Emerging targets in lipid metabolism for cancer therapy

Abstract

Cancer cells perturb lipid metabolic pathways for a variety of pro-tumorigenic functions, and deregulated cellular metabolism is a hallmark of cancer cells. Although alterations of lipid metabolism in cancer cells have been appreciated for over 20-years, there are no FDA-approved cancer treatments that target lipid-related pathways. Recent advances pertaining to cancer cell fatty acid synthesis, desaturation, and uptake, microenvironmental and dietary lipids, and lipid metabolism of tumor-infiltrating immune cells have illuminated promising clinical applications for targeting lipid metabolism. This review will highlight emerging pathways and targets of tumor lipid metabolism that may soon impact clinical treatment.

Lipid metabolism is altered in cancer cells

It has been appreciated for over 20-years that induction of de novo fatty acid synthesis (FAS) is a hallmark of cancer cells [1]. FAS is catalyzed by fatty acid synthase (FASN). In normal adult tissues, fatty FASN expression, and thereby FAS, is elevated predominantly in liver, adipose, and lactating mammary glands, indicating most normal tissues acquire lipids from circulation [1]. Conversely, cancer cells synthesize up to 90% of cellular lipids de novo in normoxic conditions [2, 3]. This distinction between non-transformed and cancer cells has made FASN an attractive therapeutic target for which clinical trials are currently ongoing [4]. However, this distinction simultaneously complicates targeting lipid metabolism for cancer therapy. Despite activating FAS, cancer cells still utilize circulating lipids, which allows for significant metabolic flexibility. Recently, the field of cancer lipid metabolism has illuminated new therapeutic strategies, including targeting fatty acid uptake through CD36 (see glossary) or lipid desaturation through stearoyl-CoA desaturase 1 (SCD1) [5, 6]. Furthermore, the field has started to clarify tumor microenvironment and organ-specific differences in lipid metabolism which can significantly impact the efficacy of lipid inhibitors [7, 8]. Dietary lipids play a role in regulating in vivo tumor metabolism, and recent studies have demonstrated how strategic dietary manipulation reduces the metabolic flexibility of cancer cells to enhance therapy [9]. Despite the field’s growth, there are no FDA-approved cancer treatments that target lipid metabolic pathways, and it is unclear which therapeutic strategies are most likely to succeed past preclinical studies.

In this review, we provide an integrated overview of emerging lipid metabolic targets related to fatty acid synthesis, fatty acid uptake, and lipid desaturation in cancer cells. In addition, we describe emerging strategies that exploit lipid metabolism to overcome cell death resistance in cancer cells. We highlight how lipids within organ-specific tumor microenvironments (TME) alter tumorigenesis through lipid-related pathways and protect or sensitize tumors to lipid-inhibitors. Furthermore, we discuss how lipids impact tumor progression and tumor infiltrating immune cells, the therapeutic potential of prescribed lipid-related diets, opportunities to enhance immune checkpoint inhibitors (ICIs), and lipid-inhibitors in clinical trials for cancer therapy.

Fatty acid synthesis

FAS is activated in cancer cells and targetable through inhibition of acetyl-CoA carboxylase (ACC) or FASN [1]. The first and rate limiting step of FAS is catalyzed by ACC, converting cytosolic acetyl-CoA to malonyl-CoA (Fig. 1) [10]. Acetyl-CoA is classically produced by ATP citrate lyase (ACLY) from cytosolic citrate; however, acetyl-CoA is also produced by acetyl-CoA synthetase enzymes (ACSS) from acetate or by acetyl-CoA transfer from mitochondria and peroxisomes to the cytosol [11–13]. FASN catalyzes the second step of FAS, elongating acetyl-CoA by 2-carbon units through sequential additions of malonyl-CoA [10]. This process requires NADPH as an electron donor and continues until palmitate is produced [10]. Palmitate is a 16-carbon saturated fatty acid (C16:0, where C16 indicates 16 carbons and 0 indicates no double bonds) and is the most abundant fatty acid in humans [10, 14]. ACC and FASN are commonly upregulated in many cancer types and support a variety of cellular functions, including membrane biosynthesis, activation of oncogenic signaling pathways, protein acylation, and escaping cell death, among others (reviewed [10, 15]). Notably, inhibition of either ACC or FASN reduces FAS. However, FASN-inhibition additionally causes malonyl-CoA accumulation, which exerts additional toxicity. First, malonyl-CoA inhibits carnitine palmitoyltransferase 1 (CPT1) on the mitochondrial outer membrane, which mediates the transfer of fatty acids to mitochondria for fatty acid oxidation (FAO), and therefore inhibits FAO (Fig. 1) [14–16]. Second, as was recently reported, malonyl-CoA inhibits mTORC1 [17]. Both ACC and FASN are potential targets for inhibiting fatty acid synthesis in cancer.

Figure 1: Fundamentals of lipid metabolism.

Cells synthesize fatty acids de novo (FAS) or acquire them through extracellular uptake. FAS occurs in the cytosol and is catalyzed by ACC and FASN enzymes, which produce palmitate, a 16-carbon SFA. MUFAs are produced by desaturation of SFAs by SCD1 or FADS2. The PUFAs, ALA and LA, are essential fatty acids and serve as building blocks for PUFA biosynthesis. Fatty acid uptake occurs through free fatty acid diffusion or CD36-mediated fatty acid uptake. When intracellular levels of palmitate are high, palmitoylated CD36 promotes preferential MUFA uptake. The composite intracellular fatty acid pool is comprised of SFAs, MUFAs, and PUFAs derived from FAS and extracellular uptake. Fatty acids are activated by ACSL enzymes and incorporated into glycerolipids and phospholipids to support membrane biogenesis and lipid storage. Fatty acid β-oxidation (FAO) occurs in the mitochondria and supports energy production by driving the TCA-cycle. Acetyl-CoA also supports the mevalonate pathway to produce cholesterol.

Abbreviations: ALA: alpha-linoleic acid, LA: linoleic acid, PUFA: polyunsaturated fatty acid, MUFA: monounsaturated fatty acid, SFA: saturated fatty acid, CTP1: carnitine palmitoyl transferase 1, FAO: fatty acid oxidation, TCA Cycle: tricarboxylic acid cycle, ETC: electron transport chain, ACSS2: acetyl-CoA synthetase enzyme 2, ACLY: ATP citrate lyase, ACAT: acetyl-CoA acetyltransferase, HMGCS: hydroxyl-3-methylglutaryl-CoA synthetase, HMGCR: hydroxyl-3-methylglutaryl-CoA reductase, ACC: acetyl-CoA carboxylase, FASN: fatty acid synthase, SCD1: stearoyl-CoA desaturase 1, FADS2: fatty acid desaturase 2, ELOV: elongation of very-long-chain fatty acids protein, G3P: glycerol-3-phosphate, GPAT: glycerol-3-phosphate acyltransferase, LPA: lysophosphatidic acid, PA: phosphatidic acid, DAG: diacylglycerol, TAG: triacylglycerol, DGAT: diglyceride acyltransferase, LD: lipid droplet, PC: phosphatidylcholine, PE: phosphatidylethanolamine, PS: phosphatidylserine, PI: phosphatidylinositol.

Acetyl-CoA Carboxylase

ACC has emerged as a controversial target for cancer therapy [10]. ACC inhibitors show excellent activity in preclinical models of hepatocellular carcinoma (HCC) and oncogenic KRAS-driven lung adenocarcinoma (LUAC) [18, 19]. Indeed, they potently inhibit FAS in both cancer types [18, 19]. However, other studies demonstrate tumor promoting effects of ACC inhibition. AMPK-dependent phosphorylation of ACC promotes cell death resistance by increasing NADPH and reducing lipid peroxidation [20–22]. Furthermore, inhibition of ACC promotes metastasis in breast cancer through increased acetylation [23]. Moreover, ACC inhibitors first trialed for fatty liver disease in humans caused unwanted hypertriglyceridemia [24]. As a result, ACC inhibitors are less poised to proceed further for cancer therapy.

Fatty acid synthase

Inhibition of FASN has emerged as a promising therapeutic strategy for oncogenic KRAS-mediated LUAC, HCC, and glioblastoma multiforme (GBM) based on recent preclinical studies and early-stage clinical trials. In LUAC, oncogenic mutations of KRAS increase FASN expression, and FASN-inhibition shows antitumor activity in KRASmut tumors in both preclinical studies and a stage I clinical trial (TVB-2640, NCT02223247) [25–27]. Notably, ~25% of LUAC patients carrying oncogenic KRAS lose their wildtype KRAS allele, making these tumors more sensitive to FASN-inhibition [28]. Mechanistically, oncogenic KRAS-driven LUAC rely on FASN to promote synthesis of saturated fatty acids (SFAs) and monounsaturated fatty acids (MUFAs) to outcompete polyunsaturated fatty acids (PUFAs) in membrane phospholipids and prevent ferroptosis (see ferroptosis section for further discussion) [26]. An ongoing stage II clinical trial (NCT03808558) for non-small cell lung cancer (NSCLC) carrying oncogenic KRAS will provide further clarity for this patient population. In HCC, the FASN-inhibitor, TVB-3664, showed promising antitumor activity in multiple mouse models as a monotherapy [29]. The combination of TVB-3664 with tyrosine kinase inhibitors (TKI) used for advanced HCC, either sorafenib or cabozantinib, led to tumor regression [29]. However, one HCC subtype is FASN-independent and resistant to TVB-3664 [29, 30]. Also, TVB-3664 did not enhance ICI, which is now first-line treatment for HCC [29, 31]. Surprisingly, patients with HCC are absent from the 6 ongoing and completed clinical trials for FASN-inhibition in cancer. Finally, TVB-2640 combined with bevacizumab (VEGF inhibitor) showed favorable antitumor activity compared to historical controls in a phase II clinical trial of GBM, which is a rapidly fatal high grade brain tumor [32]. A phase III trial testing the same drug combination is now fully enrolled (NCT05118776).

An important new study suggests that the efficacy FASN-inhibitors may be limited to brain metastases in HER2⁺ breast cancer [7]. FASN has been implicated in HER2⁺ breast cancer based on a feed-forward mechanism orchestrated between the two proteins: HER2 can phosphorylate FASN to enhance its enzymatic activity, and FASN can enhance HER2 expression and activity [33–36]. Furthermore, FASN-inhibition is therapeutic in HER2-driven mouse models and reverses Herceptin-resistance [34, 35]. However, new results challenge this broad hypothesis, as either FASN-deletion or TVB-3664 displayed minimal antitumor activity against HER2⁺ breast cancer cells grown in the mammary fat pad and the liver [37]. Importantly, FASN was indispensable for tumors grown in the brain due to restrained metabolic flexibility [37]. The lack of extracellular lipids in the central nervous system (CNS) tumor microenvironment (TME) forces brain metastases to depend entirely on endogenous lipid synthesis [37]. Consistently, the authors showed consistent upregulation of FASN in patient samples of brain metastases compared to primary and extracranial lesions [37]. These results indicate a great clinical opportunity for targeting FASN in breast cancer patients with brain metastases, but also question if FASN is a worthy target for extracranial disease. An ongoing phase II clinical trial (NCT03179904) testing TVB-2640 with Herceptin in metastatic HER⁺ breast cancer is ongoing and will help address this question.

Although TVB-2640 has reached clinical testing, more preclinical and early-stage studies are required to fully realize the potential of FASN inhibition for cancer therapy. First, with scarce clinical data, it is unclear which cancers will most benefit from FASN-targeted therapy. Two other early-stage trials are currently recruiting for metastatic castration resistant prostate cancer (NCT05743621) and resectable colon cancer (NCT02980029), which together with the previously mentioned studies will provide further preliminary evidence. Furthermore, more preclinical studies are required to determine generalizability of FASN-dependency in brain metastases, particularly for lung cancer which often disseminates to the CNS. Second, what therapy combination will best synergize with FASN-inhibition? In TMEs where lipids are plenty (i.e. not the CNS), inhibition of fatty acid uptake through CD36 has been proposed to limit compensation to FASN-inhibition [38, 39]. The combination of a PUFA-rich diet and FASN inhibition effectively increases PUFA-membrane incorporation, which could enhance ferroptosis in oncogenic KRAS LUAC (discussed further in section on ferroptosis) [40]. Finally, FASN suppresses antitumor immunity in both cancer cells and T_reg cells, which warrants further studies combining FASN inhibitors with ICIs [41, 42] (reviewed [43]).

Saturated fatty acid-induced lipotoxicity

An emerging theme in cancer biology is the mechanisms by which tumor cells maintain lipid homeostasis—the balance between SFAs, MUFAs, and PUFAs (Fig. 1–2). Palmitate is elongated by elongase enzymes (ELOV) to produce stearate (C18:0) and longer SFAs [9]. SCD1 desaturates palmitate and stearate to produce palmitoleate (C16:1n-7, where n-7 denotes the first double bond 7-carbons from the fatty acid’s methyl end) and oleate (C18:1n-9), respectively [5]. Non-oxygen dependent fatty acid desaturase 2 (FADS2) converts palmitate to sapienate (C16:1n-10) in some cells [44]. Palmitoleate, oleate, and sapienate are all MUFAs. Alpha-linolenic acid (ALA; 18:3n-3) and linoleic acid (LA; 18:2n-6) are PUFAs that cannot be synthesized intracellularly, meaning they are essential and only provided through diet and extracellular uptake [9]. ALA and LA are acted on by FADS and ELOV enzymes to make a diverse group of PUFAs (Fig. 2) [9]. Fatty acids are activated and incorporated into more complex lipids, including glycerolipids and phospholipids (Fig. 1). Glycerolipids consist of acyl groups attached to a glycerol backbone, such as monoacylglycerol (MAG), diacylglycerol (DAG), and triacylglycerol (TAG) [45]. TAGs are the main form of lipid storage and are the predominant lipid in lipid droplets, which are lipid-containing organelles bound by a lipid-monolayer [45]. Phospholipids make up membrane lipids and consist of acyl groups attached to a phosphorylated glycerol molecule, examples of which include phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylinositol (PI), among others (Fig. 1) [45]. Alterations in the lipid composition of both glycerolipids and phospholipids significantly impacts cancer cell function and resistance to cell death [9].

Figure 2: Chemical structure of commons long chain fatty acids.

Fatty acid synthesis produces palmitic acid, which can be elongated by ELOV enzymes to produce stearic acid. The SFAs, palmitic acid and stearic acid, are substrates for oxygen-dependent SCD1, producing the MUFAs, palmitoleic acid and oleic acid, respectively. Palmitic acid is also a substrate for FADS2, which produces the MUFA, sapienic acid. Polyunsaturated fatty acids are derived from essential fatty acids linoleic acid (an n-6 PUFA) or alpha-linoleic acid (n-3 PUFA). Note, this figure shows examples but is not exhaustive of all PUFAs. For omega nomenclature, n-x, x indicates where the first double bond is from the methyl end of the fatty acid.

Abbreviations: SCD1: stearoyl-CoA desaturase 1, FADS2: fatty acid desaturase 2, ELOV: elongation of very-long-chain fatty acids protein, SFA: saturated fatty acid, PUFA: polyunsaturated fatty acid, MUFA: monounsaturated fatty acid.

Mechanism of SFA-induced lipotoxicity

SFA-induced lipotoxicity occurs when SFAs accumulate sufficiently in membrane phospholipids and glycerolipids, leading to endoplasmic reticulum (ER) stress and apoptosis [5]. High levels of exogenous SFAs or inhibition of SCD1, either pharmacologically or physiologically (e.g. hypoxia), cause SFAs to accumulate in membrane phospholipids, while MUFAs oppose this process [3, 46–48]. Specifically, accumulation of glycerolipids and phospholipids with two saturated acyl groups (di-saturated) drives apoptosis [46]. MUFAs prevent synthesis of these di-saturated lipids, preventing SFA-toxicity [46]. This indicates the ratio of SFAs to MUFAs (SFA/MUFA) is an important read out of lipid homeostasis. Furthermore, the phospholipid phosphatidylinositol with two MUFA acyl groups (PI-MUFA₂) directly inhibits ER stress and apoptosis, demonstrating another way MUFAs can inhibit lipotoxicity [49]. Once di-saturated lipids accumulate in membranes, IRE1α, an ER stress kinase, senses membrane saturation through its transmembrane domain and plays a central role in either restoring the SFA/MUFA ratio or executing cell death (Fig. 3A) [46, 50, 51]. Notably, since dietary and TME SFAs can promote cancer progression [52–54], cancer cells must withstand SFA-induced toxicity to enable their tumor-promoting effects. Indeed, cancer cells employ a variety mechanisms to prevent toxic SFA-accumulation, including SCD1-induction [51], SFA-buffering in TAGs [46, 55], release of MUFAs from TAGs [56], and exogenous MUFA uptake [3, 47]. Recently, it was shown that CD36 can prevent the lipotoxicity of SFAs by preferentially increasing the uptake of MUFAs [47].

Figure 3: Saturated fatty acid-induced toxicity for cancer treatment.

A) ER stress kinases IRE1α and PERK are activated by misfolded proteins and increased membrane lipid saturation independent of their misfolded-protein domains (represented by truncated cartoon of each protein). Through a variety of mechanism, including SREBP-activation and IRE1α-dependent XBP1s-activation, adaptive ER stress signaling induces genes involved in de novo lipogenesis to promote membrane synthesis and ER homeostasis. For example, MYC-driven tumors rely on pro-survival ER stress signaling through XBP1s, and inhibition of XBP1s-activation inhibits tumor growth.

B) During the same process, SCD1 inhibition drives accumulation of SFAs, which are incorporated in membrane lipids further activating ER stress and ultimately inducing IRE1α-dependent apoptosis. In MYC-driven tumors, inhibition of SCD1 phenocopies XBP1s-inhibition.

C) In HCC, targeted activation of LXRα promotes induction of FAS genes including ACC, FASN, and SCD1. Raf inhibition disrupts its direct binding to SCD1, thereby destabilizing SCD1. When used together, saturated fatty acids accumulate to toxic levels, induce ER stress, and inhibit HCC growth in vivo.

D) A calorie restricted diet decreases lipids in the serum and tumor microenvironment as well as decreases SCD1 protein level in tumor cells. This leads to toxic SFA accumulation which inhibits PDAC growth in vivo.

E) A ketogenic diet increases serum and tumor microenvironmental lipids, so despite decreasing SCD1 protein, PDAC cells maintain a balance of the SFA/MUFA ratio through exogenous sources.

F) A ketogenic diet designed with a higher SFA/MUFA ratio causes increased SFA levels intracellularly and inhibits PDAC growth in vivo.

G) In ovarian cancer, targeting SCD1 in the presence of an SFA-rich diet promotes ER stress and inhibits peritoneal metastases.

Abbreviations: SCD1: stearoyl-CoA desaturase 1, FASN: fatty acid synthase, ACC: acetyl-CoA carboxylase, XBP1s: spliced X-box protein 1, SREBP: sterol regulatory element binding protein, LXRα: liver X receptor α, FAS: fatty acid synthesis, HCC: hepatocellular carcinoma, PDAC: pancreatic ductal adenocarcinoma, SFA: saturated fatty acid, MUFA: monounsaturated fatty acid, PERK: protein kinase RNA-like ER kinase, IRE1α: inositol-requiring enzyme type 1.

Stearoyl-CoA Desaturase 1

SCD1 is the primary monodesaturase in human cancer [5], and has emerged a therapeutic target in multiple tumor types. The lipotoxic mechanism exerted by SCD1 inhibition is best described in GBM and MYC-driven tumors: SCD1 inhibition induces a vicious cycle of SFA-accumulation and ER stress leading to apoptosis (Fig. 3B) [51, 57, 58]. Importantly, SCD1 inhibition sensitizes GBM cells to DMA damage in vitro and demonstrates a pronounced survival benefit in vivo when added to temozolomide (TMZ), an alkylating agent used for adjuvant treatment in GBM [58, 59]. In acute lymphoblastic leukemia (ALL), dissemination to the CNS requires high SCD1 expression to overcome the lipid depleted TME [60]. In HCC, SCD1 stability is dependent on the protein RAF (Fig. 3C) [61]. When combined with activators of FAS, RAF inhibition reduces SCD1, promotes SFA-toxicity, and reduces HCC growth in vivo [61]. Finally, ovarian cancer cells depend on SCD1 to prevent SFA-induced lipotoxicity [62].

Intrinsic and acquired mechanisms of resistance to SCD1 inhibition have been described in multiple types of cancer. Tumors that lose PTEN-activity by deletion or methylation often lose SCD1 expression through bystander co-deletion [59]. Surprising, these cells are intrinsically resistant to SCD1 inhibition despite retaining low level SCD1 activity [59]. Furthermore, some cancers are intrinsically resistant SCD1 inhibition through FADS2-dependent sapienate synthesis [44]. GBM cells can acquire resistance to SCD1 inhibition via robust induction of SCD1 which is reversed by inhibiting the transcription factor FOSB [59]. A second identified resistance mechanism is lipid buffering through triglyceride synthesis which is ameliorated by DGAT1/2 inhibition (see Fig. 1 for DGAT function) [58].

Dietary control of tumoral SFA/MUFA

A new study indicates strategic dietary interventions can increase the SFA/MUFA for cancer therapy [63]. Dietary intervention has long been considered as an adjuvant strategy for cancer therapy (reviewed [64]). In mouse models of LUAC and pancreatic ductal adenocarcinoma (PDAC), a ketogenic diet (KD) and calorie restricted diet (CR) both lowered SCD1 expression in tumors, likely due to decreased serum insulin [65]. However, only a CR increased the tumoral SFA/MUFA ratio and inhibited tumor growth [65]. Mechanistically, the CR lowered both systemic and TME lipids [65]. As a result, cancer cells were unable to compensate for decreased SCD1 activity through uptake of extracellular lipids (Fig. 3D) [65]. In contrast, the KD increased systemic and TME lipids, and oleate particularly (a KD is oleate rich) [65]. So, despite a moderate decrease in SCD1 expression, cells maintained their SFA/MUFA ratio from exogenous sources (Fig. 3E) [65]. Interestingly, a KD supplemented with palm oil (i.e. high in palmitate and SFAs) was sufficient to imbalance the SFA/MUFA ratio and inhibit tumor growth (Fig. 3F) [65]. These results indicate that the specific lipid composition of a ketogenic diet can be engineered for cancer therapy. Indeed, the addition of a palm oil diet to SCD1 inhibition reduced ovarian cancer peritoneal metastasis better than SCD1 inhibition alone, indicating the potential synergy of diet and lipid inhibitors (Fig. 3G) [62].

Taken together, disrupting cancer cell lipid homeostasis through SFA-induced lipotoxicity is a promising strategy for cancer therapy. SCD1 inhibitors have been generated although none have reached clinical trials for cancer [66]. Notably, YTX-7739 is a clinical-stage oral SCD1-inhibitor with good blood brain barrier penetrance that showed strong activity in GBM mouse models [58]. While inhibitor development continues, further preclinical studies are required to determine the ideal treatment combination. The combination of CD36 and SCD1 inhibition has been proposed, consistent with a role for CD36 in MUFA-scavenging (discussed further in CD36 section) [47, 67]. Interestingly, SCD1 inhibition can activate antitumor immunity by reducing ER stress in CD8⁺ T cells despite increasing ER stress in tumor cells [68]. This suggests the combination of SCD1 inhibitors with ICIs is worthy of further exploration. Finally, although diets high in SFAs may cause concern given their tumor promoting role [53, 54], tumors appear to walk a fine line when enabling this effect and resisting SFA-toxicity. Thus, diets rich in SFAs should be further explored with targeted inhibitors for therapeutic purposes.

Ferroptosis

Ferroptosis is a newly described mechanism of cell death which is regulated in part by lipid metabolism with far reaching implications for cancer (reviewed [69]). Ferroptosis is initiated by peroxidation of phospholipids containing two PUFA acyl (PL-PUFA₂) by mitochondrial reactive oxygen species (ROS) (Fig. 4) [70]. The reaction is propagated by PUFA-containing phospholipids and iron [69]. Lipid peroxidation begins in the ER and ultimately disrupts the plasma membrane leading to cell death [69, 70]. PUFA-containing phospholipids are most susceptible to peroxidation due to multiple susceptible double bonds [69]. Conversely, MUFA-containing phospholipids are resistant to peroxidation and suppress ferroptosis [69]. This dichotomy between MUFAs and PUFAs implicates many aspects of lipid metabolism in the regulation of ferroptosis (Fig. 4). Indeed, cancer cells exert multiple mechanisms to resist ferroptosis, including maintenance of lipid homeostasis [69].

Figure 4: Lipids determine sensitivity to ferroptosis.

Ferroptosis is initiated by the reaction between a PC containing two PUFAs (PC-PUFA₂) and mitochondrial ROS to form a hydroperoxide species (PC-PUFA₂-OOH). PUFA-peroxidation is propagated through PUFA-containing phospholipids, leading to plasma membrane disruption and cell death. PC-PUFA₂ synthesis is promoted by increased PUFA uptake and activation by ACSL4, and this process is opposed by MUFAs following their activation by ACSL3. Lymphatics are rich in MUFAs and can promote metastasis by protecting cells from ferroptosis. Conversely, diets rich in PUFAs can inhibit tumor growth by inducing ferroptosis. FASN and SCD1 protect cells from ferroptosis by producing SFAs and MUFAs which outcompete PUFAs in membrane phospholipids. LDs protect cancer cells from ferroptosis by providing MUFAs and sequestering PUFAs to prevent their membrane incorporation. PUFA biosynthesis catalyzed by FADS1 and ELOV5 generates AA and DHA, which are PUFAs most sensitive to peroxidation. Both the GPX4/GSH and CoQ₁₀/FSP1 pathway are endogenous inhibitors of ferroptosis. Erastin and RSL3 are examples of ferroptosis inducing agents which both act on the GPX4/GSH pathway. Blue lines indicated ferroptosis suppression; red lines indicate ferroptosis promotion.

Abbreviations: PC: phosphatidylcholine, PUFA: polyunsaturated fatty acid, ROS: reactive oxygen species, ACSL: acyl-CoA synthetase long-chain family member, MUFA: monounsaturated fatty acid, FASN: fatty acid synthase, SCD1: stearoyl-CoA desaturase 1, SFA: saturated fatty acid, LD: lipid droplets, FADS1: fatty acid desaturase 1, ELOV5: elongation of very-long-chain fatty acids protein 5, AA: arachidonic acid, DHA: docosahexaenoic acid, LA: linoleic acid, ALA: alpha-linoleic acid, GPX4: glutathione peroxidase 4, GSH: glutathione, CoQ₁₀: coenzyme Q₁₀, FSP1: ferroptosis suppressor protein 1.

Intrinsic antioxidant systems

Cells employ multiple endogenous antioxidant pathways that suppress ferroptosis and affect cancer growth [69]. The primary antioxidant system in cancers is glutathione peroxidase 4 (GPX4), which utilizes glutathione (GSH) to neutralize lipid peroxides (Fig. 4) [71]. Cells employ other intrinsic lipophilic antioxidant systems including but not limited to FSP1/CoQ₁₀ and cholesterol intermediate 7-dehydrocholesterol which can also be exploited therapeutically in vivo [72–75]. Disruption of these antioxidant systems show antitumor activity in preclinical models, although identifying susceptible cancers and tolerable orally bioavailable compounds is still a work in progress [69, 71–75].

Fatty acid activation

Fatty acids influence the composition of membrane phospholipids to regulate ferroptosis sensitivity (Fig. 4) [69]. Indeed, free fatty acids are activated by long-chain fatty acyl-CoA synthetase (ACSL), generating fatty acyl-CoAs which are substrates for membrane phospholipid incorporation (Fig. 4) [69]. Importantly, PUFAs are activated by ACSL4, and deletion of ACSL4 renders cells resistant to ferroptosis [76]. Conversely, MUFAs are activated by ACSL3, and ACSL3 deletion prevents exogenous MUFAs from inhibiting ferroptosis [77]. The role of ACSL3 is further illustrated in mouse models of melanoma [8]. Hematogenous dissemination of melanoma cells is limited by ferroptosis [8]. However, lymphatic fluid is MUFA-rich and protects metastasizing melanoma cells from ferroptosis in an ACSL3-dependent manner [8]. Thus, changes in MUFA and PUFA metabolism can dictate ferroptosis.

Fatty acid synthesis and desaturation

ACC and FASN play opposing roles in regulating ferroptosis. Inhibition of ACC decreases levels of SFAs, MUFAs, and surprisingly PUFAs [21, 22]. As a result, ACC inhibition lowers PUFA-containing phospholipids and prevents ferroptosis [21, 22]. The mechanism by which ACC inhibition decreases PUFA levels is unclear – possibly through malonyl-CoA depletion. First, depleted malonyl-CoA activates FAO, which can degrade PUFAs through FAO [78]. Second, malonyl-CoA is required by ELOV enzymes in PUFA biosynthesis, including production of arachidonic acid (AA, C20:4n-6) and docosahexaenoic acid (DHA, C22:6n-3) [69, 79]. Conversely, FASN promotes synthesis of SFAs and MUFAs which outcompete PUFAs in membrane phospholipids (Fig. 4) [26, 80]. For example, in oncogenic KRAS LUAC, FASN inhibition increases PUFA-containing phospholipids, lipid peroxidation, and ferroptosis [26]. The authors also demonstrate this effect is dependent on the ratio of exogenous SFA/PUFA [26]. When SFAs become the predominant extracellular fatty acid, PUFA-uptake and subsequent incorporation into phospholipids is reduced, and ferroptosis is prevented [26]. This illustrates how extracellular or dietary lipids dictate metabolic flexibility (discussed further below). Overall, further work is needed to firmly establish the intricacies of FAS in ferroptosis.

Recent studies have linked SCD1-dependent MUFA production to cancer cell ferroptosis resistance (Fig. 4). PI3K-AKT signaling is commonly activated in human cancer and confers resistance to GPX4 inhibition [81]. Mechanistically, PI3K-AKT-mTORC1 signaling increases mature SREPB1, which in turn elevates the transcription of FAS genes [81]. As a result, FAS genes are induced, most significantly SCD1, which generates MUFAs [81]. Importantly, inhibition of SCD1 reverses ferroptosis resistance, which is attributed to the reduction of MUFA production [81] A second ferroptosis-resistant cancer is STK11 and KEAP1 double-mutant LUAC, an extremely aggressive LUAC subtype [82]. Although each mutation individually reduces ferroptosis sensitivity, the combination exerts pronounced ferroptosis resistance [82]. A genetic screen identified SCD1 as essential in STK11/KEAP1 mutant cells, and SCD1 inhibition restored ferroptosis sensitivity [82]. Furthermore, in ovarian, breast, and colon cancer, SCD1 promoted ferroptosis resistance [83–85]. Notably, SCD1 inhibition can induce either apoptosis or ferroptosis in ovarian cancer cells [83, 86]. The introduction of exogenous SFAs drives cells towards apoptosis [83, 86]. Unfortunately, the reciprocal experiment with exogenous PUFAs were not performed.

Enzymes involved in PUFA biosynthesis promote ferroptosis sensitivity (Fig. 4). AA- and DHA-containing phospholipids are most sensitive to lipid peroxidation [69]. Although cancer cells can acquire AA and DHA exogenously, a new study shows the ability to synthesize both intracellularly confers sensitivity to ferroptosis [87]. The authors found that intestinal-type gastric cancer cells are resistant to ferroptosis due genetic silencing of FADS1 and ELOV5 by hypermethylation [87]. Conversely, mesenchymal-type gastric cancer cells, which express FADS1 and ELOV5, actively produce AA and DHA predisposing them to ferroptosis [87].

Dietary PUFAs

Dietary PUFA-supplementation is an emerging strategy to promote intratumoral ferroptosis. In colon cancer mouse models, a fish oil diet rich (rich in n-3 PUFAs) reduced tumor growth and synergized with ferroptosis inducing agents (Fig. 4) [88]. The authors noted that colon cancer cells preferentially store PUFAs in triglycerides to prevent their membrane incorporation and reduce ferroptosis [88]. The addition of DGAT1 inhibitor enhanced the antitumor effect of a PUFA diet [88]. Introduction of ALA or linoleic acid (n-6 PUFA) intraperitoneally inhibited pancreatic cancer growth through ferroptosis [89]. Finally, chia-oil derived diets (rich in ALA) showed tumor suppressive properties in a breast cancer mouse model [90]. Strategies to enhance the activity of PUFA rich diets are emerging. In a study of fatty liver disease, the combination of a PUFA diet and FASN inhibition substantially increased intracellular PUFA levels [40]. Furthermore, a KD supplemented with fish oil (KD-FO) reduced murine lung tumor growth more than other types of KDs [91]. KD-FO remarkably reduced tumoral FASN protein compared to control mice, possibly through known negative regulation of SREBP1 by PUFAs [91, 92]. Notably, a separate study demonstrated a KD alone reduces tumor growth through ferroptosis [93]. However, the KD reduced adrenal corticosterone biosynthesis which accelerated cachexia and shortened survival [93]. The addition of a glucocorticoid to the KD reversed cachexia without affecting tumoral ferroptosis, which enabled the antitumor activity of the KD to improve survival [93]. Finally, dietary PUFA supplementation enhances antitumor immunity by increasing lipid peroxidation in tumor cells [94, 95] (a review on ferroptosis and antitumor immunity can be found here [96]). Indeed, further cancer studies exploring the combination of PUFA diets with lipid inhibitors and ICIs is warranted.

Targeting fatty acid uptake for cancer therapy

Fatty acid uptake occurs through diffusion and protein-mediated mechanisms. Free fatty acids can freely diffuse through the plasma membrane [97]. Diffusion can be enhanced by fatty acid binding proteins (FABPs) and fatty acid transport proteins (FATPs) [97]. FABPs bind to fatty acids and accelerate their removal from the membrane [97]. FATPs convert fatty acids to acyl-CoAs, similar to ACLS enzymes, thereby trapping them inside the cell [97]. Fatty acid uptake is also regulated by CD36 [97]. CD36 is a transmembrane protein that binds extracellular lipids and appears to promote their uptake through endocytosis [98]. Lipoproteins and macropinocytosis can also serve as sources for intracellular lipids [99, 100]. An important consideration for fatty acid uptake is extracellular lipid availability. Since fatty acid uptake occurs independently of transporter expression, high concentrations of extracellular lipids drive their intracellular uptake [101].

CD36 in tumor cells

Inhibiting CD36-mediated fatty acid uptake has emerged as a therapeutic strategy for cancer (Fig. 5). In 2017, Pascual et al. identified a role for CD36 in metastasis of head and neck squamous cell carcinoma (HNSCC), particularly in mice fed a high fat diet (HFD) [53]. Importantly, they showed that injecting anti-CD36 antibodies, which bind to the extracellular portion of CD36 and inhibit fatty acid uptake, impairs metastasis [53]. Since, studies have confirmed CD36 as a potential therapeutic target in many other cancers, including: ovarian, prostate, breast, lung, renal, gastric, GBM, and leukemia [47, 102–109] (reviewed [6]). Recently, it was shown that CD36 is required for breast cancer metastasis in mice fed an HFD, and that systemic deletion of CD36 after tumor onset, which emulates drug therapy, inhibits metastasis [47]. Furthermore, CD36 confers resistance to HER2-targeted therapy in breast cancer [109]. Currently, there is an ongoing effort to generate a humanized CD36-inhibitory antibody (ONA-046) for clinical use [110].

Figure 5: Targeting CD36 for cancer therapy.

A) CD36 promotes tumor progression cell-autonomously through numerous mechanisms. Importantly, these depend on fatty acid uptake and storage to promote lipid homeostasis and aggressive cancer phenotypes. In CD8⁺ T cells, which inhibit tumorigenesis by killing cancer cells, CD36 promotes uptake of oxidized LDL (oxLDL) and arachidonic acid (AA) leading to ferroptosis and T cell exhaustion. In T_reg cells, which inhibit CD8⁺ T cells to promote tumorigenesis, CD36 promotes fatty acid uptake for maintenance of mitochondrial health and survival. In metastasis associated macrophages (MAM), CD36 increases fatty acid uptake and fatty acid oxidation (FAO) to drive M2 polarization leading to immunosuppression and increased liver metastases. CD36 expression in endothelial cells (E) or adipocytes (Adipo) promotes fatty acid transfer to tumor cells to increase tumor growth.

B) Since fatty acid uptake activity of CD36 requires extracellular expression, antibodies can bind to CD36 to inhibit its function. Loss of CD36 inhibits tumor progression by reducing FA uptake and promoting lipotoxicity in cancer cells. CD36-inhibition enables CD8⁺ T cells by directly preventing their exhaustion and ferroptotic cell death. It also promotes apoptosis in T_reg cells and inhibits M2 polarization in MAMs to decrease immunosuppression, further invigorating CD8⁺ T cells. Finally, loss of CD36 in endothelial cells and adipocytes decreases fatty acid transfer to cancer cells and inhibits tumorigenesis. Note: some studies utilized only genetic deletion models without inhibitory antibodies; however, CD36-targeting antibodies phenocopy CD36-deficiency in studies that use them.

The limitations and opportunities for anti-CD36 therapy are illustrated by insights into its molecular function. A recent study found that CD36 acts as an intracellular sensor for lipid saturation to prevent lipotoxic stress [47]. Mechanistically, CD36 senses intracellular saturation directly via palmitoylation, resulting in CD36-membrane localization and enhanced MUFA uptake [47]. Thus, during SFA-induced lipotoxic stress, CD36 is activated to promote MUFA uptake and restore lipid homeostasis. Indeed, CD36 is required in settings of lipotoxic stress in vitro, including adipocyte co-culture [111, 112], palmitate treatment [47, 53, 54, 113], hypoxia (which inhibits SCD1) [106], and SCD1 inhibition [67]. In vivo, an HFD drives metastasis of HNSCC and breast cancer through the action of SFAs on cancer cells [47, 52–54]. This suggests that cancers cells must withstand SFA-induced stress to enable their prometastatic effect. Consistently, when comparing the growth of primary versus metastatic tumors in mice fed either a control diet or an HFD, loss of CD36 most significantly reduces metastasis in HFD-fed mice [47, 53]. Furthermore, loss of CD36 in tumors from HFD-fed mice display increased ER stress, consistent with SFA-induced toxicity [47]. This indicates the efficacy of anti-CD36 treatment may favor metastatic cancers, TMEs with high levels of SFAs, and patients with obesity. It also suggests inducing lipotoxic stress through dietary manipulation or SCD1 inhibition can enhance anti-CD36 cancer therapy.

CD36 in non-tumor cells

Recent works indicate a tumor promoting role for CD36 in neighboring cells within the TME (Fig. 5). Three recent papers highlight that CD36 limits intratumoral CD8⁺ T cell function to reduce antitumor immunity [114–116]. Indeed, CD36 is a marker for exhausted CD8⁺ T cells [114–116]. CD36-mediated uptake of oxidized LDL or AA promotes lipid peroxidation and ferroptotic cell death in CD8⁺ T cells [114–116]. However, disruption of CD36 in CD8⁺ T cell invigorates their function and enhances antitumor immunity [114–116]. Conversely, CD36 protects intratumor T_reg cells from apoptosis, which similarly reduces antitumor immunity [117]. CD36 deletion had no effect on normal T_reg cells, which is important given the cell type’s role in preventing autoimmunity [117]. However, loss of CD36 reduced the viability of intratumoral T_reg cells, leading to activation of CD8⁺ T cells and antitumor immunity [117]. Notably, the discrepant role of CD36 between T_reg cells and CD8⁺ T cells with regard to cell survival is surprising. However, this discrepancy is clinically favorable as CD36 inhibition promotes antitumor immunity in both cell types. Importantly, the addition of PD-1 inhibitory antibodies to T_reg-specific CD36-deletion, CD36-inhibitory antibodies, or CD36-germline deletion synergized to reduce tumor growth through antitumor immunity [114, 117]. This indicates a great opportunity for anti-CD36 therapy in combination of ICIs. Furthermore, CD36 is required for metastasis-associated macrophages to promote liver metastases [118]. Finally, adipocyte- and endothelial cell-specific deletion of CD36 reduces lipid transfer to cancer cells and inhibits tumor growth in vivo [119]. Thus, there is a seemingly unanimous tumor-promoting function for CD36 within the TME, bolstering its status as an emerging anticancer target.

Concluding Remarks and Future Perspectives

Preclinical data suggests that targeting lipid metabolism can be effective for cancer therapy. FASN-inhibitors have long been sought after and have finally reached clinical trials. These studies can validate preclinical hypotheses, including the efficacy of FASN inhibition in CNS metastases and oncogenic KRAS LUAC (see outstanding questions). SCD1 has also emerged as a therapeutic target to overcome cell death resistance in multiple cancer types. We wonder how lipid inhibitors will be trialed clinically (e.g. CD36-inhibition alone or in combination with ICI or chemotherapy) and what combination of lipid inhibitors may prove beneficial (e.g. SCD1 and CD36 inhibition together). The role of diet is another important consideration, will prescribed diets prove beneficial therapeutically, and will patients tolerate them? Although exciting, there is still significant work to be done in both preclinical and clinical testing for lipid inhibitors to fully realize their potential.

Outstanding questions:

1. How effective will inhibitors of fatty acid synthesis be in the treatment of cancer, particularly in lipid-poor environments like the CNS?

2. How will ketogenic diets be incorporated into cancer therapy? Will engineering of the lipid composition of ketogenic diets be able to produce synergistic cytotoxicity when combined with chemotherapy or metabolic inhibitors?

3. How will ferroptosis modulators be applied for cancer therapy? Will SCD1-inhibitors or PUFA-rich diets increase their efficacy?

4. Will CD36-inhibitors show the same clinical efficacy that pre-clinical studies have demonstrated? Will targeting CD36 increase the efficacy of ICIs?

5. Given the metabolic plasticity of cancer cells, how can combination therapies targeting multiple lipid metabolism-related proteins, with or without dietary intervention, be used for cancer therapy?

Highlights.

1. There are no FDA-approved lipid inhibitors for cancer treatment despite promising preclinical data to suggest their utility. Only FASN inhibitors have reached clinical-stage testing.

2. New results suggest CNS resident tumors are primed for FASN and SCD1 inhibitors due the paucity of microenvironmental lipids.

3. SCD1 is an emerging therapeutic target due to the central role that monounsaturated fatty acids play in preventing cell death.

4. CD36 inhibitory antibodies are poised for a clinical breakthrough based on its seemingly unanimous pro-tumorigenic function in the TME, and CD36-inhibition may enhance the activity of immune checkpoint inhibitors.

5. Alterations in dietary lipids through prescribed diets have shown exciting activity in preclinical testing, although more optimization of diet composition and molecular targets is required.

Glossary

AMP-activated protein kinase (AMPK)

a stress kinase activated by increasing levels of AMP and ADP which can directly phosphorylate and inhibit ACC1/2 among numerous other targets

Calorie restricted diet (CR)

a diet with normal amounts of fat, protein, vitamins, minerals, but calories from carbohydrates reduced by 40%

Cluster of differentiation 36 (CD36)

aka fatty acid translocase (FAT), is a transmembrane protein that facilitates the uptake of extracellular lipids, including long chain fatty acids and oxidized LDL

Ferroptosis

iron-dependent form of non-apoptotic cell death, in which reactive oxygen species cause peroxidation of polyunsaturated fatty acids, leading to cell death. For example, loss of GPX4 or inhibition of glutathione synthesis leads to ferroptosis

Glutathione peroxidase 4

a phospholipid hydroperoxidase which inhibits ferroptosis by converting lipid hydroperoxides into lipid alcohols and requires glutathione (GSH) as a cofactor

Herceptin

aka trastuzumab, a HER2 inhibitory monoclonal antibody used for treatment of HER2+ tumors (especially breast cancer)

Human epidermal growth factor receptor 2 (HER2)

an oncogenic growth factor receptor commonly amplified in breast cancer

High fat diet (HFD)

a diet in which 60% of calories come from fat which is often used to mimic “western diets” and for diet-induced obesity in pre-clinical models

Immune check point inhibitors (ICI)

antibodies which target programmed cell death protein 1 (PD-1), programmed death-ligand 1 (PD-L1), cytotoxic T-lymphocyte associated protein 4 (CTLA-4), which are used clinically to activate the immune system to target cancer

Inositol-requiring enzyme type 1 (IRE1) α

an ER stress kinase activated by unfolded proteins and membrane lipid saturation with RNAse activity that splices XBP1-mRNA to its active form (XBP1s)

Kelch-like ECH-associated protein 1 (KEAP1)

a tumor suppressor protein that negatively regulates NRF2, a master regulator of oxidative stress

Ketogenic diet (KD)

a high-fat diet with low carbohydrate levels such that the body burns fats and induces a state of ketosis

Mammalian target of rapamycin complex 1 (mTORC1)

a protein complex that is a nutrient and energy sensor, regulates protein synthesis, and can be activated by upstream oncogenic pathways such as PI3K/AKT

MYC

a proto-oncogene transcription factor that is commonly activated in human cancers

Oncogenic KRAS

an oncogenic mutant version of RAS gene which is commonly activated in human cancers

Palmitoylation

a posttranslational protein modification made by palmitoyl-CoA

Phosphatase and tensin homolog (PTEN)

a tumor suppressor phosphatase that negatively regulates PI3K/AKT signaling

Phosphoinositide 3 kinase (PI3K)-protein kinase B (AKT)

an oncogenic signaling pathway that is commonly activated in human cancer. PI3K is negatively regulated by PTEN and activates AKT by phosphorylation. AKT is a serine/threonine kinase that activates numerous targets, including mTOR, to promote tumorigenesis

Serine/threonine kinase 11 (STK11, aka LKB1)

a tumor suppressor protein commonly mutated in lung cancer

Stearoyl-CoA Desaturase 1 (SCD1)

an oxygen-dependent mono-desaturase that converts palmitate to palmitoleate and stearate to oleate

Sterol regulatory element binding protein 1 (SREBP1)

a transcription factor and master regulator of fatty acid synthesis related genes

Treg cells

a type of CD4+ T cell that are generally immunosuppressive and inhibit antitumor immunity