been observed in various cancer types [86–88]. FAO not only supplies ATP to cancer cells but also enhances the production of NADPH and glutathione (GSH), increasing cellular resistance to drug-induced reactive oxygen species (ROS) [89, 90]. In addition to the FAO pathway, FAs can be utilized for the synthesis of monounsaturated FAs (MUFAs) and polyunsaturated FAs (PUFAs). MUFAs and PUFAs bind to the glycerol backbone and are phosphorylated to form PA [91]. Subsequently, glycerophospholipids synthesized with the involvement of PA serve as essential components of cell membranes [92]. The ratio of MUFAs to PUFAs affects membrane fluidity. Importantly, PUFAs, when incorporated into membranes under the catalysis of ACSL4, serve as major substrates for lipid peroxidation during ferroptosis [93, 94]. Elevated levels of PUFAs increase the susceptibility of cancer cells to ferroptosis, whereas upregulation of MUFA biosynthesis through SCD1 in HCC cells enhances resistance to ferroptosis [95]. Current evidence has indicated that SCD1 inhibition enhances the sensitivity of HCC cells to sorafenib treatment [96].

Increased cholesterol synthesis is a key characteristic in the progression of HCC from non-alcoholic fatty liver disease (NAFLD). Elevated cholesterol levels enhance the differentiation of cancer stem cells (CSCs) [97, 98]. Cholesterol is also a critical component of lipid membranes, and cholesterol-enriched lipid rafts increase membrane rigidity, thereby reducing the permeability of cell membranes to antitumor drugs [99]. Additionally, lipid rafts facilitate the accumulation of drug transporters such as ABCA1 and ABCG1, promoting drug efflux [100]. Moreover, lipid rafts act as platforms for multiple signaling pathways. In drug-resistant cells, cholesterol-rich lipid rafts stabilize the membrane localization of receptors, enhancing their ability to bind ligands or undergo autophosphorylation. This results in the persistent activation of signaling pathways, such as TGF-β1 and AKT, which promote HCC growth, survival, and drug resistance [101–103].

Sphingolipids are essential lipid molecules in cellular membranes and serve as signaling mediators. Integrated single-nucleus RNA sequencing and lipidomics analyses have identified dysregulation of key sphingolipid-related enzymes in HCC, including serine palmitoyltransferase (SPT), sphingosine kinase (SPK), spermine synthase (SMS) and ceramide kinase (CERK) [104]. The core of sphingolipid metabolism centers on ceramide (Cer) synthesis, a key mediator of pro-apoptotic signaling. Notably, Cer levels are reduced in HCC despite an increase in sphingolipid synthesis [105, 106], while elevated sphingolipid expression is associated with HCC progression and drug resistance [107, 108].

Amino acid metabolism in HCC

Amino acids serve as the fundamental substrates for cellular protein synthesis. In cancer, amino acid metabolism undergoes extensive reprogramming: they participate in biosynthetic processes, act as alternative fuels under nutrient deprivation, and contribute to redox homeostasis, thereby supporting cancer cell growth and chemoresistance [109]. Amino acids are generally classified into two categories: non-essential amino acids (NEAAs), such as glutamate, glutamine, serine, glycine, aspartic acid and proline; and essential amino acids (EAAs), including arginine, leucine, and methionine [110]. In this review, we focus on the metabolism of glutamine, serine, and aspartate in HCC.

Glutamine, a non-essential amino acid providing nitrogen and carbon sources, plays a critical role in the biosynthesis of various nutrients that support tumor growth and drug resistance [36]. Bioinformatics analyses have linked elevated glutamine metabolism with drug resistance and poor OS in HCC [111, 112]. CTNNB1, a frequently mutated gene in HCC, enhances glutamine production in HCC cells by upregulating glutamine synthetase (GS) [113]. Furthermore, increased expression of alanine-serine-cysteine transporter 2 (ASCT2) in HCC promotes cellular glutamine uptake [114]. Glutamine is converted to α-ketoglutarate (α-KG) by glutaminase (GLS), which then enters the TCA cycle to generate ATP. TCA cycle intermediates, such as citrate, can exit the mitochondria and serve as precursors for FA synthesis [115]. Glutamine also contributes to the synthesis of proline, arginine, and other amino acids necessary for protein production [36]. Under hypoxic conditions, glutamine-dependent proline metabolism leads to hydroxyproline accumulation, which inhibits sorafenib toxicity both in vivo and in vitro by stabilizing HIF-1α [116]. A recent study found that polypyrimidine tract-binding protein 1 (PTBP1) is highly expressed in HCC cells and promotes GLS expression. Silencing PTBP1 reduced GLS levels and decreased cisplatin resistance in HCC cells [117].

Serine/one-carbon metabolism supports the malignant growth of cancer cells [118], which enables cancer cells to enhance serine synthesis and the uptake of exogenous serine. Serine is primarily transported via SLC6A14 and SLC25A15 (also known as ASCT2) [119, 120], while serine synthesis is achieved through a glycolytic branch pathway catalyzed by phosphoglycerate dehydrogenase (PHGDH) [121]. Recent studies have revealed that enhanced serine metabolism can promote nucleotide synthesis, thereby repairing chemotherapy-induced DNA damage and conferring drug resistance to cancer [122]. Furthermore, sorafenib-resistant HCC cells exhibit enhanced serine synthesis, which promotes GSH production and augments the antioxidant system [123]. Aspartate, which is involved in the synthesis of proteins and nucleotides, plays a critical role in cancer growth and metastasis [124]. Aspartate is produced from oxaloacetate (OAA) and glutamate via catalysis by glutamic-oxaloacetic transaminase 2 (GOT2). High levels of aspartate can activate mTOR, thereby promoting the progression of HCC [125, 126].

Key molecular targets and regulatory mechanisms of metabolism-mediated MKI resistance

Metabolism in HCC is regulated by key enzymes and signaling pathways, creating an environment conducive to drug resistance. Metabolic reprogramming provides the energy and metabolic substrates necessary for the uncontrolled proliferation and survival of HCC cells, while also preserving the stem cell characteristics of the tumor. This review discusses the molecular mechanisms of MKI resistance mediated by metabolic reprogramming in HCC, which are comprehensively outlined in Table 2.

Table 2.

Molecular targets for regulating metabolism

Molecular targets	Regulating mechanism	Functions in HCC	Reference
LAT1	Promotes cellular uptake of histidine, thereby inhibiting GLUT1 expression	Facilitates histidine-mediated glycolysis suppression, thus enhancing sorafenib sensitivity	131
miR-199a	Targets and inhibits HK2	Blocks glycolysis, increases oxidative phosphorylation, and sensitizes HCC to sorafenib	138
DNAAF5	Enhances PFK1 stability by recruiting USP39	Upregulates PFKL, promotes glycolysis and cell proliferation in HCC cells, and induces sorafenib resistance	142
EGR1	Inhibits PFK1 transcription	Suppresses glycolysis, HCC growth, and sorafenib resistance	144
PFKB	Promotes the production of PFK1 agonist F2,6P	Inhibits HCC cell proliferation and sorafenib-induced apoptosis	146, 147
HDAC6	Inhibits acetylation of HSP90	Participates in sorafenib-induced glycolysis in HCC cells	151
miR-374b	Inhibits HNRNPA1 protein	Suppresses HCC glycolysis and sorafenib resistance	153
ERK1/2	Promotes nuclear localization of PKM2	Activates STAT3, β-catenin, and NF-κB signaling pathways, and promotes cell survival and drug resistance	156, 157
circUBE2D2	Acts as a miR-889-3p sponge to indirectly upregulate LDHA expression	Increases lactate levels in HCC cells and induces sorafenib resistance	159
PDK	Inhibits PDH	Suppresses oxidative phosphorylation in HCC, increases lactate production, and reduces cellular sensitivity to sorafenib	161
β-HB	Promotes PDH expression	Enhances the effect of sorafenib or regorafenib on inhibiting ERK signaling pathway and N-cadherin-vimentin proteins, and impairs HCC growth and migration	162
PDHB	Promotes transcription of SLC2A1, GPI, and PKM2	Enhances HCC glycolysis and sorafenib resistance	163
HIF-1α	Promotes transcription of GLUT1, HK2, PKM2, and LDHA	Facilitates glycolysis, tumor progression, and sorafenib resistance in HCC	167–169
RABIF	Attenuates mitophagy-induced ROS and HIF-1α	Inhibits the sensitivity of HCC cells to sorafenib	170
Gankyrin	Activates β-catenin signaling pathway, thereby upregulating c-MYC	Promotes HCC tumor growth and sorafenib resistance	173
ACYP1	Binds to HSP90 to enhance c-MYC stability	Facilitates HCC glycolysis and lenvatinib resistance	175
CPSF6	Enhances c-MYC stability	Promotes c-MYC-mediated glycolysis and sorafenib resistance, and increases VEGF expression and angiogenesis in HCC	176
ACTR	Promotes recruitment of c-MYC to glycolysis-related genes	Enhances HCC glycolysis, cell growth, and sorafenib resistance	177
T3	Inhibits PI3K/AKT signaling pathway, thereby preventing nuclear translocation of HIF-1α	Suppresses HCC glycolysis and lenvatinib resistance	181
HDAC11	Inhibits LKB1-mediated AMPK activation	Promotes HCC glycolysis and sorafenib resistance	183
BNIP3	Induces autophagy to activate AMPK	Promotes metabolic switch from oxidative phosphorylation to glycolysis in HCC cells and lenvatinib resistance	184
KDM6A	Upregulates EGFR4 expression, thereby activating PI3K/AKT/mTOR signaling	Promotes ACC1 and ACLY expression, increases fatty acid and lipid droplet formation, and leads to lenvatinib resistance	190, 191
AMPK	Inhibits ACC expression	Promotes fatty acid synthesis and sorafenib resistance	192
ACAT1	Inhibits fatty acid degradation through the GNPAT/TRIM21/FASN axis	Promotes HCC tumor growth and sorafenib resistance	193
TNFAIP8	Binds to oleic acid and upregulates ACC and FASN	Induces hepatic steatosis and promotes sorafenib and regorafenib resistance	195
SLP2	Promotes nuclear translocation of SREBP1 by stabilizing JNK2, thereby upregulating multiple lipid metabolism-related enzymes	Enhances HCC tumor growth and lenvatinib resistance	196
SPAG4	Increases SREBP1 expression and promotes its nuclear translocation	Facilitates HCC tumor growth and lenvatinib resistance	197
SLC27A4	Promotes uptake of MUFAs (monounsaturated fatty acids) by cancer cells	Inhibits ferroptosis in HCC cells and reduces sorafenib efficacy	201
HBXIP	Promotes SCD expression by co-activating transcription factor ZNF263	Inhibits ferroptosis in HCC cells and reduces sorafenib efficacy	202
URI	Inhibits SCD1 through the TRIM28/MDM2/p53 axis	Improves donafenib sensitivity in HCC	203
CREB	Regulates transcription factor CRTC3	Inhibits PUFAs production, reduces lipid peroxidation, and decreases sorafenib sensitivity	204
HNF4A-AS1	Promotes m6A modification of DECR1 mRNA	Inhibits PUFAs production, reduces lipid peroxidation, and decreases sorafenib sensitivity	206
DTX2	Inhibits HSD17B4-mediated peroxisomal β-oxidation	Inhibits PUFAs production, reduces lipid peroxidation, and decreases lenvatinib sensitivity	207
miR-23a-3p	Targets and inhibits ACSL4	Inhibits integration of PUFAs into cell membrane phospholipids, thereby reducing lipid peroxidation and sorafenib sensitivity	209
DGKH	activates the mTOR by enhancing PA synthesis	promotes HCC metastasis, stemness, and drug resistance	213
AGPAT4	Facilitates the conversion of LPA to PA	Enhances tumor growth, stemness and sorafenib resistance	215
RBM45	Promotes CPT1 expression through the PI3K/AKT/mTOR signaling pathway	Upregulates FAO and induces sorafenib resistance	218
LCAT1	Promotes CPT1 degradation	Inhibits triglyceride breakdown and FAO, leading to lenvatinib resistance	219
AKR1C3	Promotes LD accumulation	Enables HCC cells to adapt to sorafenib-induced FAO impairment and leads to sorafenib resistance	220
PPARα	Upregulates ACSL4 and FAO	Promotes cell proliferation and regorafenib resistance	222
LINC01056	Inhibits nuclear localization and transcriptional activity of PPARα	Suppresses FAO and restores sorafenib sensitivity	224
SREBP2	Inhibits cholesterol biosynthesis	Promotes HCC proliferation, cancer stem cell properties, and resistance to sorafenib and lenvatinib	229–231
LCAT	Impairs SREBP2 maturation by increasing HDL-C uptake	Inhibits HCC cholesterol biosynthesis and lenvatinib resistance	234
STARD4	Mediates mitochondrial cholesterol transport	Stabilizes mitochondrial membrane structure and promotes apoptosis resistance in HCC cells	232
ABCB1	Mediates drug efflux	Enhances resistance of HCC cells to lenvatinib	233
SPARC	Promotes stability of TRIM21	Activates PI3K-AKT signaling and induces EMT and sorafenib resistance	235
SCAP	Promotes cholesterol biosynthesis	Inhibits AMPK signaling-mediated autophagy and induces sorafenib resistance	226
SMS1	Promotes sphingomyelin synthesis	Increases sphingomyelin production and reduces ceramide content, thereby inhibiting apoptosis and inducing sorafenib resistance	241
miR-23b-3p	Targets and inhibits GLS1	Blocks glutamine metabolism and suppresses cell proliferation, invasion, migration, and sorafenib resistance	244
SMYD2	Upregulates GLS1 by stabilizing c-MYC	Promotes glutamine metabolism and sorafenib resistance	245
PPARδ	Promotes GLS1 and reductive glutamine carboxylation	Enhances sorafenib resistance in HCC	246
OGDHL	Inhibits α-KG: citrate ratio	Promotes glutamine-mediated lipid metabolism and sorafenib resistanc	247, 248
METTL3	Promotes m6A methylation of SSP genes	Promotes HCC growth and sorafenib and Lenvatinib resistance	251
IGF2BP3	Upregulates serine synthesis	Enhances antioxidant system and lenvatinib resistance	67
CFL1	Increases serine metabolism by promoting PHGDH	Promotes sorafenib resistance	252
CircRAPGEF1	Inhibits IGF2BP3-mediated degradation of ASS1 mRNA	Promotes proliferation, stemness and sorafenib	254
LINC01234	Inhibits the transcription of ASS1	Promotes cell proliferation, migration and sorafenib resistance	255

CC Hepatocellular carcinoma, LAT1 L-type amino acid transporter 1, GLUT1 Glucose transporter 1, HK2 Hexokinase 2, DNAAF5 Dynein axonemal assembly factor 5, USP39 Ubiquitin-specific protease 39, PFKL Phosphofructokinase liver type, EGR1 Early growth response 1, PFKB Phosphofructokinase kinase B, F2,6P Fructose-2,6-bisphosphate, HDAC6 Histone deacetylase 6, HSP90 Heat shock protein 90, HNRNPA1 Heterogeneous nuclear ribonucleoprotein A1, ERK1/2 Extracellular regulated protein kinases 1/2, STAT3 Signal transducer and activator of transcription 3, NF-κB Nuclear factor kappa-B, circUBE2D2 Circular RNA UBE2D2, LDHA Lactate dehydrogenase A, PDK Pyruvate dehydrogenase kinase, β-HB β-Hydroxybutyrate, PDHB Pyruvate dehydrogenase E1 β subunit, SLC2A1 Solute carrier family 2 member 1, GPI Glucose-6-phosphate isomerase, HIF-1α Hypoxia-inducible factor 1-alpha, RABIF RAB interacting factor, ROS Reactive oxygen species, ACYP1 Acylphosphatase 1, CPSF6 Cleavage and polyadenylation specificity factor 6, VEGF Vascular endothelial growth factor, ACTR Activator of thyroid and retinoic acid receptors, T3 Triiodothyronine, PI3K/AKT Phosphatidylinositol 3-kinase/Protein kinase B, mTOR Mammalian target of rapamycin, LKB1 Liver kinase B1, AMPK AMP-activated protein kinase, BNIP3 BCL2-interacting protein 3, KDM6A Lysine demethylase 6 A, EGFR4 Epidermal growth factor receptor 4, ACC1 Acetyl-CoA carboxylase 1, ACLY ATP citrate lyase, ACAT1 Acetyl-CoA acetyltransferase 1, GNPAT Glyceronephosphate O-acyltransferase, TRIM21 Tripartite motif containing 21, FASN Fatty acid synthase, SLP2 Stomatin-like protein 2, JNK2 c-Jun N-terminal kinase 2, SREBP1 Sterol regulatory element-binding protein 1, SPAG4 Sperm-associated antigen 4, SLC27A4 Solute carrier family 27 member 4, MUFAs Monounsaturated fatty acids, HBXIP Hepatitis B X-interacting protein, SCD Stearoyl-CoA desaturase, ZNF263 Zinc finger protein 263, URI Unconventional prefoldin RPB5 interactor, TRIM28 Tripartite motif containing 28, MDM2 Mouse double minute 2 homolog, CREB cAMP response element-binding protein, CRTC3 CREB-regulated transcription coactivator 3, PUFAs Polyunsaturated fatty acids, HNF4A-AS1 HNF4A antisense RNA 1, m6A N6-methyladenosine, DECR1 2,4-Dienoyl-CoA reductase 1, DTX2 Deltex 2, HSD17B4 Hydroxysteroid 17-β dehydrogenase 4, DHA Docosahexaenoic acid, ACSL4 Acyl-CoA synthetase long-chain family member 4, RBM45 RNA-binding motif protein 45, FAO Fatty acid oxidation, LCAT1 Lecithin-cholesterol acyltransferase 1, AKR1C3 Aldo–keto reductase family 1 member C3, LD Lipid droplet, LINC01056 Long intergenic non-protein coding RNA 1056, PPARα Peroxisome proliferator-activated receptor alpha, SREBP2 Sterol regulatory element-binding protein 2, LCAT Lecithin-cholesterol acyltransferase, HDL-C High-density lipoprotein cholesterol, STARD4 STAR-related lipid transfer domain containing 4, ABCB1 ATP-binding cassette subfamily B member 1, SPARC Secreted protein acidic and rich in cysteine, EMT Epithelial-mesenchymal transition, SCAP SREBP cleavage-activating protein, SMS1 Sphingomyelin synthase 1, GLS1 Glutaminase 1, SMYD2 SET and MYND domain containing 2, PPARδ Peroxisome proliferator-activated receptor delta, OGDHL Oxoglutarate dehydrogenase-like, α-KG α-Ketoglutarate, TNFAIP8 Tumor Necrosis Factor α-Induced Protein 8, DGKH Diacylglycerol kinase eta, AGPAT4 1-acylglycerol-3-phosphate O-acyltransferase 4, METTL3 Methyltransferase-like 3; CFL1, Cofilin 1

Glycolysis

Numerous studies have identified biomolecules linked to the upregulation of key glycolytic enzymes in HCC, suggesting that targeting these molecules may offer a critical strategy for regulating glycolysis. Additionally, activation of glycolysis-related signaling pathways further promotes the upregulation of glycolytic enzymes, thereby enhancing glycolytic flux.

Glycolysis-related enzymes

Aberrant expression and functional regulation of glycolytic enzymes play a central role in energy metabolic reprogramming and sorafenib resistance in hepatocellular carcinoma. This section provides a systematic review of the biological functions and regulatory mechanisms of key glycolytic enzymes, including glucose transporters GLUTs, HK2, PFK, PKM2, and LDHA (Fig. 2).

Fig. 2.

Key glycolytic enzymes and their regulatory mechanisms in HCC glycolysis. In HCC cells, glucose uptake is mediated by GLUT1. Lactate production is facilitated by HK2, PFK1, PKM2, and LDHA. PDH is downregulated in HCC cells, leading to a reduction in pyruvate entering the TCA cycle. These glycolytic enzymes are regulated by various factors

Glucose transporters (GLUTs) are high-affinity proteins located on the cell membrane that facilitate glucose entry into cells [127]. In cancer cells, the expression of GLUT1, GLUT3, and GLUT4 is significantly elevated compared to normal cells, promoting glycolysis to support tumor growth [128, 129]. Zhang et al. demonstrated that HCC exhibits increased glycolytic activity, with HCC cells preferentially absorbing glucose via GLUT1 and GLUT3. Further studies revealed that enhanced glucose uptake and glycolysis are associated with sorafenib resistance in HCC. Inhibition of GLUT1 and GLUT3 restored sorafenib sensitivity and improved its antitumor efficacy in vivo [130]. Histidine, an essential amino acid, was found to suppress GLUT1 expression, increasing HCC sensitivity to sorafenib. However, expression of histidine/large neutral amino acid transporter (LAT1), crucial for histidine uptake, is downregulated in HCC, contributing to sorafenib resistance. Overexpression of LAT1 enhanced sorafenib efficacy in HCC by promoting histidine uptake [131].

Hexokinase 2 (HK2), the most active isoenzyme in the HK family, is a primary rate-limiting enzyme in the first step of glycolysis [132]. HK2 is frequently overexpressed in various cancers, including HCC, where it catalyzes the phosphorylation of glucose to glucose-6-phosphate [133]. HK2-driven glycolysis, autophagy, and immunosuppression are critical in mediating drug resistance in cancer [134–136]. In patients with HCC, HK2 expression is significantly higher in tumor tissues compared to normal tissues and correlates with poor survival. Inhibition of HK2 reduces glucose consumption and cell viability in HCC cells, while enhancing sensitivity to sorafenib [137]. miR-199a directly targets the 3′ UTR of HK2 mRNA, suppressing HK2 expression. Overexpression of miR-199a disrupts HK2-mediated glycolysis and promotes OXPHOS, sensitizing HCC cells to sorafenib both in vitro and in vivo. These effects can be reversed by silencing miR-199 [138]. In mouse models, the incidence of liver cancer was significantly lower in HK2-knockout mice compared to those with high HK2 expression. HK2 silencing inhibited HCC cell proliferation and induced apoptosis. Furthermore, although HK2 knockdown reduced pyruvate and lactate production, the tricarboxylic acid (TCA) cycle within the OXPHOS pathway remained functional. Treatment with metformin further suppressed the growth of HK2-silenced HCC cells. Notably, HK2 silencing enhanced sorafenib sensitivity and synergized with sorafenib to inhibit HCC growth [139].

PFK, located in the cytoplasm, is a key enzyme in the glycolytic pathway. PFK1 catalyzes the conversion of fructose-6-phosphate (F6P) to fructose-1,6-bisphosphate (F1,6BP), a rate-determining step in glycolysis [140]. In HCC, PFK1 is upregulated and positively correlates with the glycolytic rate. Inhibition of PFK1 activity in HCC cells effectively reduces cell proliferation and increases the expression of apoptosis-related proteins, caspase-3 and Bax. Furthermore, PFK1 inhibition enhances HCC sensitivity to sorafenib both in vivo and in vitro [141]. PFK1 is also regulated by several biomolecules. Dynein axonemal assembly factor 5 (DNAAF5), a transcription factor involved in cytoskeletal and hydrodynamic protein complex assembly, recruits the deubiquitinating enzyme USP39, which stabilizes PFK liver type (PFKL). Overexpression of DNAAF5 promotes glycolysis, cell proliferation, and sorafenib resistance in HCC cells by indirectly upregulating PFKL. Conversely, silencing DNAAF5 reduces cell proliferation and enhances HCC sensitivity to sorafenib [142]. Early growth response 1 (EGR1) is a transcription factor that suppresses HCC development [143]. A recent study using a mouse HCC model and an in vitro human HCC organoid model investigated the role of EGR1 in HCC. The study showed that EGR1 significantly reduces the expression of PFKL, the liver-type isoenzyme subunit of PFK1, thereby inhibiting HCC growth and overcoming sorafenib resistance. Mechanistically, EGR1 binds to the promoter region of PFKL and represses its transcription, leading to a reduction in glycolysis in HCC cells [144].

PFKB, an isoenzyme of PFK2, catalyzes the conversion of fructose-6-phosphate to fructose-2,6-bisphosphate (F2,6BP), a potent activator of PFK1. Therefore, PFKB enhances PFK1 activity and glycolysis by increasing F2,6BP production [145]. A study examining the relationship between PFKB3 and sorafenib resistance in HCC cells and a subcutaneous xenograft model in nude mice revealed that high PFKB3 expression and glycolysis contribute to sorafenib resistance. Inhibition of PFKB3 promoted sorafenib-induced apoptosis [146]. Similarly, PFKB4 is upregulated in HCC cells and functions as a hypoxia-inducible gene that enhances HCC invasiveness and sorafenib resistance. Knocking out PFKB4 using the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) system significantly reduced HCC cell proliferation and restored sensitivity to sorafenib [147].

PKM2 is the final rate-limiting enzyme in the aerobic glycolysis pathway of cancer cells, catalyzing the conversion of pyruvate and ATP [148]. PKM2, an isoform of PKM, is highly expressed in cancer cells, while PKM1 is predominantly found in normal tissues. Both PKM1 and PKM2 arise from alternative splicing of PKM pre-mRNA, with their relative expression ratio critically regulating glycolysis [149]. Clinical data indicate that elevated PKM2 expression is associated with poor prognosis and reduced treatment efficacy in patients with HCC [150]. Emerging evidence suggests that upregulation of PKM2 in HCC is linked to sorafenib resistance. Under sorafenib treatment, the autophagy process degrades the p62 protein and enhances the deacetylase activity of HDAC6, preventing the inactivation of HSP90 due to hyperacetylation. HSP90, involved in protein folding and maturation, further activates PKM2 [151]. Additionally, heterogeneous nuclear ribonucleoprotein A1 (HNRNPA1), an RNA-binding protein, induces the alternative splicing of PKM2 [152]. Zhang et al. showed that sorafenib treatment upregulates both HNRNPA1 and PKM2 in HCC cells. miR-374b directly targets the 3′ untranslated region (3′UTR) of HNRNPA1, suppressing the alternative splicing of PKM pre-mRNA into PKM2, thereby inhibiting glycolysis and sorafenib resistance in HCC [153]. Treatment with PKM2 siRNA reduced HCC cell proliferation and upregulated dual-specificity phosphatase 2 (DUSP2), which dephosphorylates ERK1/2, a direct effector of the MAPK pathway. The inhibition of ERK1/2 phosphorylation by PKM2 siRNA synergized with lenvatinib to reverse lenvatinib resistance in HCC [154]. Notably, HCC cell proliferation is PKM2-dependent; inhibition of PKM2 alone suppressed cell proliferation, whereas overexpression of PKM1 had no effect [154]. A positive feedback loop exists between ERK1/2 and PKM2, where ERK1/2 promotes the non-metabolic functions of PKM2 by facilitating its nuclear translocation [155]. Specifically, ERK1/2 directly phosphorylates PKM2 at Ser37, causing its dissociation from tetramers into dimers and exposing its nuclear localization signal, thereby promoting its transport to the nucleus [156]. In the nucleus, PKM2 acts as a transcriptional coactivator for STAT3, β-catenin, and NF-κB, and the activation of these signaling pathways enhances drug resistance in cancer cells [157]. Protein arginine N-methyltransferase 6 (PRMT6), upregulated under hypoxic conditions, methylates CRAF to inhibit its binding to RAS/RAF, suppressing the ERK1/2-mediated nuclear translocation of PKM2 [156]. Inhibiting PKM2 nuclear translocation can suppress cell proliferation, promote apoptosis, and enhance sensitivity to sorafenib [158].

LDHA catalyzes the conversion of pyruvate to lactate in the cytoplasm [47]. Intracellular lactate in HCC cells serves as a fuel source or substrate for lactylation modifications, supporting cell survival and proliferation. Extracellular lactate functions as a signaling molecule, promoting angiogenesis and suppressing the tumor immune microenvironment [57]. The circular RNA circUBE2D2 is upregulated in HCC cells, promoting tumor progression and sorafenib resistance. Mechanistically, circUBE2D2 acts as a sponge for miR-889-3p, which targets and inhibits LDHA mRNA. By sequestering miR-889-3p, circUBE2D2 indirectly upregulates LDHA, increasing lactate production and enhancing sorafenib resistance in HCC cells [159]. In addition to lactate production, pyruvate can be irreversibly decarboxylated to acetyl-CoA by PDH and enter the mitochondria to participate in the TCA cycle. However, cancer cells preferentially convert pyruvate to lactate rather than utilizing OXPHOS, resulting in low PDH activity [46]. Pyruvate dehydrogenase kinase (PDK), an inhibitor of PDH, upregulates glycolysis under hypoxic conditions by suppressing PDH activity [160]. Shen et al. treated HCC cells with the PDK inhibitor dichloroacetate (DCA), which restored PDH activity, significantly enhanced OXPHOS, and reduced lactate production. This enhancement of OXPHOS sensitized the cells to sorafenib-induced oxidative stress, highlighting the critical role of PDH-mediated glycolysis/OXPHOS balance in sorafenib efficacy [161]. Low levels of the ketone body D-β-hydroxybutyrate (β-HB) in HCC cells correlate with enhanced glycolysis and lactate production. Supplementation with β-HB promotes PDH expression and inhibits glycolysis in HCC cells. Moreover, increased PDH activity potentiates the inhibitory effects of sorafenib or regorafenib on the ERK signaling pathway and N-cadherin–vimentin protein expression, thereby hindering HCC growth and migration [162]. Notably, PDHB, a key component of the E1 subunit of PDH, can also promote tumorigenesis independently of PDH. By binding to the promoter regions of SLC2A1, GPI, and PKM2, PDHB promotes the transcription of glycolytic genes and induces sorafenib resistance in HCC [163].

Glycolysis-related signaling pathway

In addition to the dysregulation of key enzyme expression, the aberrant activation of glycolysis is tightly regulated by multiple signaling pathways (Fig. 3). Extensive studies have confirmed that activation of HIF-1α, c-MYC, PI3K/AKT, and AMPK serves as a critical driver of glycolysis in cancer [164, 165]. As a solid tumor, HCC often develops under hypoxic conditions. Under hypoxia, hypoxia-inducible factor 1-alpha (HIF-1α) is upregulated and mediates metabolic reprogramming in cancer cells [166]. HIF-1α, a transcription factor, induces the expression of multiple glycolysis-related genes, including GLUT1, HK2, PKM2, and LDHA [167, 168]. In HCC, elevated HIF-1α levels are associated with tumor progression, enhanced glycolysis, and resistance to sorafenib. Knockdown of HIF-1α suppresses the expression of HK2 and GLUT1, restoring sorafenib efficacy in sorafenib-resistant HCC cells [169]. RAB-interacting factor (RABIF), a stabilizer of the pro-oncogenic RAB protein, is upregulated in HCC. Recent studies demonstrated that inhibition of RABIF attenuates the STOML2–PARL–PGAM5 axis-mediated mitophagy, reducing mitochondrial ROS production and subsequently downregulating oxidative stress-induced HIF-1α expression. The decreased HIF-1α expression resulting from RABIF inhibition suppresses the expression of HK1, HKDC1, and LDHB, ultimately enhancing sorafenib sensitivity in HCC cells [170].

Fig. 3.

Signal pathways involved in glycolysis in HCC. The PI3K/AKT, AMPK, c-MYC, and HIF-1α pathways are dysregulated in HCC, promoting the transcription HK2, GLUT1, PKM2, and LDHA. Crosstalk among these signaling pathways forms a complex network that mediates glycolysis in HCC and contributes to HCC cell resistance to MKIs

c-MYC, a core member of the MYC family, is upregulated in various cancers and promotes both cancer proliferation and metabolism [171]. Like HIF-1α, c-MYC functions as a transcription factor, enhancing the expression of glycolysis-related genes such as LDHA, PKM2, and HK2, thereby amplifying glycolytic activity [171, 172]. Gankyrin, an oncoprotein upregulated in HCC, indirectly promotes c-MYC expression by activating β-catenin signaling, leading to the upregulation of HK2, GLUT1, LDHA, and PKM2. Inhibition of c-MYC suppresses tumor growth in HCC with high Gankyrin levels and enhances the efficacy of sorafenib and regorafenib [173]. Another study similarly demonstrated that c-MYC inhibition reduces glucose consumption and lactate release in HCC, thereby enhancing the anti-HCC effects of sorafenib both in vivo and in vitro [174]. Acylphosphatase 1 (ACYP1), which is upregulated in HCC, increases LDHA expression and glycolysis, and is strongly associated with lenvatinib resistance. Overexpression of ACYP1 promotes the proliferation, invasion, and migration of HCC cells. Mechanistically, ACYP1 binds to HSP90, enhancing the stability of c-MYC and promoting glycolysis and lenvatinib resistance [175]. Cleavage and polyadenylation specificity factor 6 (CPSF6) is another stabilizer of c-MYC. CPSF6 binds to c-MYC at sites 258–360, facilitating c-MYC-mediated glycolysis and enhancing VEGF expression and angiogenesis in HCC. Depletion of CPSF6 induces c-MYC degradation, downregulates glycolysis and lactate production, and exerts a synergistic antitumor effect with sorafenib [176]. Furthermore, the activator of thyroid and retinoic acid receptors (ACTR) promotes the recruitment of c-MYC to glycolysis-related genes, significantly upregulating glycolysis. Knockdown of ACTR in HCC cells inhibits glycolysis and reverses sorafenib resistance [177].

The PI3K/AKT signaling pathway is a critical pro-oncogenic pathway that promotes cancer cell survival and metabolism [178]. It directly facilitates the phosphorylation of key glycolytic enzymes, such as HK2 and PFKFB. Additionally, PI3K/AKT regulates the expression of transcription factors like MYC and HIF-1α, further enhancing the expression of glycolysis-related genes [179, 180]. Thyroid function is linked to HCC progression, with decreased triiodothyronine (T3) levels observed in HCC. T3 supplementation has been shown to inhibit the PI3K/AKT pathway in HCC, thereby hindering mTOR-mediated nuclear translocation of HIF-1α. This process suppresses ENO2-induced glycolysis and lenvatinib resistance [181].

AMPK is a key regulator of energy metabolism in cancer and plays a significant role in HCC initiation, progression, treatment, and prognosis [182]. Research on AMPK’s role in HCC glycolysis remains controversial. In HCC cells, the overexpression of histone deacetylase 11 (HDAC11) reduces sensitivity to sorafenib. HDAC11 inhibits acetylation at the LKB1 promoter region, decreasing LKB1 expression. LKB1 is a key activator of AMPK, and by suppressing the LKB1/AMPK axis, HDAC11 promotes glycolysis in HCC [183]. Conversely, in lenvatinib-resistant HCC cells, BCL2-interacting protein 3 (BNIP3) induces mitophagy, activating AMPK/ENO2 signaling. This metabolic shift from OXPHOS to glycolysis contributes to lenvatinib resistance [184]. The dual role of AMPK in glycolysis may be influenced by the complex TME. For instance, AMPK inhibits glycolysis by suppressing the mTOR pathway [185]. However, under hypoxic conditions, AMPK promotes the stabilization and transactivation of HIF-1α, inducing the expression of glycolysis-related genes [186]. Notably, a study by Guo et al. found that in HCC, the duration of AMPK activation is dependent on cellular glycolysis levels. Inhibition of glycolysis shortens AMPK activation time [187]. Thus, whether a feedback mechanism exists between AMPK and glycolysis warrants further investigation.

Lipid metabolism

Fatty acid metabolism

Fatty Acid Synthesis

FA levels in HCC are critical for the response to pharmacological treatment. A study analyzing FA degradation (FAD)-related genes through transcriptomics and lipidomics in 41 patients with HCC found that low FAD gene expression correlated with better responses to sorafenib, while high FAD expression was associated with poor sorafenib efficacy [188]. This highlights the significant role of FA metabolism in mediating MKI resistance in HCC. Increasing evidence suggests that targeting FA metabolism, particularly de novo FA synthesis, can impact HCC growth, metastasis, and drug resistance. ACLY, a key rate-limiting enzyme in FA synthesis, is abnormally overexpressed in HCC and closely linked to sorafenib resistance. Clinical data indicate that patients with HCC exhibiting high ACLY expression have a significantly reduced response to sorafenib. Further research has indicated that lipid metabolism is more active in sorafenib-resistant HCC cells, and ACLY knockout enhances sorafenib sensitivity both in vivo and in vitro [189]. KDM6A, which is upregulated in HCC tissues and associated with poor prognosis, promotes FGFR4 expression, activating the PI3K/AKT/mTOR signaling pathway. This activation induces lipid metabolic reprogramming and contributes to lenvatinib resistance in HCC [190]. Mechanistically, activation of the PI3K/AKT pathway increases the expression of ACC1 and ACLY, thereby promoting FA synthesis and lipid droplet (LD) formation, which in turn contributes to HCC drug resistance [191]. ACC is another key rate-limiting enzyme in FA biosynthesis. Lally et al. demonstrated that mutations in the AMPK phosphorylation sites of ACC1 and ACC2 prevented their inhibition by the AMPK signaling pathway, leading to enhanced FA synthesis and liver lesions in mice. Similarly, mutations in ACC1 in HCC cells increased FA synthesis and promoted sorafenib resistance [192].

FASN is upregulated in HCC cells and contributes to enhanced FA metabolism in CSCs. Recent studies have demonstrated that FASN inhibition reduces HCC cell proliferation, invasion, and migration, while attenuating CSC-like properties. Notably, FASN inhibition increases HCC sensitivity to sorafenib [193]. ACAT1 is upregulated in HCC cells under PA stimulation, where it acetylates GNPAT at lysine 128, thereby enhancing GNPAT expression. GNPAT inhibits TRIM21, a protein that mediates the degradation of FASN. Through the GNPAT/TRIM21/FASN axis, ACAT1 promotes FA synthesis in HCC and induces sorafenib resistance. Combined inhibition of ACAT1 and sorafenib significantly delays tumor formation in mice [193]. Furthermore, FASN can directly interact with HIF-1α to facilitate its nuclear translocation, protecting HIF-1α from ubiquitination and proteasomal degradation. This interaction promotes glycolysis in HCC cells and contributes to resistance to sorafenib-induced ferroptosis [194]. Tumor Necrosis Factor α-Induced Protein 8 (TNFAIP8) is upregulated in HCC and promotes autophagy-mediated survival of HCC cells by interacting with the ATG3-ATG7 protein complex. TNFAIP8 specifically binds to oleic acid (OA), facilitating LD formation and hepatic steatosis. Additionally, TNFAIP8 is associated with the expression of multiple fatty acid metabolism-related enzymes, and inhibition of TNFAIP8 results in downregulation of ACC and FASN. Furthermore, TNFAIP8 regulates ethanol-induced hepatic steatosis and confers HCC resistance to sorafenib and regorafenib [195].

SREBP1 is a critical regulator of lipid metabolism in HCC, driving the expression of enzymes such as ACC, FASN, and SCD1. The expression of SREBP1 in patients with HCC correlates with tumor size, distant metastasis, and OS. Inhibition of SREBP1 expression has been shown to reverse lenvatinib resistance in HCC cells [196]. SREBP1 is regulated by the SLP2/JAK2 axis. SLP2 binds to the C-terminal region of JNK2, affecting its ubiquitin–proteasome degradation pathway and stabilizing JNK2. JAK2 then promotes the translocation of SREBP1 into the nucleus, facilitating de novo lipogenesis [196]. Additionally, the N-terminal region of sperm-associated antigen 4 (SPAG4) binds to lamin A/C, which increases SREBP1 expression and promotes its nuclear translocation and transcriptional activity. This enhances lipogenesis and contributes to lenvatinib resistance in HCC [197].

The ratio of PUFAs/MUFAs and lipid peroxidation

Induction of ferroptosis serves as a key mechanism by which MKI eliminates HCC cells. Ferroptosis is characterized by the accumulation of Fe²⁺, which catalyzes ROS generation and promotes lipid peroxidation, leading to membrane degradation. Inhibition of lipid peroxidation can block erastin-induced ferroptosis in HCC [198, 199]. PUFAs are primary targets of lipid peroxidation, and increasing MUFA levels, rather than PUFAs, enhances cancer cell resistance to ferroptosis. Solute carrier family 27 member 4 (SLC27A4) is significantly upregulated in HCC tumor tissues and associated with FA metabolism. SLC27A4 can be upregulated via the toll-like receptor 4/nuclear factor (TLR4)//NF-κB pathway mediated by nicotinamide phosphoribosyltransferase (NAMPT) [200]. It promotes MUFA uptake in cancer cells, leading to high levels of MUFA-containing phosphatidylcholines and phosphatidylethanolamines in HCC. Knockdown of SLC27A4 sensitizes cancer cells to ferroptosis, enhancing sorafenib efficacy against HCC [201]. The oncoprotein hepatitis B X-interacting protein (HBXIP) is upregulated in HCC and impedes sorafenib-induced ferroptosis. HBXIP promotes SCD expression by co-activating the transcription factor ZNF263, increasing MUFA levels to suppress ferroptosis. Inhibition of HBXIP enhances sorafenib sensitivity and increases malondialdehyde (MDA) production and GSH depletion [202]. The unconventional prefoldin RPB5 interactor (URI) is associated with high SCD1 expression in donafenib-resistant HCC cells through inhibition of p53. URI interacts with TRIM28 and prevents p53 degradation via the TRIM28/MDM2 axis, as p53 represses SCD1 transcription. Inhibition of SCD1 increases donafenib sensitivity in organoids from patients with p53 wild-type HCC and in xenograft models in nude mice [203]. Conversely, enzymes that promote PUFA production are inhibited in sorafenib-resistant HCC. The cAMP response element-binding protein (CREB) suppresses PUFA synthesis by regulating the transcriptional co-activator 3 (CRTC3), contributing to sorafenib resistance. Knockdown of CRTC3 increases PUFA abundance, promoting lipid peroxidation and enhancing sorafenib sensitivity [204]. 2,4-Dienoyl-CoA Reductase 1 (DECR1) suppresses PUFA accumulation in cancer [205]. Zhao et al. reported that lncRNA HNF4A-AS1, which interacts with METTL3, is downregulated in HCC. Overexpression of HNF4A-AS1 promotes m6A modification of DECR1 mRNA, leading to YTHDF3-mediated degradation of DECR1 and increased PUFA levels, thereby contributing to sorafenib resistance [206]. The activated JAK/STAT3 pathway in HCC upregulates expression of the E3 ubiquitin ligase Deltex 2 (DTX2). DTX2 ubiquitinates the SCP domain of HSD17B4 at lysine 645, resulting in the degradation of HSD17B4, a peroxisomal β-oxidation enzyme. By inhibiting HSD17B4-mediated β-oxidation, DTX2 downregulates docosahexaenoic acid (DHA)-containing PUFA expression, conferring lenvatinib resistance. Knockdown of DTX2 or DHA supplementation reverses lenvatinib resistance [207]. ACSL4 promotes cellular lipid peroxidation by incorporating PUFAs into membrane phospholipids [208]. miR-23a-3p is significantly upregulated in HCC and confers resistance to ferroptosis by binding to the 3′ UTR of ACSL4 mRNA and suppressing its expression. Inhibition of miR-23a-3p restores ACSL4 expression and rescues sorafenib efficacy in HCC cells [209].

PA

In addition to its role in glycerophospholipid synthesis, PA serves as a crucial signaling lipid within cells. Activation of phospholipase D (PLD), which promotes PA production, facilitates the nuclear translocation of Yes-associated protein (YAP), thereby conferring drug resistance to cancer cells [210]. Furthermore, PA can stimulate the activation of the mTOR signaling pathway by modulating growth factors and nutrients, such as amino acids [211, 212]. A recent study has revealed that diacylglycerol kinase eta (DGKH) is transcriptionally activated by E1A binding protein p300 (EP300) and is enriched in poorly differentiated HCC. Upregulated DGKH activates the mTOR signaling pathway by enhancing PA synthesis, thereby promoting HCC metastasis, stemness, and drug resistance [213]. Lysophosphatidic acid (LPA) refers to a small phospholipid containing a single fatty acid substituent attached to a glycerol backbone [214]. Ma et al. identified a metabolic protein named 1-acylglycerol-3-phosphate O-acyltransferase 4 (AGPAT4), which facilitates the conversion of LPA to PA. This process activates the mTOR/S6K/S6 signaling pathway, thereby enhancing tumor growth and stemness in HCC and conferring sorafenib resistance to HCC cells. Targeted inhibition of AGPAT4 expression in the liver via AAV8 virus restores sensitivity to sorafenib [215].

FAO

During energy deprivation, FAO serves as a crucial pathway for energy supply in cancer cells. Thus, in addition to contributing to the synthesis of other lipids, FAs accumulated in cancer cells are extensively directed into the FAO pathway. In HCC, OXPHOS and ROS production are suppressed, while FAO is upregulated. This process is regulated by NANOG, which is activated through Toll-like receptor 4 (TLR4) signaling. Furthermore, FAO supports the self-renewal of tumor-initiating stem-like cells (TICs) and contributes to resistance to sorafenib [216].

Carnitine palmitoyltransferase 1 (CPT1), located on the outer mitochondrial membrane, mediates the entry of FAs into the mitochondria and acts as a key rate-limiting enzyme in FAO [217]. The expression of RNA-binding motif protein 45 (RBM45) is positively correlated with body mass index in patients with HCC. In response to lipid accumulation, RBM45 activates the PI3K/AKT/mTOR signaling pathway, which in turn activates CPT1, upregulating FAO and contributing to sorafenib resistance in HCC [218]. Conversely, lecithin–cholesterol acyltransferase 1 (LCAT1) interacts with caveolin 1 (CAV1) and CPT1, inhibiting CAV1-mediated PRKACA phosphorylation and promoting the ubiquitination and degradation of CPT1. In HCC, LCAT1 deficiency enhances triglyceride breakdown and upregulates FAO, resulting in lenvatinib resistance [219]. Long-term sorafenib treatment suppresses FAO in HCC cells, which is associated with ROS generation and ferroptosis. Additionally, AKR1C3-dependent LD accumulation regulates lipid metabolism to adapt to sorafenib-induced impairment of FAO, thereby contributing to sorafenib resistance [220]. Interestingly, sorafenib treatment also promotes the expansion of myeloid-derived suppressor cells (MDSCs) by upregulating peroxisome proliferator-activated receptor α (PPARα), contributing to an immunosuppressive TME. Inhibition of PPARα attenuates the effect of sorafenib on MDSCs and restores HCC sensitivity to sorafenib [221]. Furthermore, PPARα binds to retinoicic acid X receptors (RXRs) to form a complex, which is associated with the upregulation of ACSL4 expression. ACSL4 promotes FAO in HCC cells by participating in the conversion of long-chain unsaturated fatty acids to Acetyl-CoA. Inhibition of ACSL4 reduces HCC cell proliferation by blocking FAO and enhances the sensitivity of HCC to regorafenib [222, 223]. LncRNA LINC01056 is significantly downregulated in sorafenib-resistant HCC and is associated with enhanced FAO. Overexpression of LINC01056 suppresses FAO and restores sorafenib sensitivity in HCC. Mechanistically, LINC01056 directly interacts with PPARα, inhibiting its nuclear localization and transcriptional activity, thus reducing the transcription of FAO-related genes [224]. Notably, one study found that nuclear accumulation of SREBP1 in HCC cells negatively regulates the transcription of medium-chain acyl-CoA dehydrogenase (ACADM), thereby suppressing ACADM-mediated FAO. These findings highlight a complex regulatory network in lipid metabolism within HCC cells, maintaining a balance between FA synthesis and FAO flux [225].

Cholesterol metabolism

Analysis of a cohort of 78 patients with HCC treated with sorafenib following hepatectomy revealed that hypercholesterolemia is associated with MKI resistance in patients with HCC [226]. The intracellular free cholesterol level in HCC cells has been proposed as a potential biomarker for MKI resistance [227]. SREBP2, a key regulator of cholesterol metabolism [228], is cleaved in the endoplasmic reticulum. Upregulated SREBP2 promotes cholesterol biosynthesis, contributing to MKI resistance in HCC. Elevated cholesterol levels also facilitate HCC proliferation and maintain stemness [229]. Inhibition of SREBP2 reduces cholesterol synthesis in HCC cells, suppressing tumor growth and restoring sensitivity to sorafenib and lenvatinib [230, 231]. Furthermore, SREBP2 enhances the transcription of STARD4 (STAR-related lipid transfer domain containing 4), which increases STARD4-mediated mitochondrial cholesterol transport. Cholesterol within mitochondria stabilizes mitochondrial membrane structure, inhibits mitochondrial permeability transition pore (mPTP) opening, and reduces the release of mitochondrial components (such as cytochrome c), thereby conferring resistance to apoptosis [232]. Dysregulated cholesterol metabolism is also linked to the expression of drug efflux transporters. In lenvatinib-resistant HCC cells, ABCB1 expression is upregulated, significantly increasing lenvatinib efflux. This upregulation is dependent on cholesterol metabolism activation and lipid raft formation [233]. Lecithin–cholesterol acyltransferase (LCAT), an upstream regulator of SREBP2, is upregulated by estrogen and promotes high-density lipoprotein cholesterol (HDL-C) uptake by increasing LDLR and SCARB1 expression. HDL-C accumulation impairs SREBP2 maturation, inhibiting SREBP2-mediated cholesterol synthesis and lenvatinib resistance [234].

Substantial evidence supports the role of cholesterol as a key mediator in signal transduction within HCC cells. High expression of SPARC (secreted protein acidic and rich in cysteine) enhances TRIM21 stability through interaction with its E3 ligase activity, promoting cholesterol accumulation. Elevated cholesterol levels activate the PI3K/AKT signaling pathway, inducing epithelial-mesenchymal transition (EMT) and sorafenib resistance [235]. Inhibition of SREBP2-mediated cholesterol synthesis suppresses both the PI3K/AKT and STAT3 pathways, inducing apoptosis and cell cycle arrest in HCC cells [236]. Sorafenib treatment activates AMPK, which, in conjunction with CCAAT/enhancer-binding protein delta (CEBPD), induces excessive autophagic cell death in HCC cells [237]. The cholesterol sensor SCAP is significantly upregulated in HCC, promoting proliferation, invasion, and migration of HCC cells. Upregulated SCAP, accompanied by intracellular cholesterol deposition, inhibits AMPK pathway activation, a key regulator of autophagy. By promoting cholesterol accumulation, SCAP inhibits AMPK-mediated autophagy and contributes to sorafenib resistance [226]. Cholesterol metabolism also plays a critical role in the crosstalk between HCC and the TME. Sema3C upregulates HMGCR expression in hepatic stellate cells (HSCs) via the NRP1/ITGB1–NF-κB axis, promoting cholesterol biosynthesis and enhancing TGF-β1 sensitivity. This process drives cancer-associated fibroblast (CAF) transformation and extracellular matrix (ECM) remodeling, supporting CSC stemness. Inhibition of Sema3C downregulates HMGCR-mediated cholesterol synthesis, reverses CAF-dependent drug resistance, and enhances sorafenib efficacy [238].

Sphingolipid metabolism

Sphingolipids are essential components of the HCC cell membrane and function as autocrine and paracrine mediators. The balance between sphingomyelin and Cer metabolism plays a pivotal role in regulating HCC growth and drug resistance [108]. Sphingomyelin is hydrolyzed by sphingomyelinase to produce Cer, and the accumulation of Cer disrupts mitochondrial function, causing oxidative stress and cell death in HCC cells [239]. Sorafenib treatment in HCC cell lines upregulates dihydroceramide, which mediates cell death. In contrast, inhibiting dihydroceramide promotes drug resistance without affecting tumor growth. These findings underscore the critical role of sphingolipid metabolism in regulating HCC sensitivity to MKIs [240]. Another study revealed that sphingomyelin synthase 1 (SMS1) catalyzes the conversion of Cer to sphingomyelin. Sorafenib treatment upregulates SMS1 expression in HCC, promoting tumor growth and migration while reducing sorafenib toxicity. Inhibition of SMS1 suppresses sphingomyelin synthesis, increases Cer levels, promotes apoptosis, and enhances sorafenib sensitivity in HCC cells [241].

In summary, lipid metabolic reprogramming is a key mechanism underlying the resistance of HCC to MKIs such as sorafenib and lenvatinib. This process primarily involves fatty acid, cholesterol, and sphingolipid metabolic pathways. Aberrant key molecules in lipid metabolism collectively constitute a metabolic resistance network that enables HCC cells to evade the therapeutic effects of MKIs (Fig. 4).

Fig. 4.

FA and cholesterol metabolism in HCC. HCC cells exhibit increased FA uptake and de novo FA synthesis. FAs are converted into acetyl-CoA under the mediation of ACSL1/4 and CPT1, which then enters the TCA cycle. Conversely, citrate derived from the TCA cycle is transported from the mitochondria to the cytoplasm, where it is converted into acetyl-CoA to participate in fatty acid and cholesterol synthesis. In MKI-resistant HCC cells, the synthesis of PUFAs is reduced, while SCD1-mediated production of MUFAs is enhanced, collectively promoting resistance to ferroptosis in HCC cells. Additionally, SREBP-2-mediated cholesterol synthesis increases lipid rafts and ABC transporters on the cell membrane, thereby facilitating drug efflux and activation of drug resistance signaling pathways

Amino acid metabolism

Glutamine metabolism

Glutamine (Gln) metabolism plays a pivotal role in HCC progression and MKI resistance (Fig. 5). A retrospective study involving 431 patients with HCC found that patients with GS expression had poorer OS, more aggressive tumor features, and activation of tumor-promoting genes, compared to patients with GS-negative HCC, who showed better responses to sorafenib [242]. The Gln/α-KG metabolic pathway is central to Gln metabolism. GLS1/2 catalyzes the conversion of Gln to glutamate, which is further converted by glutamate dehydrogenase (GDH1) to α-KG and ammonia, providing essential nutrients for tumor growth [243]. In sorafenib-resistant HCC cells, downregulation of miR-23b-3p promotes cell proliferation, invasion, migration, and sorafenib resistance. MiR-23b-3p binds to the 3′ UTR of GLS1 mRNA, inhibiting GLS1 expression and hindering Gln metabolism [244]. In contrast, SET and MYND domain-containing protein 2 (SMYD2), a methyltransferase, promotes Gln metabolism by stabilizing c-Myc and upregulating GLS1 expression. Knocking down SMYD2 restores sorafenib sensitivity in HCC cells by inhibiting GLS1 [245]. The α-KG produced by Gln metabolism enters the TCA cycle to generate ATP and also promotes lipid metabolism in HCC. PPARδ is upregulated and targets GLS in sorafenib-resistant HCC cells, promoting reduced Gln carboxylation. This reduction in Gln carboxylation produces isocitrate through the α-KG-derived pathway, providing a substrate for FA synthesis [246]. Similarly, a study by Dai et al. showed that downregulation of the mitochondrial multi-enzyme OGDH complex (OGDHL) increased the α-KG: citric acid ratio, promoting lipid metabolism and sorafenib resistance. Furthermore, α-KG can activate the mTOR signaling pathway, which enhances the transcription of key enzymes involved in de novo lipid synthesis [247, 248]. Additionally, Gln serves as a precursor for synthesizing GSH, a critical factor in maintaining cellular redox balance. Inhibition of GS reduces Gln synthesis, significantly lowering GSH production, which induces oxidative stress in HCC cells and increases sensitivity to lenvatinib-induced cell death [249].

Fig. 5.

Glutamine metabolism in HCC. The synthesis and uptake of glutamine are increased in HCC cells. The accumulated glutamine is catalyzed by GLS to produce glutamate, which participates in GSH synthesis and the TCA cycle. This process provides energy for HCC, balances oxidative stress, and upregulates lipid metabolism

Serine metabolism

Previously, a study identified PHGDH in the serine synthesis pathway (SSP) as a key factor regulating resistance to MKIs. Sorafenib treatment induces HCC cells to activate the SSP by upregulating PHGDH, thereby counteracting the toxicity of sorafenib. Conversely, knocking out PHGDH blocks the SSP, reduces the production of α‑KG, serine, and NADPH, and consequently increases ROS levels and sensitivity to sorafenib [250]. Subsequent research elucidated the role of post‑transcriptional modifications in SSP activation. Chen et al. revealed that MKI‑resistant HCC cells exhibit enhanced N6‑methyladenosine (m6A) methylation of mRNAs encoding crucial SSP enzymes, including PHGDH, phosphoserine aminotransferase 1 (PSAT1), and phosphoserine phosphatase (PSPH). Mechanistically, METTL3 mediates the m6A modification of these SSP‑related genes, while IGF2BP3, functioning as an m6A‑specific reader, inhibits the degradation of their mRNAs. The upregulation of these SSP enzymes promotes serine synthesis, leading to increased protein expression and antioxidant production. This study also investigated the effects of a highly specific m6A inhibitor (STM2457) on HCC driven by Tp53 knockout and Myc overexpression in mice, as well as on MKI‑resistant HCC cells. The results demonstrated that STM2457 significantly suppresses HCC growth and restores sensitivity to sorafenib and lenvatinib by inhibiting the m6A‑mediated activation of the SSP [251]. Another study found that lactate accumulation‑induced lactylation of IGF2BP3 upregulates serine synthesis, resulting in increased levels of methylation substrates such as S‑adenosylmethionine (SAM). SAM promotes m6A methylation of PCK2 and NRF2, leading to an enhanced antioxidant system and lenvatinib resistance in HCC cells [67]. Cofilin 1 (CFL1) expression is elevated in sorafenib‑resistant HCC patients and is associated with poor prognosis. CFL1 increases serine metabolism by promoting PHGDH expression, thereby resisting sorafenib‑induced oxidative stress. Co‑delivery of CFL1 siRNA and sorafenib via nanoparticles (NPs) effectively enhances anti‑tumor activity with favorable safety [252].

Aspartate metabolism

Argininosuccinate synthase 1 (ASS1), a rate-limiting enzyme in the urea cycle, catalyzes the condensation of citrulline and aspartate to form arginine [253]. Several ncRNAs regulate aspartate metabolism by suppressing ASS1. CircRAPGEF1 is enriched in liver cancer stem cells. Methyltransferase-like 3 (METTL3) promotes N⁶-methyladenosine (m⁶A) modification of circRAPGEF1, enabling IGF2BP3 to bind to the UGGAC motif of circRAPGEF1 via its KH domain. This interaction not only enhances the stability of circRAPGEF1 but also prevents IGF2BP3-mediated degradation of ASS1 mRNA, leading to aspartate accumulation and activation of the mTOR downstream effectors S6K/CAD. Overexpression of circRAPGEF1 promotes stemness and proliferation of HCC cells through the aspartate/S6K/CAD axis. Importantly, this process also confers sorafenib resistance in HCC cells, whereas nanoparticle-delivered circRAPGEF1 siRNA effectively restores the anti-tumor activity of sorafenib [254]. Similarly, LINC01234 was found to be upregulated in HCC and associated with ASS1 suppression. Mechanistically, LINC01234 binds to the ASS1 promoter and inhibits its transcription. Suppression of ASS1 reduces aspartate consumption, resulting in elevated aspartate levels that activate the mTOR signaling pathway, thereby promoting HCC cell proliferation, migration, and sorafenib resistance. Inhibition of LINC01234 significantly impedes tumor growth in nude mice and restores the sensitivity of HCC cells to sorafenib [255].

Targeted metabolism as a therapeutic strategy for reversing MKI resistance

As research on HCC metabolic reprogramming progresses, several compounds have been identified through experimental data that target metabolic pathways in HCC and demonstrate anti-tumor activity (Table 3). Additionally, certain drugs traditionally used for other systemic diseases have recently been found to restore MKI sensitivity in HCC by modulating metabolic processes (Table 4). This section systematically reviews existing studies and summarizes potential compounds targeting metabolism, as well as drugs related to tumor metabolism, for potential repurposing.

Table 3.

Agents targeting metabolism to reverse the MKI resistance of HCC

Compounds	Targets	Functions in HCC	Reference
Agents as glycolysis inhibitor
2-DG	HK2	Synergize with sorafenib to inhibit tumor growth	257, 258
Rg3	HK2	Reduce glucose consumption and lactate production, and coordinately inhibit tumor growth with sorafenib	137
PB2	PKM2	Inhibit proliferation and invasion of HCC cells, induce apoptosis, and increase sorafenib sensitivity	158
Shikonin	PKM2	Inhibit proliferation and invasion of HCC cells, induce apoptosis, and increase sorafenib sensitivity	259
EGCG	PFK	Enhance the inhibitory effect of sorafenib on tumor growth in HCC	141
Isoacteoside	PDHB	Synergize with sorafenib to inhibit tumor growth	163
Betulin	SREBP-1	Increase the sensitivity of HCC cells to sorafenib treatment	260
Genistein	HK2 and GLUT1	Promote the sensitivity of HCC cells to cell death and enhance the anti-tumor effect of sorafenib on HCC	169
2ME2	Inhibits HIF-1α and HIF-2α	Promotes the sensitivity of HCC cells to sorafenib treatment	261
NaBu	c-MYC/HK2	Block glycolysis and restore the anti-tumor activity of sorafenib	174
Curcumin	PI3K/AKT, IL-6/JAK/STAT3 and IL-1β/NF-κB	Inhibit EMT of HCC cells and improve sorafenib efficacy	262
Histidine	GLUT1, HK2 and STAT3	Enhance sorafenib sensitivity	131
Niclosamide	IGF-1R	Inhibit cellular glycolysis and cancer stemness, and enhance sorafenib-induced mitochondrial membrane potential dysfunction and cell death	263
RPS	FASN, CPT1, GLUT1, Myc, Akt, mTOR and LDHA	Enhance sorafenib sensitivity	264
Dau	miR-199a/HK2 and PKM2	Restore the sensitivity of HCC to sorafenib	138
Agents targeted for lipid metabolism
UA	ING5/PI3K/AKT	Downregulate the expression of ACC1 and ACLY, and reverse sorafenib resistance in HCC cells	191
ND-654	ACC	Improve the anti-tumor activity of sorafenib	192
Cerulenin	FASN	Reduce the stem cell-like characteristics of HCC and synergize with sorafenib to inhibit tumor growth	193
TVB3664	FASN	Its combination with cabozantinib or sorafenib showed synergistic effects in HCC	265
LDA	ACSL4	Enhances the anti-tumor activity of regorafenib	222
Betulin	SREBP-2	Inhibit HCC tumor growth and enhance lenvatinib-induced cell death	231
Emodin	SREBP-2	Inhibit HCC tumor growth and enhance sorafenib-induced cell death	236

2-DG 2-Deoxy-D-glucose, HK2 Hexokinase 2, Rg3 Ginsenoside Rg3, PB2 Proanthocyanidin-B2, PKM2 Pyruvate kinase M2, EGCG Epigallocatechin-3-gallate, PFK Phosphofructokinase, PDHB Pyruvate dehydrogenase E1 β subunit, SREBP-1 Sterol regulatory element-binding protein 1, GLUT1 Glucose transporter 1, NaBu Sodium butyrate, PI3K/AKT Phosphatidylinositol 3-kinase/Protein kinase B, IL-6/JAK/STAT3 Interleukin-6/Janus kinase/Signal transducer and activator of transcription 3, IL-1β/NF-κB Interleukin-1β/Nuclear factor kappa-B, EMT Epithelial-mesenchymal transition, IGF-1R Insulin-like growth factor 1 receptor, RPS Rhizoma Paridis saponins, FASN Fatty acid synthase, CPT1 Carnitine palmitoyltransferase 1, mTOR Mammalian target of rapamycin, LDHA Lactate dehydrogenase A, Dau Dauricine, UA Ursolic acid, ING5 Inhibitor of growth family member 5, ACC1 Acetyl-CoA carboxylase 1, ACLY ATP citrate lyase, ACC Acetyl-CoA carboxylase, SREBP-2 Sterol regulatory element-binding protein 2, 2ME2 2-Methoxyestradiol, LDA Liquidambaric acid

Table 4.

Repurposed drugs for metabolic targeting

Drugs	Targets	Functions in HCC	Reference
Simvastatin	PKM2	Inhibit the proliferation of HCC cells, and promote cell apoptosis and sorafenib sensitivity	267
Aspirin	PFKFB3	Restore sorafenib-induced cell death	146
Orlistat	FASN	Decrease the fatty acid level and glucose uptake in HCC cells, and enhance the toxicity of sorafenib to HCC cells	268
Maprotiline	CRABP1/ERK/SREBP2	Inhibit the proliferation and metastasis of HCC cells, and enhance sorafenib sensitivity	230
Erlotinib	EGFR	Inhibit ABCB1-mediated drug efflux, thereby improving lenvatinib resistance	233
Tigecycline	Complex I/IV	Synergize with sorafenib to inhibit tumor growth	269, 270

PKM2 Pyruvate kinase M2, PFKFB3 6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3, FASN Fatty acid synthase, CRABP1 Cellular retinoic acid-binding protein 1, ERK Extracellular regulated protein kinase, SREBP2 Sterol regulatory element-binding protein 2, EGFR Epidermal growth factor receptor, ABCB1 ATP-binding cassette subfamily B member 1

Compounds targeting metabolic molecular targets

Targeting glycolysis

2-deoxy-D-glucose (2-DG) is a well-known glycolysis inhibitor that specifically targets the activity of HK2 [256]. The combination of 2-DG and sorafenib produces a synergistic effect, significantly enhancing the inhibition of HCC cell proliferation, invasion, and migration compared to monotherapy. Furthermore, this combination also effectively inhibits tumor growth in mouse models [257, 258]. Ginsenoside Rg3 (Rg3), a terpenoid compound, notably reduces HK2 expression in HCC cells, leading to decreased glucose consumption and lactate production. Additionally, Rg3 exhibits synergistic anti-tumor effects when combined with sorafenib, inhibiting HCC cell activity [137]. The natural compounds proanthocyanidin-B2 (PB2) and Shikonin inhibit glycolysis and lactate production in HCC cells by modulating PKM2. These compounds suppress HCC cell proliferation and invasion, while inducing apoptosis. Importantly, co-treatment with PB2 or Shikonin and sorafenib significantly increases the chemosensitivity of HCC to sorafenib [158, 259]. Epigallocatechin-3-gallate (EGCG), an active ingredient in green tea, inhibits PFK activity, inducing metabolic stress and HCC cell death by affecting the oligomeric structure of PFK. EGCG also enhances sorafenib's tumor growth inhibitory effects in HCC mice, suggesting its potential as an adjuvant to reverse sorafenib resistance [141]. Isoacteoside inhibits glycolysis and tumor growth in HCC by reducing PDHB activity. Combining Isoacteoside with sorafenib yields superior synergistic anti-tumor activity compared to monotherapy in a mouse xenograft model, with no significant weight changes or toxic effects observed in the mice [163]. Betulin, an SREBP-1 inhibitor, suppresses glucose metabolism and tumor metastasis in HCC cells by downregulating SREBP-1. Additionally, Betulin treatment enhances the sensitivity of HCC cells to sorafenib [260].

Additionally, several compounds target signaling pathways associated with glycolysis. Genistein, a natural flavonoid compound, directly downregulates HIF-1α in HCC cells, inhibiting the expression of HK2 and GLUT1. Genistein treatment reduces aerobic glycolysis, enhances HCC cell sensitivity to cell death, and boosts the anti-tumor effect of sorafenib both in vitro and in vivo [169]. 2-Methoxyestradiol (2ME2) inhibits HIF-1α and HIF-2α, and downregulates the expression of VEGF, LDHA, and cyclin D1. 2ME2 and sorafenib act synergistically to inhibit the proliferation and angiogenesis of HCC and induce apoptosis in HCC cells [261]. Sodium butyrate (NaBu), the sodium salt of the short-chain FA butyrate, inhibits HK2 by downregulating c-Myc, thereby suppressing glycolysis and proliferation while promoting apoptosis in HCC cells. Furthermore, NaBu restores sorafenib's anti-tumor activity in both in vitro and in vivo models by inhibiting glycolysis in sorafenib-resistant cells [174]. Several other compounds inhibit various metabolic targets. For instance, curcumin suppresses HIF-1α and LDH-mediated aerobic glycolysis via the PI3K/AKT signaling pathway, while inhibiting FASN and lipid synthesis, thereby reducing tumor growth. Additionally, curcumin inhibits cellular EMT by regulating the IL-6/JAK/STAT3 and IL-1β/NF-κB pathways. When added to mouse food, curcumin activates immune function, downregulates EMT, and reverses metabolic disorders, significantly improving sorafenib efficacy [262]. Exogenous histidine therapy decreases the expression of tumor markers related to glycolysis (GLUT1 and HK2), inflammation (STAT3), angiogenesis (VEGFB and VEGFC), and stemness (CD133). Combined histidine and sorafenib therapy significantly enhances sorafenib sensitivity [131]. Niclosamide, an insect repellent approved by the US FDA, has been shown to enhance the anti-tumor effects of sorafenib. Niclosamide inhibits IGF-1R phosphorylation induced by sorafenib, downregulating IGF-1R-mediated glycolysis and cancer cell stemness. In vitro, the combination of niclosamide and sorafenib enhances sorafenib-induced mitochondrial membrane potential dysregulation and cell death [263]. Rhizoma Paridissaponens (RPS) modulates glycolysis and lipid metabolism, counteracting the adverse effects of sorafenib treatment. The combination of sorafenib and RPS demonstrates significantly higher anti-tumor activity compared to monotherapy. Mechanistically, RPS reduces the mRNA levels of FASN, CPT1, GLUT1, Myc, Akt, mTOR, and LDHA, while increasing p53 expression, thus enhancing sorafenib sensitivity in HCC cells [264]. In HCC cells, Dauricine (Dau) increases sorafenib-induced apoptosis and maintains consistent sensitivity to sorafenib. Dau dose-dependently increases miR-199a, which targets and inhibits HK2 and PKM2, thereby reducing glycolysis in HCC cells. Restoring sorafenib sensitivity through Dau/miR-199a offers a novel potential therapeutic strategy for HCC [138].

Targeting lipid metabolism

Ursolic acid (UA), a pentacyclic triterpenoid compound, inhibits ING5 transcription by reducing the binding of serum response factor (SRF) and Yin Yang-1 (YY1) to the ING5 promoter region. By suppressing ING5, UA impedes ING5-mediated activation of the PI3K/AKT signaling pathway, downregulating the expression of ACC1 and ACLY, and reducing LD formation. UA exhibits both anti-tumor activity and the ability to reverse sorafenib resistance in HCC cells [191]. ND-654, a novel liver-specific ACC inhibitor, suppresses the expression of ACC1 and ACC2 by mimicking AMPK-mediated dysregulation of ACC phosphorylation. When combined with sorafenib, ND-654 significantly improves the survival rate of tumor-bearing rats [192]. Chen et al. utilized cerulenin, a FASN inhibitor, to reduce FA production in HCC tumor stem cells. This intervention decreased CSC-like characteristics, including stemness gene expression, spheroid formation, tumorigenicity, and metastatic potential. The combination of cerulenin with sorafenib exhibited a significant synergistic effect, particularly against HCC tumor stem cells [193]. The other FASN inhibitor, TVB3664, has been investigated in preclinical studies involving HCC cell lines and HCC mouse models induced by phosphatase and tensin homolog (PTEN) loss and c-MET activation. TVB3664 as a monotherapy exhibited moderate efficacy, whereas its combination with cabozantinib or sorafenib showed synergistic effects and markedly suppressed the activation of the AKT/mTOR signaling pathway [265]. Liquidambaric acid (LDA), a pentacyclic triterpenoid compound derived from various plants, can block the cell cycle and promote apoptosis in HCC cells. LDA downregulates the expression of RXRα, thereby inhibiting the formation of the PPARα–RXRα heterodimer and suppressing ACSL4-mediated FAO. LDA enhances the anti-tumor activity of regorafenib both in vivo and in vitro [222]. Additionally, emodin and betulin inhibit the expression of SREBP-2, thereby reducing cholesterol synthesis in HCC cells. Downregulation of cholesterol suppresses the AKT/mTOR and STAT3 pathways, inhibiting tumor growth and enhancing sorafenib-induced cell death [231, 236].

Targeted metabolic repurposing of existing drugs

Many clinically used drugs with established efficacy and safety profiles have been shown to benefit patients with HCC. Statins, a class of cholesterol-lowering drugs, inhibit hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase, blocking cholesterol synthesis. Recent studies have reported that statin use reduces the incidence of HCC in patients with chronic liver disease [266]. Furthermore, simvastatin targets PKM2 and inhibits PKM2-mediated glycolysis. Simvastatin treatment reduces HCC cell proliferation, increases apoptosis, and enhances sorafenib sensitivity [267]. Aspirin, a nonsteroidal anti-inflammatory drug (NSAID), has also been shown to reduce glycolytic activity in HCC cells by inhibiting PFKFB3 expression. In studies involving two sorafenib-resistant HCC cell lines and a subcutaneous tumor model in nude mice, aspirin significantly restored sorafenib-induced cell death. Moreover, the combination of aspirin and sorafenib did not cause weight loss, liver injury, or inflammation in the mice [146].

Orlistat, a classic FASN inhibitor used for obesity and overweight management, significantly reduces cellular FA levels and glucose uptake in HCC cells. The combination of orlistat and sorafenib enhances sorafenib toxicity toward HCC cells, induces G1 cell cycle arrest, and promotes apoptosis [268]. Maprotiline, a tetracyclic antidepressant, binds directly to cellular retinoic acid-binding protein 1 (CRABP1), inhibiting the ERK–SREBP2 signaling pathway and attenuating cholesterol biosynthesis. Maprotiline suppresses the proliferation and metastasis of HCC cells and increases their sensitivity to sorafenib [230]. Dysregulated cholesterol metabolism and lipid raft activation are key characteristics of lenvatinib-resistant HCC cells. The EGFR-activated STAT3 pathway upregulates cholesterol metabolism and promotes lipid raft-dependent upregulation of ABCB1, enhancing the exocytosis of lenvatinib. Combination therapy with the EGFR inhibitor erlotinib and lenvatinib significantly reduces ABCB1 levels in HCC cells, overcoming lenvatinib resistance both in vitro and in vivo [233].

Tigecycline, an antibiotic primarily used against multidrug-resistant bacteria, binds to the 30S ribosomal subunit in bacteria, inhibiting protein translation. Studies have shown that due to the structural similarity between mitochondrial ribosomes and bacterial ribosomes, tigecycline can suppress mitochondrial protein synthesis and impair the activities of mitochondrial complexes I and IV. This leads to mitochondrial dysfunction and restricts the turnover of electron acceptors required for glutamine oxidation. Consequently, this process reverses sorafenib resistance in HCC cells, producing a synergistic effect when combined with sorafenib [269, 270].

Metabolic reprogramming of HCC and TIME

Metabolic reprogramming in cancer cells modulates immune cell responses within the TME through the release of metabolites or by inducing local nutrient depletion. Interestingly, enhanced glucose metabolism observed in sorafenib-resistant HCC cells promotes PD-L1 expression, thereby attenuating the immune response in HCC [271]. On the other hand, modulating the state of CAFs and tumor-associated macrophages (TAMs) can restore HCC sensitivity to sorafenib [272–274]. This suggests that targeting metabolic reprogramming may yield dual benefits from immunotherapy and MKI therapy. In this section, we discuss how dysregulated glycolysis, lipid metabolism, and amino acid metabolism drive immunosuppression (Fig. 6), which may represent an effective strategy to enhance the efficacy of combined MKI and immunotherapy.

Fig. 6.

The role of metabolic reprogramming of HCC in TIME. In HCC cells, glycolysis, lipid metabolism, and amino acid metabolism are enhanced. The metabolic reprogramming in HCC leads to the release of various metabolites into the TME, such as lactate, bile acids (predominantly palmitic acid), PUFAs, and cholesterol. These metabolites inhibit the activation of CD8⁺ T cells, CD4⁺ T cells, MAIT cells, and NKT cells. Conversely, they promote the activation of Treg cells, M2-TAMs, N2-neutrophils, and CAFs, thereby suppressing the immune response against HCC. Additionally, glucose deficiency contributes to the exhaustion of CD8⁺ T cells and drives γδ T cells toward a pro-tumorigenic phenotype, while glutamine deficiency restricts the activity of NK cells

Aberrant glycolysis promotes immunosuppression

The extrusion of excessive lactate by HCC cells leads to an acidic TME, which profoundly suppresses immune responses. High lactate levels inhibit the activation and proliferation of CD8 + T cells and promote apoptosis, thereby contributing to immunosuppression within the TME [275]. In murine hepatocellular carcinoma models, increased glucose flux and lactate production were observed alongside a highly exhausted tumor-infiltrating T-cell population [276]. Concurrently, lactate upregulates PD-1 expression on regulatory T (Treg) cells and activates the TGF-β/SMAD3 signaling pathway in Treg cells, further suppressing effector T cells [277, 278]. Overexpression of MCT4, which mediates lactate secretion on the HCC cell surface, promotes Treg cell accumulation and reduces HCC responsiveness to immune checkpoint inhibitors (ICIs) [279]. Conversely, targeted inhibition of MCT4 reduces TME acidification and increases the secretion of C-X-C motif chemokine ligands 9 and 10 (CXCL9/CXCL10), resulting in enhanced CD8 + T cell infiltration and cytotoxicity [280]. Furthermore, intracellular lactate in HCC cells enhances H3K18la, which binds to the PD-L1 promoter and directly increases PD-L1 expression in HCC cells [281]. HCC-derived exosomal SLC16A1-AS1 increases the stability of SLC16A1 in TAMs. Elevated SLC16A1 augments lactate uptake by TAMs and induces M2 polarization via activation of the c-Raf/ERK signaling pathway [282]. Lactate also enhances H3K18la and nuclear protein 1 (NUPR1) lactylation in TAMs, activating the ERK and JNK signaling pathways within TAMs, which leads to their M2 polarization. M2-polarized TAMs inhibit CD8 + T cell enrichment in the TME, resulting in diminished responses of HCC to immunotherapy [283, 284]. Lactate additionally enhances the activation and recruitment of CAFs. CAFs promote PD-L1 expression on HCC cells and CD8 + T cell apoptosis, while inhibiting cell proliferation, through the expression of Salmonella pathogenicity island 1 (SPI1) [285, 286]. Utilizing single-cell RNA sequencing and a prospective clinical cohort (NCT04649489), Xu et al. found that serum galectin-4, elevated in patients with HCC resistant to combined atezolizumab and bevacizumab therapy, inhibits LDHA degradation. LDHA-mediated glycolysis and lactate increase promote N2 polarization of tumor-associated neutrophils and suppress CD8 + T cell infiltration and cytotoxicity [287].

Accelerated glucose metabolism by HCC cells depletes glucose in the TME. Importantly, following activation, CD8 + T cells also consume glucose via glycolysis to rapidly generate energy and release increased levels of IFN-γ, IL-2, and TNF-α, thereby competing with HCC cells for glucose [288, 289]. Evidently, cancer cells are the predominant winners in this competition in most instances. In contrast, Treg cells exhibit higher expression of MCT1 and LDH and can utilize lactate to sustain themselves via OXPHOS, conferring a survival advantage in the low-glucose, high-lactate TME [290, 291]. HCC cells deficient in Smad4 enhance CXCL10 secretion and activate the mTOR/LDHA pathway-mediated glycolysis in CD8 + T cells, leading to increased CD8 + T cell infiltration and enhanced immune effector function [292].

Regulatory role of lipid metabolism in TIME

An increase in palmitic acid-mediated lipid metabolism has been observed in TIME-suppressed HCC. Concurrently, alterations in the lipid composition of tumor tissues were noted, characterized by the accumulation of long-chain acylcarnitines (LCACs) and a reduction in lysophosphatidylcholines (LPCs). These changes contribute to T cell exhaustion and the infiltration of M2 macrophages and CAFs [293, 294]. The precise mechanisms by which lipid metabolism regulates the TIME remain incompletely understood. Recent evidence suggests that lipid metabolites derived from HCC cells may be involved in immune responses. Firstly, the upregulation of FASN and ACSL4 in HCC enhances fatty acid metabolism, leading to an increase in free fatty acids within the TME. These free fatty acids, predominantly palmitic acid (PA/C16:0), can reduce the infiltration and activation of CD8⁺ T cells. This phenomenon may be attributed to palmitic acid enhancing its palmitoylation within T cells, thereby activating the STAT3 pathway and mediating CD8⁺ T cell exhaustion [295, 296]. Furthermore, ACSL4 drives the synthesis of PUFAs in HCC. PUFAs accumulate in mucosal-associated invariant T (MAIT) cells, leading to MAIT dysfunction through lipid peroxidation-mediated ferroptosis. Although PUFA accumulation was not observed in CD8⁺ T cells or natural killer (NK) cells, the increase in PUFAs indeed impairs the anti-tumor activity of CD8⁺ T cells [297, 298]. The liver is the primary site for cholesterol synthesis. In HCC, the ZDHHC3 gene drives enhanced cholesterol synthesis. In obese HCC patients, significant cholesterol accumulation was observed, which suppresses tumor immunity mediated by natural killer T (NKT) cells and CD4⁺ T cells [97, 299]. Deficiency in SIRT5 leads to increased cellular bile acid (BA) synthesis, particularly of taurine-conjugated bile acids. High levels of BAs are associated with an increase in M2-type tumor-associated macrophages (TAMs) and maintain an immunosuppressive TME during the initiation of HCC [300].

Certain immune cells in HCC also exhibit reprogramming of lipid metabolism. γδ T cells, through T cell receptor (TCR)-mediated signaling, activate the PI3K/Akt/mTOR and c-Myc pathways. This process promotes enhanced glycolytic activity while attenuating FAO. However, with tumor progression and glucose deprivation, γδ T cells sustain themselves via fatty acid metabolism and transform into a pro-tumorigenic subset expressing PD-1 and CTLA-4 [301]. In TAMs, increased fatty acid uptake and synthesis lead to the accumulation of triglycerides and LDs. TAMs laden with LDs highly express TREM2, PD-L1, CD206, and CD163, which can significantly suppress the cytotoxic antitumor activity of CD8⁺ T cells and recruit Treg cells [302]. Concurrently, abnormally enhanced FAO is observed in TAMs and drives their polarization toward an M2 phenotype. M2-type TAMs, through enhanced FAO, increase the secretion of IL-1β, exert immunosuppressive effects, and promote HCC metastasis [303, 304].

Amino acid deficiency in the TIME

The competitive uptake of amino acids by cancer cells reduces the availability of nutrients for immune cells, which may contribute to the suppression of the TIME. For instance, TAM-derived circPETH-147aa is transported to HCC cells via extracellular vesicles and enhances the stability of SLC43A2 mRNA. The upregulation of SLC43A2 facilitates the sequestration of leucine and methionine by HCC cells, leading to methionine and leucine deprivation in CD8⁺ T cells and a consequent decline in their immune activity [305]. Furthermore, glutamine deficiency in the TME reduces the expression of perforin (PRF1) and granzyme B (GZMB) in CD8⁺ T cells, impairing their antitumor function. Glutamine is also associated with c-Myc protein levels in NK cells. The activation and survival of NK cells depend on c-Myc-mediated oxidative phosphorylation. Upon glutamine withdrawal, the loss of c-Myc protein results in retarded NK cell growth and compromised cellular responses [306, 307].

Clinical prospects and challenges

In recent years, advancements in technologies such as gene editing, stable isotope tracing, and metabolomics have enabled precise dissection of the tumor metabolic network and its product dynamics [308]. Substantial evidence indicates that metabolic reprogramming plays a key role in reducing HCC's sensitivity to MKIs, particularly sorafenib and lenvatinib. Various molecules involved in tumor metabolic reprogramming have been identified as critical factors in HCC progression and drug resistance, highlighting metabolism as a potential therapeutic target to overcome MKI resistance. As discussed in the previous section, several agents can modulate metabolic reprogramming in HCC cells, and combining these agents with MKIs may synergistically suppress HCC progression. Additionally, the regulation of metabolism-related genes through gene-editing technologies such as CRISPR-Cas9, siRNA, and small activating RNA (saRNA) offers a promising therapeutic approach [309–311]. For example, miR-494, highly expressed in HCC, promotes glycolysis by targeting the catalytic subunit of glucose-6-phosphatase (G6pc), contributing to sorafenib resistance. Combining antagomiR-494 with sorafenib enhances sorafenib-induced cytotoxicity in HCC cells [312]. Advancements in targeted delivery systems now allow precise and stable transport of combination therapies involving MKIs to HCC cells. For instance, various nanoparticles have been developed to co-encapsulate sorafenib/lenvatinib alongside other metabolism-targeting compounds or RNA modifiers, yielding synergistic anti-tumor effects [313–315]. One study created RGD pentapeptide-modified high-density lipoproteins loaded with sorafenib and anti-miRNA21 to reverse sorafenib resistance in HCC [316]. Xu et al. constructed a galactose-modified lipid-polyplex (Gal-SLP) encapsulating both sorafenib and USP22 shRNA. Sorafenib-induced ROS promoted the release of USP22 shRNA, which enhanced sorafenib efficacy by downregulating glycolysis in HCC cells [317].

However, despite promising results in basic research, the translation of metabolism-targeting strategies into clinical applications faces challenges such as model discrepancies and toxicity control. Current in vivo studies on HCC metabolism largely rely on patient-derived xenograft (PDX) mouse models, which may lack a normal human immune microenvironment. One study found that, compared to primary cancers, PDX models only partially restored altered glycolysis and TCA cycle-related pathways after multiple passages [318]. Moreover, because non-cancerous cells, including immune cells, also have metabolic vulnerabilities, drugs targeting metabolism may affect tumor cell metabolism and other components within the TME, potentially leading to immunosuppression or severe adverse effects [319]. In early studies targeting glycolysis for cancer treatment, infusion of the glycolytic inhibitor 2-DG caused severe adverse reactions in patients [320]. In a Phase I clinical trial, David's team used the lipid-based miR-34a mimic MRX34 in 85 patients with HCC. Although considerable clinical efficacy was observed, all 85 patients experienced significant adverse events, resulting in four deaths and the termination of the trial [321]. Therefore, mitigating the toxic effects of metabolism-targeting agents on non-tumor tissues remains a critical challenge that must be addressed before these strategies can be applied clinically.

Conclusion

Metabolic reprogramming, a hallmark of HCC, plays a pivotal role in mediating resistance to MKIs, offering crucial insights for overcoming current challenges in systemic HCC therapy. This article systematically reviews the dysregulation of glycolytic, lipid metabolic, and amino acid metabolic pathways in HCC, emphasizing the central roles of key metabolic enzymes (such as GLUT1, HK2, FASN, and GLS1) and signaling molecules in driving MKI resistance. These insights provide a foundation for identifying potential therapeutic targets. Preclinical studies targeting these metabolic abnormalities have demonstrated that various agents can reverse drug resistance, presenting novel strategies for HCC treatment. We also summarize the roles of metabolic reprogramming in the TIME, which is associated with the efficacy of MKIs and MKI-immunotherapy combination therapy. However, the precise targeting of HCC’s unique metabolic phenotypes and translating these findings into clinical benefits remain significant challenges for both basic research and clinical practice.

Abbreviations

IGF2BP3

Insulin-like growth factor 2 mRNA-binding protein 3

PCK2

Phosphoenolpyruvate carboxykinase 2

NRF2

Nuclear factor erythroid 2-related factor 2

HKDC1

Hexokinase domain-containing 1

NADPH

Nicotinamide adenine dinucleotide phosphate

TGF-β1

Transforming growth factor-β1

MCTs

Monocarboxylate transporters

DCA

Dichloroacetate

GPI

Glucose-6-phosphate isomerase

ENO2

Enolase 2

AMPK

AMP-activated protein kinase

FGFR4

Fibroblast growth factor receptor 4

JNK2

C-Jun N-terminal kinase 2

SPAG4

Sperm-associated antigen 4

ZNF263

Zinc finger protein 263

TRIM28

Tripartite motif containing 28

MDM2

Mouse double minute 2 homolog

m6A

N6-methyladenosine

YTHDF3

YTH N6-methyladenosine RNA binding protein 3

PRKACA

CAMP-dependent protein kinase catalytic subunit alpha

SCAP

SREBP cleavage-activating protein

IGF-1R

Insulin-like growth factor 1 receptor

USP22

Ubiquitin-specific protease 22

SIRT5

Sirtuin 5

Authors’ contributions

JL wrote the majority of the manuscript and created the figures. YH and JW L contributed to the manuscript writing. MS and YX assisted in figure and table preparation and manuscript revision. FD revised the manuscript. GH revised the manuscript and supervised the process. Data authentication is not applicable. All authors have read and approved the final version of the manuscript.

Funding

No funding was received.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.