Considering that fatty acids are substrates to PtdCho synthesis, the increased ratio of PtdCho biosynthesis in cancer can be in part a response to the enhanced fatty acid synthesis frequently observed in cancer cells (24). Additionally, it is intuitive thinking that one of the causes of cholinic phenotype is an enhanced ability of importing free extracellular choline by tumor cells. Of note, choline does not cross cell membranes by passive diffusion, being dependent on a choline transport system composed of four transporter families categorized according to their affinity for choline, high-affinity choline transporter 1, choline transporter-like proteins, polyspecific organic cation transporters, and organic cation/carnitine transporters. Several studies have underlined increased expression of each subtype of choline transporter in different human cancer cell lines in comparison with normal counterparts (11, 25). Given the fact that choline uptake is a rate‐limiting step in phospholipid metabolism and a prerequisite for cancer cell proliferation, the inhibition of choline transporters in cancer cells results in lower levels of intracellular choline accompanied by cell death induction (26, 27). As a consequence of decreased intracellular choline levels, there is also a reduction of PtdCho and PCho levels. In this context, cancer cells hydrolyze sphingomyelin as a compensatory response to maintain the generation of PtdCho and PCho. However, sphingomyelinase‐catalyzed hydrolysis of sphingomyelin also generates apoptosis-inducing factor ceramide, which activates caspase-3 and results in apoptotic cell death induction (28). Notably, the increased capacity of the cancer cell to import extracellular choline is a major contributor to the cholinic phenotype. However, further molecular characterization is needed to define what orchestrates the different combinations of choline transporters and how they lead to enhanced choline transport in cancer cells to drive the discovery of potential cancer targets.

The PtdCho cycle is composed of a network of enzymes whose expression and activity can be modulated by genetic alterations present in cancer cells. Increased intracellular levels of PCho in cancer cells are mainly attributed to upregulation of ChoK enzyme and also partially derived from phosphatidylcholine-specific phospholipase C (PC-PLC) and phosphatidylcholine-specific phospholipase D (PC-PLD) enhanced activity.

Accordingly, the ChoKα isoform, which catalyzes the phosphorylation of Cho to PCho, is upregulated in epithelial ovarian (13), breast (25), bladder (29), and colon (22) cancer cells. In addition to this metabolic function, in vitro and in vivo pieces of evidence show that ChoKα overexpression contributes to tumor progression, metastasis, and aggressiveness (29). Moreover, ChoKα overexpression has a prognostic significance and predicts poor prognosis of colorectal cancer (30), early-stage non-small cell lung cancer (31), and hepatocellular carcinoma patients (32). Importantly, PCho can also be generated by the breakdown of PtdCho through PC-PLC activity. PC-PLC is upregulated in ovarian and breast cancer cells of different subtypes and accounts for 20%–50% of intracellular PCho production (15, 33). Moreover, in vitro inhibition of PC-PLC activity resulted in cell proliferation arrest in both cancer models (13, 34). In particular, a decrease of migration and invasion potential, together with a loss of mesenchymal traits, was observed after treatment of breast cancer cells with a PC-PLC inhibitor (15). Additionally, PC-PLD hydrolyzes PtdCho in PA and free choline, which can reenter the CDP pathway and generate choline intermediates. PC-PLD expression is elevated in diverse cancer types, such as gastric (35), breast (36, 37), epithelial ovarian (13), and melanoma (38). Evidence shows that PC-PLD regulates multiple tumor cell events, such as cell transformation, proliferation, survival, and migration (39).

Mechanistic investigation revealed that PtdCho metabolic enzyme activation is dependent on oncogenic signaling pathways, mainly the oncogene ras that affects the activities of ChoK, PC-PLC, and PC-PLD enzymes. Glunde et al. (11) describe the oncogenic signaling pathways involved in the regulation of choline metabolism enzymes (11). In this sense, ras-transformed cells exhibit increased ChoK activity accompanied by increased levels of its product, PtdCho (40–42). Further investigation revealed that in mammalian cells, the mechanism of ChoK regulation by ras implies the involvement of two ras effectors, RAL GTPase guanine nucleotide dissociation stimulator (RALGDS) and Phosphoinositide 3-kinase (PI3K) signaling (41). Several studies underline that the oncogene ras also regulates the activity of PC-PLD enzyme (39, 43, 44). Moreover, the enzyme PC-PLC is a downstream target of Ras, and its activation plays an important role in inducing Ras-mediated mitogenic signaling (45, 46). Thus, oncogene-driven activation of ChoK, PC-PLD, and PC-PLC enzymes increases PtdCho synthesis and degradation, leading to the accumulation of energy-rich molecules and providing sources for cancer cell proliferation.

Cancer choline metabolism is also modulated by harsh tumor microenvironment (TME) conditions, mainly hypoxic and nutritional stress. A large number of studies indicate increased levels of tCho-containing compounds in cancer cells (47). It is interesting to stress out that these choline metabolites were observed to be heterogeneously distributed in tumor sections (48–50). This is of particular interest, as Glunde et al. (51) demonstrated in vivo that hypoxic areas of human prostate tumor xenografts contain increased tCho levels (51). Additionally, they reported that in vitro exposure of prostate cancer cells to hypoxia generates increased levels of PCho and tCho as well as increased expression of ChoK. They also provided mechanistic insights into how hypoxia induces choline metabolism by demonstrating that Hypoxia Inducible Factor-1, HIF-1 directly binds to ChoKα promoter region. Similarly, intermittent hypoxia also upregulates ChoK in rat pheochromocytoma PC-12 cells (52). In contrast, opposing evidence shows a decrease in choline levels in cancer cells exposed to hypoxic conditions without loss of cell viability (53). A hypoxia-mediated inhibition of choline phosphorylation has also been demonstrated in cancer cells (53, 54) as a result of a decrease in ChoK expression and activity mediated by HIF-1α (55).

Importantly, Glunde et al. (51) observed that not all tumor areas with high tCho levels were colocalized with hypoxic regions, indicating that other tumor environment conditions can also modulate choline metabolism. Acidosis can elicit opposite effects in PtdCho synthesis once in vitro evidence shows that it inhibits ChoK but enhances CCT (CTP:phosphocholine cytidylyltransferase) enzyme activity with a net increase of PtdCho pool in low pH conditions (56). However, another in vitro evidence shows that acidosis decreases PCho levels, reinforcing that ChoK is inhibited by low pH, but also increases GPC levels, indicating enhanced PtdCho degradation (57). Of note, GPC production from PtdCho catabolism is mediated by lysophospholipase, and phospholipase A2 catalyzed reactions with the generation of free fatty acids. It is interesting to note that acidosis can inhibit glycolysis in human cancers (58) and, as a consequence, result not only in a reduction of ATP generation but also in decreased amounts of acetyl-CoA, which feeds the tricarboxylic acid (TCA) cycle for aerobic respiration. Thus, these pieces of evidence drive us to suggest that during acidosis, decreased PtdCho synthesis and increased PtdCho breakdown allow cancer cells to fuel beta-oxidation of fatty acids as a source of acetyl-CoA ( Figure 3 ).

In this sense, ample evidence supports the notion that in aerobic conditions, cancer cells exhibit increased choline uptake to activate anabolic metabolic pathways to sustain their high proliferative rates. On the contrary, hypoxic conditions have been reported to diminish choline uptake while enhancing glucose and acetate in cancer cells (53). Kamphorst et al. (59) demonstrated that during hypoxia, acetate is the major additional source of carbon donor for acetyl-CoA, and the generation of this precursor for fatty acid biosynthesis allows cancer cells to maintain lipogenesis and proliferation under hypoxic conditions (59). Interestingly, the in vitro study by Yoshimoto et al. (60) revealed that the rate of acetate incorporation in tumor cells under hypoxia is superior to the rate observed in normal cells, and the metabolic fate of acetate in hypoxic cells was preferentially incorporated into PtdCho (60). Importantly, they also showed that this increased acetate uptake and lipid incorporation were positively correlated with growth activity. Thus, cancer cells might have these metabolic adaptations to maintain the anabolism of fatty acids, which require acetyl-CoA units to compensate for inhibition of glycolytic ATP and acetyl-CoA production by activation of fatty acids as a source of energy.

There is a correlation between survival and cancer cell migration under stress conditions, suggesting that cancer cells in addition to surviving and suppressing cell death also trigger a migration phenotype to escape from stressful regions (61). This raises the question of how migration is triggered by stressful conditions and if enhanced choline metabolism could be a linker of these phenomena.

Oxygen- and nutrient-deprived areas arise as a consequence of inadequate blood supplies during solid tumor growth and impose an energetic pressure on cancer cells. To survive, cancer cells need to first suppress cell death induced by these stress conditions and ultimately provide means for obtaining energy. One possibility is that cancer cells with increased amounts of choline compounds have the advantage of obtaining energy and building blocks from the degradation of these lipids. Alternatively, cancer cells in parallel induce migration to sites where nutrition could be found, and choline compound storage can contribute to this process by their breakdown products. In this sense, Zheng et al. (62) revealed that under stress caused by serum withdrawal, MDA-MB-231 human breast cancer cells exhibited increased PC-PLD enzyme activity concomitant to an enhanced migration and invasion potential (62). Compelling evidence gained from PLD2 overexpression (63), an isoform of PC-PLD gene, in low-invasive breast cancer cells resulted in the conversion of these cells into a highly aggressive phenotype with increased capacity of lung metastasis formation, which was inhibited by two different small-molecule inhibitors of PC-PLD activity (63). Animals deficient in another PC-PLD gene isoform, PLD1, or treated with a small-molecule inhibitor of PLD1 activity, exhibited reduced tumor growth, angiogenesis, and metastasis (63). Aberrant expression of both PC-PLD isoforms has been detected in different cancers and linked to cancer cell survival and a pro-metastatic phenotype through activation of different signaling pathways revised in Yao et al. (64). Thus, a stressful tumor environment can drive PtdCho degradation, in particular through PC-PLD hydrolysis, which contributes to a cancer cell migration program. These data indicate PC-PLD as a potential target for cancer therapy and point toward a small-molecule dual inhibitor of PLD1 and PLD2 as a promising strategy.

PtdCho metabolism has been linked to opposing cellular events, cell proliferation, and cell death. Noticeably, the cell cycle controls PtdCho homeostasis to avoid an excess or deficit of membranes. Essentially, PtdCho metabolism is modulated during the cell cycle and is characterized by a high rate of PtdCho degradation and resynthesis in the G1 phase, a reduced PtdCho turnover that leads to doubling PtdCho amounts in the S phase, and a cessation of PtdCho metabolism in G2/M phases (65). In the opposite direction, the first evidence of a direct link between cell death and PtdCho synthesis was a report showing that Chinese hamster ovary (CHO) cells with a mutation in the CCT enzyme resulted in PtdCho depletion and concomitant apoptosis induction (66). The molecular mechanism by which PtdCho depletion is sensed and transduced to cell death has not yet been fully elucidated; however, evidence shows that inhibition of PtdCho synthesis triggers apoptosis through a mechanism that involves the activation of an endoplasmic reticulum, ER stress response (67, 68). Moreover, PtdCho is a substrate for sphingomyelin (SM) synthesis, and the final step of this biosynthetic route involves the exchange of the phosphocholine head group from PtdCho to ceramide. Yen et al. (69) demonstrated that in parallel to the intracellular decrease of PtdCho, SM levels also decrease and the apoptosis inducer ceramide accumulates (69).

More recently, studies have implicated lipid metabolism in the non-apoptotic cell death process of ferroptosis. This process is characterized by the accumulation of iron-dependent lethal lipid peroxides (LPOs) that can be generated from the oxidation of phospholipids, such as arachidonoyl and adrenoyl, by the catalysis of acyl-CoA synthetase long-chain family member 4 (ACSL4), LPCAT3, and 15-lipoxygenase (15-LOX/ALOX15) (70, 71). A detailed underlying mechanism of ferroptosis in cancer biology was reviewed by Li and Li (72). Chemotherapy and mainly ionizing radiation (IR) therapy generate reactive oxygen species (ROS) that can target lipid peroxidation and cause ferroptosis induction. Of interest, IR was reported to induce ferroptosis, and inhibition of ACSL4 enzyme activity reverted IR-induced ferroptosis and promoted radioresistance (73, 74). While oxidative metabolites from arachidonoyl and adrenoyl can generate "find me signals" and elicits an antitumor response, ferroptotic cancer cells have increased PTGS2 gene expression, which encodes cyclooxygenase 2 (COX-2) to produce prostaglandin E2 (PGE2), a major pro-inflammatory factor (75). These data suggest that ferroptosis and lipid metabolism may be involved in resistance to therapy. Further research to expand the understanding of the unique features of ferroptosis will unveil the therapeutic windows to precisely target this process.

Importantly, PtdCho depletion can indirectly interfere in cell viability once it is an important source of lipid mediators that are known to regulate cell growth, such as PA and DAG. Accordingly, a balance between mitogenic and antimitogenic lipid mediators derived from PtdCho can dictate the fate of cells toward cell proliferation, arrest, or death (76). This is of particular interest in oncology once several antitumoral drugs, including chemotherapy and radiotherapy, induce an increase in proapoptotic ceramide levels and parallel mitogenic DAG levels (77).

PtdCho turnover (catabolism) is mediated by phospholipases (A2, C, and D), generating both choline-containing phospholipids (e.g., PhoC, GPC, and choline), that can be reutilized for PtdCho biosynthesis and lipid mediators that regulate multiple protumoral signaling pathways. The list of dysregulated bioactive lipids that have been shown to contribute to tumor biology includes AA, eicosanoids, DAG, PA, lysophosphatidic acid (LPA), platelet-activating factor (PAF), ceramide, sphingosine, and other lysosphingolipids (120). This list continues to grow, and here we highlight PtdCho-derived lipid mediators emerging as lipid second messengers involved in resistance to therapy ( Figure 4 ).

More than 50% of PtdCho synthesized in the Kennedy pathway is remodeled through the Lands cycle (121). In this remodeling pathway, PtdCho is hydrolyzed by phospholipase A2 (PLA₂), resulting in the production of lysophosphatidylcholine (lyso-PC) and AA. Once released, AA participates in the biosynthesis of eicosanoids such as prostaglandins and leukotrienes. Notably, AA is metabolized through the enzyme COX-2 into the terminal product PGE2. Elevated levels of COX-2 and PGE2 are frequently observed in many cancers and are associated with cancer initiation, progression, and resistance to therapy. Mechanistically, the activated COX-2/PGE2 pathway leads to therapy resistance mainly through affecting the TME by inducing epithelial–mesenchymal transition (EMT), suppressing anticancer immunity, and regulating cancer stem cell (CSC) homeostasis (122). Several reports with diverse cancer cell lines have shown that EMT is promoted by COX-2-induced PGE2 production, of which inhibition reverts this phenomenon (123–125). Moreover, EMT promoted by COX-2/PGE2 axis confers resistance to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor erlotinib (126).

Under physiological conditions, another PtdCho-derived lipid mediator generated from PLA2 activity, lyso-PC, is rapidly metabolized or reacylated to avoid the cytolytic induction caused by high intracellular concentrations due to its amphipathic property. The reacylation of lyso-PC is performed by lysophosphatidylcholine acyltransferases (LPCATs) by adding fatty acids at the sn-2 position to yield PtdCho, which rapidly gets recycled by the Lands cycle. These cycles of deacylation and reacylation of PtdCho modify the fatty acid composition of the phospholipids de novo generated in the Kennedy pathway and produce membrane asymmetry and diversity. Intracellular Lyso-PC concentration is also regulated by its hydrolysis through the enzymatic activity of autotaxin (ATX), an enzyme with lysophospholipase D activity, that generates LPA. Several cancers exhibit the activity of ATX enhanced and consequent increased levels of LPA (127) that is associated with cancer development and poor prognosis (128). Several reports have exploited the ATX–LPA signaling on cancer cell protection against chemotherapy and radiotherapy. Minami et al. (129) showed that LPA signaling via LPA receptor (LPAR5) regulates the resistance to cisplatin and dacarbazine in a melanoma cell line (129). Additionally, ATX–LPA signaling was reported to protect colon cancer cells from cisplatin and 5-fluorouracil-induced apoptosis (130) and to decrease cisplatin cytotoxic effect in human ovarian cancer cells (131). Inhibition of ATX activity reverts the protective effect of LPA on Taxol-induced apoptosis in breast cancer cells (132). The mechanism involved in ATX–LPA axis attenuation of chemotherapy-induced cell death includes the activation of PI3K–Akt survival pathway (132) and stabilization of nuclear factor E2-related factor 2 (Nrf2) transcription factor. Nrf2 increases the transcription of multidrug-resistant transporters and antioxidant genes, counteracting the chemotherapy-induced oxidative damage (133). Similarly, ATX inhibition enhances the radiotherapy-induced apoptosis in breast cancer cells (134) and attenuates radiation-induced survival, invasion, and angiogenesis in glioblastoma cells (135). LPA-mediated therapy resistance could also be attributed to its role in regulating tumor-associated macrophage (TAM) formation and tumor immunity (136–138).

In addition, to convert lyso-PC into PtdCho, LPCAT enzymes reacetylate lyso-PC and generate another lipid that is a potent cellular mediator, platelet-activating factor (PAF). Four enzymes (LPCAT1–LPCAT4) constitute the LPCAT family and, despite all LPCAT members being involved in lyso-PC conversion into PtdCho, only LPCAT1 and LPCAT2 are known to play an important role in PAF production (139, 140). We have demonstrated in vitro that the simultaneous silencing of all four LPCAT transcripts by modular nucleic acid nanoparticles resulted in lyso-PC (lyso-PAF 16:0 and 18:0) accumulation and enhanced the radiation cytotoxic effect in melanoma cells. We suggest that interfering in LPCAT-mediated signaling disturbs the generation of lyso-PAF, and PAF and contributes to cancer cell sensitization (141). PAF generally refers to alkyl-PAF, the most active form of PAF; however, abundant amounts of an acyl analog of PAF (acyl-PAF) is concomitantly generated with the alkyl PAF species. For a long time, acyl-PAF was considered an inactive PAF analog (142). Intriguingly, Chaithra et al. (143) have demonstrated in vitro and in vivo that acyl-PAF dampens PAF-R signaling and suppresses the action of alkyl-PAF (143). Accordingly, besides exerting their actions through a single PAF receptor (PAF-R), this pair of lipid mediators has opposite effects as inflammatory set-point modulators. The acyl-PAF has been neglected in PAF biology studies in the oncology field, and it is crucial to address the complex interplay between PAFR, alkyl-PAF, acyl-PAF, and their common catabolic enzyme PAF acetylhydrolase (PAF-AH) to unravel the role of PAF/PAFR signaling pathway. PAF is implicated in cancer progression by triggering inflammation and promoting proliferation, survival, metastasis, angiogenesis, and immune-suppressive responses (144). Elevated levels of PAF or increased PAF-R expression was observed in response to various stimuli, including therapeutic agents (145). As outlined, cisplatin increases PAF-R expression, and its inhibition by a PAF-R antagonist resulted in the chemosensitization of melanoma (146) and ovarian (147) cancer cells in vitro and in vivo. Additionally, PAF is generated following the treatment of B16F10 melanoma cells with chemotherapic agents such as etoposide, cisplatin, and melphalan. Importantly, elevated levels of PAF and oxidized lipids with PAF-R agonist activity were detected after the treatment with these drugs as a result of their common ability to induce ROS (148). Similar results were obtained after exposure of melanoma and cervical cancer cells to irradiation (149, 150). Interesting, in a murine melanoma model with a dual injection of B16F10 cells, treatment of one tumor with irradiation or chemotherapy augmented the growth of the untreated tumor in a PAF-R-dependent manner (150) (148). This evidence offers important insight into the systemic role of PAF and PAF-R agonists on negative regulation of therapy efficacy. In this sense, increased activation of PAF–PAFR axis impacts chemo/radioresistance through inducing immunosuppression by modulating regulatory T cells (Tregs) in a COX-2–dependent process (148). These results have driven the investigation of PAF/PAF-R axis in the tumor repopulation phenomenon.

The most prominent consequence of anticancer therapy is a massive induction of cell death frequently associated with a residual number of surviving tumor cells with the capacity to repopulate the tumor. The molecular mechanism involved in tumor repopulation has been investigated, and in 2011, Huang et al. (151) showed that in radiotherapy-induced apoptotic cancer cells, activated caspase-3 activates cPLA-2 and results in increased levels of PGE2, which as mentioned above can trigger protumoral signaling pathways and stimulate the growth of surviving tumor cells culminating in tumor repopulation (151). Interestingly, compelling evidence has indicated that PAF is at least partially responsible for this mitogenic effect of dying cells. Bachi et al. (152) reported that co-injection of apoptotic cells and a subtumorigenic dose of melanoma cells promote the tumor growth, and this phenomenon was inhibited by PAF-R antagonists (152). The following study showed that irradiated TC-1 cells promote the in vitro proliferation of TC-1 viable cells that was diminished by PAF-R antagonist treatment. In the same study, in an in vivo repopulation assay with a model of a human carcinoma cell line expressing (KBP) or not (KBM) PAF-R, the co-injection of live KBP cells and irradiated-induced dying KBM cells resulted in faster tumor growth compared with co-injection of a mixture of live and irradiated KBM cells (153).

Another catabolic route of PtdCho is mediated by PC-PLC enzyme that hydrolyzes PtdCho into PCho and DAG. The latter is probably the best-studied second messenger in cancer biology. It has been shown that the transformation of cells with oncogenes, such as ras, results in a prolonged and persistent elevation in DAG levels. Moreover, DAG activates the protein kinase C (PKC) pathway that is involved in several protumoral pathways, including cell cycle progression, tumorigenesis, and metastatic dissemination (154). Another lipid mediator, PA, is generated from the PtdCho hydrolysis mediated by PC-PLD. The mammalian target of rapamycin (mTOR) was reported as the main target of PA in cancer cells (155). The stability and activity of mTOR complexes depend on interaction with PA and result in signals for cancer cell survival (156). PA interacts with mTOR in a manner that is competitive with the mTOR inhibitor rapamycin, and as a consequence, elevated PC-PLD activity, frequently observed in tumors, confers rapamycin resistance (157). Upregulation of PLD2 was observed in multidrug-resistant colon and breast cancer cells, suggesting that PC-PLD could provide a survival signal involved in therapy resistance (158).

In line with the experimental observations mentioned above, it is not unreasonable to assume that cancer cells exhibit a prominent PtdCho degradation. The fact that PtdCho degradative enzyme activity (46, 159, 160) and PtdCho-derived mediator levels are frequently found elevated in tumors (161), cancer cells with increased PtdCho degradation hijack PtdCho-derived lipid mediators to favor tumor progression and enhance therapy resistance.

Lipid droplets (LDs) are predominantly formed by triacylglycerol (TAG) and PtdCho. In the last years, the role of these organelles in cancer has been well recognized, and more recently, some groups have made an effort to understand their role in tumor resistance. LD formation and accumulation were found in some drug-resistant cell lines, raising the possibility that these organelles might confer resistance to therapy (162–165). The association of choline metabolism and LD was demonstrated by the presence of active LPCAT1 and LPCAT2 in LDs by Moessinger et al. in 2011 (166). Recently, Cotte et al. (167) demonstrated that 5-fluorouracil and oxaliplatin-induced lipid droplet formation in colorectal cancer cell lines was supported by the enzyme LPCAT2 (167). Moreover, it was observed that LPCAT2-dependent lipid droplets conferred resistance to chemotherapy in these cells, and this effect could be reversed by inhibition of LD biogenesis, indicating the potential of LPCAT2 as a target to increase chemotherapy efficacy.

Some studies conducted in normal cells showed the relationship between choline and DNA. Due to the presence of three methyl groups in the nitrogen atom, choline can donate these methyl groups to the formation of S-adenosylmethionine (SAM) that is the main methyl donor for the epigenetic alteration in DNA and histones (168, 169). In 2004, Niculescu et al. (170) observed an increase in cyclin-dependent kinase inhibitor 3 (CDKN3) levels in choline-deficient neuroblastoma cells due to its hypomethylation leading to a reduction in proliferation, revealing that choline can interfere with tumorigenesis as a modulator of DNA methylation (170). A few years later, this finding was corroborated by studies using rodent models that showed that diets low in choline led to an increase in spontaneous hepatocarcinoma (revised in 171). Additionally, Kovacheva et al. demonstrated that choline deficiency was responsible for DNA methyltransferase 1 (DNMT1) overexpression due to its hypomethylation, which led to a global DNA hypermethylation in rats (172). These studies suggest that choline might contribute to methyl metabolism and DNA methylation and gene regulation in carcinogenesis and tumor progression.

Furthermore, it is known that the effect of cytotoxic therapy is mainly dependent on nuclear DNA damage extension and the DNA repair capacity of tumor cells to remove these lesions, and choline metabolism can interfere in this process. In 2007, Mori et al. (173) observed that under ChoK knockdown, the death ratio of 5-fluorouracil-treated breast cancer cells increased and, at the molecular levels, this effect was accompanied by a decrease in the expression of some DNA repair-related genes such as RAD23 that is known to participate in the nucleotide excision repair pathway (173). Using a rodent model to study carcinogenesis, choline deficiency was found to be correlated with the silencing of some tumor suppressor genes including the DNA repair genes BRCA1 and hMLH1, indicating that this metabolite also modulates DNA stability (174).

In the last years, genome sequencing from tumor cells led to the identification of oncogenic mutations that are responsible for tumor cell survival and growth. Some of these mutations were found to be druggable, and blockade of the signaling pathways governed by them had improved cancer treatment in these cases. A noticeable example is the receptor tyrosine kinases (RTKs) that are often constitutively activated in different tumor types. Interestingly, some choline metabolites seem to participate in some of these oncogenic pathways controlled by RTKs. A possible correlation between RTK and choline metabolism was demonstrated by Pisanu et al. (175). The authors verified an increase in PC-PLC activity and PCho content in Human epidermal growth factor receptor 2, HER2--overexpressing ovarian cancer cells. Previously, Paris et al. (176) showed PC-PLC accumulation in the plasma membrane of HER2-overexpressing breast tumor cells (176). PC-PLC inhibition caused a downregulation in HER2 levels due to HER2 internalization that impaired its return to the cell membrane and the activation of HER2 signaling pathways. Moreover, the authors demonstrated that PC-PLC is physically associated with both HER2 and EGFR, and blockade of PC-PLC was able to reduce cell proliferation even in trastuzumab-resistant cells. These results provide evidence that PC-PLC is a promising target to counteract the oncogenic effect of HER2 amplification mainly in breast and ovary malignancies. Concerning the other molecules from choline metabolism, in 2012, it was shown that EGFR interacts with choline kinase α2 (ChoKα2) (177). The authors observed that c-Src-dependent phosphorylation sites of CHKA2 are necessary for EGF-dependent cell growth, suggesting that ChoKα may be an effective target for the treatment of tumors that overexpress EGFR and c-Src. In prostate cancer, the enzyme ChoKα was proposed to be a chaperone for androgen receptor, since its transcriptional activity was dependent on ChoKα. The inhibition of these choline kinases caused a decrease in cell proliferation in vitro, tumor growth, and metastasis in vivo, demonstrating its potential as a target for prostate cancer treatment (178). A recent work by Lin et al. (179) also described the association between ChoKα and EGFR in hepatocarcinoma (179). The authors observed that the pro-metastatic effect of ChoKα is mediated by its binding to EGFR, promoting its dimerization and AKT activation. Additionally, ChoKα overexpression promoted resistance to EGFR-targeted drugs both in vitro and in vivo, and the dual inhibition of ChoKα/mammalian target of rapamycin complex 2, mTORC2 might overcome the resistance to EGFR-targeted therapy in these tumors.

Still, under this context, PAF metabolite also interacts with EGFR in cancer cells. In ovarian tumor cells, PAF increased EGFR phosphorylation via PLCβ, intracellular Ca2+, Src, and the ADAM-mediated release of EGFR ligand HB-EGF, showing the interaction between PAFR and EGFR signaling pathways (147). More recently, one of the enzymes responsible for PAF production, LPCAT1, was also shown to be required for EGFR signaling. In GBM cells, EGFRvIII altered cell lipid composition through LPCAT1 that is, in turn, upregulated by EGFR. Knockdown of LPCAT1 was able to reduce tumor growth in vivo, indicating that targeting LPCAT1 can be a promising strategy to treat or reduce tumor recurrence in amplified EGFRvIII GBMs (180). These studies demonstrate the potentially actionable role of the choline-related enzyme in cancer treatment.

Autophagy is a catabolic mechanism that plays an important role in the lysosomal degradation of protein aggregates, macromolecules, and damages organelles to recycle cellular components and sustain cell metabolism. This dynamic cellular self-digestion has a dual role in cancer cells, acting as a tumor suppressor or tumor promoter, depending on cancer type and stage. Cancer cells display activation of diverse processes to overcome stress, among them, autophagy. This program provides metabolic needs and helps cancer cells to sustain tumor viability and promote drug resistance (181, 182). While autophagy has been plenty studied, the role of lipids in this process is in its early stages, in part, due to the technical challenge of working with lipids. Phospholipids derived from the Kennedy pathway play an important role in the first phases of autophagy. PtdEth acts as an anchor of the microtubule-associated light protein light chain 3 (LC3), essential to cargo selection and autophagosome biogenesis (183). Recently, choline phospholipids (ChoPL), composed of phosphatidylcholines, sphingomyelin, and lysophosphatidylcholines, were reported in the autophagosome assembly. In this study, Andrejeva et al. (184) demonstrated that autophagy induced by anticancer drugs, followed by the incorporation of ¹³C-labeled choline, resulted in a high de novo synthesis of ChoPL in cancer cells. Moreover, to investigate the mechanism responsible for this process, they used MT58 cells that hold a temperature-sensitive mutation in the rate-limiting enzyme of PtdCho synthesis, CTP: phosphocholine cytidylyltransferase α1 (CCTα1). They showed that the loss of CCTα1 activity revokes autophagy and impairs cells to sustain autophagosome formation for extended periods of autophagy (184). By this, novel studies have also shown the importance of a second human CTP: phosphocholine cytidylyltransferase, CCTβ3 enzyme. In short-term starved cells, CCTβ3 is recruited to the autophagosome membrane to activate PtdCho synthesis and induce omegasome expansion. Despite that, CCTβ3 did not cause a meaningful upregulation of autophagy. However, opposite effects were observed in cells submitted to long periods of starvation, indicating that CCTβ3 is critical in the PtdCho synthesis to sustain prolonged autophagy (185). Additionally, the induction of autophagy in CCTβ-null cancer cells was significantly suppressed, and such effect was reversed by rescued expression of CCTβ3. Interestingly, the re-expression of CCTβ3 increased cell survival after starvation, indicating the relevance of PtdCho metabolism to autophagy activation and the subsequent impact in survival and resistance to therapy of cancer cells. In this sense, it was demonstrated that treatment with ChoK inhibitors, as B-3D and EB-3P, in liver cancer cells caused the reduction of autophagy components and induced apoptosis (186). Thus, it is likely that elevated levels of choline phospholipids observed in cancer cholinic phenotype may sustain drug-induced cytoprotective autophagy, which favors therapy resistance.

The modification in lipid cellular composition observed in cancer cells has also been noticed in the extracellular vesicles (EVs) secreted by them. EVs are nanostructures delimited by a lipid bilayer that carry a range of biologically active macromolecules like RNAs, DNA, protein, lipids, and cytokines. These spherical structures can bind to the plasma membrane or be engulfed by recipient cells, leading to a reprogramming that affects their functionality in the TME. Exosomes, a type of EVs that originated from multivesicular bodies, with bioactive lipids, such as PGE2α, PGE1, and PGE2, are secreted by macrophages and tumor cells into the TME (221).

sPLA2, cPLA2, iPLA2, COX-1, COX-2, AA, and PGE2 were already identified in tumor-derived EVs (204, 222–224). Concerning PGE2, EVs carrying this prostaglandin were shown to be associated with immune escape (204) and release of pro-inflammatory cytokines responsible for MDSC recruitment in breast cancer microenvironment (225, 226). In addition, the blockade of PGE2/EP4 signaling reduced the secretion of EVs by basal mammary epithelial stem cells while promoting the release of EVs and CSC-associated proteins from transformed mesenchymal breast cancer cells, modulating tumor progression. Although the relationship between choline metabolism and EVs is still an unexplored field, one might propose that lipid metabolism indeed affects the production and secretion of EVs, as well as interactions with the recipient cells, and more evidence showing the consequences of altered choline metabolism in tumor-derived EVs and their effects in the TME is a matter of time (227).