The controversial role of metabolic reprogramming in anti-tumor therapy resistance

Abstract

Anti-tumor therapies, including chemotherapy, immunotherapy, targeted therapy and endocrine therapy highlight an increasingly important role, but drug resistance limits their efficacy. In the nutrient-poor tumor microenvironment (TME), tumor cells resist the anti-tumor therapies by turning into ‘resistance continuum’. Metabolic reprogramming serves an essential character in the cellular ‘resistance continuum’. The process of metabolic reprogramming is heterogeneous, dynamic and complex in drug-resistant tumor cells. Some resist the killing effects of anti-tumor drugs by enhancing glycolysis, some by heightening lipid metabolism, some by boosting amino acid metabolism, and some even by potentiating OXPHOS which is contrary to the traditional Warburg effect. These metabolic reprogrammed processes could help tumor cells to maintain intracellular redox homeostasis and evade ferroptosis under anti-tumor treatments. This review has attempted to describe the common mechanisms in drug-resistant tumor cells under chemotherapy, immunotherapy, targeted therapy and endocrine therapy from the perspective of metabolic reprogramming, and summarizes recent advances in research on targeting metabolism to overcome drug resistance.

Introduction

Over the past decades, advances in precision oncology have provided cancer patients with a variety of advanced and personalized anti-tumor therapies, but drug resistance limits the effectiveness. Although researchers have attempted various approaches to overcome drug resistance, the outcomes have been unsatisfactory. As tumor cells adapt to the dynamically changing tumor microenvironment (TME), they gradually alter their state trajectory and develop drug resistance. The cellular ‘resistance continuum’ involves stepwise combinations of gene expression programs, phenotypic plasticity, stress adaptation, and metabolic reprogramming [1]. It is worth mentioning that metabolic reprogramming plays a critical role in the acquisition of drug resistance in tumor cells. When faced with treatment-induced stress, tumor cells form adaptations by altering various metabolic processes, including glycolysis, oxidative phosphorylation (OXPHOS), lipid metabolism, amino acid metabolism, and evading ferroptosis to resist the cytotoxic effects of anti-cancer drugs.

Tumor cells adapt to the nutrient-poor TME by reprogramming metabolism, which is often driven by alterations in proto-oncogenes and tumor suppressor gene, thereby supporting their proliferation, invasion, metastasis, evasion of the immune surveillance, and development of drug resistance [2, 3]. Tumor cells change their primary energy production from OXPHOS to glycolysis in order to meet their metabolic demands of rapid growth. Even in the presence of oxygen, tumor cells switch their metabolism from OXPHOS to glycolysis, known as the Warburg effect [4]. Although Warburg pointed out that glycolysis plays a dominant role in the metabolic reprogramming of tumor cells, a growing number of studies in recent years have revealed that other metabolic pathways such as OXPHOS, fatty acid oxidation (FAO), ferroptosis, also contribute to the drug-resistance in anti-tumor treatments. In the nutrient-poor TME, tumor cells alter their metabolic processes and metabolites by depriving them of necessary nutrients. Meanwhile, tumor cells modulate the function of tumor-infiltrating lymphocytes (TILs) through metabolic reprogramming and altered metabolites, constructing an immunosuppressive TME that allows them to escape immune surveillance and resist anti-tumor therapy [5–7]. Metabolic reprogramming and immune modulation interact to facilitate tumor cell proliferation, metastasis, immune escape, and drug resistance.

Due to the complex tumor heterogeneity, in the duration of treatment with anti-tumor drugs, drug-resistant tumor cells are gradually screened out, escaping the killing of anti-cancer drugs, resulting in tumor recurrence or progression. In this process, some tumor cells enhance glycolysis, some strengthen OXPHOS, some increase lipid metabolism, and some suppress ferroptosis. Additionally, it was reported that some drug-resistant tumor cells enter a dormant state similar to embryonic developmental arrest, allowing them to survive under chemotherapeutic agents, during which cell proliferation and metabolic processes are inhibited [8]. Therefore, it is necessary to clarify the metabolic reprogramming process of tumor cells during anti-tumor therapy resistance and identify specific metabolic reprogramming targets to enhance anti-tumor efficacy.

This article aims to elucidate the common mechanisms of tumor cell drug resistance under chemotherapy, immunotherapy, targeted therapy and endocrine therapy from the perspective of metabolic reprogramming, and to review the research progress of targeting metabolic therapies in anti-tumor treatment.

Role of glycolysis during drug resistance

Role of glycolysis during chemotherapy resistance

The recent studies have indicated that the metabolic reprogramming processes, particularly glycolysis, enable tumor cells to overcome the cytotoxic effects of anti-tumor drugs and develop chemoresistance [9]. Chemotherapy, which could inhibit the glycolytic process of tumor cells, enhance the efficacy of immune cells and impair tumor cells. However, under the treatment of chemotherapy, drug-resistant tumor cells with increased glycolytic flux are progressively selected. Chong et al. found that doxorubicin-resistant breast cancer cells activate glycolysis and the tricarboxylic acid (TCA) cycle compared to non-resistant cells [10]. Lin et al. reported that colon cancer cells enhance glycolysis and the pentose phosphate pathway (PPP) by up-regulating the POU2F1-ALDOA axis, reducing their sensitivity to oxaliplatin and promoting tumor progression [11].

Hypoxia is one of the important factors linking chemotherapy resistance and glycolysis in tumor cells. The mismatch between rapidly growing tumor cells and the relative lack of vasculature exacerbates hypoxia, and the hypoxic TME assists tumor cell progression, metastasis, and drug resistance. Hypoxia inducible factor-1 (HIF-1α) of the hypoxic TME plays a vital role in anti-tumor therapy resistance [12]. The high-expressed HIF-1α induces the transcriptional overexpression of numerous hypoxia-responsive genes associated with metabolic reprogramming in tumor cells [13]. Ai et al. reported that the high-expressed HIF-1α confer drug resistance to tumor cells through the glycolytic pathway [14]. Inhibiting HIF-1α could increase the sensitivity of resistant ovarian cancer cells to cisplatin by shifting the aerobic glycolysis to OXPHOS, resulting in cell death due to excessive reactive oxygen species (ROS) production [14]. Zhao et al. demonstrated that in the glucose-deprived TME, tumor cells enhanced the transcription of HIF-1α and lactate dehydrogenase (LDH) by increasing glucose-regulated proteins 78 (GRP78) expression, which subsequently contributed to their metabolic reprogramming and the drug resistance [15]. Shukla et al. showed that HIF-1α stabilization mediated through MUC1, a large type 1 transmembrane protein, regulated an enhanced glycolytic flux, resulting in glucose addiction in pancreatic cancer cells. This favors increased pyrimidine biosynthesis, enhancing the intrinsic levels of deoxycytidine triphosphate (dCTP), reducing the adequate level of gemcitabine through molecular competition, resultly contributing to drug resistance. The combination therapy of inhibiting HIF-1α or pyrimidine biosynthesis with gemcitabine could significantly reduce tumor load [16]. Similarly, Shigeta et al. demonstrated that gemcitabine-resistant urothelial carcinoma cells weakened the cytotoxic effect of chemotherapy by increasing dCTP production [17]. Moreover, they reported an increase in reduced glutamine metabolism and stabilization of Hif-1α expression by isocitrate dehydrogenase 2 (IDH2) in gemcitabine-resistant tumor cells [17]. IDH2-mediated metabolic reprogramming also enhances antioxidant defenses by increasing nicotinamide adenine dinucleotide phophate (NADPH) and glutathione (GSH) production, thereby helping tumor cells become cross-resistant to cisplatin. Down-regulation or pharmacological inhibition of IDH2 restores chemosensitivity.

Studies have shown that epigenetics participated in chemotherapy resistance through glycolysis. Duan et al. demonstrated that pemetrexed-resistant lung cancer cells increased LDHA expression by upregulating aldo-keto reductase family 1 B10 (AKR1B10) which is associated with metastatic relapse, promotes glycolysis, enhances the Warburg effect. Consequently, the increased lactate promotes histone lactylation (H4K12la), activating the transcription of the cell-cycle-related gene CCNB1, accelerating DNA replication and the cell cycle [18]. Besides, N6-methyladenosine (m6A) could exacerbate the Warburg effect, inducing tumor cell growth, metastasis, and drug resistance [19]. M6A is the most abundant modifier in human mRNAs, which regulates mRNA splicing, decay, and translation [20]. Zhang et al. reported that 5-fluorouracil (5-FU)-resistant colon cancer cells up-regulate m6A by increasing methyltransferase-like 3 (METTL3), which stabilizes HIF-1α mRNA and enhances LDHA expression, thereby facilitating tumor cell entry into glycolysis and contributing to drug resistance [21].

Role of glycolysis during immune checkpoint blockade (ICB) resistance

Immune checkpoint blockade (ICB), represented by anti-programmed death-1 (anti-PD-1) and anti-programmed death ligand-1 (anti-PD-L1), has shown remarkable progress in anti-tumor treatments, providing new opportunity for cancer patients. However, the response rate is only 10–30%, limiting the anti-tumor efficacy [22]. ICB resistance is usually thought to be related to low or no expression of PD-1/PD-L1 in tumor tissue, or lack of sufficient immune cells for effective tumor antigen processing, presentation, and immunologic cytotoxicity, especially CD8 + T lymphocytes. Additionally, gene mutations and intestinal flora have been reported to be involved in the ICB resistance [23]. However, the mechanism involved in ICB resistance from metabolic reprogramming is poorly studied.

It was reported that highly tumor-derived lactate was associated with PD-L1 over-expression, low CD8 + TILs and the resistance to ICB [24]. Guo et al. demonstrated that up-regulated aerobic glycolysis in tumor cells promoted the dissociation of hexokinase2 (HK2) from mitochondria and bound to cytosolic inhibitor of nuclear factor-κB alpha (IkBα) in glioblastoma cells [25]. HK2 acts as a protein kinase and phosphorylates IkBα, resulting in IkBα degradation and nuclear factor-κB (NF-κB) activation-dependent PD-L1 expression for tumor immune escape [25]. The HK inhibitor combined with the anti-PD-1 could counteract tumor immune escape and enhance the anti-tumor efficacy of ICB. Besides, Wu et al. demonstrated that the tumor-intrinsic glycolysis pathway confers resistance to T cell-mediated killing. Inhibition glycolysis by targeting Glut1 led to the accumulation of intracellular ROS, which sensitized tumor cells to T cell-mediated bystander killing through tumor necrosis factor-alpha (TNF-α) [26]. Treatment with the Glut1-specific inhibitor enhanced tumor sensitivity to ICB and synergized with anti-PD-1 therapy in mouse models [26].

In the nutrient-poor TME, tumor cells predate the limited nutrients available to immunocytes, altering their metabolic processes, suppressing the anti-tumor immunity and contributing to ICB resistance. A retrospective study showed that anti-PD-1-resistant gastric cancer patients had higher IL-4 levels than immunotherapy-responsive patients, and further studies demonstrated that IL-4 could help macrophages to enhance glycolysis, lactate production, and up-regulate the expression of FcγRIIB through the phosphatidylinositol 3 kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) pathway, which could lead to CD8 + T-cells exhaustion and drug resistance [27]. Kumagai et al. reported that lactate in the highly glycolytic TME induced PD-1 expression in regulatory T (Treg) cells through monocarboxylate transporter 1 (MCT1), resulting in the impairment of ICB [28]. They also demonstrated that targeting LDHA and MCT1 potentiated the effectiveness of ICB in treating intrahepatic tumors. He et al. found that serum amyloid A (SAA) could activate glycolysis through the LDHA/signal transducer and activator of transcription 3 (STAT3) pathway, up-regulate the expression f PD-L1 on neutrophils, and release oncostatin M, thus impairing T cell function in hepatocellular carcinoma (HCC) [29]. The inhibition of STAT3 or SAA alleviated neutrophil-mediated immunosuppression and augmented the anti-tumor efficacy of ICB in vivo.

Role of glycolysis during targeted therapy resistance

Metabolic reprogramming is a predominant factor contributing to tumor cell resistance to targeted therapy. Epidermal growth factor receptor (EGFR) mutations enable non-small cell lung cancer (NSCLC) cells to increase glucose uptake and initiate glycolysis by enhancing glycolytic enzymes, such as glucose transporters (GLUTs) and phosphofructokinase-1 (PFK-1), facilitating their rapid growth in the nutrient-poor TME [30, 31]. In addition to inhibiting EGFR tyrosine kinase phosphorylation, the tyrosine kinase inhibitor (TKI) gefitinib downregulates the expression of GLUTs in EGFR-mutant cells and inhibits tumor cell growth. However, the gefitinib-resistant NSCLC cells increase their GLUT expression, increase glucose uptake, and maintain tumor cell growth. Suzuki et al. suggested that the activity of GLUT1-mediated glucose metabolism could be a key factor determining the responsiveness of NSCLC cells to EGFR TKIs and that co-inhibition of GLUT1 may be a way to ameliorate resistance to EGFR TKIs in NSCLC cells [32].

Furthermore, changed cytokines and chemokines in the TME are involved in the resistance to targeted therapy through glycolysis. The epidermal growth factor (EGF) secreted by tumor-associated stromal cells, synergistically promotes the process of epithelial-mesenchymal transition (EMT) through the EGFR/PI3K/HIF-1α pathway with glycolysis, leading to TKI resistance in NSCLC cells [33]. Additionally, it was reported that cancer-associated fibroblasts (CAFs) stimulated hepatocyte growth factor (HGF) overexpression by uptaking lactate through the NF-κB pathway, which then activated the cellular-mesenchymal to epithelial transition factor (c-MET) pathway, leading to the development of non-EGFR-dependent TKI resistance in NSCLC cells [34, 35].

Role of glycolysis during endocrine therapy resistance

Using the RNA-Seq data method, Verma et al. detected androgen deprivation therapy (ADT, enzalutamide) resistant prostate cancer cells. They found that enzalutamide-resistant prostate cancer cells progress and develop drug resistance through glycolysis involving the Solute Carrier Genes SLC12A5, SLC25A17, and SLC27A6 [36].

Role of mitochondrial metabolic reprogramming during drug resistance

Mitochondria, as one of the essential organelles for cellular OXPHOS to produce adenosine triphosphate (ATP) and energy for biological activities, engage in metabolic reprogramming and drug resistance in tumor cells. Pyruvate generated by glycolysis enters the mitochondria and proceeds oxidative decarboxylation to produce acetyl coenzyme A (acetyl-CoA) in the presence of the pyruvate dehydrogenase complex. Acetyl-CoA enters the TCA cycle and condenses with oxaloacetate to form citric acid, which eventually produces carbon dioxide and water as well as reduced coenzymes such as NADH and FADH in a sequence of reactions. These reduced coenzymes transfer electrons to the electron transport chain (ETC), driving protons to pump out of the mitochondrial matrix and forming a proton gradient. This gradient drives ATP synthase to synthesize ATP when protons return. Acetyl-CoA is one of the essential metabolites in the tumor cells, which not only participates in many metabolic processes, powering the TCA cycle, entering in OXPHOS, and supporting the synthesis of fatty acids (FAs), but also enables the acetylation of histone, involved epigenetics in tumor cells, and contributes to tumor growth, invasion, and drug resistance [37].

Tumor cells are not only resistant to drugs through enhanced glycolysis as described above, but also through inhibition of mitochondrial-derived OXPHOS. Gu et al. showed that renal clear cell carcinoma (RCC) restrains mitochondrial metabolism and contributes to drug resistance by down-regulation of SIRT3, whereas RCC with overexpressed SIRT3 further improved the cytotoxic effects when combined with anti-tumor drugs (Resveratrol, Everolimus, and Temsirolimus) [38]. Furthermore, it was reported that the NADPH oxidase isoform (NOX4) serves as a mitochondrial energetic sensor, linking metabolic reprogramming to drug resistance in tumor cells [39]. Shanmugasundaram et al. demonstrated that suppressed ATP levels in mitochondria bind to NOX4, and the programmed NOX4-derived ROS inhibits P300/CBP-associated factor (PCAF)-dependent acetylation and lysosomal degradation of the pyruvate kinase-M2 isoform (PKM2), which promotes tumor cell proliferation, migration, and drug resistance [39].

Reverse Warburg effect during drug resistance

However, not all tumor cells develop drug resistance by enhancing glycolysis or inhibiting OXPHOS. It has been shown that some tumor cells resist the cytotoxic effects of anti-tumor agents by enhancing OXPHOS [40]. Fantin et al. reported that targeting glycolysis potentiated the anti-tumor efficacy of LDH inhibitors, and tumor cells switch their metabolism to OXPHOS. It suggests that the metabolic conversion of tumor cells to glycolysis under hypoxia is merely a preference but not unidirectional reprogramming [41]. Bonuccelli et al. proposed a ‘reverse Warburg effect’ opinion, suggesting that CAFs produce lactate through glycolysis, which is then supplied to neighboring tumor cells through the paracrine pathway, activating mitochondria, increasing OXPHOS and promoting tumor cell growth [42]. Matamala Montoya et al. demonstrated that bone-marrow stromal cells (BMSCs) skew the metabolic phenotype of multiple myeloma cells shifting towards a drug-resistant phenotype, with increased OXPHOS, serine synthesis pathway (SSP), TCA cycle, and GSH synthesis [43]. Different subpopulations of tumor cells employ distinct metabolic reprogramming processes to adapt to the nutrient-poor TME, supporting their growth, proliferation, and metastasis. The ‘Warburg effect’ and the ‘reverse Warburg effect’ are equally crucial for tumor progression and drug resistance [44]. Yan et al. found that acute myeloid leukemia cells increased mitochondrial ETC activity and energy production by up-regulating the expression of Rho-associated coiled-coil kinase 2 (ROCK2), which contributed to the development of resistance to the chemotherapy drug doxorubicin [45]. Cruz-Bermúdez et al. reported that NSCLC cells enhance their mitochondrial OXPHOS and resistance to cisplatin by transcriptional coactivator peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α). Inhibition of PGC-1α or mitochondrial OXPHOS could enhance chemotherapeutic drug sensitivity [46].

In addition to chemotherapy, it was demonstrated evidence of enhanced reliance on mitochondrial OXPHOS in oncogene-addicted cancer cells manifesting acquired resistance to targeted therapies [47]. Lin et al. discovered that TKI-resistant lung cancer cells enhance their OXPHOS process through the IGF2BP3-COX6B2 pathway, which attenuated the sensitivity of anti-tumor therapy [48]. The animal models have demonstrated that inhibition of the OXPHOS process in cancer cells significantly inhibited the resistance to gefitinib [48]. Waldschmidt JM et al. showed that dabrafenib-resistant BRAF mutated myeloma cells exhibited a metabolic pattern with OXPHOS as the primary energy source. The combination of BRAF inhibitor and IACS-010759, a complex I inhibitor of the mitochondrial ETC, could significantly improve the anti-tumor efficacy [49]. Liu et al. demonstrated that the transcription factor early growth response gene 1 (EGR1) was participated in the resistance of B-cell malignancies to Ibrutinib. EGR1-mediated PDP1 activation, a phosphatase that dephosphorylates and activates the E1 component of the pyruvate dehydrogenase complex, promotes the transition of cancer cells to OXPHOS and increases the ATP production, providing enough energy for the progression of ibrutinib-resistant lymphoma cells [50]. Targeting the OXPHOS process inhibited the proliferation of ibrutinib-resistant lymphoma cells in mouse model [50]. Chen et al. found that METTL1 was up-regulated in anlotinib-resistant oral squamous cell carcinoma (OSCC). Enhanced METTL1 promotes global mRNA translation and stimulates OXPHOS through m7G tRNA modification. Bioenergetic analysis revealed that METTL1 guided the metabolic switch from glycolysis to OXPHOS in anlotinib-resistant cells [51]. Blocking the OXPHOS process attenuated the effect of METTL1 on anlotinib resistance [51].

Proteomic studies combined with metabolic profiling studies revealed that tamoxifen-resistant estrogen receptor α (ER) positive breast cancer cells increase mitochondrial biogenesis by upregulating NQO1 (NAD(P)H dehydrogenase [quinone] 1) and GCLC (glutamate-cysteine ligase catalytic subunit) to assist in drug resistance [52]. Marco Fiorillo et al. demonstrated that FoxO3a (the Forkhead box class O3a) was able to counteract the increased oxygen consumption rate and extracellular acidification rate observed in tamoxifen-resistant breast cancer cells by reducing their energetic activity and glycolytic rate, thereby resensitizing cancer cells to tamoxifen and resulting in inhibited tumor growth [53].

Role of OXPHOS in quiescent tumor cells during drug resistance

Among the rapidly growing tumor cells, a subset of relatively quiescent cells that are resistant to conventional anti-tumor therapy demonstrate metabolism heterogeneity and contribute to tumor recurrence [54]. These relatively quiescent tumor cells are far from blood vessels, and due to the deprivation of glucose nutrients, they are predominantly metabolized by OXPHOS-supplied energy instead of glycolysis [54]. The quiescent breast cancer cells exhibited elevated tumor necrosis factor receptor-associated protein 1 (TRAP1) and an OXPHOS-enhanced phenotype [54]. Saleem et al. demonstrated that CD210, a dual inhibitor of mitochondrial TRAP1 and cytoplasmic heat shock protein90 (HSP90), effectively inhibited both the OXPHOS process in quiescent tumor cells and the glycolytic process in rapidly proliferating tumor cells, resulting in the elimination of doxorubicin-resistant breast cancer cells in the animal model [55].

Role of ATP and downstream metabolites during drug resistance

Significantly elevated ATP in the TME binds to its specific receptor, the purinergic receptor, and rewires tumor cells’ glycolysis, lipid metabolism, and amino acid metabolism to facilitate their growth, metastasis, and drug resistance [56]. Besides, tumor cells enhance the synthesis of immunosuppressive adenosine by up-regulated ectonucleoside triphosphate diphosphohydrolase-1 (CD39) and ecto-5’-nucleotidase (CD73). Accumulation of adenosine in the TME engages in tumor cell progression and immunosuppression by recruiting and activating immunosuppressive Tregs and inhibiting CD8 + T lymphocytes [22]. Adenosine inhibits the anti-tumor efficacy of CD8 + T lymphocytes by inhibiting MHC-I expression with increased IL-10, inhibiting KCa3.1 channels, and interacting with its surface-expressed A2A receptor [22]. Lu et al. reported that the binding of CD39 expressed on macrophages to CD73 expressed on tumor cells elevated adenosine levels in the TME, ultimately leading to anti-PD-1 resistance [57]. The inhibition of adenosine expression or blockade of adenosine-associated signaling pathways improved ICB resistance and enhanced therapeutic efficacy [57–59].

Role of lipid metabolic reprogramming during drug resistance

In addition to glycolysis and OXPHOS, tumor cells take advantage of lipid metabolic reprogramming to develop drug resistance. Lipid metabolism is fundamental to the biological activity of cells, including the provision of energy, the synthesis of membrane, the construction of signaling molecules (such as prostaglandin E2 (PGE2)), and the participation in epigenetics through FA acetylation [60]. In the nutrient-poor TME, the activation of proto-oncogenes and inactivation of tumor suppressor genes enhance glycolysis in the tumor cell as well as enabling them to enter into lipid metabolism and supply sufficient energy through FAO. This process includes increased lipid uptake, enhanced lipogenesis, improved FAO, and enriched lipid storage [61]. Meanwhile, lipid metabolic reprogramming contributes to the construction of an immunosuppressive TME. Targeting the reprogrammed lipid metabolism has the potential to be a prospective intervention to breakthrough anti-tumor resistance and improve therapeutic efficacy.

Role of lipid metabolic reprogramming during chemotherapy resistance

By leveraging the Cancer Genome Atlas dataset, Tadros et al. discovered that lipid metabolism was the most important metabolic pathway associated with poor response to gemcitabine in pancreatic cancer patients [62]. They reported that the increased fatty acid synthase (FASN)-mediated lipogenesis promoted gemcitabine resistance in pancreatic cancer patients. The application of FASN inhibitor orlistat significantly increased the responsiveness to gemcitabine and overcame the resistance to gemcitabine in the animal model of pancreatic cancer [62]. Furthermore, Ma et al. demonstrated that TGFB2 was a key gene in gemcitabine-resistant pancreatic cancer cells, which is post-transcriptionally stabilized by METTL14-mediated modification of m6A, followed by up-regulation of sterol regulatory element binding factor 1 (SREBF1) and its downstream adipogenic enzymes through the PI3K-AKT pathway that promotes lipid accumulation and confers resistance to gemcitabine in pancreatic cancer cells [63]. The pancreatic cancer-resistant mouse model confirmed that the TGFB2 inhibitor, imperatorin, braked through gemcitabine resistance and significantly improved efficacy [63]. Moreover, Cotte et al. demonstrated that lipid synthesis catalyzed by lysophosphatidylcholine acyltransferase 2 contributed to colorectal cancer cell resisting 5-FU and oxaliplatin [64]. Tan et al. reported that cisplatin-resistant ovarian cancer cells displayed enhanced FA uptake, increased FA synthesis, concurrent with diminished glucose uptake, suggesting a shift from glucose- to FA-dependent metabolism [65]. In response to cisplatin-induced oxidative damage, increased FA uptake promoted cancer cell survival by enhancing β-oxidation [65]. They also showed that blocking β-oxidation by erdafitinib, a fibroblast growth factor receptor (FGFR) TKI, could inhibit proliferation and induce apoptosis in cisplatin-resistant cancer cells by decreasing malonyl-CoA production and FA synthesis [65]. Erdafitinib combined with cisplatin or carboplatin synergistically suppressed ovarian cancer proliferation in vitro and in vivo [65].

In addition to increased FA synthesis, uptake, and oxidation, specific fatty acids and their metabolites are engaged in chemoresistance of tumor cells. Tsai et al. demonstrated that glioblastoma increased the expression of prostaglandin-endoperoxide synthase 2 and the synthesis of PGE2 from arachidonic acid through specifcity protein 1 (Sp1), and activated the FAO and TCA cycle in mitochondria to augment ATP production, contributing to the resistance to temozolomide [66]. The blockade of PGE2 action by EP1 (PGE2 receptor) antagonist ONO-8713 exhibited a potentially therapeutic effect on temozolomide-resistant glioblastoma in vivo and in vitro [66]. Besides, several studies have shown that increased PGE2 contributes to cancer cell survival, progression, and immune escape in the TME. On the one hand, PGE2 can activate PGE2 receptors to promote cancer cell proliferation [67, 68]. On the other hand, PGE2 acts as an immunosuppressor in TME, decreasing the activity of dendritic cells, NK cells, and Th1 cells, but stimulating immunosuppressive cells, such as M2 macrophages and Treg cells [69, 70].

Role of lipid metabolic reprogramming during ICB resistance

In addition to chemotherapy resistance, lipid metabolism reprogramming is involved in ICB resistance. It was reported that cholesterol accumulated in tumor cells promoted PD-L1 expression to resist ICB, and that statin hypolipidemic drugs improved the efficacy of ICB therapy [71, 72]. Moreover, Lei et al. found that accumulated cholesterol in tumor cell membranes made ‘cell softening’, subsequently impeding T cell-mediated cell killing, and that depletion of cholesterol to harden tumor cells enhanced the cytotoxicity of TILs and improved the efficacy of ICB in mice [73]. Furthermore, lipid metabolic reprogramming regulates the anti-tumor efficacy of immune cells by altering downstream pathways. The abundance of cholesterol in TME contributes to the accumulation of cholesterol in CD8 + TILs, which consequently induces endoplasmic reticulum stress, leading to T-cell exhaustion and increased expression of immune checkpoints, and reduction of cholesterol in CD8 + T cells enhances the anti-tumor efficacy [74]. Besides, Yuan et al. demonstrated that the low-density lipoprotein receptor (LDLR) promoted the effector function of cytotoxic T lymphocytes by (1) mediating cholesterol uptake involved in T-cell initiation and clonal expansion, and (2) interacting with the T-cell receptor (TCR) complex to regulate their recycling and signaling [75]. However, they found that TME downregulates CD8 + T cell LDLR level and TCR signaling via tumor cell-derived proprotein convertase subtilisin/kexin type 9 (PCSK9) which binds to LDLR and prevents the recycling of LDLR and TCR to the plasma membrane thus inhibits the effector function of CTLs [75]. Additionally, the cholesterol metabolite produced by tumor cells, cholesterol sulfate, is involved in ICB resistance by forming a chemical barrier in the TME that prevents CD8 + T cell infiltration [76].

Besides, bioinformatics study indicated that FASN level is relevant to immune infiltration and ICB response in NSCLC patients and bladder cancer patients [77, 78]. Chen et al. reported that MK1775 inhibited FASN-mediated FA synthesis to inhibit FAO in macrophages through the PI3K/AKT/mTOR pathway [79]. Consequently, MK1775 inhibits tumor proliferation by impairing lipid interactions among tumor cells, macrophages, and CD8 + TILs, thus enhancing the efficacy of anti-PD-1 therapy [79].

Role of lipid metabolic reprogramming during targeted therapy resistance

Antiangiogenic drugs (AAD) are widely applied in anti-tumor therapy, however, drug resistance limits their efficacy [80]. Apart from the lack of oxygen caused by the mismatch between rapid tumor growth and limited vascular oxygen supply, the density of blood vessels surrounding tumor cells under AAD treatment decreased significantly, further exacerbating the lack of oxygen [81]. On the one hand, the hypoxia-exacerbated TME produces cytokines and growth factors to help tumor cells bypass AAD treatment [82]; on the other hand, it alters the composition and efficacy of immune cells, which consequently favoring tumor invasion [83]. Additionally, lipid metabolic reprogramming of tumor cells has an instrumental effect in resistance to targeted therapy. It was reported that AAD-induced tumor hypoxia initiated the FAO process and increased the uptake of free FA that stimulated cancer cell proliferation and the development of drug resistance [84]. Inhibition of carnitine palmitoyl transferase 1 A (CPT1), the critical rate-limiting enzyme for FAO, dramatically interfered with free FA-induced cell proliferation, enhanced the therapeutic efficacy of AAD and potentiated its anti-tumor effects [84]. The combination of CPT1 inhibitor and AAD potently suppressed tumor proliferation [84].

Sorafenib, one of AAD, plays a vital role in anti-tumor therapy of HCC, but drug resistance limits its efficacy. Dai et al. reported that lipid metabolic reprogramming under the down-regulated oxoglutarate dehydrogenase-like (OGDHL), one of the rate-limiting components of the key mitochondrial multi-enzyme OGDH complex, contributed to the resistance of HCC cells to sorafenib, tumor progression, and recurrence [85]. On the one hand, the suppressed OGDHL reduces the activity of OGDHC, increasing the ratio of α-KG to citric acid, subsequently promoting the reductive carboxylation of α-KG for adipogenesis through the reverse TCA cycle in HCC. On the other hand, repressed OGDHL activated the mTORC1 pathway in an α-KG-dependent manner, boosting the transcription of stearoyl-CoA desaturase 1 (SCD1) and FASN for de novo lipogenesis [85]. The overexpression of OGDHL can enhance the sensitivity to sorafenib and improve the efficacy of targeted therapies in HCC [85]. Besides, Ding et al. demonstrated that HCC cells induce lipid metabolism reprogramming and develop resistance to sorafenib by enhancing SCD1 expression through the up-regulated unconventional prefoldin RPB5 interactor (URI) [86].

EGFR-mutated cancer cells up-regulate the expression of multiple key enzymes of lipid metabolism and consequently promote lipid metabolism reprogramming [87, 88]. Zhang et al. demonstrated that EGFR stabilized SCD1 through Y55 phosphorylation, thereby enhancing monounsaturated FA synthesis to promote lung cancer cell growth [87]. Pan et al. have shown that under the treatment of EGFR-TKIs, drug-resistant tumor cells accumulated cholesterol in lipid rafts, which facilitates the interaction between EGFR and Src and results in the reactivation of EGFR/Src/Erk signaling pathway, leading to SP1 nuclear translocation and ERRα re-expression [89]. Subsequently, the re-expression of ERRα maintains tumor cell progression by adjusting the detoxification process of ROS [89]. They also reported that the combination of cholesterol-lowering drug lovastatin and the ERRα inverse agonist XCT790 assisted NSCLC cells to overcome resistance to gefitinib and osimertinib in vivo and in vitro [89]. Besides, Van den Bossche et al. found that cetuximab-resistant head and neck squamous cell carcinoma cells exhibit a peroxisome proliferator-activated receptor alpha (PPARα)-mediated lipid metabolism reprogramming, characterised by enhanced FA uptake and FAO, while glycolysis remains untouched [88]. In the cetuximab-resistant mouse model, the combinatory treatment with the PPARα inhibitor and cetuximab significantly reduced tumor growth compared with anti-EGFR therapy alone [88].

Except for HER2 mutations, abnormal activation of HER2 downstream or bypass signaling pathways, and immune suppression [90], HER-2 positive breast cancer cells become resistant to Trastuzumab by reprogramming their metabolism. Lipid metabolic reprogramming plays an important role in breast cancer cells, especially considering their proximity to adipocytes in a fatty acid-rich environment [91]. It has been reported that HER-2 positive breast cancer cells are resistant to anti-HER2 therapy by enhancing lipid metabolism with increased FA uptake, de novo synthesis, storage, and FAO [92, 93]. Duan et al. demonstrated that secondary trastuzumab-resistant HER-2 positive breast cancer cells potentiated the activity of unsaturated fatty acids (UFAs) such as α-linoleic acid, linoleic acid, arachidonic acid and excessive PGE2 [94]. Especifically, the transcriptional activities of key genes participating in lipid metabolism is regulated by variant promoter H3K27me3 and H3K4me3 modifications in addition to promoter-enhancer interactions [94]. Nandi et al. have shown that the combination of anti-HER2 with the deficiency of CPT1 significantly perturbed breast cancer cell growth, enhanced apoptosis, and reduced lung metastasis [95].

Role of lipid metabolic reprogramming during endocrine therapy resistance

Although ER-positive breast cancer patients have a relatively longer survival time and better prognosis, about one-third of patients are still resistant to endocrine therapy represented by tamoxifen and fulvestrant [96]. The resistance mainly results from ER alpha mutation/amplification, mTOR/AKT pathway activation, cell cycle protein D1 activation, and metabolic reprogramming. ER-positive breast cancer has a heightened dependence on glycolysis compared to other breast cancers [97, 98], and the inhibition of estrogen secretion by endocrine therapy leads to an initial inhibition of glycolysis [99] and a progressive accumulation of lipids [100]. Ahn et al. had shown that endocrine therapy-resistant ER-positive breast cancer cells support their resistance through enhanced metabolism by AMPK-FAO-OXPHOS [96]. The enhanced FAO was associated with CPT1, and CPT1 knockdown or treatment with FAO inhibitors in vivo and in vitro remarkably enhanced the effectiveness of ER-positive breast cancer cells to endocrine therapy [96].

Role of amino acid metabolic reprogramming during drug resistance

Apart from glycolysis, OXPHOS and lipid metabolism, amino acid metabolic reprogramming facilitates the adaptation of tumor cells to the nutrient-poor TME and the development of drug resistance. Targeting amino acid metabolic reprogramming could improve the therapeutic efficacy of tumor cells by overcoming resistance to anti-tumor treatments, including chemotherapy, immunotherapy, and targeted therapy.

Role of glutamine metabolism reprogramming during drug resistance

Glutamine, as the main growth-supporting substrate, not only provides a carbon source but also serves as a nitrogen source for the de novo biosynthesis of nitrogen-containing compounds such as nucleotides and non-essential amino acids [101]. In the hypoxic TME, most metabolically reprogrammed tumor cells cannot enter OXPHOS through glycolysis, but they can utilize glutamine to participate in alternative pathways to help them synthesize energy and biomolecules [102]. The high requirement for glutamine in cancer cells was first proposed by Harry Eagle in the 1950s [103]. Some studies contradicting the Warburg Effect suggest that cancer cells retain functional mitochondria that could meet their energy needs by glutaminolysis or lipolysis, demonstrating unanticipated metabolic heterogeneity [104, 105]. Glutamine not only links to the TCA cycle through the production of α-KG by glutaminase and glutamate dehydrogenase (GLUDs), but also supplies the synthesis of other non-essential amino acids, FAs, and nucleotide.

It was reported that glutamine metabolism reprogramming was correlated with drug resistance in tumor cells [106]. Soumoy et al. demonstrated that melanoma cells with mutant BRAF, NRAS, or cKIT become resistant to targeted drugs through a series of metabolic reprogramming, including glutaminolysis, glycolysis and OXPHOS [107]. Park et al. have shown that pancreatic carcinoma cells build a hypoxic TME and resist chemotherapy by increasing glutamine catabolism, vivo and vitro experiments demonstrated that inhibiting glutamine catabolism can improve the efficacy of chemotherapy [108]. Various lncRNAs, including the abhydrolase domain containing 11 antisense RNA 1, LINC00857, nuclear paraspeckle assembly transcript 1, XLOC_006390, were found to be involved in chemotherapeutic resistance by regulating the expression of amino acid transporters and glutamine metabolizing enzyme, leading to glutamine metabolic reprogramming [106].

Moreover, glutamine metabolic reprogramming in tumor cells is linked to acquired drug resistance and EMT, which interacts with each other to promote tumor cell progression. Wang et al. reported that drug-resistant NSCLC cells highly relied on glutamine utilization with GLUD1 and mainly fluxed into OXPHOS along with higher EMT-mediated migration and invasion capacity [109]. They also proved that GLUD1-derived α-KG generation and the consequent accumulation of ROS mainly induce tumor cell migration and invasion through the activation of Snail. Vivo and vitro experiments demonstrated that the application of the GLUD1 inhibitor, R162, could counteract the drug resistance and EMT-induced cell migration and invasion of NSCLC cells [109].

Role of serine metabolism reprogramming during drug resistance

Serine metabolism reprogramming has been reported as one of the reasons for resistance of tumor cells to various chemotherapeutic drugs and targeted agents, such as gemcitabine, doxorubicin, 5-FU, sorafenib, and erlotinib [110, 111]. Metabolomics analyses have revealed that gemcitabine-resistant bladder cancer cells up-regulate serine synthesis by increasing phosphoglycerate dehydrogenase (PHGDH) expression to facilitate their proliferation [111]. The PHGDH inhibitor, NCT503, impaired synthesis of glucose-derived serine and induced apoptosis in bladder cancer cells [111]. Moreover, Zhang et al. demonstrated that under exposure to doxorubicin, triple-negative breast cancer cells augmented the synthesis of serine by increased PHGDH and subsequently conversion of serine to GSH, which counteracted doxorubicin-induced ROS formation. In mouse models, the inhibition of PHGDH enhanced the responsiveness of breast cancer cells to doxorubicin and improved its anti-tumor efficacy [112]. Besides, PHGDH, as the crucial enzyme for serine synthesis, increased the activity of a protein kinase RNA-like ER kinase (PERK) cascade signaling, which promotes macrophage polarization toward an immunosuppressive M2 type and accordingly contributes to the immune escape [113, 114]. However, Rossi et al. discovered that the loss of PHGDH non-catalytically potentiated cancer dissemination and metastasis [115]. In detail, the inhibited PHGDH reduces its interaction with the glycolytic enzyme phosphofructokinase, and subsequently activates the hexosamine-sialic acid pathway to provide precursors for protein glycosylation, including enhanced sialylation of integrin αvβ3, which augments tumor cell migration and invasion [115].

Role of reduced glutathione (GSH) metabolism repogramming during drug resistance

GSH is produced from glutamine, cysteine and glycine through γ-glutamylcysteine synthetase and glutathione synthetase. Reduced GSH can remove intracellular free radicals and protect cells from oxidative damage. Under long-term oxidative stress, tumor cells produce more ROS than normal cells. Anti-tumor agents produce further oxidative stimulation in cancer cells, causing irreversible damage to cellular lipids, proteins, and DNA, leading to tumor cell death. However, long-term chemotherapeutic drug could enhance the antioxidant system in cancer cells, resulting in their chemoresistance. GSH and its related metabolic enzymes constitute the most critical antioxidant defense system in the organism, protecting the tumor cells from the attack of ROS, which provides conditions for chemoresistance. Additionally, GSH binds to chemotherapeutic drugs entering the cell to form complexes for transferring to the outside of the cell, decreasing the intracellular chemotherapeutic drug concentration and interfering with their DNA damage [116]. Asare-Werehene et al. reported that cisplatin-resistant ovarian cancer cells produced higher GSH than chemosensitive cells [117]. Yu et al. demonstrated that cisplatin-resistant cancer cells maintain their resistance not by increasing DNA damage repair, but by enhancing their reducing potential, such as GSH [118]. This is associated with inactivating KEAP1 mutations or reductions in KEAP1 RNA correlated with downstream Nrf2 activation [118].

Lenvatinib is one of the first-line treatments in advanced HCC. Lu et al. reported that high levels GSH and S-adenosylmethionine (SAM) strengthened the antioxidant defense system in lenvatinib-resistant models [119]. They showed that accumulated lactate under glycolysis induced lysine lactylation of IGF2BP3, subsequently improving the supply of methylated substrates, such as SAM, for m6A methylation of PCK2 and NRF2 mRNAs [119]. The lactylated IGF2BP3-PCK2-SAM-m6A loop sustains increased levels of PCK2 and Nrf2, strengthening redox homeostasis and inducing lenvatinib resistance in HCC [119]. Targeting IGF2BP3 or glycolysis inhibitor 2-DG could rescue the sensitivity of HCC cells to lenvatinib in vivo [119]. Wang et al. showed that EGFR-mutant NSCLC cells, exposed to EGFR TKIs, exhibit reprogramming of branched-chain amino acid mediated by H3K9 demethylation, leading to elevated GSH and decreased ROS, consequently inducing drug resistance [120]. ROS-inducing drugs combined with TKI defeated this drug resistance in mouse models [120]. Furthermore, Zhang et al. demonstrated that AKR1B1 interacted with and activated STAT3 to up-regulate the cystine transporter solute carrier family 7 member 11 (SLC7A11) in TKI-resistant NSCLC cells, which results in increased cystine uptake and transport to the GSH biosynthesis, ROS detoxification, and TKI drug resistance [121]. The anti-diabetic drug epalrestat, as an AKR1B1 selective inhibitor, could reverse the efficacy of EGFR TKIs in resistant cells and delayed resistance in lung cancer patient-derived xenograft mice [121].

Role of arginine metabolism repogramming during drug resistance

Fang et al. reported that the chemoresistant oesophageal squamous cell carcinoma (OSCC) activated the insulin-like growth factor-I receptor (IGF1R) pathways and up-regulated the expression of metabolic enzymes argininosuccinate synthase (ASS1) and pyrroline-5-carboxylate reductase 1 (PYCR1) through c-MYC [122]. Mechanistically, elevated ASS1 potentiates arginine metabolism involved in biosynthesis, while PYCR1 increases proline metabolism to maintain redox balance, enabling OSCC cells to proliferate under cisplatin treatment, and the IGF1R inhibitor linsitinib potentiated the anti-tumor effectiveness of cisplatin [122].

Role of ferroptosis during drug resistance

In addition to the altered GSH under amino acid metabolic reprogramming is involved in evading ferroptosis, it has been reported that reprogrammed glycolysis, OXPHOS, and lipid metabolism could help tumor cells to maintain intracellular redox homeostasis and evade ferroptosis under anti-tumor treatments [123–125]. It seems that ferroptosis occupies a central position in the metabolic reprogramming process of anti-tumor resistance. Ferroptosis is a recently discovered form of regulated cell death characterised by excessive accumulation of iron within cells, triggering lipid peroxidation, leading to an imbalance in intracellular redox homeostasis and inducing oxidative modification of phospholipid membranes, subsequently resulting in cellular membrane damage and cell death [126, 127]. Specifically, acyl-CoA synthetase long chain family (ACSL) catalyzes the binding of polyunsaturated fatty acids (PUFAs), such as arachidonic acid, to coenzyme A to produce the corresponding PUFAs-CoA, which then is integrated into phospholipids by lysophosphatidylcholine acyltransferase 3 (LPCAT3), forming phospholipids PUFAs (PUFAs-PLs). PUFAs-PLs undergo a series of peroxidation reactions to produce phospholipid hydroperoxides, which subsequently initiate ferroptosis. Extrinsic and intrinsic pathways can activate ferroptosis. The extrinsic pathway is triggered by the suppression of cell membrane transporters such as the cystine/glutamate transporter (also known as system xc-, consisting of transporter SLC7A11 and transmembrane regulator SLC3A2) or by activation of the iron transporters serotransferrin and lactotransferrin) [128]. The intrinsic pathway is initiated by the inhibition of antioxidant enzymes (such as GSH peroxidase 4 (GPX4)) [128].

The metabolic reprogramming of cancer cells facilitates their drug resistance by evading ferroptosis. Studies have shown that enhancing ferroptosis is effective in overcoming resistance to anti-tumor therapies, including targeted therapy, chemotherapy and immunotherapy [129]. Poursaitidis et al. reported that EGFR-mutant NSCLC cells are highly dependent on cysteine and sensitive to ferroptosis induced by SLC7A11 inhibition or cysteine predation [130]. Yang et al. showed that ferroptosis-resistant HCC cells up-regulated the expression of S100 calcium-binding protein P (S100P), facilitated lysosomal degradation of acetyl-CoA carboxylase alpha, which is indispensable for de novo biosynthesis of lipids, thus down-regulated lipid biosynthesis, inhibited ferroptosis, promoted tumor progression, and developed resistance to sorafenib [131]. The removal of S100P significantly inhibited HCC progression in the animal experiment [131]. Pan et al. demonstrated that pancreatic ductal adenocarcinoma (PDAC) activated the transcription factor YY1 which regulates family with sequence similarity 60 member A (FAM60A) expression, leading to downstream suppression of PPAR and ACSL1/4 and activation of GPX4 pathways, protecting it escaping from ferroptosis, promoting tumor progression and chemoresistance [132]. FAM60A knockdown sensitized PDAC cells to gemcitabine treatment [132].

Furthermore, ferroptosis in tumor cells is involved in immunoregulation in the TME. On the one hand, ferroptosis in tumor cells could trigger or activate anti-tumor immunity [129]. On the other hand, CD8 + T cells promote cancer cell ferroptosis by secreting interferon gamma (IFNγ), (1) inhibiting SLC7A11 and SLC3A2 expression to limit cystine uptake [133], and (2) enhancing ACSL4 expression to promote PUFA-PLs [134]. However, TME inhibits ferroptosis through metabolic reprogramming to help tumor cells resist ICB. Freitas-Cortez et al. demonstrated that anti-PD-1-resistant tumor cells up-regulate fatty acid-binding protein 7 (Fabp7), which lowers the transcription of ferroptosis-inducing genes like Lpcat3 but raises the transcription of ferroptosis-protective genes such as Bmal1, triggering defensive metabolic reprogramming that protect cells from ferroptosis induced by CD8 + T lymphocyte and resist to anti-tumor immunotherapy [135]. Additionally, high-expressed Fabp7 in CD8 + T cells could interfere with the expression of circadian clock genes and induce apoptosis through p53 stabilization [135]. Targeting Fabp7 could enhance immunotherapy effectiveness by re-sensitizing anti-PD-1-resistant tumors to ferroptosis with increased CD8 + T cell infiltration [135]. Lin et al. reported that tumor cells uptake itaconate via SLC13A3, an itaconate transporter, from tumor-associated macrophages (TAMs), thereby activating the NRF2-SLC7A11 pathway and escaping from immune-mediated ferroptosis [136]. They also showed that treatment with the SLC13A3 inhibitor sensitizes tumor cells to ferroptosis, curbs tumor progression, and bolsters ICB effectiveness [136]. Furthermore, Ma et al. found that CPT1A interacts with L-carnitine from TAMs to trigger ferroptosis resistance and CD8 + T cell exhaustion in lung cancer [137]. CPT1A inhibits the ubiquitination and degradation of c-Myc to enhance its own expression, and this CPT1A/c-Myc positive feedback loop strengthens the cellular antioxidant ability by triggering the NRF2/GPX4 pathway and decreases the level of PUFAs-PLs through suppressing ACSL4, thereby avoiding ferroptosis [137]. Targeting CPT1A potentiated the anti-tumor response of ICB through enhanced ferroptosis in tumor-bearing mice [137]. Additionally, macrophages can engulf ferroptotic cells through Toll-like receptor 2. However, Luo et al. discovered that under exposure to ferroptosis therapy, phospholipid peroxidation in macrophages weakened their phagocytosis ability to eradicate ferroptotic tumor cells, eventually generating anti-tumor resistance [138]. Besides, it was reported that neutrophils induce ferroptosis in tumor cells by transferring myeloperoxidase-containing granules [139]. And CAFs inhibit tumor cell ferroptosis by up-regulating DLEU1 or secreting miR-522 [140, 141].

Discussion

Tumor metabolic reprogramming, genetic signatures, immune regulation, and the stromal microenvironment regulate each other and together form a complex and tightly interconnected network that engages in tumor cell proliferation, immune escape, drug resistance, angiogenesis, immortality, invasion, and metastasis. Under intrinsic or extrinsic pressure, tumor cells gradually adapt to the altering TME through metabolic reprogramming (Fig. 1) and changing genetic characteristics (including proto-oncogenes, tumor suppressor genes, and epigenetics). Anti-tumor drugs could inhibit these metabolic reprogramming processes of tumor cells and restrain their proliferation (Fig. 2). Under the anti-tumor treatments, tumor cells gradually reprogram their metabolism to help them avoid cytotoxicity and develop drug resistance (Fig. 3). The process of metabolic reprogramming in these drug-resistant tumor cells is heterogeneous, dynamic and complex. Some resist the killing effects of anti-tumor drugs by enhancing glycolysis, some by heightening lipid metabolism, some by boosting amino acid metabolism, some by strengthening intracellular reducing potential through the inhibition of ferroptosis, and some even by potentiating OXPHOS which is contrary to the traditional Warburg effect. Moreover, the process of tumor cells resisting anti-tumor treatment is a dynamic process, and the mechanisms and metabolic reprogramming processes of early resistance and late resistance are different [142]. Furthermore, this metabolic reprogramming process may involve multiple metabolic alterations of various macromolecules, not a single molecule or even a particular type of metabolism. In this paper, we reviewed the metabolic reprogramming process of various drug-resistant tumor cells under chemotherapy, immunotherapy, targeted therapy, and endocrine therapy, as well as summarized current preclinical studies (Table 1) of targeting metabolic therapies, in an attempt to break the bottleneck of anti-tumor resistance and prolong the survival time of cancer patients from the perspective of metabolic reprogramming.

Fig. 1.

The metabolic reprogramming process of tumor cells in the nutrient-deficient TME

Fig. 2.

The metabolic reprogramming process of tumor cells under anti-tumor therapy

Fig. 3.

The metabolic reprogramming process of drug-resistant tumor cells

Table 1.

The preclinical studies of targeting metabolic therapies to overcome anti-tumor resistance

Metabolic pathway	Drug resistance	Mechanism	Key Finding	References	Year
Glycolysis	Chemotherapy resistance	The high-expressed HIF-1α confers drug resistance to tumor cells through the glycolytic pathway	Inhibiting HIF-1α could increase the sensitivity of resistant ovarian cancer cells to cisplatin by shifting the aerobic glycolysis to OXPHOS.	[14]	2016
		HIF-1α stabilization mediated through MUC1 regulated an enhanced glycolytic flux, enhancing the intrinsic levels of dCTP which could reduce the adequate level of gemcitabine through molecular competition	The combination therapy of inhibiting HIF-1a or pyrimidine biosynthesis with gemcitabine could significantly reduce tumor load	[16]	2017
		1)Gemcitabine-resistant urothelial carcinoma cells weakened the cytotoxic effect of chemotherapy by increasing dCTP production. 2)Increased aerobic glycolytic with PPP processes in gemcitabine-resistant tumor cells, companied with increasing NADPH and GSH production by IDH2, helping tumor cells become cross-resistant to cisplatin	Down-regulation or pharmacological inhibition of IDH2 restores chemosensitivity of gemcitabine in the murine urothelial carcinoma xenograft model.	[17]	2023
	ICB resistance	Up-regulated aerobic glycolysis in tumor cells promoted the dissociation of HK2 from mitochondria and bound to cytosolic IkBα in human glioblastoma cells. HK2 acts as a protein kinase and phosphorylates IkBα, resulting in IkBα degradation and NF-κB activation-dependent PD-L1 expression for tumor immune escape	The HK inhibitor combined with the anti-PD-1 could counteract tumor immune escape and enhance the anti-tumor efficacy of ICB.	[25]	2022
		Tumor-intrinsic glycolysis pathway confers resistance to T cell-mediated killing. Inhibition glycolysis by targeting Glut1 led to the accumulation of intracellular ROS, which sensitized tumor cells to T cell-mediated bystander killing through TNF-α	Treatment with the Glut1-specific inhibitor enhance tumor sensitivity to ICB and synergized with anti-PD-1 therapy in mouse models.	[26]	2023
		Treg cells uptake lactic acid in the highly glycolytic TME via MCT1 and robustly express PD-1, resulting in the impairment of PD-1 blockade therapy	Targeting LDHA and MCT1 potentiated the effectiveness of ICB in treating intrahepatic tumors.	[28]	2022
		SAA could activate glycolysis through the LDHA/STAT3 pathway, increase the expression f PD-L1 on neutrophils, and release oncostatin M, thus impairing T cell function in HCC	The inhibition of STAT3 or SAA alleviated neutrophil-mediated immunosuppression and augmented the anti-tumor efficacy of ICB in vivo.	[29]	2024
Mitochondrial Metabolic Reprogramming	Targeted therapy resistance	TKI-resistant lung cancer cells enhance their OXPHOS process through the IGF2BP3-COX6B2 pathway, which attenuated the sensitivity of anti-tumor therapy	The inhibition of the OXPHOS process in the cancer cells significantly inhibited gefitinib resistance.	[48]	2023
		EGR1-mediated PDP1 activation, a phosphatase that dephosphorylates and activates the E1 component of the pyruvate dehydrogenase complex, promotes the transition of cancer cells to OXPHOS and increases the ATP production, providing enough energy for the progression of ibrutinib-resistant lymphoma cells	Targeting the OXPHOS process inhibited the proliferation of ibrutinib-resistant lymphoma cells in mouse model.	[50]	2023
	Quiescent tumor cells	The quiescent breast cancer cells exhibited elevated TRAP1 and an OXPHOS-enhanced phenotype	CD210, a dual inhibitor of mitochondrial TRAP1 and cytoplasmic HSP90, effectively inhibited both the OXPHOS process in quiescent tumor cells and the glycolytic metabolic process in rapidly proliferating tumor cells, resulting in the elimination of doxorubicin-resistant breast cancer cells in the animal model.	[55]	2025
Lipid Metabolic Reprogramming	Chemotherapy resistance	Increased FASN-mediated lipogenesis promotes gemcitabine resistance in pancreatic cancer patients	The application of FASN inhibitor orlistat significantly increased the responsiveness to gemcitabine and overcame the resistance to gemcitabine in animal models of pancreatic cancer.	[62]	2017
		TGFB2 is post-transcriptionally stabilized by METTL14-mediated modification of m6A, followed by up-regulation of SREBF1 and its downstream adipogenic enzymes through the PI3K-AKT pathway that promotes lipid accumulation and confers resistance to gemcitabine in pancreatic cancer cells	The pancreatic cancer-resistant mouse model confirmed that the TGFB2 inhibitor, imperatorin, breaks through gemcitabine resistance and significantly improves efficacy.	[63]	2024
		Cisplatin-resistant ovarian cancer cells display enhanced FA uptake, increased FA synthesis, concurrent with diminished glucose uptake, suggesting a shift from glucose to FA-dependent anabolic and energy metabolism. The increased FA uptake facilitates cancer cell survival under cisplatin-induced oxidative stress by enhancing β-oxidation	Blocking β-oxidation by erdafitinib, a FGFR TKI, combined with cisplatin or carboplatin synergistically suppresses OC proliferation in vitro and growth of patient-derived xenografts in vivo.	65]	2022
		Glioblastoma increased the expression of prostaglandin-endoperoxide synthase 2 and the synthesis of PGE2 from arachidonic acid through sp1, and thereby activated the FAO and TCA cycle in mitochondria to augment ATP production, contributing to the resistance to temozolomide	The blockade of PGE2 action by EP1 (PGE2 receptor) antagonist ONO-8713 exhibits a potentially therapeutic effect on temozolomide-resistant glioblastoma in vivo and in vitro.	[66]	2022
	Targeted therapy resistance	The suppressed OGDHL reduces the activity of OGDHC, increasing the ratio of α-KG to citric acid, subsequently promoting the reductive carboxylation of α-KG for adipogenesis through the reverse TCA cycle in HCC. And repressed OGDHL activated the mTORC1 pathway in an α-KG-dependent manner, boosting the transcription of SCD1 and FASN for de novo lipogenesis	The overexpression of OGDHL can enhance the sensitivity to sorafenib and improve the efficacy of targeted therapies in HCC.	[85]	2020
		Cetuximab-resistant head and neck squamous cell carcinoma cells exhibit a PPARα-mediated lipid metabolism reprogramming, characterised by enhanced FA uptake and FAO	The combinatory treatment with the PPARα inhibitor and cetuximab led to a greater reduction of tumor growth compared to anti-EGFR therapy alone in the cetuximab-resistant mice model.	[88]	2025
		Under the treatment of EGFR-TKIs, drug-resistant tumor cells accumulated cholesterol in lipid rafts, which facilitates the interaction between EGFR and Src and results in the reactivation of EGFR/Srk signaling pathway, leading to SP1 nuclear translocation and ERRα re-expression. Subc/Ersequently, the re-expression of ERRα maintains tumor cell progression by adjusting the detoxification process of ROS	The combination of cholesterol-lowering drug lovastatin and the ERRα inverse agonist XCT790 assisted NSCLC cells to overcome resistance to gefitinib and osimertinib in vivo and in vitro.	[89]	2022
		HER-2 positive breast cancer cells are resistant to anti-HER2 therapy by enhancing lipid metabolism with increased FA uptake, de novo synthesis, storage, and FAO	The combination of anti-ErbB2 with the deficiency of Cpt1a significantly perturbed breast cancer cell growth, enhanced apoptosis, and reduced lung metastasis.	[95]	2024
	Endocrine Therapy Resistance	The endocrine therapy-resistant ER-positive breast cancer cells could support their resistance through enhanced metabolism by AMPK-FAO-OXPHOS	CPT1 knockdown or treatment with FAO inhibitors in vivo and in vitro remarkably enhanced the effectiveness of ER-positive breast cancer cells to endocrine therapy.	[96]	2024
Animo Acid Metabolic Reprogramming	Glutamine	Pancreatic carcinoma cells build a hypoxic TME and resist chemotherapy by increasing glutamine catabolism	Vivo and vitro experiments demonstrated that inhibiting glutamine catabolism can improve the efficacy of chemotherapy.	[108]	2023
		Drug-resistant NSCLC cells highly relied on glutamine utilization with GLUD1 and mainly fluxed into OXPHOS along with higher EMT-mediated migration and invasion capacity. GLUD1-derived α-KG generation and the consequent accumulation of ROS induce tumor cell migration and invasion through the activation of Snail	Vivo and vitro experiments demonstrated that the application of the GLUD1 inhibitor, R162, could counteract the drug resistance and EMT-induced cell migration and invasion of NSCLC cells.	[109]	2022
	Serine	Under exposure to doxorubicin, triple-negative breast cancer cells augmented the synthesis of serine by increased PHGDH and subsequently conversion of serine to GSH, which counteracted doxorubicin-induced ROS formation	In mouse models, the inhibition of PHGDH enhanced the responsiveness of breast cancer cells to doxorubicin and improved its anti-tumor efficacy.	[112]	2016
	Reduced GSH	Accumulated lactate under glycolysis induced lysine lactylation of IGF2BP3, subsequently improving the supply of methylated substrates, such as SAM, for m6A methylation of PCK2 and NRF2 mRNAs. The lactylated IGF2BP3-PCK2-SAM-m6A loop sustains increased levels of PCK2 and NRF2, strengthening redox homeostasis and inducing lenvatinib resistance in HCC	Treatment with liposomes carrying siRNAs targeting IGF2BP3 or the glycolysis inhibitor 2-DG restored lenvatinib sensitivity in vivo.	[119]	2024
		EGFR-mutant NSCLC cells exposed to EGFR TKIs exhibit reprogramming of branched-chain amino acid mediated by H3K9 demethylation, leading to elevated GSH and decreased ROS, consequently inducing drug resistance	ROS-inducing drugs combined with TKI defeated this drug resistance in mouse models.	[120]	2019
		AKR1B1 activated STAT3 to up-regulate the cystine transporter SLC7A11 in TKI-resistant NSCLC cells, which results in increased cystine uptake and transport to the GSH biosynthesis, ROS detoxification, and TKI drug resistance	The anti-diabetic drug epalrestat, as an AKR1B1 selective inhibitor, could reverse the efficacy of EGFR TKIs in resistant cells and delayed resistance in lung cancer patient-derived xenograft mice.	[121]	2021
	Arginine	The chemoresistant OSCC activated theIGF1R pathways and promoted the expression of metabolic enzymes ASS1 and PYCR1 through c-MYC. Elevated ASS1 potentiates arginine metabolism involved in biosynthesis, while PYCR1 increases proline metabolism to maintain redox balance, enabling OSCC cells to proliferate under cisplatin treatment	The IGF1R inhibitor linsitinib potentiated the anti-tumor effectiveness of cisplatin.	[122]	2023
Ferroptosis		Ferroptosis-resistant HCC cells up-regulated the expression of S100P, facilitated lysosomal degradation of acetyl-CoA carboxylase alpha, which is indispensable for de novo biosynthesis of lipids, thus down-regulated lipid biosynthesis, inhibited ferroptosis, promoted tumor progression, and developed the resistance to sorafenib	The removal of S100P significantly inhibited HCC progression in the animal experiment.	[131]	2025
		PDAC activated the transcription factor YY1 which transcriptionally regulates FAM60A expression, leading to downstream suppression of PPAR and ACSL1/4 and activation of GPX4 pathways, in order to protect it escaping from ferroptosis, promoting tumor progression and chemoresistance	FAM60A knockdown sensitized PDAC cells to gemcitabine treatment.	[132]	2024
		CD8 + T cells promote cancer cell ferroptosis by secreting IFNγ which inhibit SLC7A11 and SLC3A2 expression to limit cystine uptake	The depletion of cystine in combination with ICB synergistically enhanced T cell-mediated anti-tumor immunity and induced ferroptosis in tumor cells in mouse models.	[133]	2019
		Anti-PD-1-resistant tumor cells up-regulate Fabp7, which lowers the transcription of ferroptosis-inducing genes like Lpcat3 but raises the transcription of ferroptosis-protective genes such as Bmal1, triggering defensive metabolic reprogramming that protect cells from ferroptosis induced by CD8 + T lymphocytes and evade anti-tumor immunotherapy. Furthermore, cancer cells increase Fabp7 expression in CD8 + T cells, disrupting circadian clock gene expression and triggering apoptosis through p53 stabilization	Targeting Fabp7 could enhance immunotherapy effectiveness by re-sensitizing anti-PD-1-resistant tumors to ferroptosis with increased CD8 + T infiltration.	[135]	2025
		Tumor cells uptake itaconate via SLC13A3, an itaconate transporter, from TAMs, thereby activating the NRF2-SLC7A11 pathway and escaping from immune-mediated ferroptosis	Treatment with the SLC13A3 inhibitor sensitizes tumor cells to ferroptosis, curbs tumor progression, and bolsters ICB effectiveness.	[136]	2024
		CPT1A interacts with L-carnitine from TAMs to trigger ferroptosis resistance and CD8 + T cell exhaustion in lung cancer. CPT1A inhibits the ubiquitination and degradation of c-Myc to enhance its own expression, and this CPT1A/c-Myc positive feedback loop strengthens the cellular antioxidant ability by triggering the NRF2/GPX4 pathway and decreases the level of PUFAs-PLs through suppressing ACSL4, thereby avoiding ferroptosis	Targeting CPT1A potentiated the anti-tumor response of ICB through enhanced ferroptosis in tumor-bearing mice.	[137]	2024

In the nutrient-deficient TME, tumor cells promote rapid growth through metabolic reprogramming, shifting metabolism from OXPHOS to glycolysis and lipid metabolism, and helping them escape immune surveillance by depriving immune cells of nutrients and suppressing their anti-tumor immune efficacy.

Anti-tumor therapeutic agents inhibit the glycolytic process of tumor cells and enhance anti-tumor immunity.

Under the anti-tumor treatments, tumor cells gradually reprogram their metabolism to help them avoid cytotoxicity and develop drug resistance. The process of metabolic reprogramming in these drug-resistant tumor cells is complex. Some resist the cytotoxic effects of anti-tumor agents by enhancing glycolysis, some by heightening lipid metabolism, some by boosting amino acid metabolism, some by strengthening intracellular reducing potential through the inhibition of ferroptosis, and some even by potentiating OXPHOS which is contrary to the traditional Warburg effect. Additionally, tumor cells acquire drug resistance by suppressing the efficacy of immune cells.

Although preclinical studies and clinical trials have been attempted, the heterogeneity of tumor cells, drug toxicity and side effects make targeting metabolic reprogramming therapy limited. Current clinical trials are mostly focused on phase I regarding safety and tolerability [143, 144], and there is a lack of breakthrough findings. Finding effective metabolic reprogramming biomarkers and appropriate supplementary therapies may provide more precise treatment for drug-resistant cancer patients, and effectively prolong the survival time of cancer patients.

Researches have shown that exosomes carry important bioactive substances that facilitate the drug resistance of tumor cells by reprogramming their metabolism [145]. The current research has focused on exosomes as a potential target for improving anti-tumor resistance, which deserves further investigation in the future. In the complex TME, on the one hand, exosomes secreted by tumor cells can reshape the metabolism of stromal cells and promote tumor proliferation; on the other hand, exosomes derived from tumor-associated stromal cells contribute to the growth of tumor cells by regulating their metabolism [146]. Furthermore, it was reported that exosomes derived from drug-resistant tumor cells could migrate to drug-sensitive tumor cells, reprogramming their metabolism and facilitating drug resistance in susceptible tumor cells [147]. Deng et al. demonstrated that circ_0001610, derived from the exosomes of oxaliplatin-resistant cells, promoted colon cancer cells to reshape cellular metabolism towards OXPHOS and conferred them stem-like properties, leading to chemoresistance [148]. Besides, exosomes associated with tumor-associated immune cells, particularly macrophages, are engaged in the metabolic reprogramming and drug resistance of tumor cells [149].

Besides, stem cells in tumor cell populations confer resistance to chemotherapy and immunotherapy through metabolic reprogramming. Inhibition of stem cells or cancer stemness might further improve the efficacy of targeting metabolic therapy. He J et al. demonstrated that cytotoxic chemotherapy reprogrammed metabolism by inhibiting the expression of Glutathione S-transferases Mu (GSTM) family members and recruiting breast cancer stem cells (BCSCs), leading to the recurrence and metastasis of triple-negative breast cancers [150]. Xu et al. reported that the inhibition of tyrosine phosphatase SHP-1 in leukaemia stem cells increased their glycolysis and OXPHOS with the upregulation of phosphofructokinase platelet through the AKT-β-catenin pathway, consequently enhancing the sensitivity of acute myeloid leukaemia cells to chemotherapy and vulnerability to immunosurveillance [151]. Furthermore, autophagy [152] and HSPs [153] are involved in the metabolic reprogramming of tumor cells, supporting their adaptation to the nutrient-poor TME and the development of drug resistance. Targeting autophagy and HSPs may also be attempted to improve the efficacy of targeting metabolic therapies and deserves further exploration.

Abbreviations

TME

Tumor microenvironment

OXPHOS

Oxidative phosphorylation

PPP

Pentose phosphate pathway

TCA

Tricarboxylic acid

HIF-1α

Hypoxia inducible factor1

ROS

Reactive oxygen species

LDHA

Lactate dehydrogenase

GRP78

Glucoseregulated proteins 78

dCTP

Deoxycytidine triphosphate

IDH2

Isocitrate dehydrogenase 2

NADPH

Nicotinamide adenine dinucleotide phophate

GSH

Glutathione

AKR1B10

Aldo-keto reductase family 1 B10

FAO

Fatty acid oxidation

m6A

N6-methyladenosin

5-FU

5-fluorouracil

METTL3

Methyltransferaselike 3

ICB

Immune checkpoint blockade

PD-1

Programmed death1

PD-L1

Programmed death ligand1

HK2

Hexokinase2

IkBα

Inhibitor of nuclear factorB alpha

NF-κB

Nuclear factorB

TNF-α

Tumor necrosis factoralpha

PI3K

Phosphatidylinositol 3 kinase

AKT

Protein kinase B

mTOR

Mammalian target of rapamycin

Treg

Regulatory T

MCT1

Monocarboxylate transporter 1

SAA

Serum amyloid A

STAT3

Signal transducer and activator of transcription 3

HCC

Hepatocellular carcinoma

EGFR

Epidermal growth factor receptor

TKI

Tyrosine kinase inhibitors

NSCLC

Nonsmall cell lung cancer

GLUTs

Glucose transporters

PFK-1

Phosphofructokinase1

BCAA

Branchedchain amino acid

α-KG

Alphaketoglutarate

CAFs

Cancerassociated fibroblasts

HGF

Hepatocyte growth factor

c-MET

Cellularmesenchymal to epithelial transition factor

EGF

Epidermal growth factor

EMT

Epithelial mesenchymal transition

ADT

Androgen deprivation therapy

ATP

Adenosine triphosphate

acetyl-CoA

Acetyl coenzyme A

ETC

Electron transport chain

FA

Fatty acids

RCC

Renal clear cell carcinoma

NOX4

NADPH oxidase isoform

PCAF

P300/CBP-associated factor

PKM2

Pyruvate kinaseM2 isoform

BMSCs

Bonemarrow stromal cells

SSP

Serine synthesis pathway

PGC-1α

Peroxisome proliferatoractivated receptor gamma coactivator1 alpha

ROCK2

Rho-associated coiled-coil kinase 2

MCL

Mantle cell lymphoma

DLBCL

Diffuse large Bcell lymphoma

OSCC

Oral squamous cell carcinoma

AI

Aromatase inhibitor

ER

Estrogen receptor

NQO1

NAD(P)H dehydrogenase [quinone] 1

GCLC

Glutamatecysteine ligase catalytic subunit

FoxO3a

The Forkhead box class O3a

mTORC1

Mammalian target of rapamycin complex 1

TRAP1

Tumor necrosis factor receptorassociated protein 1

HSPs

Heat shock proteins

PGE2

Prostaglandin E2

FASN

Fatty acid synthase

FGFR

Fibroblast growth factor receptor

SREBF1

Sterol regulatory element binding factor 1

Sp1

Specifcity protein 1

LDLR

Lowdensity lipoprotein receptor

TCR

T-cell receptor

PCSK9

Proprotein convertase subtilisinkexin type 9

AAD

Antiangiogenic drugs

CPT1

Carnitine palmitoyl transferase 1 A

OGDHL

Oxoglutarate dehydrogenaselike

SCD1

StearoylCoA desaturase 1

URI

Unconventional prefoldin RPB5 interactor

UFAs

Unsaturated fatty acids

HNSCC

Head and neck squamous cell carcinoma

PPARα

Peroxisome proliferatoractivated receptor alpha

GLUD

Glutamate dehydrogenase

SAM

S-adenosylmethionine

SLC7A11

Solute carrier family 7 member 11

PHGDH

Phosphoglycerate dehydrogenase

PERK

Protein kinase RNAlike ER kinase

IGF1R

Insulinlike growth factorI receptor

ASS1

Argininosuccinate synthase

PYCR1

Pyrroline5carboxylate reductase 1

ACSL

AcylCoA synthetase long chain family

PUFAs

Polyunsaturated fatty acids

LPCAT3

Lysophosphatidylcholine acyltransferase 3

PUFAs-PLs

Phospholipids polyunsaturated fatty acids

GPX4

Glutathione peroxidase 4

S100P

S100 calcium-binding protein P

PDAC

Pancreatic ductal adenocarcinoma

FAM60A

Family with sequence similarity 60 member A

IFNγ

Interferon gamma

Fabp7

Fatty acidbinding protein 7

TAMs

Tumorassociated macrophages

TILs

Tumorinfiltrating lymphocytes

MDSCs

Myeloidderived suppressor cells

Teff

Effector T cells

Tm

Memory T cells

SREBP-1

Sterol regulatory elementbinding protein 1

CSF-1R

Colony stimulating factor1 receptor

PSCs

Pancreatic stellate cells

GM-CSF

Granulocytemacrophage colonystimulating factor CXCL12 CXC motif ligand 12

CTLA-4

Cytotoxic Tlymphocyteassociated protein 4

ITGB4

Integrin b4

GSTM

Glutathione Stransferases Mu

BCSCs

Breast cancer stem cells

Author contributions

ZYB and YY conceived this review. This review were performed by XYL, WYQ, FQ, YY and ZYB collectively. XYL, WYQ and FQ wrote the manuscript. ZYB and YY polished this review. All authors read and approved the final manuscript.

Funding

This work was financially supported by the Joint General Program project of Liaoning Province, China (Grant No. 2023-MSLH-134).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This article does not contain any studies with human participants or animals performed by any of the authors.

Competing interests

The authors declare no competing interests.