Sculpting tumor microenvironment with immune system: from immunometabolism to immunoediting

Metabolic features of cancer cells and their impacts on infiltrating immune cells

Tumor cells engage aerobic glycolysis to support their metabolic demand and adenosine triphosphate (ATP) production even when sufficient oxygen is present, a phenomenon known as the Warburg effect. Although the efficiency of ATP production of aerobic glycolysis is lower than oxidative phosphorylation (OXPHOS), increased glucose uptake and consumption compensate the inefficient ATP production rates and the engagement of aerobic glycolysis supports the biosynthesis of macromolecules used for cell proliferation. Thus, as a result of elevated glucose consumption in tumor cells, the TME becomes a glucose‐deprived microenvironment in some tumor types 32, 33, 34. However, metabolic reprogramming from OXPHOS to aerobic glycolysis is also required for the metabolic demands of T cell proliferation and effector functions 35. During T cell activation, T cells engage metabolic switch to meet bio‐energetic demands for following proliferation and differentiation. Upon T cell receptor (TCR) activation, naive T cells, which rely on OXPHOS for ATP production, switch to anabolic metabolism, including elevated aerobic glycolysis, the pentose–phosphate pathway and glutaminolysis 36, 37. TCR activation also induces mitochondrial biogenesis and fission, leading to looser cristae structures and lower respiratory functions 38. Thus, glucose deprivation in the TME can disarm T cell‐mediated anti‐tumor responses, including suppressing interferon (IFN)‐γ production, cytotoxicity and motility in T cells 32, 33, 39, 40. In support of this, several studies have demonstrated that increasing glucose consumption via up‐regulated glycolysis activity in cancer cells indeed represses glucose uptake and utilization in tumor‐infiltrating lymphocytes (TILs), thereby impeding anti‐tumor immunity 32, 33. Furthermore, glucose deprivation can also up‐regulate PD‐1 expression, leading to an exhausted feature in TIL 41. In addition to glucose deprivation, hypoxia, which is another physiological factor in the immunosuppressive TME, also has considerable impacts on TIL proliferation and function 24. Hypoxia‐induced HIF‐1α has been shown to promote expression of co‐inhibitory receptors and PD‐L1 in TILs and cancer cells, respectively 42. Intriguingly, hypoxia also impacts T cell metabolism by reducing the use of OXPHOS and enhancing glycolysis in effector T cells, which may further promote effector functions and anti‐tumor immunity 43. However, the combination of glucose deprivation and hypoxia in the TME can drastically weaken anti‐tumor immunity, as immune cells may not be able to sustain metabolic flexibility while dealing with these two contradictory conditions. Mechanistically, declined glucose availability in the TME cannot support hypoxia‐induced metabolic switch. This conflict between bio‐energetic demands induced by glucose deprivation and hypoxia may further lead to the impairment of mitochondrial fitness in CD8⁺ T cells (Fig. 1). Recent studies have demonstrated that CD8⁺ TILs are functionally impaired due to toxic levels of reactive oxygen species (ROS) and accumulated dysfunctional mitochondria 44, 45, which has been also reported in the lymphocytic choriomeningitis virus (LCMV)‐c13 chronic viral infection system 46. Of note, these dysfunctional mitochondria not only influence energy production, but also lead to alterations of metabolites, which may be involved in cellular signaling transduction and epigenetic regulations.

In line with increased glucose consumption in tumor cells, lactate is massively produced and released from tumor cells via monocarboxylate transporters (MCT)‐1 and ‐4, which simultaneously facilitate the transport of lactate anion and proton 47, 48. This further leads to the accumulation of lactate and increased acidity in the TME. The high abundance of lactate results in plastic remodeling of cytokinome and suppresses proliferation and mobility of cytotoxic T lymphocytes and NK cells, but supports formation of pro‐tumorigenic TAMs and T_regs 49, 50. Mechanistically, lactate stimulates pro‐tumorigenic M2 polarization of TAMs by stabilizing HIF‐1α 51. Moreover, lactate may support T_reg differentiation and survival in a MCT4‐dependent manner 52, 53. Amino acids, such as glutamine, tryptophan and arginine, also play critical roles in T cell differentiation, activation and effector functions 54, 55, 56, 57, 58, 59. However, lactate could promote glutamine uptake, which may lead to depriving glutamine in the TME 60. Moreover, over‐expression of glutaminase in tumor cells and elevated expression of IDO in tumor and immune modulatory cells also further dampen T cell anti‐tumor immunity by depriving glutamine and tryptophan in the TME 61. Additionally, a high abundance of arginase‐1 in the TME also results in cell cycle arrest and impairment of antigen‐specific T cell responses by restricting arginine availability 62. Overall, the metabolic features of tumor and tumor‐infiltrating immune modulatory cells could create metabolically hostile conditions that are detrimental for mounting effective anti‐tumor immunity.

Genomic mutation, neoantigen and immunoediting determines immunogenicity of tumor cells and beyond

Emerging evidence has shown that alterations in the expression and activity of TP53, RAS, MYC, PTEN and CTNNB1 in cancer cells dampen anti‐tumor immunity by altering cytokine and chemokine profiles in the TME 63, 64. Moreover, driver mutations leading to constitutive activation of the mitogen‐activated protein kinase (MAPK) pathway and other pro‐tumorigenic signaling cascades have been revealed to facilitate the establishment of immunosuppressive TME by recruiting MDSCs and T_reg cells and reducing major histocompatibility complex (MHC) class I expression and up‐regulating expression of PD‐L1 and bioenergetic activity in tumor cells 65. Furthermore, genome instability results in the high frequency of genomic mutations and contributes to aberrant functions of oncogenes and tumor suppressors, which support tumor progression and determine cancer immunogenicity. Genomic instability in cancer cells also results in the generation of neopeptides, which are produced from the mutated peptide sequence, including non‐synonymous mutations, frame‐shifting, intron retaining and fusion genes 66, 67, 68, 69. Some of these tumor‐specific neopeptides can be processed and presented by antigen‐presenting cells as neoantigens, which will be further recognized by T cells. Therefore, identifying tumor neoantigens appears to represent ideal targets for boosting cancer immunogenicity and cancer therapy. The identified neoantigens can be subsequently utilized in the development of neoantigen‐based cancer vaccines 70. Moreover, the higher neoantigen burden in tumors has been shown to be correlated with the clinical benefit of immune check‐point blockade in melanoma and non‐small‐cell lung cancer patients 71, 72. Together, these findings indicate that somatic mutations in oncogenes or tumor suppressor genes involved in certain pathways may alleviate cancer immunogenicity. Based on these findings on how host anti‐tumor immunity can be engaged and hampered, immunoediting has been proposed to select tumor cells with the highest ability to impair anti‐tumor immunity to survive during tumorigenesis 3, 73. Generally, it is believed that cancer immunoediting consists of three phases: elimination, equilibrium and escape. Elimination is a consequence of immunosurveillance, which includes innate and adaptive immune responses to cancer cells. In this phase, the developing tumors are successfully eradicated by immune cells before they become clinically apparent. However, sometimes rare tumor cell variants may not be destroyed completely in the elimination phase and further proceed into the equilibrium phase, which maintains a balance between prolonged elimination of cancer cells and the growth of resistant tumor variants. When cancer cells circumvent immune recognition or the functional dormancy may be broken, they progress from the equilibrium phase into the escape phase, which is acknowledged as one of the hallmarks of cancer. In the escape phase, cancer cells escape by reduction of immunoselection pressure and establishment of immunosuppressive TME. Interestingly, as described previously, a majority of tumor cells in the clinically apparent stage utilize aerobic glycolysis, and this feature could give them superior ability to disarm anti‐tumor immunity by impairing the metabolic fitness of tumor‐infiltrating T and NK cells. During tumor initiation, transformed cells harboring different genomic mutations generate various clones, which lead to heterogeneity in the hallmarks of cancer, including metabolic preference. Therefore, we postulate that tumor cells with heterogeneity of metabolic preference may evolve into glycolysis‐dominant cancer as a result of the specialized format of immunoediting, in which the metabolic cross‐talk between tumor and immune cells force the survival and expansion of tumor cells with elevated glycolytic activity, and possibly other unexplored metabolic features (Fig. 2). In this model, based on our understanding of how metabolic cross‐talk disarms anti‐tumor immunity, we hypothesize that the host immune system not only eradicates tumor cells expressing immunogenic antigens, but also metabolically unfit cancer cells, which fail to abolish the metabolic fitness of anti‐tumor immune cells. Moreover, tumor clones with up‐regulated glycolysis also generate more immunosuppressive features such as limited glucose, but increase lactate availability in the TME, which further recruits and supports immunosuppressive cells, as mentioned previously. Consequently, aerobic glycolysis becomes the dominant metabolic preference in most cancer cells, as observed in murine and human cancers. Most importantly, if this model is true, we may find special metabolic targets that allow us to better exploit host anti‐tumor immunity by reorienting the TME from ‘metabolic harassment’ circumstance into a metabolically supportive microenvironment.