Impact of Immunometabolism on Cancer Metastasis: A Focus on T Cells and Macrophages

Abstract

Despite improved treatment options, cancer remains the leading cause of morbidity and mortality worldwide, with 90% of this mortality correlated to the development of metastasis. Since metastasis has such an impact on treatment success, disease outcome, and global health, it is important to understand the different steps and factors playing key roles in this process, how these factors relate to immune cell function and how we can target metabolic processes at different steps of metastasis in order to improve cancer treatment and patient prognosis. Recent insights in immunometabolism direct to promising therapeutic targets for cancer treatment, however, the specific contribution of metabolism on antitumor immunity in different metastatic niches warrant further investigation. Here, we provide an overview of what is so far known in the field of immunometabolism at different steps of the metastatic cascade, and what may represent the next steps forward. Focusing on metabolic checkpoints in order to translate these findings from in vitro and mouse studies to the clinic has the potential to revolutionize cancer immunotherapy and greatly improve patient prognosis.

Significant progress has been done in using immunotherapy to boost the antitumor responses of the immune system (Rosenberg and Restifo 2015; Yang 2015). While these therapies have demonstrated success in a small subsets of tumor types (Ascierto et al. 2016; Herbst et al. 2016), they are still largely ineffective in most solid tumors (Klebanoff et al. 2016). It is believed that cancer cells are able to generate an immunosuppressive environment which has a significant effect on immune function (Whiteside 2006; Joyce and Fearon 2015; Scharping et al. 2016). Cancer cells are well known for their distinct capability to alter their metabolic profile in order to support their high demands for energy to fuel their rapid proliferation (Renner et al. 2017). In doing so they deplete nutrients available in the tumor microenvironment (TME) and secrete metabolites which contribute to weak antitumor immune response as well as recruitment and support of the immunosuppressive members of the immune system. In addition, this metabolic alteration of cancer cells is believed to be the major reason cancer cells are able to develop resistance to therapy in general (Barsoum et al. 2014; Allard et al. 2016; Kawalekar et al. 2016). Understanding how cancer cell metabolism can influence metabolic processes and differentiation of immune cells could lead the way to exploiting these metabolic changes to improve antitumor immunity using immunotherapy.

Because metastasis is the leading cause of cancer-related mortality and the strongest predictor of poor outcome in patients (Chaffer and Weinberg 2011; Ferlay et al. 2015), it is important to understand how we can target metabolic checkpoints at different steps of metastasis. Reports have shown that different metabolic pathways are reprogrammed in cancer cells at different steps along the metastatic cascade in order to allow for cancer cell survival and proliferation (Massagué and Obenauf 2016). Metastasis starts from the primary tumor. From there cancer cells invade the surrounding tissue and enter the circulation. If able to survive in circulation, where they can be most easily recognized by immune cells, they arrive at the premetastatic niche (PMN) which is often primed by factors secreted from the primary tumor in order to ensure their survival. Cancer cells in a distal organ can now enter a state of dormancy which could last several years or lead to colonization of the metastatic niche, resulting in micrometastases and eventually outgrowth of the secondary tumor, thereby completing the metastatic cascade. In order for immune cells to control and eliminate cancer cells, they must adapt to the ever-changing environment surrounding cancer cells throughout their transition through the metastatic cascade.

Cancer cells are exposed to immune cells at each step of the metastatic cascade providing an opportunity for the immune system to recognize immunogenic cancer cells and limit their proliferation. One major cell type of the immune system able to kill cancer cells are antigen-specific CD8⁺ T cells (Fridman et al. 2012). Upon recognition of cancer cells, antigen-specific CD8⁺ T cells are activated, followed by rapid proliferation and secretion of cytolytic granules, containing perforin and granzyme B, leading to the death of the target cell. Cancer cells, however, are known to utilize mechanisms to suppress CD8⁺ T cell function and recruiting immunosuppressive regulatory T cells (Treg) to the tumor in order to protect them from CD8⁺ T cells and other killer and effector cells of the immune system (such as NK cells) (Kitamura et al. 2015). Additionally, the TME can support immunosuppressive differentiation of T cells at each step of the metastatic cascade.

While some experts suggest that T cells are the main players in antitumor immune responses (Joyce and Fearon 2015; Tietze et al. 2017), others suggest that macrophages play a key role (Leek et al. 1996; Biswas and Mantovani 2010; Komohara et al. 2016). Macrophages heavily contribute to all the steps of metastasis and are often the major immune cell type within the TME (Nielsen and Schmid 2017). Macrophages are also present in the PMN, where they prepare the site for the arrival of newly disseminated cancer cells and promote their survival (Nielsen and Schmid 2017). Once cancer cells reach the site of metastasis, macrophages are there to promote the metastatic niche environment which promote cancer cell survival and proliferation (Nielsen and Schmid 2017). Macrophages are also believed to be the key players in the development of resistance to traditional chemotherapy (Nielsen and Schmid 2017). Whether one cell is more important of the other in abating metastasis will all depend on the step of the metastatic cascade, the tumor type, the (pre) metastatic niche, and as well, the therapeutic regimen the patient will undergo from tumor diagnosis.

Here, we provide an overview of the emerging field of cancer immunometabolism in order to pinpoint metabolic check points of immune cells which can be targeted to improve cancer-specific immune responses and develop new therapies for this deadly disease as such or in combination with standard immunotherapies. We will focus on the metabolism of T cells (Fig. 1) and macrophages (Fig. 2) throughout the metastatic cascade, and outline the contrast between immunosuppressive and antitumor effector subtypes in each of these two immune cell lineages.

Figure 1.

T cell metabolic pathways in the different steps of the metastatic cascade. (A) T cell metabolism in the TME. Elevated arginase levels in the TME lead to the inhibition of CD8⁺ T cell response. Tryptophan depletion, in part due to excess IDO production, in the TME also leads to compromised CD8⁺ T cell effector functions. Furthermore, tryptophan metabolism leads to elevated levels of kynurenine further inhibiting effector CD8⁺ T cells. Inhibition of the ACAT1 enzyme, affects cholesterol esterification and results in enhanced CD8⁺ T cell effector function. HIF-1α is known to reduce Treg differentiation through proteosomal degradation of Foxp3. HIF-1α is also associated with mTORC1 activation, increased glycolysis, and Treg decline. TSC1 is a negative regulator of mTORC1 and therefore important for Treg differentiation. Activation of AMPK is crucial for the shift in metabolism from glycolysis to fatty acid oxidation (FAO), therefore, favoring Treg differentiation. Glycolysis leads to excess lactate production, which is toxic for T_E cells and favors Treg cell polarization. AMPK and PD-1 activation lead to CPT1A activation which is a key enzyme in FAO (B) T cell metabolism and the premetastatic niche. Priming the PMN with 27HC has been shown to indirectly lead to reduced numbers of CD8⁺ T cells at metastatic sites. Inhibition of CYP27A1 enzyme targeting 27HC synthesis has been linked to reduced metastasis. Stat3 expression on myeloid cells in the PMN leads to inhibition of CD8⁺ T cells. Expression of STAT3 on myeloid cells in the lymph node PMN results in the inhibition of CD8⁺ T cells. (C) Immunometabolism and metastasis initiation. To the best of our knowledge, there are no metabolism specific factors known. (D) Immunometabolism and cancer cell in circulation. High levels of the arginase enzyme secreted by circulating cancer cells has been shown to inhibit CD8⁺ T cells. (E) Metastatic niche. Prolyl hydroxylase domain (PHD) proteins expressed by T cells suppresses the activity of T_E cells and promotes Treg cells thereby promoting metastasis in the lung. Deletion of Spns2 in mice leads to reduced metastasis to the lung and increased T_E cell numbers.

Figure 2.

Tumor-associated macrophage (TAM) metabolic pathways in the different steps of the metastatic cascade. (A) TAM metabolism in the TME. TAMs polarization occurs as a plethora of states with both anti- and protumoral features, and with different metabolic phenotypes. TAMs are known to support cancer in the early stages by displaying a glycolysis-related inflammatory function, followed by a switch to a more oxidative phosphorylation (OXPHOS) related function in the later stages of tumor progression. The switch to glycolysis in TAMs is under the control of the Akt–mTOR–HIF-1–PKM2 axis. The sustained glycolysis enhances lactic acid accumulation in the TME, which leads to the acquisition of a protumoral function of TAMs. Furthermore, TAMs rely on uptake and oxidation of extracellular lipids, but also express high levels of fatty acid synthase (FASN) and PPARγ to favor tumor progression and metastasis. Immunosuppressive and protumoral TME is also sustained by PGE2 and iron release, tryptophan and arginine metabolism. Glutamine metabolism is another crucial pathway, indeed, the activity of glutamine synthetase (GS) in TAMs promotes the immune escape, angiogenesis, and metastatic dissemination. (B) TAM metabolism and the premetastatic niche. To the best of our knowledge, there are no information about metabolic remodeling in TAMs. (C) Immunometabolism and metastasis initiation. In a tumor context, in which glutamine tissue levels are not homogeneous, due to the presence of cells with different glutaminolytic capacity, TAMs sense glutamine deprivation and induce GS, which metabolically reprograms TAMs toward a proangiogenic and prometastatic phenotype. In a hypoxic TME, macrophagic arginine metabolism leads to nitric oxide (NO) production, which promotes tumor blood vessel normalization and endothelial cell (EC) activation. Furthermore, REDD1 up-regulation in hypoxic TAMs, which inhibits mTOR, lowers TAM glucose consumption and thus favors ECs glycolytic hyperactivation and dysfunctional blood vessel formation, which ultimately leads to metastasis initiation. (D) Immunometabolism and cancer cell in circulation. Knowledge about the metabolic crosstalk between CTC and macrophages is scarce. (E) Metastatic niche. NADPH oxidase (NOX) 1 and 2 deficiency impair the macrophagic protumoral function during tumor development, leading to decreased metastasis. Furthermore, the macrophages at the metastatic niche might influence the availability of specific nutrients: this is the case of REDD1 up-regulation in hypoxic conditions and GS up-regulation in response to extracellular glutamine deprivation.

INTRINSIC CELLULAR METABOLISM OF IMMUNE CELLS

T Cells

Once T cells complete their development and differentiation in the thymus they enter the circulation as nave CD4⁺ or CD8⁺ T cells. These cells are small in size and have lower metabolic activity than activated T cells. Energy ensuring survival and homeostatic proliferation of naïve T cells is provided mainly by OXPHOS and FAO (Table 1; O'Neill et al. 2016).

Table 1.

Intrinsic metabolism of T cells and macrophages

	Cell subset	Preferred metabolism	Main factors
General T cells	Naïve T cells	OXPHOS, FAO	AMPK
	Activated T cells	↑glucose and glutamine uptake, ↑PPP and OXPHOS, FAS, aerobic glycolysis.	mTORC1
CD8+ T cell subset	TE	Glycolysis and glutaminolysis	↑GLUT1, mTORC1 → HIF1α, mTORC/SREBP pathway
	TM	OXPHOS, FAO	AMPK, ↑LAL
Treg cells	tTreg	Glycolysis (metabolically similar to TE)	mTORC1
	pTreg and iTreg	FAO and OXPHOS (metabolically similar to TM)	AMPK
Macrophages	Classically activated (“M1”)	Glycolysis, interruption of the TCA cycle	↑PFKFB3, IL-1β, and HIF1α
	Alternatively activated (“M2”)	↓Glycolysis, ↑FAO and OXPHOS	↑PPARγ

Upon encounter with their cognate antigen peptide, presented in the context of MHC class I or MHC class II molecules on the surface of target cells, CD8⁺ and CD4⁺ T cells (respectively) undergo activation followed by extensive proliferation leading to differentiation into effector T cells. This change in the cell activation status is accompanied by a shift in metabolic processes. A mass spectroscopy analysis by Wang et al. (2011) showed that activated T cells accumulate higher levels of metabolites involved in anabolic processes after anti-CD3/28 activation compared to resting T cells. In addition to ATP, activated T cells also require metabolic intermediates necessary to provide building blocks for new fast expanding cells. Increase in glucose and glutamine uptake, aerobic glycolysis (Warburg effect), pentose phosphate pathway (PPP), OXPHOS, and fatty acid synthesis (FAS) are all characteristics of activated T cell metabolism (Maciolek et al. 2014). As more building blocks are required in these highly proliferating cells, FAO is down-regulated while FAS increased (Chapman et al. 2017). A major player in the coordination of metabolism in activated T cells is mTORC1 compared to AMPK in resting T cells (Pearce and Pearce 2013). These general metabolic routes are in common to all different subsets of CD8⁺ and CD4⁺ T cells, however, the different subsets of T cells can exhibit unique metabolic profiles which are associated with their distinct functions. In the following paragraphs, we will focus on the specific metabolic pathways of cytotoxic CD8⁺ T cells and CD4⁺ regulatory T cells (Treg) which have contrasting roles in the immune system's response to cancer. Intrinsic metabolism of the different subsets of T cells and macrophages is outlined in Table 1.

Cytotoxic CD8⁺ T Cells

CD8⁺ T cells are essential for the immune clearance of infections and cancer. Following activation as a result of antigen encounter and costimulatory signals, CD8⁺ T cells change their metabolism from OXPHOS to glycolysis allowing them to rapidly proliferate. Activated CD8⁺ T cells can then become short-lived effector T cells (T_E) or memory precursor effector T cells leading to long-lived memory T cells (T_M). These two CD8⁺ T cell subsets differ in their preferred metabolic pathways. T_E cells rely mostly on glycolysis and glutaminolysis while T_M cells favor FAO and OXPHOS (as is true for resting T cells) (van der Windt et al. 2012; Sukumar et al. 2013). T_E up-regulate the expression of the glucose transporter (GLUT1) which results in a higher influx of glucose into the cell and increased glycolysis ultimately leading to inflammatory cytokine production and clonal expansion (Schurich et al. 2016). Glucose uptake is further reinforced by TCR signaling which leads to activation of mTORC1 and the expression of HIF-1α (Finlay et al. 2012). In addition to glucose, glutamine is also crucial for T_E cell function and survival as metabolic intermediates can be generated by the mitochondria via glutamine metabolism (Carr et al. 2010). FAS is induced upon activation in T_E through the mTOR/SREBP pathway in order to meet the high demand for lipids required for rapid proliferation and to support inflammatory effector functions (Kidani et al. 2013; Lee et al. 2014).

In contrast to T_E, T_M cells favor FAO and OXPHOS for their source of energy to maintain survival. Compared to T_E cells, T_M cells have a more pronounced mitochondrial content and therefore an enhanced capacity to rapidly produce ATP via OXPHOS upon their restimulation (Chapman et al. 2017). T_M cells also express a higher level of lysosomal acid lipase (LAL) and are, therefore, better equipped to catabolize lysosomal lipids, explaining the enhanced intrinsic preference of T_M cells for FAO compared to T_E cells (O'Sullivan et al. 2015). While mTORC1 stimulates T_E differentiation and suppresses T_M cell generation, the opposite is true for AMPK (Araki et al. 2009; Pearce et al. 2009; Rao et al. 2010). The generation of these two distinct subsets can be explained by asymmetric cell division upon antigen stimulation. The daughter cell proximal to the APC is more likely to differentiate into T_E cell and the APC distal daughter cell is more likely to differentiate into T_M cell. This is due to the asymmetrical distribution of the metabolic components in the process of cell division (Chang et al. 2007). The daughter cell closest to the APC tends to inherit higher levels of mTORC1 and c-Myc as well as some amino acid transporters (Pollizzi et al. 2016; Verbist et al. 2016). On the other side, the daughter cell distal to the APC generally displays lower mTORC1 activation, higher mitochondrial content and increased AMPK activity resulting in higher FAO (Blagih et al. 2015; Pollizzi et al. 2016). In addition, the distal daughter cell has also been shown to up-regulate survival factors which determine the long-term survival phenotype of T_M cells (Pollizzi et al. 2016).

Targeting CD8⁺ T cell metabolism has been proven to be a very promising strategy to improve their fitness in a tumor context but also in other pathologies. Recently, a report by Wu et al. (2019). demonstrated the importance of serine-palmitoyl-transferase subunit SPTLC2 in CD8⁺ T cell function against viral infection in Hereditary sensory and autonomic neuropathy (HSAN) patients.

Regulatory T Cells (Treg)

In addition to conventional CD4⁺ T cells, those expressing αβ TCR and differentiate into effector and memory phenotypes, there are several other distinct CD4⁺ T cell subsets that arise during the development of the immune system. One of these “nonconventional” T cells are regulatory T cells (Treg). The role of Treg cells in the immune system is to suppress the immune response to self-antigen. However, Treg cells also have deleterious consequences when recruited to tumor sites where they exert immunosuppressive pressure on antitumor immune cells resulting in a diminished antitumor immune response and poor prognosis. In vivo most Treg cells arise in the thymus (tTreg). A small subset of Treg cells can also be induced in the periphery from naïve CD4⁺ T cells (pTreg). These pTreg cells are induced locally, especially at mucosal sites, to induce immune tolerance in tissues. Treg cells can also be induced in vitro (iTreg) when they somewhat mimic the phenotype of in vivo derived pTreg cells. All three Treg types rely on the expression of Foxp3 and depend on IL-2 for maintenance. The main difference between the different Treg cell subtypes is that in tTreg cell Foxp3 is expressed constitutively while expression of Foxp3 in pTreg and iTreg cells is induced after antigen mediated activation (Shevach and Thornton 2014). Considering this significant difference, it is not surprising that these Treg cell subtypes also differ in their metabolic pathway dependence. tTreg cells tend to utilize similar metabolic pathways as T_E cells while pTreg and iTreg cells are more metabolically similar to T_M cells (Newton et al. 2016). In vivo developed tTreg cells, therefore, depend on glycolysis for their proliferation and function and pTreg (and iTreg) cells utilize exogenous lipids for FAO and glucose-derived pyruvate for OXPHOS. Aside for their different metabolic profile, there is no known marker to distinguish the two different in vivo derived Treg subtypes (tTreg and pTreg) (Shevach and Thornton 2014; Szurek et al. 2015). It is therefore difficult to determine which subtype is more significantly involved in tumor mediated immunosuppression. One can speculate that since pTreg cells can be induced locally at the tumor site, this subtype has a higher impact in the tumor context.

A recent study by Acharya et al. (2019) reports on the importance of mevalonate in the differentiation of Treg cells. In brief, mevalonate treatment increases Treg cell differentiation in vitro and enhanced their suppressive function compared to untreated Treg cells. Additionally, mevalonate was shown to boost pTreg cells in vivo.

Macrophages

Macrophages are ubiquitous phagocytic cells of the innate immune system. These cells can be polarized in different ways, express specific surface markers and acquire particular functional states, depending on the stimulating factors such as cytokines and other signals. The in vitro phenotypes of activated macrophages are conventionally divided in the “M1” and “M2” categories (Mantovani et al. 2002; Condeelis and Pollard 2006), in line with the traditional concept of binary polarization. Classically activated macrophages, also known as M1 macrophages, are polarized by inflammatory signals, such as interferon gamma (IFNγ) and lipopolysaccharides (LPS), and exhibit bactericidal, immunostimulatory, and antitumoral activities. In contrast, alternatively activated macrophages or M2 macrophages are polarized by anti-inflammatory signals, such as IL10, IL4, and IL13, and are generally involved in the resolution of inflammation, involved in immunosuppression, tumor invasion, tumor growth, angiogenesis, and metastasis (Lewis and Pollard 2006; Qian and Pollard 2010).

Proinflammatory Macrophages

The different polarization states are associated with peculiar metabolic programs, at least in vitro. The main feature of proinflammatory, classically activated macrophages resides in the production of ATP, which derives more from glycolysis than from mitochondrial metabolism. This peculiarity represents an advantage for macrophages activated in hypoxic regions of tumors (Nizet and Johnson 2009). Proinflammatory macrophages increase glycolysis by inducing the expression of the PFKFB3 isoform (Rodríguez-Prados et al. 2010). Beside glycolysis, their metabolism is characterized by interruption of the TCA cycle. TCA cycle intermediates escape mitochondria and accumulate into the cytosol. Among these, succinate up-regulates IL-1β by promoting HIF1α stabilization (Tannahill et al. 2013). High cytosolic succinate levels favor posttranslational lysine succinylation on proteins, leading to significant changes protein function (Xie et al. 2012). Indeed succinylation of pyruvate kinase M2 (PKM2) promotes its translocation into the nucleus, where it interacts with HIF1α to boost IL-1β transcription (Wang et al. 2017).

Succinate displays signaling roles at the extracellular level, since it can accumulate in the extracellular milieu (Rubic et al. 2008) where it binds to the succinate receptor SUCNR1/GPR91, a G-protein-coupled macrophage surface sensor for extracellular succinate (He et al. 2004). In response to inflammatory signals like LPS, macrophages activate a GPR91-mediated signal transduction that fosters inflammation through IL-1β production (Littlewood-Evans et al. 2016) and extracellular succinate can contribute to this in an autocrine manner. Succinate scavenging by other cells is then protective against neuroinflammation in an in vivo model of experimental autoimmune encephalomyelitis (EAE) (Peruzzotti-Jametti et al. 2018).

Citrate efflux from mitochondria due to up-regulation of the mitochondrial carrier SLC25A1 (Infantino et al. 2014; Palmieri et al. 2015) contributes, together with the PPP, to cytosolic NADPH generation, the latter being fundamental to sustain iNOS for NO and lipid synthesis (Palmieri et al. 2015). Finally, LPS increases the expression of the immunoresponsive gene 1 (IRG1), which catalyzes the production of itaconic acid, a known inhibitor of the glyoxylated pathway in bacteria (Michelucci et al. 2013). LPS-activated macrophages rely then on itaconic acid to support antimicrobial activity. In a negative feedback loop, itaconic acid is also known to inhibit inflammation and inflammasome activation by inducing Nrf2, a transcription factor involved in the anti-inflammatory and antioxidant functions (Mills et al. 2018). Radical species production is another metabolic feature of classically activated macrophages, namely, ROS and NO, important for bactericidal activity. NADPH oxidase is the main source of ROS, which relies on NADPH production by the PPP and the malic enzyme (Infantino et al. 2014; Palmieri et al. 2015). NO is produced by the inducible nitric oxide synthase (iNOS). A significant portion of ROS is produced by mitochondria, that respond to IRG1 expression by increasing FAO and OXPHOS (Hall et al. 2013).

Alternatively Activated Macrophages

Alternatively activated macrophages display also specific metabolic features, but can be distinguished in vitro depending on the cytokines used for differentiation. In general, glycolysis is repressed whereas FAO is enhanced (Vats et al. 2006) and this associates to mitochondrial biogenesis (O'Neill and Hardie 2013) and increased OXPHOS (Vats et al. 2006), which perfectly suits the energy and biosynthetic demands of alternatively activated macrophages. In contrast with classically activated macrophages, the redox state of alternatively activated macrophages is not directed toward ROS, NO, and NADPH synthesis. Indeed, PPP is inhibited (Nagy and Haschemi 2015) and the levels of GSH are reduced. Regarding glutamine metabolism, IL-4 activated macrophages enhance glutamine flux into oxoglutarate, which is important for the engagement of FAO and the epigenetic reprogramming of M2 genes (Liu et al. 2017). In IL-10 macrophages, glutamine synthesis is enhanced through up-regulation of GS (Palmieri et al. 2017), which is necessary for nucleotide and uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) synthesis for support of protein folding and trafficking (Wellen and Thompson 2012; Jha et al. 2015) especially the lectin/mannose receptors, which, in their highly glycosylated form, are among the most typical alternative polarization markers (Sica and Mantovani 2012). Glutamine synthetase expression senses also glutamine depletion, and its expression polarizes these cells toward a M2-like status (Palmieri et al. 2017).

IMMUNOMETABOLISM IN THE TUMOR MICROENVIRONMENT

The tumor microenvironment is composed of tumor cells as well as infiltrating immune cells, endothelial cells (EC), fibroblasts, secreted factors, and cytokines as well as extracellular matrix (ECM) proteins in and surrounding the primary tumor. A characteristic of malignant tumor cells is an altered metabolism, specifically increased glucose consumption and lactate production. As cancer cells proliferate they generate an increasingly immunosuppressive environment thereby restricting T cell infiltration and effector function. This immunosuppressive environment is further exacerbated by the recruitment of immunosuppressive cells, such as Treg cells and certain macrophage subtypes. Immune cells adapt to this environment and ultimately promote tumor progression. Evidence suggests that nutrient competition within the TME is a major driver for tumor progression as the increased consumption of glucose by cancer cells directly limits nutrient availability to T cells (Chang et al. 2015; Ho et al. 2015).

T Cells

CD8⁺ T Cells

Effector CD8⁺ T cells are critical for antitumor immunity. However, their function is suppressed by the TME. Therefore, boosting CD8⁺ T cell function within the TME represents a major interest in the cancer immunotherapy. Increased concentrations of extracellular lactic acid secreted by cancer cells can suppress CD8⁺ T cell effector function and can block the secretion of lactic acid by the immune cells leading to cell death due to elevated intracellular lactic acid levels (Fischer et al. 2007). CD8⁺ T effector cells rely heavily on glucose as the source of energy, similarly to cancer cells. However, cancer cells are able to uptake glucose with superior efficiency compared to T cells. Consequently, this lack of glucose available to T cells is responsible for suboptimal IFNγ production and cytolytic function of CD8⁺ T cells (Chang et al. 2015). Checkpoint inhibitor therapy with antibodies against CTLA-4, PD-1, and PD-L1, can restore glucose availability in the TME and therefore restore T cell effector function (Chang et al. 2015). However, this therapy has only been shown effective in a small subset of cancer types and patients.

Amino acids, specifically arginine, tryptophan, glutamine, and cysteine, are also important for CD8⁺ T cell function. Therefore, the depletion of these amino acids from the TME can have devastating effects on the function of tumor infiltrating lymphocytes (TILs). Rodriguez et al. (2004) demonstrated that the depletion of arginine in the TME due to excess arginase I production can inhibit antigen-specific T cell response in the TME (Rodriguez et al. 2004).

Tryptophan depletion in the TME can also lead to compromised CD8⁺ T cell function and subsequently tumor progression (Uyttenhove et al. 2003). The enzyme indoleamine-2,3-dioxygenase (IDO) is the rate-limiting enzyme converting tryptophan into kynurenin. Excess production of IDO creates a shortage of environmental tryptophan, an amino acid required for full T cell activation (Carbotti et al. 2015). Metabolism of tryptophan in the TME leads also to the production of kynurenine which further inhibits antitumor effector T cells and induces the transcription of PD-1 (Liu et al. 2010, 2018). Considering these findings, pharmacological targeting of IDO should lead to improved CD8⁺ T cell function in the TME. There are currently several ongoing clinical trials assessing IDO inhibitors to determine its effect on improved antitumoral immune responses (examples in Table 2).

Table 2.

Examples of T cell and macrophage metabolic targets in metastasis

Target	Role in metastasis	Tumor type	Pharmacologica l inhibitor	Preclinical evidence	Clinical evidence	Therapeutic setting
Atg5	Deletion of Atg5 increases CD8+ T cell fitness in the TME through an autophagy dependent regulation of CD8+ T cell metabolism	Studies done in breast cancer mouse model	None known	(DeVorkin et al. 2019)	None	Undetermined
PHD	Inhibition of PHD limits tumor metastasis in the lungs	Lung metastasis	DMOG (pan-PHD inhibitor)	(Clever et al. 2016)	None	With ATC
Spns2	Reduced metastasis when inhibited	Effect of this gene in metastasis confirmed in breast, colon and lung cancer	Unknown	(Van der Weyden et al. 2017)	None	Undetermined
27HC synthesis	Reduced lung metastasis after inhibition of 27HC synthesis	Breast cancer	Inhibition of CYP27A1 enzyme	(Baek et al. 2017)	None	Undetermined
Arginase 1	Reduced cancer proliferation	MC38 colon adenocarcino ma mouse model	2 (S)-amino-6-boronohexanoic acid (ABH)	(Arlauckas et al. 2018)	None	No synergic effect with anti-PD1 therapy
	Increased arginase concentration in circulation inhibits T cell activity	Reduced plasma arginine levels are shown in AML, breast, pancreatic and colon patients	Arginase inhibitors: PDL-NMMA/L-NOHA	(Mussai et al. 2018)	None	Undetermined
Cycloxygen ase 2 (COX-2) and microsomal PGE2 synthase	Reduced immune suppression	Bladder cancer	Celecoxib (COX2 inhibitor) and CAY10526 (mPGE2 inhibitor)	(Prima et al. 2017)		Possible synergic effect with anti-PD-1 therapy
Glutamine synthetase (GS)	Reduced lung metastasis after GS inhibition	Lung carcinoma murine model	Genetic ablation in TAMs Principal known inhibitor: methionine sulfoximine (MSO)	(Palmieri et al. 2017)	None	Undetermined
LDH-A	Reduced immunosuppression	K-Ras murine model of lung carcinoma	Genetic ablation in TAMs	(Seth et al. 2017)	None	Potential synergic effect with anti-PD1 therapy
Toll-like receptor 9	Counteracted inhibitory activity of cancer cells on macrophages	PDAC murine model	CpG nucleotide (TLR9 agonist)	(Liu et al. 2017)	None	Undetermined
Indoleamine 2,3-dioxygenase (IDO)	IDO inhibition in macrophages reduces immune suppression	Metastatic solid tumors	IDO inhibitor Indoximod (Phase I)	(Vacchelli et al. 2014)	(Soliman et al. 2014) NCT02752074 (ends April 2019)	Phase I: combined with docetaxel Phase II: combined with immune checkpoint inhibitors
			Epacadostat (Phase II–III)
ACAT1	ACAT1 inhibition leads to enhanced CD8+ T cell function	Mouse melanoma model with lunch metastasis	ACAT inhibitor Avasimibe	(Yang et al. 2016)	None	Possible synergism with anti-PD-1therapy

Cholesterol within the plasma membrane of CD8⁺ T cells is essential for clustering of TCR and the formation of immunological synapses (Zech et al. 2009). Yang et al. (2016) showed that the modulation of cholesterol metabolism via inhibition of the ACAT1 enzyme, responsible for cholesterol esterification, leads to enhanced effector function of CD8⁺ T cells in mouse tumor models. This was further enhanced when combined with PD-1 inhibitions. Another recent report demonstrates the importance of cholesterol in facilitating CD8⁺ T cell exhaustion in the TME via an ER-stress-XBP1 dependent mechanism (Ma et al. 2019). Taken together, modulation of cholesterol metabolism could contribute to enhanced CD8⁺ T cell function and reduced exhaustion in cancer and therefore warrants further investigation.

Regulatory T Cells

Treg cells are present in many different tumor types where they play a major immunosuppressive role. Since Treg cells rely mainly on FAO, instead of glycolysis, for survival and function they are naturally conditioned for survival in the otherwise harsh TME. Additionally, the presence of IDO and subsequently high levels of kynurenine, which are deleterious for effector T cell function, favor Treg differentiation. Additionally, IDO expression in glioma leads to increased recruitment of Treg cells in the tumor (Wainwright et al. 2012; Vacchelli et al. 2014). This metabolic adaptation can be linked to the ability of the Treg transcription factor Foxp3 to suppress glycolysis and enhance OXPHOS (Angelin et al. 2017).

In contrast with T_E cells which need high glucose levels in order to proliferate, Treg cells can survive in low glucose environments with high lactate concentration as the TME and the addition of l-lactate in the medium of naïve CD4⁺ T cells under Treg polarization conditions even enhances Treg cell formation. It has been shown that Foxp3 suppresses Myc and glycolysis, enhances NAD:NADH ratio and OXPHOS. Foxo1 and PTEN also lead to down-regulation of glycolysis in Treg cells, offering a favorable advantage to Treg cells to survive in the acidic environments such as the TME. Lactate dehydrogenase (LDH) is the enzyme that catalyzes the conversion of pyruvate to lactate and back. With this respect, LDH inhibitors offer a promising target to rewire cancer cell metabolism, control extracellular pH and enhance the immune response by protecting T_E cells from the toxic effects of lactate (Xie et al. 2014; Angelin et al. 2017).

Hypoxia-induced factor 1 (HIF-1) transcription factor is thought of as a metabolic checkpoint for the differentiation of Treg and Th17 cells (Shi et al. 2011). The expression and function of HIF-1 promotes IL-17 expression and therefore Th17 cell differentiation (Shi et al. 2011). On the other hand, HIF-1α subunits have been shown to decrease Treg development by inducing proteasomal degradation of Foxp3 (Dang et al. 2012). Supporting this, Dang et al. (2012) showed that HIF-1α deficient mice had reduced Th17 cell differentiation and increased Treg cell numbers. HIF-1 is linked to activation of mTORC1 and increased glycolysis and subsequently Treg impairment (Shi et al. 2011). Tuberous sclerosis 1 (TSC1), a negative regulator of mTOR, is important for Treg differentiation from naïve CD4⁺ T cells as demonstrated by Park et al. (2013) who reported that the loss of TSC1 in T cells leads to a decreased Treg cell differentiation in favor of Th17 cell differentiation. Targeting TSC1 in the TME can, therefore, represent a therapeutic approach in order to impair the immunosuppressive environment and result in enhanced infiltration of cytotoxic CD8⁺ T cells.

Expansion of Treg cells in TME coincides with the activation of AMPK, which acts as a nutrient deprivation sensor and is responsible for the metabolic shift from glycolysis to FAO, therefore leading to inhibition of effector T cell activation and an increase of Treg cell as well as memory T cell differentiation. Additionally, AMPK and PD-1 activation leads to increased expression of carnitine palmitoyl-transferase 1A (CPT1A) which is a key enzyme of FAO (Patsoukis et al. 2015). AMPK activation also promotes OXPHOS and subsequently suppresses mTOR and HIF-1 signaling. This suppression of mTOR leads to a reduced amino acid uptake and FAS which further impairs effector T cell differentiation (Pearce and Pearce 2013). It is important to note that the inhibition of mTOR in T cells has also been linked to the promotion of memory T cell differentiation (Araki et al. 2009) and has been correlated to their protection from apoptosis (Eikawa et al. 2015). Additionally, in cancer cells, mTOR inhibition has been shown to lead to apoptosis of cancer cells (Saxton and Sabatini 2017). Considering all these reports it is clear that mTOR inhibitor treatment can be promising depending on the dominant immune cell population in a specific tumor type.

Macrophages

It is well established that tumorigenesis relies strongly on the cellular components of the TME. An important component of the TME are macrophages, also known as TAMs. Although cancer progression skews TAMs toward a polarization state resembling in some ways that of in vitro skewed M2 macrophages (Condeelis and Pollard 2006; Qian and Pollard 2010), it is now evident that TAMs polarization occurs as a plethora of states (Aras and Zaidi 2017) with both anti- and protumoral features, and the resulting effects ultimately promote tumor growth regulation, angiogenesis, invasion, and metastasis (Condeelis and Pollard 2006; Franklin and Li 2014). The dynamic functional state of TAMs is substantiated by strong evidence suggesting that the antitumoral polarization during tumor initiation switches to an immunosuppressive protumoral state during progression and metastasis formation (Franklin and Li 2014). In general, TAMs are known to support cancer in the early stages by displaying a glycolysis-related inflammatory function, followed by a switch to a more OXPHOS related function in the later stages of tumor progression (Boscá et al. 2015). This metabolic phenotypic switch is underlined by a variety of functional signals present in the TME during the different phases of carcinogenesis. Below are examples on how the metabolic/signaling program is multifaceted during cancer development. The switch to glycolysis in TAMs is under the control of the Akt–mTOR–HIF-1 axis. Signaling factors such as DAMPs activated PI3K–Akt, lead to up-regulation of the glycolytic pathway (Schmid et al. 2011) through stabilization of HIF1α, and the concomitant escape of succinate and citrate from the TCA cycle, the latter being utilized for the cytosolic AcCoA synthesis (Krawczyk et al. 2010; Huang et al. 2014). This phenomenon is associated with inflammation (Schmid et al. 2011). Glycolysis is further enhanced by PKM2, which is present in a dimeric inactive form in TAMs, and potentiates HIF1α in sustaining glucose metabolism (Palsson-McDermott et al. 2017).

These same metabolic features are potentially double-faceted in the complex scenario of the TME. For instance, hypoxia induced-glycolysis enhances lactic acid accumulation in the TME, which leads to the acquisition of a protumoral function of TAMs, promoting tumorigenesis through the release of immunosuppressive and proangiogenic mediators such as IL-10, VEGF, and ARG1 (Colegio et al. 2014). Moreover, the phosphoinositide-3-kinase (PI3K) signal can be overwritten by mTOR, which, once activated, promotes an immune suppressive phenotype in macrophages (Byles et al. 2013). However, targeting REDD1, a physiological inhibitor of mTOR, promotes glycolysis thus decreasing the availability of glucose for EC (Wenes et al. 2016), leading to vessel normalization, to reduced hypoxia and to the inhibition of metastasis formation. Finally, the same PKM2 can modify its conformation into an active tetrameric structure, leading to an M2 like phenotype of TAMs (Palsson-McDermott et al. 2017).

Regarding amino acid metabolism, modes of channeling in distinct metabolic routes produce different functions in TAMs. Arginine metabolism can be associated to the production of NO, which in TAMs results in an antitumoral function (Stuehr and Nathan 1989). However, arginine flux toward polyamines favors a protumoral state of TAMs, with ornithine and collagen production and enhanced proliferation and remodeling (Chang et al. 2001; Ho and Sly 2009). Glutamine metabolism is another crucial pathway for TAM function. Glutamine is traditionally a metabolic fuel for inflammatory macrophages (Murphy and Newsholme 1998), which display high expression of glutaminase. However, glutamine synthesis from glutamate, which occurs through the activity of GS, is a metabolic feature of alternatively activated macrophages induced by IL-10 and promotes the immunosuppressive, proangiogenic and metastatic function of TAMs (Palmieri et al. 2017). Tryptophan metabolism can occur through IDO, which is peculiar of alternatively activated macrophage and in TAMs is responsible for the acquisition of immunosuppressive microenvironment (Zhao et al. 2012b).

Macrophage display peculiar features with respect to lipid metabolism, which is oriented toward synthesis in proinflammatory (Infantino et al. 2014; Palmieri et al. 2015), while uptake (Huang et al. 2014) and oxidation of extracellular lipids are favored in anti-inflammatory macrophages (Odegaard and Chawla 2011). TAMs express high levels of FASN and PPAR, which promotes the effect of tumor growth. In particular, PPARγ mediates alternatively activated macrophage polarization to favor tumor progression and metastasis.

Distinct polarization of macrophages also influences iron metabolism. Classically activated macrophages uptake and store iron, whereas alternatively activated tend to export iron, and make it available for tumor progression (Cairo et al. 2011). In TAMs, iron activates prolyl hydroxylases thus promoting HIF1α destabilization (Koskenkorva-Frank et al. 2013; Wilkinson and Pantopoulos 2013).

IMMUNOMETABOLISM IN THE PREMETASTATIC NICHE

Primary tumor cells secrete a plethora of growth factors and cytokines that induce the formation of a microenvironment in distant organs which is favorable for their growth and metastasis. The PMN is defined as the predetermined microenvironment which attracts the colonization of cancer cells ahead of their arrival. This is distinct from the metastatic niche which is shaped from circulating tumor cells (CTCs) following their arrival in this favorable environment (Peinado et al. 2017). The key components of the PMN are tumor-derived secreted factors, extracellular vesicles, bone-marrow-derived cells, suppressive immune cells and host stromal cells. The preparation of this microenvironment starts from the recruitment of bone-marrow-derived cells. Growth factors and cytokines released from the cancer cells as VEGFA, TGF-β, G-CSF, and the chemokine CCL2 promote the recruitment of hematopoietic progenitor cells, immature myeloid cells, macrophages, T effector cells and also immunosuppressive cells such as myeloid-derived suppressor cells (MDSCs), TAMs, Treg cells, and tumor-associated neutrophils into distant secondary sites (Liu and Cao 2016). The secretome of the immature myeloid cells in combination with the presence of immunosuppressive immune cells prepares the PMN and promotes the recruitment of CTCs.

In the research of premetastatic and metastatic niche, the predominant model is breast cancer and the majority of the knowledge about metabolism at the metastatic site concerns this type of cancer and its metastatic sites, although with most of the emphasis on pulmonary metastasis. It has been shown that the metabolic profile of cancer cells in the primary tumor can dictate the site of metastasis. Dupuy et al. have described the different metabolic profiles of primary tumor breast cancer cells and show that glycolysis-dependent cells metastasize to the liver and the ones relying more on OXPHOS metastasize to the bone and lung. However, very little is known about the metabolic profile of the PMN and how by changing the metabolic profile of the PMN we can prevent the colonization of the CTCs. It has been shown that primary breast cancer cells release exosomes containing mir-122 which leads to the down-regulation of GLUT1 and pyruvate kinase (PKM) in the brain PMN and subsequently suppresses glucose uptake by the stroma cells in the PMN of brain and lung (Fong et al. 2015). This leads to the increase of the available glucose in the PMN which can affect the recruitment and the fitness of immune cells that are attracted in the PMN for release of cytokines and growth factors. Moreover, Joo et al. (2014) have shown that MDA-MB-231 breast cancer cells injected in nude mice release nucleotides such as ATP or UTP that activate the P2Y2 receptor expressed by the cancer cells and also by monocytes, thus inducing the preparation of the PMN formation in the lung, by promoting lysyl oxidase secretion, collagen crosslinking, and monocyte recruitment.

Tumor-derived extracellular vesicles (TEVs) released from tumor cells have also been implicated in the promotion cancer formation and metastasis through their effect on normal cells. Type I interferon receptor-IFN-inducible cholesterol 25-hydroxylase (IFNAR1-CH25H) pathway is important for the resistance of normal cells to the effect of TEVs released from melanoma cells, restricting the formation of lung metastasis. Ortiz et al. (2019) show that low expression of CH25H in leukocytes of melanoma patients was correlated with poor prognosis. Also, transgenic mice incapable of down-regulating the IFNAR1 and CH25H were resistant in TEV-induced development of metastasis indicated the important role of this pathway in the procedure.

The metabolic environment of the PMN is influenced also by the metabolic state of the immune cells that infiltrate the area and vice versa the metabolic environment of the PMN dictates the infiltration and the proliferation of the immune cells that arrive there.

T Cells

CD8⁺ T Cells

Antigen-specific CD8⁺ T cells are able to kill the myeloid cell in the premetastatic lymph nodes. However, the expression of Stat3 on myeloid cells at the premetastatic lymph nodes has been shown to lead to the inhibition of CD8⁺ T cell-mediated killing and protection of the myeloid cells (Zhang et al. 2015). A study by Baek et al. (2017) has proven the association of 27-hydroxycholesterol (27HC) production to increased metastatic burden in several murine breast cancer models. The results of this study suggest that the involvement of 27HC in this process is due to its effect on myeloid cell function which is further associated with increased numbers of neutrophils and γδ T cells. Importantly, treatment of mice with 27HC resulted in reduced numbers of CD8⁺ T cells at metastatic sites. Furthermore, the results of this study indicate that the prometastatic effect of 27HC is due to its ability to prime the PMN. Inhibition of the CYP27A1 rate-limiting step enzyme in the 27HC synthesis, was shown to reduce metastasis in these mouse models. Therefore, 27HC inhibition should be further explored for its potential to reduce priming of the PMN and increase CD8⁺ T cell presence at this step of the metastatic cascade.

Unfortunately, there is still a major gap of knowledge in the field of immunometabolism pertaining to CD8⁺ T cells within the PMN. It would be clinically advantageous to explore possibilities of exploiting these important immune cells in their control of cancer metastasis before its establishment in the PMN.

Treg Cells

Treg cells highly infiltrate various different tumor types and a comprehensive meta-analysis including more than 15,000 cancer cases shows that Foxp3⁺ Treg cells correlate with poor overall survival in cervical, renal, melanoma, and breast cancer. However, Treg number was correlated with improved survival in colorectal, head and neck, and esophageal cancer (Shang et al. 2015). Overall, the accumulation of Treg cells in the circulation and in the primary tumor has been associated with metastasis. The accumulation of Treg cells in the PMN can be promoted by TAMs through CCL22 production and by tumor-evoked regulatory B (Breg) cells through TGF-β secretion (Curiel et al. 2004). Olkhanud et al. (2009) show that secretion of TGF-β from Breg cells is necessary for the recruitment of Treg cells in the lungs and promotion of metastasis of breast cancer cells in the 4T1 model (Sun et al. 2010). Also, Treg cells are recruited to and accumulate in the bone of prostate cancer patients with bone metastasis through the CXCL12/CXCR4 signaling where they contribute to bone deposition (Zhao et al. 2012a). Furthermore, it has been shown that 4T1 tumor cells require CXCR4⁺ Treg cells in order to metastasize to the lungs as CXCR4⁺ Treg cells accomplish their immunosuppressive role by killing NK cells through the release of β-galactoside-binding protein (Olkhanud et al. 2009). The exact mechanism by which Treg cells promote metastasis is not well described and it is dependent on the tumor type and stage. Tumor infiltrating Treg cells can promote vascularization of ovarian cancers and migration of the cancer cells from the primary tumor to the blood circulation through the production of VEGF-A and depletion of Treg cells was shown to inhibit this process (Facciabene et al. 2011). Tan et al. (2011) show that Treg cells can promote mammary breast cancer cell metastasis through the RANK–RANKL signaling as they are the main population produces RANKL. The combination of antibodies against RANKL and CTLA-4 can impair the tumor growth and metastasis in murine melanoma models (Moreno Ayala et al. 2019).

Treg cells can also promote the development of metastasis indirectly by suppressing the antitumor immune responses. Additionally, tumor cells produce immunosuppressive cytokines TGF-β and IL-10 which further enhance Treg cell accumulation. As mentioned before, Treg and T_E cells rely on IL-2 for their proliferation and maintenance. Treg cells express IL-2 receptor chain α (CD25) and their contribution to IL-2 concentration in the TME is low. Therefore, they compete with tumor-suppressive effector CD8⁺ T cells for IL-2 availability in the TME. Also, Treg cells can kill T effector cells directly by releasing perforin and granzyme B and also through the Fas–FasL ligand axis. Furthermore, Treg cells express CTLA-4 which binds to the CD80/CD86 on the surface of the antigen-presenting cells and inhibits their binding to the costimulatory signal CD28 expressed on T_E cells which is necessary for their activation. CD28 activation leads to PI3K stimulation and subsequently activation of mTORC2 and glucokinase (GCK) expression. Glucokinase has been shown to promote glycolysis in Treg cells, cytoskeleton rearrangement and therefore improves Treg migration to the inflammatory site providing GCK as a promising target for autoimmune diseases and also cancer (Kishore et al. 2017).

Macrophages

Emerging evidence points at TAMs as main modulators of the microenvironment that favors tumor metastasis (Luo et al. 2006; Sangaletti et al. 2008; Noy and Pollard 2014; Afik et al. 2016). Several in vivo studies confirm that cancer cells survival is dependent on macrophage recruitment (Gil-Bernabe et al. 2012; Sharma et al. 2015b) and that macrophage recruitment by the primary tumor initiates an immunosuppressive PMN (Sharma et al. 2015b). However, although the PMN has been investigated by means of genomic and proteomic approaches (Hiratsuka et al. 2006; Erler et al. 2009; Kowanetz et al. 2010; Peinado et al. 2017), very little is known about metabolic remodeling in the PMN by macrophage recruitment.

An interesting approach has been described by Kim et al. (2019) in which a microfluidic platform that incorporates EC and ECM scaffolds was developed to evaluate the distinct role of recruited monocytes and macrophages in establishing the premetastatic niches. This demonstrates that a macrophage-induced remodeling occurs prior metastasis formation. Additionally, it might represent a suitable approach to unravel the metabolic mechanism underpinning the premetastatic niche formation.

IMMUNOMETABOLISM AND METASTASIS INITIATION

Cancer cells evade immune recognition in part by recruiting immunosuppressive cells such as Treg cells and certain macrophage populations (Janssen et al. 2017). Tumor cells are also able to alter the TME in a way to promote blood vessel formation (angiogenesis) which facilitates the shedding of tumor cells into the circulation. To the best of our knowledge, there are no distinct metabolic features of metastatic tumors compared to noninvasive tumor types. Instead, cancer cells take advantage of different metabolic alterations in their invasive process which differ depending on the origin of the primary tumor (Peinado et al. 2017; Elia et al. 2018).

T Cells

There is minimal knowledge connecting T cell metabolism with tumor cell extravasation and metastasis initiation. Therefore, in order to understand future steps in the study of T cell metabolism and metastasis, it is beneficial to look at the knowledge available on cancer cell metabolism and how it can affect the metabolism of T cells revealing potential immunotherapy targets. Metabolic changes in cancer cells have been shown to be responsible for the changes in the activation of signaling pathways resulting in increased invasiveness of cancer cells through the degradation of the ECM, modification of cell–cell interactions and even stimulation of cancer cell motility. For example, the expression of HK2 and PKM2 results in increased glycolysis as well as lactate production and amino acid metabolism (Lunt et al. 2015) and has been shown to lead to increased motility and invasion of cancer cells (Yu et al. 2015; Botzer et al. 2016; Lu et al. 2016). Considering that effector T cells such as cytotoxic CD8⁺ T cells also rely on glycolysis for their effector function it is reasonable to infer that this enhanced need for glycolysis in invasive cancer cells even further limits CD8⁺ T cell response. Additionally, increased fatty acid metabolism and OXPHOS has been linked to a more invasive phenotype of cancer cells (Elia et al. 2018). The heterogenous metabolic pathways utilized by invasive cancer cells could be explained by the high energy demands associated with motility and other characteristics contributing to a highly invasive phenotype. Due to the diverse metabolic profile of invasive cancer cells, it is difficult to pinpoint to therapeutic targets in order to specifically block the first step of the metastatic cascade. Studies connecting the metabolism of invasive cancer cell and tumor infiltrating T cells would lead to new insights on immunometabolism and could point to new immunotherapeutic targets in the metastatic setting.

Macrophages

Knowledge on the contributing role of TAMs in fostering cancer cell invasion and metastasis initiation comes from the seminal studies on how metabolic pathways in immune cells affect the formation of blood vessels of tumor vascular structures (Mazzone et al. 2018; Pearce and Pearce 2019). Significant cues are provided by studies on hypoxic conditions, which shape TAM metabolism and function (Laoui et al. 2014) to favor angiogenesis, fostering the shedding of tumor cells into the circulation. Hypoxia appears to significantly modify gene expression in TAMs, and this might instate interesting metabolic crosstalk mechanisms between TAMs and EC. REDD1 expression in hypoxic TAMs is paradigmatic: its up-regulation in hypoxic TAMs, which inhibits mTOR, lowers TAM glucose consumption and thus favors ECs glycolytic hyperactivation and dysfunctional blood vessel formation, which ultimately leads to cancer cell dissemination and metastasis (Wenes et al. 2016). Additionally, the effect of radical species on TAM function affect angiogenesis. As described in macrophages, in a hypoxic TME, arginine metabolism leads to NO production, which displays an immune suppressive function (Doedens et al. 2010). However, NO in normoxic TAMs promotes tumor blood vessel normalization and EC activation (Klug et al. 2013). This evidence highlights the context specificity by which TAM metabolism can affect vessel formation, leading to cancer cell dissemination and metastasis (Rivera and Bergers 2013). This concept is additionally substantiated by evidence on TAM glutamine metabolism. Glutamine is a typical energy fuel for proinflammatory macrophages, although evidence on the anti-inflammatory role of glutamine also exists. In a tumor context, in which glutamine tissue levels are far from homogeneous (Castegna and Menga 2018) due to the presence of cells with different glutaminolytic capacity, TAMs sense glutamine deprivation and induce GS, which metabolically reprograms TAMs toward a proangiogenic phenotype with increased metastasis (Palmieri et al. 2017). This suggests that glutamine competition fosters TAMs function toward a proangiogenic behavior by a mechanism involving GS expression (Castegna and Menga 2018). This is an example of direct metabolic crosstalk between different players in the TME. However, GS expression in TAMs is dependent on stimuli other than glutamine availability, since it can be up-regulated by IL-10 (Palmieri et al. 2017) even in a glutamine-rich environment. Another example of the cytokine-mediated proangiogenic effect on TAMs comes from the discovered effect of IL-6 secretion by ECs on TAMs. IL-6 induces peroxisome proliferator-activated receptor gamma (PPAR-γ), that regulates lipid uptake and glucose metabolism and also promotes an alternative macrophage activation by HIF-2α induction, which practically leads to microvascular endothelial cell proliferation (Takeda et al. 2010).

A recent report by our group describes a population of macrophages expressing podoplanin (PoEMs) (Bieniasz-Krzywiec et al. 2019). These cells are located near lymphatic vessels where they contribute to lymphangiogenesis, lymphoinvasion, and metastatic dissemination of cancer cells by aiding in their intravasation. In breast cancer patients, the association of PoEMs with the tumor lymphatic vessels correlates with lymph node and distant metastasis. This population of TAMs is characterized by high expression of glucose uptake and anaerobic glycolysis-related genes. Given this distinct metabolic gene expression, it would be interesting to explore if metabolic reprogramming therapies could target these cells and alter their ability to contribute to the dissemination of cancer cells.

IMMUNOMETABOLISM AND CIRCULATING TUMOR CELLS

Only a small fraction of CTCs are able to survive in circulation and they do so by displaying higher levels of OXPHOS, increased ATP production as well as pyruvate and oxygen consumption, which is unlike the famous Warburg metabolism of cancer cells (Caneba et al. 2012). Metastasis is generally not an efficient process as the success rate of cancer cells that were shed from the primary tumor to develop into metastasis is very low (Strilic and Offermanns 2017). It is generally believed that CTCs are met with different immune surveillance in the circulation than what they encounter in the protective TME (Leone et al. 2018). Immune cells that encounter CTCs possess both pro- and antitumoral properties. A better understanding of the interactions of these different cells in circulation could lead to the development of new therapeutics targeting metastasis before it has a chance to develop.

T Cells

In patients, it is generally reported that higher numbers of infiltrating lymphocytes correlated with an increase survival rate and decreased metastasis occurrence (Gooden et al. 2011). Additionally, in breast cancer patients with detectable circulating cancer cells, Gruber et al. (2013), found an increase proportion of First Apoptosis Signal (FAS) expression on CD4⁺ T helper cells indicating an initiation of apoptosis in these cells. It remains unclear if this increased apoptosis of T cells in patients with circulating cancer cells is due to immunosuppression properties of the cancer cell in circulation or an effect of the primary tumor (Leone et al. 2018).

The detachment of cancer cells from the ECM leads to the accumulation of ROS which leads to cell death (Hawk and Schafer 2018). Invasive cancer cells have acquired the ability to regain the homeostasis of ROS in order to survive this oxidative stress once in circulation. Cancer cells do this by optimizing ROS scavenging in the mitochondria, reducing ROS production via different pathways and up-regulating antioxidant mechanisms (Elia et al. 2018). Oxidative stress has been shown to inhibit T_E function by down-regulating Th1 response through reduction of IFNγ production (Frossi et al. 2008).

Arginine levels are reported to be lower in both patients with solid tumors (Vissers et al. 2005) and those with blood cancers, such as acute myeloid leukemia (Mussai et al. 2018), compared to healthy controls. In the case of AML, this has been shown to be due to increased secretion of arginase II from cancer cells (Mussai et al. 2013). Additionally, reports by Mussai et al. (2013, 2018), demonstrate that these increased arginase levels have a significant effect on T cell proliferation and function. In a study using blood samples from AML patients the researchers observed that although circulating T cell demonstrated the capacity to be activated in response to antigen, thus epigenetic modifiers could enhance their immune function, however, this effect was not observed (Mussai et al. 2018). In the same study, a coculture of human T cells and purified AML blasts demonstrated that AML blasts were able to inhibit T cell proliferation. T cell proliferation was restored upon treatment with arginase inhibitors L-NMMA and L-NOHA (Mussai et al. 2018).

Macrophages

Knowledge about the relationship between immunometabolism and CTCs is scarce. An interesting strategy to address the functional features of the CTC-macrophage interaction has been described by Hamilton et al. (2016). In this study, PBMCs were exposed to CTC cultures obtained from the blood of small-cell lung cancer (SCLC) patients with relapsing disease, or to their corresponding conditioned media (CM). The induction of monocyte–macrophage differentiation in these conditions was compared to the differentiation obtained by the metastatic SCLC cell line. Interestingly, CM of the SCLC CTCs induced the monocyte–macrophage differentiation with a higher frequency compared to the metastatic cell line itself, indicating the CTCs are able to release differentiation factors much more effectively. From this study, it emerges that SCLC CTCs recruit and educate a distinct type of macrophages with specific functionalities toward invasion, immune protection, extravasation, and cachexia (Hamilton et al. 2016). Furthermore, this approach, although with the obvious limitations of an in vitro study, could be a suitable strategy to uncover elements of metabolic crosstalk between CTC and macrophages useful to understand the complex scenario of metastasis formation and identify novel pharmacological targets.

IMMUNOMETABOLISM AND THE METASTATIC NICHE

Local immunity is an important feature of metastatic sites because CTCs must evade local immune responses for successful metastasis (Massagué and Obenauf 2016). As it is mentioned before the preference of specific tumor types to metastasize to specific organs and not to others, known as organotropism, is a result of the expression of specific ligand–receptor interactions between the tumor cells from the primary tumor and the stromal cells at the premetastatic site. The environment that the tumor cells encounter once arriving at the PMN will dictate whether or not they will invade the particular tissue. The metabolic profile of the cancer cells in the secondary organs can differ from their metabolism in the primary site as it depends on the nutrient availability (Elia et al. 2018).

T Cells

Not much is known about the metabolism of T cells specifically in the metastatic niche. Clever et al. (2016) showed that of the oxygen-sensing PHD proteins expression by T cells in the lung is important in the maintenance of local tolerance against harmless antigens, but promotes metastasis of CTCs in the lung by suppressing the Th1 CD4⁺ T cell and CD8⁺ T cell function and promoting Treg cells. It was shown that genetic deletion or pharmacologic inhibition of PHD proteins in T cells limits tumor colonization of the lung and improves the efficacy of adoptive cell transfer (ACT) immunotherapy.

Tumor cells overexpress IDO which has a strong immunosuppressive role as is it deleterious for T_E cells, favors Treg cells differentiation and leads to MDSC expansion and activation which is Treg-dependent. Inhibition of IDO not only inhibits cancer cell growth but also enhances the immune response by promoting antitumor immune cells and inhibiting immunosuppressive populations (Holmgaard et al. 2015). Additionally, IDO has been shown to promote the creation of the lung metastatic niche in breast cancer model (Smith et al. 2012).

A genome wide in vivo screen performed by van der Weyden et al. (2017) identified novel host regulators of the metastatic disease. This study revealed 19 novel genes, which when deleted in mice have the ability to alter the metastatic potential of cancer cells. The authors highlight the role of sphingosine-1-phosphate (S1P) transporter homolog 2 (Spns2) in metastasis. The deletion of Spns2 lead to reduced lung metastasis as well as increased T_E and NK cells numbers in the lung. The negative effect on metastasis of this gene was confirmed in breast, colon, and lung cancer, therefore, highlighting the Spns2 gene as a promising new cancer drug target.

Macrophages

Accumulating evidence suggests that the presence of specific nutrients dictates particular metabolic reprogramming in cancer cells at the metastatic site. It is conceivable that this very same environmental-imposed metabolic rewiring also involves TAMs, which are known to play a crucial role in metastasis formation and are also very plastic both metabolically and functionally. With respect to the metastatic niche, metabolic competition among different cell type might play a crucial role in modulating macrophage function. The available evidence points at two distinct scenarios: (1) macrophages at the metastatic niche might influence the availability of specific nutrients in the TME, leading to functional changes in other cells: this is the case of REDD1 up-regulation that lowers glucose consumption in TAMs, making the source available to sustain other cells, among which cancer cells (Wenes et al. 2016) and (2) macrophages at the metastatic niche might undergo a phenotype rewiring after sensing nutrient deprivation: this is the case of GS up-regulation in response to extracellular glutamine deprivation, which not only forces TAM toward a prometastatic phenotype, but also promotes glutamine secretion for cancer cell's use (Palmieri et al. 2017). More evidence with this respect is strongly awaited.

A very interesting contribution for the study of the metastatic niche comes from the previously mentioned work by van der Weyden et al. (2017), where 19 novel genes were identified as modulators (once disrupted in mouse) of the host control of metastasis. Although very little is known about the function of many of these genes, their involvement, if expressed by macrophages, in shaping niche metabolism to favor metastasis formation could be speculated. Among the antimetastatic genes is the one encoding for PI3K that generates phosphatidylinositol 3,4,5-trisphosphate (PIP3), which recruits PH domain-containing proteins to the membrane, including AKT1, an mTOR activator. This is not in line with the phenotype of Akt1-deleted macrophages, in which inducible NO synthase and IL-12β secretion are enhanced, together with enhanced bacteria clearance (Kuijl et al. 2007; Arranz et al. 2012) and response to LPS (Androulidaki et al. 2009; Arranz et al. 2012). Among the prometastatic genes are several NOX subunits, in particular cybB and cybA, the enzymatic activity of which is mainly involved in ROS production in macrophages. This result is not expected considering that NOX activity is distinctly associated with proinflammatory macrophages (Xu et al. 2016). However, it is consistent with the recent findings that NOX1 and NOX2 deficiency impairs macrophage differentiation and the occurrence of M2-type TAMs during tumor development, leading to decreased metastasis (Xu et al. 2016). Another interesting prometastatic target is the heat shock protein 90 (HSP90) aa1 (or α), which is known to function as extracellular modulator of inflammation in cancer-associated fibroblasts through the activation of the NF-kB (RELA) and STAT3 transcription programs including the proinflammatory cytokines IL-6 and IL-8 (Bohonowych et al. 2014; Zuehlke et al. 2015). Incidentally, the mitochondrial isoform of HSP90, the TNF receptor-associated protein 1 (TRAP1), is a known metabolic modulator of succinate dehydrogenase activity leading to a pseudohypoxic state associated to succinate accumulation in cancer cells (Masgras et al. 2017). All these findings highlight once again the limitations of the in vitro “M1”- and “M2-like” metabolic classification in explaining the complex metabolic scenario of metastasis formation.

OPPORTUNITY FOR IMPROVEMENT OF CANCER IMMUNOTHERAPY—FOCUS ON METABOLISM IN THE CONTEXT OF A METASTATIC DISEASE

It is important to note that most current immunotherapeutic approaches are being tested in patients with metastatic disease, however, it is difficult to delineate if tested immunotherapeutic approaches act specifically on any part of the metastatic cascade. In this section, we attempt to summarize some of the potentially promising therapeutic approaches focusing on the metabolic rewiring of T cells and/or macrophages in the metastatic setting. A recent extensive bioinformatic analysis of metabolic profiles from 900 tissue samples from seven different cancer types identified common metabolic changes associated with different steps of progression of aggressive metastatic disease (Reznik et al. 2018). This study highlights the clinical significance of systemic changes in metabolite concentration due to tumor progression and demonstrates the importance of exploiting this knowledge to develop novel therapeutics with metabolic targets for the treatment of cancer.

T Cells

The metabolic rewiring of the T cells is a very promising strategy which can boost the efficiency of cancer immunotherapy. One approach, adoptive cell transfer (ACT), consists of the isolation of tumor-specific T cells from the tumor of the patient, ex vivo expansion and reinfusion into the patient combined with administration of high doses of IL-2. A second strategy consists of the generation of genetically engineered T cells in order to express the chimeric-antigen receptor (CAR T cells) or a tumor-antigen-specific T cell receptor (TCR) which are expanded and injected into the patient (Kishton et al. 2017). Both approaches rely on the in vitro expansion of the T cells before the injection and offer a great opportunity to induce changes in the cell metabolism in order to improve the fitness of the cells, prolong their activation status and delay the following exhaustion of the injected T cells in the TME. The metabolic needs of T cells during ACT can differ significantly (O'Sullivan et al. 2019). Ex vivo expansion of these cells involves supplying IL-2 and a variety of amino acids to ensure survival, activation, and proliferation. At this step aerobic glycolysis is important. However, it is important to remember that long-term stimulation and proliferation will lead to T cell exhaustion. After ACT the activated T cells must adopt their metabolism in the constantly changing TME. At the same time the formation of T_M cells is critical (the specific metabolic need of T_M cells was described in a previous section of this report). Considering these different metabolic needs of T cells throughout the ACT process, we must, therefore, tailor any metabolic alterations in these cells depending on the step of the process and on the tissue in which we want to intervene (different primary tumors vs. different metastatic niches), thus making this strategy more challenging.

Another promising strategy of cancer immunotherapy relies on the fact that metabolic rewiring can play a key role in immune cell function. This therapeutic strategy would involve the coadministration of immune checkpoint inhibitors and drugs aimed at metabolic targets. These potential therapeutic options are summarized below and outlined in Table 2. Sukumar et al. (2013) show that inhibition of glycolysis during the in vitro expansion of CD8⁺ T cells prior to ACT enhance their potential to become long-lived memory cells and therefore enhance the efficacy of cancer immunotherapy. Furthermore, arginine supplementation during T cells expansion leads to enhanced OXPHOS and better antitumor function (Geiger et al. 2016).

The glucose competition between cancer cells and T cells leads to low glucose availability in the TME. Due to unavailability of glucose, T cells down-regulate the aerobic glycolysis and the phosphoenolpyruvate (PEP) enzyme. Phosphoenolpyruvate is an important regulator of the NFAT signaling and T effector functions and also is an inhibitor of SERCA. It has been shown that overexpression of PEP carboxykinase 1 (PCK1) enzyme which converts oxaloacetate (OAA) into PEP in CD4⁺ and CD8⁺ T cells prior to ACT leads to increased PEP production, more efficient effector function and inhibition of the tumor growth and prolonged survival in a murine melanoma tumor model (Ho et al. 2015). PD-1 administration inhibits glycolysis and amino acid metabolism of T cells, promotes lipolysis and FAO by inducing the expression of CPT1 leading to prolonged survival of T cells (Patsoukis et al. 2015).

Targeting CD8⁺ T Cells

A recent study by DeVorkin et al. (2019), reports on the role of an autophagy gene Atg5 in CD8⁺ T cells response in the TME. When Atg5 was deleted in CD8⁺ T cells, these cells acquired an effector memory phenotype and increased IFN-γ and TNF-α production. This study demonstrated that the impairment of autophagy in T cells can shift the metabolic advantage in the TME in favor of T cells, therefore, this strategy has the potential to enhance immunotherapy approaches in cancer.

4-1BB (CD137):4-1BBL is one of the costimulatory signals that enhance CD8⁺ T cell proliferation. The anti-4-1bb treatment has been shown to stimulate both glycolysis and FAO and prolonging the survival of the activated CD8⁺ T cells (Choi et al. 2017).

Targeting Treg Cells

The complete depletion of Treg cells leads to the development of autoimmune diseases, therefore, targeting specific tumor infiltrating Treg cells (TI-Treg) is essential. The specific targeting of TI-Treg cells demands a better understanding of the differences between TI-Treg and Treg cells. The differences in metabolic profiles could be key for the development of new pharmacologic approaches. The depletion of TI-Treg cells, or even better their rewiring from immunosuppressive to immune-promoting cells, will be able to improve the response to therapy and lead to better disease outcome. One should also keep in mind that the metabolic requirements during CD4⁺ naïve T cells polarization to Treg cells can be different from the metabolic requirements of proliferating Treg cells in the tumor. The most efficient approach should implement the inhibition of the proliferation of TI-Treg cells and the reverse of their suppressive phenotype. Plitas et al. (2016) attempted to answer how TI-Treg cells differ from other Treg subtypes by RNA-seq analysis. They did so by comparing the transcriptomic profile of Treg cells derived from the breast tumor, normal breast parenchyma and Treg cells from peripheral blood of breast cancer patients. They found that CCR8 is highly expressed in tumor-derived Treg cells compared to controls and is strongly associated with poor prognosis. Additionally, Moreno Ayala et al. (2019) summarizes the important pathways for Treg cells and the differences between TI-Treg and Treg cells.

Until now the most common drug to target Treg cells has been the anti-CD25 antibody. Treatment of human in vivo derived Treg cells with daclizumab, the FDA approved anti-CD25 antibody, leads to Foxp3 down-regulation and induction of IFN-γ production (Shimizu et al. 1999). However, CD25 is not expressed only by TI-Treg cells and the global depletion of Treg cells can lead to severe side effects due to the development of autoimmunity.

PI3K/Akt/mTOR signaling pathway is activated after CD28 stimulation in conventional T cells. However, in Treg cells, there is a reduced activity of PI3K/Akt (Huynh et al. 2015). This pathway is inhibited by Phosphatase and Tensin homolog (PTEN) and PH-domain leucine-rich-repeat protein phosphatase (PHLPP), both highly expressed in Treg cells even after IL-2 stimulation in contrast to T_E cells. Pharmacologic inhibition of PTEN leads to a decrease of Treg cells in the tumor context (Sharma et al. 2015a). Recently it was shown that PI3Kδ isoform is specifically important for TI-Treg cells and not T conventional cells and that the inhibition of PI3Kδ in combination with immunotherapy results in a decrease in suppressive Treg cells and an increase in CD8⁺ T cells in the tumor context (Ahmad et al. 2017).

Glucocorticoid-induced TNF receptor (GITR) is a member of the TNF receptor superfamily that is highly expressed at CD4⁺CD25⁺Foxp3⁺ Treg cells and at low levels at naïve and T memory cells (Nocentini and Riccardi 2009). Anti-GITR mAb specifically inhibits TI-Treg cells without affecting the systemic Treg cells and Treg cells derived from tumor-draining lymph nodes. At the same time, this treatment enhances the CD8⁺ T cell memory pool offering a great opportunity to specifically target TI-Treg cells without compromising the function of effector cells (Coe et al. 2010). Intravenous or intratumoral injection of anti-GITR lead to tumor regression in advanced murine tumors and coadministration of anti-GITR and anti-CTLA-4 had a synergistic effect (Ko et al. 2005). At the moment a clinical trial with the GWN323 anti-GITR mAb monotherapy or combined with anti-PD1 is ongoing in patients with advanced solid tumors and lymphomas (NCT02740270). Also, a clinical trial with TRX518 anti-GITR mAb in patients at stages III and IV malignant melanoma or other solid tumors is ongoing (NCT01239134).

As previously mentioned murine in vitro-induced Treg cells are generated upon TGF-β stimulation of naïve CD4⁺ T cells and rely more on FAO than glycolysis compared to T_E cells (Michalek et al. 2011). In contrast, human-derived Treg cells isolated from blood are highly glycolytic but require FAO in addition to glycolysis for their proliferation in vitro (De Rosa et al. 2015). This is similar to what is known for CD8⁺ T cells which require both glycolysis and FAO for proliferation.

Macrophages

Given their genetic stability, compared to cancer cells, targeting TAMs might represent a promising strategy to treat cancer, especially combined with more traditional approaches, such as chemotherapy. Several TAM specific approaches have been tested in mouse and some specific TAM therapies have been tested in clinical trial. The different approaches in mice, extensively reviewed in Poh and Ernst (2018), comprehend (1) promoting macrophage depletion; (2) impairing macrophage recruitment; and (3) reprogramming macrophage function. Targeting metabolism is expected not to interfere with macrophage viability but to polarize TAMs toward a more antitumoral and proinflammatory phenotype, which will eventually result in reduced cancer growth and metastasis development. One of these studies, the IDO1 inhibition in combination with checkpoint inhibitors, has reached clinical Phases 1 and 2, but it did not always show additional benefit to the anti-PD-1 antibody use (Soliman et al. 2018) (NCT02752074).

The opportunities for a metabo-therapeutic approach to revert TAM phenotypes comes from the evidence dissecting the specific metabolic features of the differently polarized macrophages. High-resolution intravital imaging with single-cell RNA seq has been used to uncover profiles of immunosuppressive macrophages in mice, and Arginase 1 emerged as a distinctive marker (Arlauckas et al. 2018). Arginase 1 inhibition has been tested alone or in combination with anti-PD1 therapy. Although there are no additive or synergistic benefits reported in combining these treatment strategies in the used murine models, a statistically significant effect early after treatment with the Arginase inhibitor was present (Arlauckas et al. 2018). Furthermore, genetic or pharmacological targeting of the COX2/mPGES1/PGE2 pathway involved in the regulation of PD-L1 expression in tumor-infiltrating myeloid cells has been reported to reduce PD-L1 expression in these cells, providing an opportunity to interfere with immune suppression in tumor host (Prima et al. 2017). A recent study supports the concept of targeting stromal LDH-A to block tumoral immune escape. Myeloid-specific deletion of LDH-A promoted the accumulation of antitumoral TAMs, expressing CD86 and MCP-1, which reduced angiogenesis and the number of PD-L1+ cancer cells, while fostering the antitumoral T cell immunity via induction of IL-17 and IFNγ-producing CD8⁺ T (Seth et al. 2017). Similarly, GS targeting is a promising pharmacological target to revert TAM phenotype, due to the reduced angiogenic, immunosuppressive effect of GS specific deletion in macrophages, which leads to decreased metastasis (Palmieri et al. 2017). Furthermore, targeting the ABC transporter responsible for cholesterol efflux from macrophages has been shown to induce an IL-4 reprogramming able to reduce ovarian cancer progression (Goossens et al. 2019).

Except for TAMs and Treg cells, polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) have a positive impact in tumor initiation and progression. As it was recently shown by Veglia et al. (2019), the Fatty Acid Transport Protein 2 (FATP2) is overexpressed in PMN-MDSCs compared to neutrophils and plays an important role in the suppressive phenotype of these cells. Pharmacologic inhibition of FATP2 alone leads to tumor suppression in different murine tumor models, and this therapeutic effect is synergically stronger when FATP2 blockade is combined with anti-CTLA4, anti-CSF1R or anti-PD1 depending on the tumor model.

Besides a direct pharmacological action on specific enzymatic activities, less specific approaches have been evaluated. Activation of the Toll-like receptor 9 with a CpG oligodeoxynucleotide has been reported to reprogram macrophage metabolism toward de novo lipid biosynthesis, promoting antitumor activity, including engulfment of CD47⁺ cancer cells (Liu et al. 2019). Macrophage-associated V-set and Ig domain-containing 4 (VSIG4) has been proposed as an M1 metabolic checkpoint inhibitor (Liao et al. 2014). Targeting VSIG4 could potentially repolarize TAMs toward M1-like macrophages. Although the metabolic effects of its targeting are still elusive, specific targeting of VSIG4 may prove to be a novel, efficacious strategy for the treatment of lung carcinoma (Liao et al. 2014).

CONCLUDING REMARKS

Metabolism is a biological process essential for every cell in the body. Extensive studies show that metabolic changes are responsible for different immune cell phenotypes which determine the immune response to different pathogens and cancer. Boosting the antitumor immune response has emerged as a promising therapeutic approach to combat cancer and targeting immunometabolism will likely represent an important direction in this field. As highlighted in this review, cellular metabolism has the potential to modulate the immune response to cancer cells at different stages of the metastatic cascade. Therefore, there is potential for a significant impact of utilizing drugs targeting metabolic targets in the immunotherapeutic regimens. In this setting patients would only have to be treated for a short time allowing for the rewiring of the TME, creating a window of opportunity with a favorable environment for antitumor immune cells to recognize and eliminate cancer cells.

It is important to study the individual steps in the metastatic cascade and identify therapeutic targets we can exploit in order to stop cancer at different steps of the disease. Despite this need, there are still large gaps in our knowledge of immunometabolism as it relates to the control of cancer and the metastatic cascade. The important next step in immunometabolism to focus on are (1) involvement of immune cells in the PMN priming, (2) immune cell-mediated killing of cancer cells in circulation, and (3) immune cell clearance of cancer cells arriving at the metastatic niche and the elimination of immunosuppressive cells accumulating at this site. In this review, we summarize the knowledge currently available on the immunometabolism of T cells and macrophages as it relates to metastasis. We highlight several potential therapeutic targets identified so far, however, the path to the clinic remains obscure. Additionally, it would be important to test the efficacy of any new therapeutics in combination with available checkpoint inhibitors as several studies indicate a potential synergistic effect. There are also technical constraints to consider which make these studies more difficult to carry out. Therefore, in vivo assays allowing for the study of immunometabolism at different steps of the metastatic cascade are urgently needed. Overall, it is clear that future therapeutics targeting immunometabolism has the potential to greatly improve the lives of patients living with this largely uncurbable disease.