HKII is considered as the first rate-limiting enzyme in the glycolytic pathway (73). HKII was shown to be upregulated in breast cancer tissues and found to contribute to PTX resistance (74, 75). Even if the major role of GLUT-1 and HKII in cellular metabolism is highly documented their expression after DOX or taxane treatment has not been systematically evaluated. Based on the studies of Zhou R et al. several chemotherapeutic agents may have a significant and direct effect on GLUT-1 and HKII expression (76). Of the HKII inhibitors 3-bromopyruvate (3-BrPA), and lonidamine (LND) have been used in preclinical experiments (36). 3-BrPA enhances drug accumulation by inactivating ABC transporters restoring the cytotoxic effects of DOX. It was shown that 3-BrPA used in combination with DOX significantly inhibited the growth of subcutaneous tumors in multiple myeloma mice (36, 63). The anticancer effect of 3-BrPA was also demonstrated on hepatocellular carcinoma both in in vitro and in in vivo studies and consequently, this drug has been approved by the FDA (63). Inhibition of HKII by LND enhanced the effects of DOX in rituximab-resistant lymphoma cell lines (RRCL) and used in combination with platinum and paclitaxel presented a good tolerability (77, 78).

However metformin, one of the most frequently used drugs in the treatment of type II diabetes mellitus demonstrated antitumor activity combined with or following other therapeutic agents in vitro and in vivo. The detailed mechanism of action of metformin in tumors is not fully elucidated but based on recent studies is mediated through regulation of AMP kinase (AMPK)/mammalian target of rapamycin (mTOR) and insulin/IGF-1 signaling pathways (65). Studies present several aspects of metformin activity in breast carcinomas like metformin downregulation of protein and lipid synthesis, modulation of mitochondrial respiration, induction of stem cell death etc. have been presented and discussed. The authors also outline that many of these anti-cancer effects are molecular subtype-specific being most potent in triple-negative breast carcinomas (79, 80). Nonetheless a study performed on 320,000 persons diagnosed with incident diabetes mellitus has not found any association between metformin treatment and the incidence of major cancers (81).

The first step of the pentose phosphate pathway (PPP) is catalyzed by the G6PD enzyme. A relatively high number of studies have interrogated the role of G6PD in different cancer types but how the overexpression of G6PD contributes to the development of different tumors is largely unknown (82). Yang HC et al. in a very recent study have demonstrated that deregulated G6PD status and oxidative stress are highly related and are involved in cancer progression (83). Another study analyzing MDA-MB-231 cells implanted as orthotopic xenografts has found that the loss of G6PD modestly decreased primary site growth and the ability of breast cancer cells to colonize the lung (84). A relatively high number of G6PD inhibitors were documented in the last years but sometimes with contradictory results. 6-aminonicotinamide, was found to decrease mammosphere formation and aldehyde dehydrogenase activity (85, 86). Ghergurovich JM et al. presented that the most known G6PD antagonist, dehydroepiandrosterone, does not inhibit robustly G6PD in cells and as a consequence they have identified a small molecule (G6PDi-1) that more effectively inhibits G6PD and later the PPP pathway (68).

Our understanding of PFK enzymes and their roles in cancer has developed significantly in the last years. The formation of fructose-1,6-bisphosphate from fructose-6-phosphate is catalyzed by PFK-1 in an irreversible step and therefore PKF-1 is considered as one of the most important rate-limiting enzyme in the process of glycolysis and is regulated by several effectors like fructose 2,6-bisphosphate (F2,6-BP), AMP and ATP. Moreover F2,6-BP levels are strongly associated with the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK-2, PFKFB) which display tissue-specific pattern. Four known isozymes of PFKFB are described (PFKFB1, PFKFB2, PFKFB3 and PFKFB4) with different kinase-to-phosphatase activities (87).

Of the PFKFB family members, PFKFB3 and PFKFB4 in particular, are overexpressed in breast cancers and in numerous other malignancies (87). PFKFB3 presents the highest kinase activity among the four isoforms and its inhibition resulted in the suppression of the growth of tumor cells by downregulating the glycolytic flux (74, 88).

Considering that standard chemotherapy inevitably is associated with the development of chemoresistance, the observation made by Kotowski K. et al. that PFKFB3 inhibition therapy concurs with carboplatin and paclitaxel in therapy-resistant cell lines of gynecological cancers to reduce tumor weight presents an important step in further therapeutic approach (87). Resveratrol (a natural product found in various plants) nowadays is also considered as a potential anti-tumoral drug that reduces glucose metabolism and viability in cancer cells. Gomez LS et al. have demonstrated that resveratrol decreases cell viability, glucose consumption and ATP content in the human breast cancer cell line MCF-7 (66).

Phosphoglycerate-kinase (PGK) catalyzes the first substrate-level phosphorylation in glycolysis while producing 3-PG. PGK1 activity is highly responsible for maintaining energy homeostasis and serine biosynthesis. Clinically, PGK1 was overexpressed in many types of tumors. The exact molecular mechanisms for PGK1-involved drug resistance are not fully clarified but were found to be upregulated in radioresistant astrocytomas and cisplatin-resistant ovarian cancers. Moreover, inhibition of PGK1 increased the sensitivity of gastric adenocarcinoma cells to chemotherapeutic drugs 5-fluorouracil and mitomycin-C (89–91). Sun S et al. by analyzing PGK1 at protein and mRNA level in breast carcinoma cases have found that high PGK1 expression is associated with worse overall survival and concluded that PGK1 may be considered as a predictive biomarker of chemoresistance to paclitaxel treatment in breast cancer (92).

Proteomics profiling revealed high enolase-1 (ENO-1) expression in ER+ breast carcinomas. The high ENO-1 status was correlated with poor prognosis and with positive nodal status (93). Qian X et al. by analyzing gastric cancers have found that elevated levels of ENO-1 proteins (or downregulation of ENO-1 targeting miR-22) were associated with shorter overall survival and ENO-1 is a novel biomarker to predict drug resistance and overall prognosis in gastric cancer. Based on their data targeting ENO-1 by chemical inhibitors or upregulating miR-22 could be valuable to overcome drug resistance (94). The ENO functions in different cancers and as a potential cancer biomarker are summarized in a very recent study by Almaguel FA et al. (95).

The conversion of phosphoenolpyruvate to pyruvate with generation of ATP is catalyzed by PK. Till now four PK isoforms are described: PKM1, PKM2, PKR, and PKL (96). Apart from regulating glucose metabolism, PKs are involved in the modulation of gene expression, cell cycle regulation and cell–cell communication (97). Both PKM1 and PKM2 protein expression showed high levels in TNBC and targeting PK inhibited the proliferation of TNBC MDA-MB-231 and MDA-MB-436 cells by involving the NFκB signaling pathway (98). In a study involving 296 invasive breast carcinoma samples a correlation between PKM2 expression and the prediction of chemosensitivity to epirubicin and 5-fluorouracil was demonstrated (99). Moreover, in ER+ breast carcinoma models using MCF-7 and T47D cells PKM2 enhanced chemotherapy resistance by promoting aerobic glycolysis (100). It was also shown that PKM2 expression correlated mostly with cisplatin resistance in breast carcinomas (74).

A large amount of literature has been published regarding LDH expression in different cancers, its effect on cancer development and progression, and its diagnostic and prognostic significance. Lactate production is an important phenomenon in the cancer microenvironment and is used as a mechanism of tumor escape from the immune response.

LDH is a tetrameric enzyme composed of two different subunits LDHA (M) and LDHB (H) (encoded in humans by LDHA and LDHB genes, which can assemble into five different isoenzymes as LDH1 or LDHB (H4), LDH2 (M1H3), LDH3 (M2H2), LDH4 (M3H1), and LDH5 or LDHA (M4) and is involved in the conversion of pyruvate to lactate. Deregulated levels of LDHs have been reported in multiple tumors (101, 102). In cancer cells, LDHA plays an important role in rapid conversion of pyruvate to lactate, minimizing in this way the pyruvate entry into TCA cycle in the mitochondria (103). LDHA is also involved in tumor cells proliferation, angiogenesis, tumor cells invasion and migration (103–105).

Based on the review by Mishra D et al. inhibition of LDHA and LDHB is of great interest and is unlikely to cause any possible side effect (103).

Data presenting the role of LDHA in drug resistance described that in chondrosarcoma inhibiting LDHA increased cancer cell sensitivity to DOX (106), in breast cancer cells, led to re-sensitization to paclitaxel (107). A link between LDHA and paclitaxel resistance was described by Varghese E et al. (74). Oxamate, an inhibitor of LDHA, combined with paclitaxel-induced apoptosis in paclitaxel-resistant breast carcinoma (MDA-MB-435 and MDA-MB-231) cells by inhibiting cellular glycolysis (107).

Inhibition of LDHA by small molecule inhibitors or its knockdown by siRNAs or shRNAs decreases tumorigenicity (107). LDHA seems to be a safe therapeutic target, as patients with hereditary LDHA gene deficiency only show symptoms (exertional myoglobinuria) after strenuous exercise but not under ordinary circumstances (108). In accordance with gene deficiency, LDHA inhibition was not proven to be harmful to normal cells (109). Studies showed that LDHB is up-regulated in triple-negative breast cancer (110). Recently, a highly selective LDHB inhibitor was identified (111).

In this pathway NADPH is generated by the oxidation of G6P via an alternative pathway to glycolysis. In PPP ribose-5-phosphate-R5P is synthesized as an essential precursor in nucleotide biosynthesis. PPP is made up of two branches: the oxidative and the non-oxidative. In the PPP oxidative phase, G6P is converted into ribulose 5-phosphate and CO₂, leading to the synthesis of NADPH.The non-oxidative arm of the PPP composed of a series of reversible reactions generates pentose phosphates for ribonucleotide synthesis (112).

PPP activation has been described in different types of cancer and was associated with metastasis, angiogenesis, and response to chemotherapy and radiotherapy (113). Giacomini I et al. have found that in particular, the oxidative branch of PPP seems to be involved in cisplatin resistance (114). They also suggest that the possibility to selectively deliver into the same cancer cell an anticancer drug and a PPP inhibitor or drug affecting the glucose seems a strategy to overcome the problem of drug resistance (114). Goldman A et al. demonstrated that after taxane treatment metabolic reprogramming of breast cancer cells was observed and characterized by increased glycolytic and oxidative respiration and glucose flux through the PPP pathway (115). The association of PPP with cancer resistance to ACs has been described in several studies (16, 116). Polimeni M et al. (116) by comparing a DOX-resistant human colon cancer cell line (HT29-DX) with a DOX-sensitive (HT29 cells) have found increased PPP and G6PD activity in DOX-resistant cell lines (116).

Alternatively known as the tricarboxylic acid cycle (TCA) or Krebs cycle is a hub of the metabolic system. Acetyl-CoA, a common product of carbohydrate, fatty acid and amino acid metabolism enters TCA cycle that oxidizes the acetyl group of acetyl-CoA to two molecules of CO₂ with the generation of three NADH, one FADH₂ and one ATP equivalent GTP.

The main enzymes involved in TCA cycle are: 1. citrate synthase (catalyzes the condensation of acetyl-CoA and oxaloacetate to yield citrate), 2. aconitase (isomerizes citrate), 3. isocitrate dehydrogenase (catalyzes the oxidative decarboxylation of isocitrate to α-ketoglutarate with the coupled reduction of NAD⁺ to NADH and CO₂ release), 4. α-ketoglutarate dehydrogenase (decarboxylates α-ketoglutarate to succinyl-CoA, releasing CO₂ and NADH, 5. succinyl-CoA synthetase (converts succinyl-CoA to succinate with the coupled synthesis of a GTP), 6. succinate dehydrogenase (catalyzes the dehydrogenation of succinate yielding fumarate and FADH₂, 7. fumarase (catalyzes the hydration of fumarates to yield malate) and 8. malate dehydrogenase (reforms oxaloacetate) (117) ( Figure 4 ).

NADH and FADH₂ are important products of citric acid cycle. Their reoxidation by O₂ through the mediation of the electron-transport chain during OXPHOS completes the metabolic breakdown (117).

Several cancer types are characterized by drastic changes in TCA cycle enzymes compared to normal tissues. Accordingly components of the TCA cycle may be exploited therapeutically for the treatment of disease. Due to the importance of the TCA cycle in normal cell development, high toxicity of this approach have to be considered. By performing metabolomics profiling Ning Shen et al. have found significant changes in the TCA cycle and reactive oxygen species (ROS) related pathways in sensitive TNBC cells compared to resistant TNBC cells (118). Applying small molecule inhibitors to disturb the enhanced TCA for cancer treatment start to evolve CPI-613 (inhibiting both prolyl hydroxylases and α-ketoglutarate dehydrogenase complex- as a key regulator enzyme of TCA) is being tested in phase I and II clinical trials, as a single agent or in combination with standard chemotherapy to treat cancers (119). As many tumors utilize glutamine as a source for TCA cycle, suppression of glutaminolysis by small molecule inhibitors seems also an attractive approach to target tumors. CB-839 disrupts the conversion of glutamine to glutamate and alters TCA cycle, glutathione production, and amino acid synthesis (119).

Oxidative phosphorylation is an active metabolic pathway in many cancers and there is great interest in targeting mitochondrial OXPHOS in cancer. Evans KW et al. performed RNA sequencing from pre-treatment biopsies of TNBC patients who received neoadjuvant chemotherapy. They demonstrated that the top canonical pathway associated with worse outcome showed higher expression of OXPHOS signature. They also found that inhibiting OXPHOS may be a novel approach to enhance efficacy of several targeted therapies in TNBCs and have validated the antitumor efficacy of the combination of palbociclib, a CDK4/6 inhibitor, and IACS-10759 (a selective OXPHOS inhibitor that blocks complex I) in vitro and in vivo (120). Another study has established that TNBC cell lines (MDA-MB-468 and MDA-MB-231) were highly dependent on OXPHOS when compared to HR+ cell line MCF-7 (121). Considering the ubiquitous necessity of OXPHOS in healthy cellular metabolism, a major barrier to mitochondrial-targeted drugs is the need for cancer-cell selectivity.

Fatty acid β-oxidation (FAO) is a primary bioenergetic source by which fatty acids are broken down in a multistep process. Fatty acids enter cells through fatty acid transport proteins like: fatty acid translocase (FAT/CD36), tissue-specific fatty acid transport proteins (FATP), and plasma membrane-bound fatty acid-binding protein (FABPpm) (142, 143). In breast cancer loss-of-function studies have demonstrated that CD36 is critical in fatty acid uptake (144). Compared to CD36, FATPs and FABPpm have received far less attention (145).

Fatty acid transportation across the mitochondrial membrane is controlled by several enzymes. Once inside the cell, a CoA group is added to the fatty acid by acyl-CoA synthetase, forming long-chain acyl-CoA (146). The conversion of the long-chain acyl-CoA to long-chain acylcarnitine by carnitine palmitoyltransferase 1 (CPT1) is the step by which fatty acid is transported across the inner mitochondrial membrane, the process controlled by carnitine translocase (CAT). CPT2 enzyme converts back the long-chain acylcarnitine to long-chain acyl-CoA that enters the fatty acid β-oxidation pathway (146) ( Figure 4 ).

The transport of fatty acyl-CoAs into the mitochondria by CPT1 proteins is considered a rate-limiting step in the mitochondrial fatty acid β-oxidation pathway is (147). All three known isoforms of CPT1 enzyme (CPT1A, CPT1B, and CPT1C) have been recognized as important players in drug resistance. Using an integrated genomic strategy, Gatza ML et al. identified CPT1A as an important player also in cell proliferation especifically in hormone receptor-positive breast carcinomas (148).

The β-oxidation is a cyclic process responsible for the mitochondrial disintegration of long-chain acyl-CoA to acetyl-CoA. This process is controlled by following four main enzymes: acyl-CoA dehydrogenase, enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and ketoacyl-CoA transferase (thiolase) (149). In each cycle, FAD-dependent dehydrogenation and NAD-dependent oxidation leads to the formation of FADH₂ and NADH.

Complete oxidation of the produced acetyl-CoA, NADH, and FADH₂ is accomplished by the TCA and OXPHOS ( Figure 4 ) (146, 150).

Many types of cancer exhibit a high activity of FAO such as TNBC, KRAS mutant lung cancer, hepatitis B-induced hepatocellular carcinoma etc. (136, 151, 152). The mechanisms of FAO activation in different tumor cells are under serious debate and different mechanisms have been proposed to explain drug-induced FAO activation and the role played by FAO in drug resistance (153).

Studies have reported that prominent oncoproteins play important role in the activation of several FAO enzymes. Overexpression of c-Myc in TNBC resulted in FAO enzymes and metabolic intermediates upregulation whereas inhibition of FAO blocked Myc-driven tumorigenesis (154). Another study identified CPT1B as a downstream target of the JAK/STAT3 pathway in breast cancer (135, 136).

Hoy AJ et al. discuss several aspects of the association of tumor fatty acid metabolism and therapy resistance. In recurrent breast carcinomas they described enhanced CPT1B mRNA expression compared to tumours that did not recur (135). In patients with pancreatic and gastric cancers higher CPT1A expression was associated with chemoresistance and with shorter overall survival (135, 155, 156).

Studies suggest that pathways like PI3K/AKT/mTOR, JAK/STAT3 play a pertinent role in lipid metabolism regulation (157–159). Wang T et al. have found that leptins upregulate STAT3 and FAO activity and this metabolic switch promotes cancer stemness and chemoresistance. In in vivo conditions blocking FAO re-sensitized resistant cells to chemotherapy (136, 159). Another study described that inhibition of FAO by mercaptoacetate and etomoxir sensitizes paclitaxel-resistant lung adenocarcinoma cells (160).

Targeting FAO came into the focus of researches as a chemosensitization strategy given its key role in promoting tumor cell survival via energy generation. More and more studies suggest that addition of FAO inhibitors completely or partially inhibited drug resistance of cancer cells (160–162).

Substantial efforts have been documented to develop strategies to target fatty acid biosynthesis since new and new studies have demonstrated that activation of de novo fatty acid synthesis is specific to some cancerous tissues (163).

In a standard way fatty acids are synthesized through the fatty acid synthesis cycle ( Figure 5 ). The substrate for FA synthesis is cytoplasmic acetyl-CoA, which is obtained through different mechanisms (163). The synthesis of fatty acids from acetyl-CoA and malonyl-CoA involves two main steps: 1. carboxylation of acetyl-CoA by acetyl-CoA carboxylase (ACC) to form malonyl-CoA (ATP dependent) and 2. decarboxylation of the malonyl group in the condensation reaction catalyzed by the multifunctional FASN containing substrate shuttling domain ACP and six catalytic domains: malonyl/acetyl-CoA-ACP transacylase, β-ketoacyl-ACP synthase, β-ketoacyl-ACP reductase, β-hydroxyacyl-ACP dehydrase, enoyl-ACP reductase and thioesterase (163, 164).

The enzymes playing role in FA biosynthesis are under the control of sterol regulatory element-binding proteins (SREBPs) (165). The three isoforms of SREBPs (SREBP1a, -1c, and -2) are encoded by two different genes (166). The importance of SREBP-1 in lipid metabolism and tumor prognosis is discussed in several papers (167, 168). In vitro studies performed on MDA-MB-231 and MCF7 breast cell lines revealed that the suppression of SREBP-1 significantly inhibited cell migration and invasion (168). Targeting SREBP could therefore be an efficient strategy to halt tumor growth (163). Fatostatin, a nonsteroidal diarylthiazole derivative inhibits the activation of SREBP1 and additionally inhibits cell proliferation (169).

Reduced fluidity of lipid bilayers in the membranes are also characteristics of chemoresistant cancer cell lines. Membrane lipid composition has gained high importance in cancer research. The reduced membrane fluidity influence or disrupt drug uptake via passive diffusion or endocytosis (135). Plasma membrane cholesterol was reported to be elevated in AC-resistant MCF-7 breast cancer cells (199). The increased expression of sphingolipid metabolizing enzymes, specifically ceramide transport protein, sphingosine kinase 1 and 2 have been correlated with resistance to DOX and paclitaxel (39).

Studying the cholesterol metabolic reprogramming in cancer came in the focus of researches in the last years based in part on its important role in maintaining cellular homeostasis (providing essential hormones) as well as considering its role in forming lipid rafts, an indispensable cell membrane structure in cancer cells. As many tumor-related proteins are located in lipid rafts this cell component is an important platform for oncogenic signaling pathways. Considering other aspects like the metabolites obtained by de novo modification of cholesterol (mevalonic acid, farnesyl pyrophosphate and geranylgeranyl pyrophosphate) also playing an important role in cancer growth and/or oncogenic signaling pathways it is reasonable to consider the significant role played by cholesterol in tumor metabolism. Cholesterol deficiency in cell membranes has been shown to inhibit cancer progression (157) and cholesterol metabolic reprogramming in cancer cells is strongly linked to several aspects of drug resistance (157, 200).

Cholesterol is mostly synthetized through the mevalonate pathway ( Figure 5 ) or acquired from the circulation via LDL receptor (LDLR)-mediated endocytosis. In the mevalonate pathway 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) is reduced to mevalonate by HMG-CoA-reductase enzyme considered as the primary control site for cholesterol biosynthesis (201). The enzyme activity and protein level is controlled by multiple regulatory mechanisms, its competitive inhibitors are among the most widely prescribed medications, collectively known as statins. Further a series of enzymatic reactions convert mevalonate to cholesterol (199, 202, 203).

The role of cholesterol in drug resistance has been demonstrated in several cancer types including breast cancer (199) and is mostly related to AC resistance given the role of the cell membrane permeability in AC drugs. It was shown that decreased cholesterol levels resulted in the increased uptake of DOX (204).

Some mechanisms through which cholesterol regulates drug resistance are lipid rafts, ABC transporters, drug uptake. These mechanisms are highly discussed in a recently published review of Yan A et al. (199).

The “membrane-lipid therapy” based on the modulation of membrane lipid composition is considered as an effective therapeutic strategy in several diseases. Preta G. mentions that instead of modifying membrane cholesterol/sphingolipids content new schemes like modulation of membrane bilayer properties (fluidity and elasticity) by inducing changes in the organization of lipid rafts are preferred (205).

The accumulation of lipid droplets is a less well-studied aspect of chemoresistant cancer cell lines (135). Lipid droplets are considered to directly contribute to chemoresistance by providing an extra source of lipids for FAO when nutrient stress occurs, or may play a role in hydrophobic drug sequestration (206).

Sirois I et al. in a very impressive study by analyzing a novel MDA-MB-436 cell-based model of chemoresistance characterized by a unique and complex morphologic phenotype with numerous lipid droplets and a set of primary chemoresistant TNBCs have identified several metabolic vulnerabilities. These include a dependence on perilipin family member perilipin 4 (PLIN4), a protein coating of the observed lipid droplets that play important role in fat mobilisation, expressed both in experimental conditions (in the chemoresistant TNBC cells) and in chemoresistant tumors treated with neoadjuvant DOX-based chemotherapy. Their results call attention to a novel mechanism of chemotherapy resistance that may have therapeutic consequences in the treatment of drug-resistant cancer (207).