Increasing scrutiny has been devoted in recent years to the role of metabolism in cancer [1]. As a result, metabolism reprogramming, deregulation, and metabolic flexibility have become recognized as critically important cancer hallmarks [2,3]. It has become clear that cancer cells hijack signaling circuits to rewire their metabolism and energy production as an adaptation to the need of enabling and sustaining rapid proliferation and continuous growth. Cancer cells have to survive in harsh tumor microenvironment, exposed to oxidative and other stress (like lactate-mediated acidification) in nutrient-poor and often hypoxic conditions, while maintaining the capacity to orchestrate metastasizing spread and adaptive resistance to cancer treatment. Reprogrammed metabolism in cancerous lesions is also one of the crucial obstacles to cancer immunotherapy as it promotes dysfunction in immune cells in the tumor microenvironment [4].

Tumors display a fundamental energy supply transformation from oxidative phosphorylation to glycolysis. In normal cells, glucose is catabolized to pyruvate that converted to acetyl-CoA fuels tricarboxylic acid (TCA) cycle capable (by generation of NADH and FADH) of providing mitochondrial respiratory chain with electrons for energy production with a high efficiency (up to 36 ATP molecules from one glucose molecule) [2]. In contrast, most of pyruvate produced by glycolysis in cancer cells is reduced to lactate owing to upregulated levels of lactate dehydrogenase in these cells [5]. The conversion of pyruvate to lactate lowers the levels of reactive oxygen species in cancer cells (that are typically elevated) and thereby lessening their exposure to oxidative stress, which helps their survival [2]. However, this leaves cancer cells with a considerably lower efficiency of energy generation (only 2 ATP molecules attained per one glucose molecule) [6]. This problem is in most tumors solved by markedly intensified glycolysis, achieved by upregulated levels of transporters for glucose and elevated expression of the majority of glycolytic enzymes (mediated by activity of oncogenes Ras, Myc and HIF-1) [7].

Prevalence of higher basal levels of reactive oxygen species evidenced in cancer cells in comparison to normal cells (resulting from oxidants/antioxidants imbalance) serve at low and moderate levels as signal transducers for activating cell proliferation, angiogenesis, and migration/invasion [8].

Other most striking changes in tumor bioenergetics, serving to meet increased bioenergetic and biosynthetic demand, include the intensification of glutaminolytic flux, amino acid and lipid metabolism, enhancement of mitochondrial biogenesis, upregulation of pentose phosphate pathway, and the use of intracellular anabolic pathways for de novo macromolecules generation [2,7,9].

Another adaptation in cancer cells is the induction of nutrient scavenging mechanisms for sustaining their continued proliferation [10]. For the same purpose, these cells were shown to co-opt signaling pathways and transcriptional networks for securing the metabolic flux through intermediary metabolism; the latter include PI3K-Akt-mTORC1 and MYC [1]. The intermediary metabolism is the call phrase for highly integrated reactions network providing cells with metabolic energy, reducing power and biosynthetic intermediates [11], and is employed in malignant cells to support tumor mass production [1].

Most impactful cancer-associated metabolic transformation is arguably the reprogrammed/accelerated lipid metabolism. It is recognized to have a fundamental role in cancer pathogenesis and is of critical importance for the resistance of tumors to chemotherapy and other forms of cancer treatment [12]. Lipids (including phospholipids, glycolipids, sterols/cholesterol esters, and fatty acids or triglycerides) are not only primary constituents of cell membranes but are engaged as essential participants in nearly all cellular functions, including signal transduction, recognition and communication in cell-cell interaction, inflammation (with notable roles of prostaglandins, and leukotrienes), protein regulation (like palmitoylation and myristylation), and cell energy provision. Lipid metabolism regulates the balance between the synthesis and breakdown of these molecules. Cellular lipid homeostasis is continuously maintained by the involvement of lipophagy (a form of selective autophagy) and liposomal lipolysis that is controlled by signaling pathways regulating lipid hydrolysis on transcriptional and post-transcriptional levels [13]. Lipophagy is closely involved in lipid metabolism alterations in cancer cells and directly impacts metastasis dissemination process [14]. When their concentrations rise beyond physiological levels, intracellular lipids express toxic effects (known as lipotoxicity) [15]. In such situations, lipids can change membrane fluidity, act as detergents, disturb pH homeostasis, precipitate overactivation of signaling cascades, boost formation of reactive oxygen species, and can become oxidized into lipid peroxides [13]. Lipotoxicity risk is particularly relevant in cancer cell metabolism [16].

Lipids also play pivotal roles as influential intercellular signaling molecules contained in cargo discharged from extracellular vesicles (small membranous structures formed as nanosized lipid bilayers) released from almost all types of cells [17]. Extracellular vesicles released from malignant cells have critical roles in cancer progression as they help shaping tumor microenvironment and promote invasive and metastasizing progression [18].

The abnormal lipid metabolism in cancer cells features exceptional build-up of lipid droplets, extensive stimulation of lipogenesis (de novo synthesis of fatty acids and their conversion to triglycerides) paralleled by reduced reliance on dietary and liver-synthesized lipids, increased overall cellular lipid content, and fatty acid oxidation [19]. In the latter case fatty acids (originated from sources outside the cell or liberated from internal lipid resources) become oxidized within the mitochondria to serve as catalytic fuel for the generation of energy for cancer cells. Lipid metabolism reprogramming also exhibit other cells present in tumor microenvironment including stromal and immune cells, which additionally influences tumor functional phenotypes and alters immune control [20].

Lipid droplets in cancer cells are not just organelles serving as inert reservoirs of neutral lipids but have many important functions. This includes securing energy for proliferation, invasion, and metastasis, conferring resistance to cell death by lightening lipotoxic stress, preserving redox homeostasis, delivering signaling molecules, and enabling communication of cancer cells with tumor microenvironment [21]. The latter can be facilitated by direct release of lipid droplets from cells within extracellular vesicles [13].

Among more than 150 proteins residing on the surface of lipid droplets, the three major classes are perilipins, neutral lipases, and acyl-CoA synthases [21]. Dynamic interaction between these proteins enables rapid adjustments for promptly responding to changing nutritional/metabolic and stress conditions [22]. This is consistent with the findings of recent studies revealing that lipid droplets in cancer cells are not homogenous but are exhibiting differences in lipid concentrations and composition [23].

Distinct varieties of lipid usage and processing are exhibited in different types of cancer [24]. For instance, prostate cancers rely primarily on β-oxidation of fatty acids for energy production, while pancreatic cancers show preference for the increased uptake of exogenous fatty acids [25]. Breast cancers, in turn, resort to high intensity of de novo lipogenesis for securing fatty acids for cellular membranes formation and production of lipid signaling molecules [24]. Hence, distinct cancers feature various adaptations of lipid metabolism. Moreover, in situ tumors differ in their metabolic requirements compared to metastatic lesions.

The master transcription factor orchestrating lipid metabolism re-programing in cancerous lesions is HIF (hypoxia-inducible factor) [26], which is in turn regulated by bioactive sphingolipid activity [27]. To perform this role HIF acts as a transcription factor for hundreds of target genes regulating glucose and lipid metabolism as well as controlling genes that accelerate tumorigenesis, cancer angiogenesis, metastasis and invasion by either upregulating or downregulating their expression [28]. It upregulates the expression of genes mediating lipid uptake and synthesis while downregulating the activity of lipid oxidation genes [29]. Hallmark tumor-specific transformation in substituting oxidative phosphorylation for intensified glycolysis (the Warburg effect) is mediated by HIF-1α [26]. Among the genes regulated by HIF-1 that are controlling lipid functions are those mediating β-oxidation, lipogenesis, lipolysis, and lipid droplet formation [26,30]. The subunit HIF-2α was discovered to raise cellular iron and promote ferroptosis, regulated cell death pathway triggered by excessive lipid peroxidation [31].

Lipid signaling typically comprise signal transduction cascades involving cell membranes [32]. It is triggered as a response by which interior molecules of the cell react to a signal external to the cell and represents a process that allows the cell to react to the events occurring in their local environment. The primary messenger (chemical/physical signal from the outside of the cell) usually remains outside the responding cell bound to surface receptors exposed on the cell membrane. These receptors are linked to sensors that react to the ligand binding by activating lipid cleaving enzymes located in the membrane or at the intracellular surface of the lipid bilayer. Fragments cleaved from lipid molecules become secondary messengers serving as intracellular signals binding to downstream intracellular enzymes to extend the cascading processes [33-35]. Recent advances in the perception of signal transduction redefined in the context of extensive crosstalk among multiple signaling pathways include a revelation that one lipid messenger can become interconverted into another with the production of secondary metabolites of lipid messengers that themselves possess bioregulatory functions [33].

Hyper oxidative stress, associated with the excessive intracellular accumulation of reactive oxygen species, is a highly potent instigator of important cell signaling activities. Although sometimes prominent in untreated cancer cells (and associated with peroxidation of cellular phospholipids [36]), it can be greatly potentiated with some cancer therapy treatments, particularly photodynamic therapy (PDT) [37]. Two principal lipid signaling networks controlling nods of actions determining tumor responses to treatments like PDT will be elaborated below.

SCAP/SREBP pathway

Transcription factors family recognized as master regulators of lipid homeostasis are known as sterol regulatory element binding proteins (SREBPs) [38,39]. These key physiological regulators of lipid synthesis engaged as upstream controllers of lipid metabolism are required for transcriptional activation of lipogenic gene expression [20]. It is well known that the activity of SREBPs is highly upregulated in various cancers [40]. Elevated intracellular cholesterol levels in cancer cells are maintained by SREBP-mediated activation of LDL receptor-controlled cholesterol uptake and suppressing its export via ABCA1 transporter. Together with mTOR pathways SREBPS activity is critical for accumulation of lipid droplets in cancer cells [20]. The important roles of SREBPs in many critical functional cellular processes include ER stress responses, cell apoptosis, autophagy, phagocytosis, and innate immune responses of macrophages [41]. Genes encoding NLRP inflammasome proteins are directly activated by SREBP-1a isoform, and this inflammasome is responsible for the upregulation of caspase-1 (that elicits programmed cell death by apoptosis or pyro ptosis). On the other hand, caspase-1 is known to trigger the export of SREBPs from the ER [42].

The three isoforms of this transcription family SREBP-1a, SREBP-1c, and SREBP-2 have each different roles in lipid synthesis [39]. The SREBP-1c isoform is responsible for fatty acid synthesis and energy storage, SREBP-1a for genes required for global lipid synthesis and growth, while SREBP-2 is relatively specific for cholesterol synthesis. They are all synthetized as inactive precursors anchored to the ER membranes. For their maturation these precursors translocate to the Golgi to undergo a sequential two-step cleavage process mediated by membrane-bound site-1 protease (S1P) and site-2 protease (S2P). This releases the transcriptionally active SREBP forms with NH₂-terminal domain, which enter the nucleus as homodimers and bind to sterol regulatory element (SRE) sequences in DNA to stimulate the transcription of target lipogenic genes. Upon completing their function, nuclear SCREBP forms are degraded by ubiquitin E3 ligase-mediated proteasome system [40].

Signal transduction leading to SREBPs-mediated transcription is triggered by sensing the decline in the cellular cholesterol levels by a cholesterol-sensing protein named SREBP cleavage-activating protein (SCAP) situated in the ER [43]. Newly synthesized SREBP precursors remain in the ER by forming a complex with SCAP, which is kept in this location by binding to the ER-resident insulin-induced gene proteins (INSIGs). The SCAP/INSIG association is strengthened by binding cholesterol molecules. With the decline in cholesterol levels SCAP dissociates from INSIG, and this will facilitate the incorporation of SCAP/SREBP into coatomer II-coated vesicles that transport these complexes from the ER to the Golgi. The activation of SREBPs is regulated by a negative feedback loop whereby newly synthesized cholesterol and lipids enhance the binding of SCAP and INSIG to retain the SCAP/SREBP complexes in the ER [40].

We have hypothesized that hyper oxidative stress induction in tumor cells by PDT provokes a pronounced activation of SREBPs for regulating both pro-survival and programmed death responses in these cells and tested this by investigating the effect of SREBP inhibition. One SREBP inhibitor that has been extensively tested in cancer cells is the diary thiazole derivative fatostatin A, which binds to SCAP and blocks the translocation of the SCAP/SREBP complex to the Golgi [40]. The survival of PDT-treated mouse SCCVII tumor cells were significantly reduced in presence of fatostatin A, which demonstrates that SREBPs have a critical role in the cancer response to PDT [44]. This is supported by the findings from the same study showing that the additional killing effect in PDT-treated tumor cells can be obtained by further impairing SREBP activity with combining fatostatin A with INF-4Es (selective caspase-1 inhibitor).

Sphingolipid signaling cascades

In addition to phospholipids and cholesterol, the role of the major component of cellular membrane lipid bilayers is held also by sphingolipids [45]. While glycerol backbone-structured phospholipids are distributed throughout the cell membrane, sphenoid base backbone-built sphingolipids tend to concentrate in membrane domains known as lipid rafts [46]. The major classes of sphingolipids are ceramides, sphingomyelins, cerebrosides, and gangliosides [45]. While the main controller of lipid synthesis genes is the SREBP family, the enzymes responsible for sphingolipid synthesis are not regulated by SREBPs, but rather the opposite, the activity of SREBPs is controlled by sphingolipids at the post-transcriptional level [46].

In addition to playing a central role in in the structural stability of membranes of cells and cellular organelles, sphingolipid family includes a number of highly bioactive lipids with pivotal roles in important cellular signaling pathways. The main signal transduction cascade starts when the membrane sphingomyelin is hydrolyzed into ceramide (regarded as the main lipid second messenger) and sphingosine-1-phosphate (S1P), which both have critical but opposite roles in cellular responses to stress [47]. While S1P acts as s pro-survival factor and ligand to G protein-coupled receptors, ceramide triggers intrinsic and extrinsic apoptotic pathways via receptor-independent signals.

Tumors are characterized by disrupted sphingolipid metabolism and abnormal profile with altered expression of various sphingolipid species and their metabolic enzymes [48]. Key sphingolipids profoundly impact tumor biology and are implicated in cancer development. They strongly influence tumor cell conduct, stromal cell activity including immune cell behavior, and affect tumor aggressiveness, angiogenesis, and extracellular matrix remodeling [48].

Sphingolipid signal messengers have a critical impact in tumor cell signaling elicited following PDT treatment [49,50]. The levels of many sphingolipids in cancer cells become highly elevated after diverse stress-inducing insults including tumor PDT, which can be attributed to the induction of de novo synthesis instigated by the upregulation of the involved enzymes such as dihydroceramide desaturase [51]. The resultant rise in tumor concentrations of principal sphingolipids, including ceramide and S1P, which have decisive roles in cell survival or death as well as in immune surveillance [52]. In particular, C(18)-ceramide occupies a principal role in programmed cell death pathways signal transduction directing to mitochondrial apoptosis or alternatively to other death pathways including necroptosis and mitophagy [53]. In contrast, blocking the conversion of ceramide into sphingosine was found to dampen the activity of immunoregulatory effectors (both lymphoid and myeloid) dedicated to preventing tumor immune rejection [49]. We have shown that ceramide, S1P, and sphingosine, which become exposed and/or released from PDT-treated cell surface, possess the capacity to act as damage-associated molecular patterns (DAMPs) [49,54,55]. As DAMPs, these sphingolipid molecules induce inflammasome formation and influence the activity of inflammatory and immune cells, promote NF-κB signaling, and orchestrate immunogenic cell death (ICD) in affected cancer cells [49]. Ceramidase inhibitor LCL521 and a variety of other agents targeting sphingolipid activity were demonstrated to have a strong impact on antitumor immune activity and therapeutic outcome of tumor PDT [49,56-58]. Notably, the positive therapeutic impact on tumor PDT outcome demonstrated with adjuvant LCL521 treatment can be at least partially attributed to the effective selective restriction of the activity of two principal immunoregulatory cell populations (regulatory lymphoid cells – Tregs and myeloid-derived suppressor cells – MDSCs) [56].

Lipid peroxidation is one of regular events in normal lipid metabolism, particularly in cancer cells. Cellular lipids are continuously exposed to oxidative attack by reactive oxidative species (ROS), particularly at sites of carbon-carbon double bonds in unsaturated phospholipids, which leads to the abstraction of allylic hydrogen and formation of carbon-centered lipid radical (L*) [59] (Figure 1). This triggers the propagating process in which L* rapidly reacts with oxygen and forms a lipid peroxy radical (LOO*) that then abstracts a hydrogen from another lipid molecule creating another L* and lipid hydroperoxide (LOOH) in the continuing chain reaction. Presence of proximal chain-breaking antioxidants including low-level nitric oxide (NO), α-tocopherol, and butylated hydroxytoluene (BHT) can suppress this primary process of lipid peroxidation [60]. Hence, LOOH are the primary products of the lipid peroxidation reactions that are considered an integral process (and common feature) of lipid metabolism [36]. Invariably, ROS interacting with lipids in this process are also normal byproducts of cellular metabolic activity such as aerobic respiration. It was estimated that a typical cell produces every second around 50 hydroxyl radicals [59]. Cells with higher metabolic rates (with notable examples of cancer cells) produce greater rates of ROS and experience higher levels of lipid peroxidation. On the other hand, elevated labile iron pool inherently harbored by many malignancies [61] causes cancer cells to be predisposed for Fenton-catalyzed redox reactions resulting in the formation of hydroxyl radical with consequent intensified lipid peroxidation activity [62].

Mechanisms for controlling lipid peroxidation within cells rely on well-developed enzymatic and non-enzymatic antioxidant systems that can be mobilized by adaptive stress response [63]. Among the most prominent is cysteine/glutamate antiporter/glutathione peroxidase 4 (GPX4) axis [36]. The best demonstration that lipid hydroperoxide removal is essential for mammalian life was obtained using the lethal phenotype of GPX4-knockout mice, which cannot survive past embryonic day-8 [64]. The plausible explanation is that excessive lipid peroxidation can result in irreparable cell membrane damage with associated damage to proteins and nucleic acids. This can trigger multiple cell death pathways, with specific lipid peroxidation products steering the progression of different programmed cell death pathways, including ferroptosis, necroptosis, pyroptosis, lethal autophagy, and apoptosis [36].

Of particular importance is that, out of various aldehydes that can be formed as secondary lipid peroxidation products, malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE) act as key signaling molecules. Depending on the generated levels, they may exert significant pro-survival or pro-death impact on cells (particularly tumor cells) undergoing lipid peroxidation upon sustaining oxidative stress [63]. Among different products of lipid peroxidation, MDA is considered most mutagenic while 4-HNE appears most toxic.

Upon its formation 4-HNE can, depending on metabolic situation in the cell, promote cell survival or death. In pro-survival role, 4-HNE can play important role as signaling molecules stimulating the expression of genes for transcription factors such as Nrf2 (to enhance detoxifying antioxidant capacity of cells) or AP-1 and NF-κB for key survival pathways in sub lethally stricken cancer cells [59]. At elevated levels, 4-HNE forms abundant amounts of adducts with proteins and DNA with eventual cytotoxic and genotoxic consequences and induction of ferroptosis, apoptosis or other forms of programmed cell death [65].

As a product of lipid peroxidation that can be assayed spectrophotometrically, MDA has been used as biomarker of both lipid peroxidation and oxidative stress [66]. High capacity of reaction with multiple biomolecules with adducts formation is also exhibited by MDA. Reports on MDA-mediated signaling activity include i) the engagement in a signaling pathway controlling glutamine synthetase activity and generation of glutamate [67], ii) signaling cascade regulating islet glucose-stimulated insulin secretion dominated by Wnt pathway [68], and iii) controlling the expression of specificity protein-1 (Sp1) gene and the levels of Sp1 and Sp3 proteins that are important transcription factors for genes involved in chromatin remodeling [69].

Therefore, while the primary lipid peroxidation products have at elevated levels notable detrimental effects on their own, they are inherently unstable and tend to break down to form secondary lipid peroxidation products. These end-products of chain lipid peroxidation are potentially more harmful and cytotoxic [59].

The process of lipid peroxidation may last for hours or even days after its initiation, particularly in hypoxic tumor regions [70]. Such extended lifetimes facilitate translocation of lipid peroxide molecules to other intercellular sites and infliction of damage to biomolecules distant to the initial lipid radical formation. In addition to undergoing damage-enhancing one-electron reduction or damage-attenuating two-electron reduction at membrane or lipoprotein sites of origin, LOOHs can also become engaged in such reactions upon translocating to other membrane/lipoprotein sites. Phospholipid- and cholesterol-derived hydroperoxides (PLOOHs and ChOOHs, respectively) are known to be capable of this translocation [71]. Most of the studies dealing with this process have been carried out with photodynamically-generated ChOOHs in PDT model systems [54,72,73]. Spontaneous as well as protein mediated ChOOH transfer between membranes has been described. Specific positional ChOOHs such as free radical-derived 7α-OOH and ¹O₂-derived 5α-OOH transfer between membranes at different rates, the former moving much faster than the latter [72]. Spontaneous intermembrane transfer of PLOOHs, although faster than that of parent PLs, is still much slower than that of all ChOOHs [72]. The first evidence for the cytotoxic potential of translocated ChOOHs was obtained about 20 years ago, using liposome-bearing ChOOHs as donors and GPx4-null breast cancer COH-BR1 cells as acceptors [72]. Cell viability decreased exponentially during liposome/cell incubation, and this was prevented by desferrioxamine, an iron chelator/redox inhibitor, indicating that free radical damage initiated by one-electron turnover of translocated ChOOHs was responsible. Transfer proteins are known to play an important role in lipid metabolism and membrane biogenesis. Two well-known intracellular examples are (i) sterol carrier protein-2 (SCP-2), a non-specific Ch transporter [72,73], and (ii) steroidogenic acute regulatory (StAR)-family proteins [74]. In each case, model studies have described the potential pathological effects of ChOOHs, which might accompany parent Ch in subcellar delivery, e.g., to mitochondria [73,75]. Such effects could include (a) reduced testosterone output of Leydig cells due to loss of Cyp11A1 function [76] and (b) deficient Ch homeostasis in vascular macrophages due to impaired reverse Ch transport, potentially leading to atherogenesis [75]. Some of these ChOOHs might arise from antitumor PDT action, although no negative effects specifically attributed to them have been described up to now. It is important to point out that oxidative stress generated PLOOHs might also be disseminated via their transfer proteins. However, much less is known about this than that pertaining to ChOOHs, and much remains to be learned about the negative consequences of PLOOH transfer.

Lipids are increasingly recognized as fundamental participants in cancer-associated metabolic reprogramming. They have pivotal roles as key cellular signal transduction molecules governing critical cell responses in tumor microenvironment, particularly following therapy-related oxidative stress. Signaling activity controlled by SCAP/SREBP pathway and sphingolipid signaling cascades are principal signal transduction networks determining oxidative stress responses in cancer cells, particularly those inflicted by cancer therapy. Peroxidation of cellular lipids triggered by ROS, which is accentuated in cancer cells and additionally elevated in these cells by therapy-induced oxidative stress, is a fundamental event for generating prominent potent cell signaling molecules. While primary lipid peroxidation products such as phospholipid and cholesterol hydroperoxides (PLOOHs, CHOOHs) are important participants, various aldehydes formed as secondary lipid peroxidation products are potentially more impactful signaling molecules with potent mutagenic and/or cytotoxic effects.

One of the critical unresolved issues of lipid peroxidation is the actual impact of relatively long persisting lipid peroxidation products translocated to distant intercellular sites. Better understanding of this largely ignored and neglected phenomenon, including deciphering its possible role in lethal ferroptosis, should lead to a more precise identification of its relevance for therapeutic outcomes.

Both authors have no conflict of interest to declare.

The research of AWG was supported by USPHS Grant CA70823 from the National Cancer Institute and by Grant 5520347 from the Advancing a Healthier Wisconsin Research and Education Foundation.

Both authors contributed equally to the manuscript.