Cancer immunotherapy is shifting paradigms in cancer care. T cells are an indispensable component of an effective antitumor immunity and durable clinical responses. However, the complexity of the tumor microenvironment (TME), which consists of a wide range of cells that exert positive and negative effects on T cell function and survival, makes achieving robust and durable T cell responses difficult. Additionally, tumor biology, structural and architectural features, intratumoral nutrients and soluble factors, and metabolism impact the quality of the T cell response. We discuss the factors and interactions that modulate T cell function and survive in the TME that affect the overall quality of the antitumor immune response.
It is well-established that the immune system plays a major role in tumor control. Presentation of tumor-associated antigens (or lack thereof) can activate the immune system to recognize and eliminate dysfunctional and aberrant tumor cells. From decades of research, we appreciate the complexity of the immune machinery, which involves all aspects of the immune system from innate to adaptive responses, that acts as a defense mechanism against tumor growth. Dedicated and persistent efforts by many investigators have ushered in a new era for the clinical management of cancer and has brought the immune system to the forefront of clinical medicine for patients with cancer. Therapies targeting all components of the immune system, also known as immunotherapy and once significantly underappreciated, are now the standard of care for many patients with different types of tumors. The preponderance of evidence supports the CD8 T cell as the major subset of immune cells that is critically important for antitumor immunity and clinical efficacy in preventing disease progression and extending overall survival.
While the lymphoid compartments generally act as the sites for the initiation and development of antitumor immunity, the TME exerts significant influence on the quality of the immune response (Figure 1). In addition to the TME recruiting immune cells and supporting de novo T cell activation, it is generally an immunosuppressive environment. This suppressive environment poses challenges and limits the clinical efficacy of immunotherapy. In this review, we will discuss how T cells function and survive in the TME with an emphasis on CD8 T cells. We will also address challenges in maintaining an effective T cell response and opportunities to improve the clinical efficacy of antitumor immunity.
Antigens derived from tumor cells may be first encountered in the interstitial fluid that drains from lymphatic vessels and into lymph nodes. Proteins and peptides are taken up by immune scavengers and specialized antigen presenting cells (APCs) such as dendritic cells (DCs). These APCs process and present the proteins and peptides derived from the tumor cells and the TME on major histocompatibility complexes (MHC) I and II by direct priming or through a process known as cross-presentation (Huang et al., 1994; Wolkers et al., 2001; Hargadon et al., 2006; McDonnell et al., 2010). Although macrophages (Pfeifer et al., 1993; Rock et al., 1993), neutrophils (Tvinnereim et al., 2004), B cells (Ke and Kapp, 1996), and DCs (Chung et al., 2005; McDonnell et al., 2010) can cross-present, the process is most efficiently carried out by DCs expressing CD8a (den Haan et al., 2000; Belz et al., 2004; Allan et al., 2006) and CD103 (Bedoui et al., 2009; del Rio et al., 2007). Recognition of the MHC-peptide complexes by naïve T cells and engagement of costimulatory receptors by ligands on the activated APCs results in a signaling cascade that leads to T cell activation, clonal expansion supported by interleukin 2 (IL-2), and the differentiation into effector T cells (Smith-Garvin et al., 2009). Cancer cells that have metastasized into lymph nodes can also directly present antigens to T cells, but this interaction is generally suboptimal in activating naïve T cells due to the lack of co-stimulatory ligand expression by the cancer cells. The TME can also serve as a site for the initial activation of naïve T cells. In experiments using mice that lack lymph nodes and adoptive transfer of naïve T cells, investigators have shown that the TME can support de novo naïve T cell activation and differentiation into effector T cells (Yu et al., 2004; Thompson et al., 2010).
T cell activation is negatively regulated by inhibitory molecules, the most extensively studied being cytotoxic T lymphocyte-associated antigen 4 (CTLA-4). It is constitutively expressed on T cells and binds to the same ligands (CD80 and CD86) as CD28 (Linsley et al., 1991; Azuma et al., 1993; Freeman et al., 1993; Hathcock et al., 1993; Linsley et al., 1994). The affinity of CTLA-4 for CD80 and CD86 is significantly higher than that of CD28 and the general dogma is that CTLA-4 competes with CD28 to modulate the amplitude of TCR signaling during CD4 and CD8 T cell activation (Linsley et al., 1994; Egen and Allison, 2002; Riley et al., 2002; Schneider et al., 2002; Parry et al., 2005; Schneider et al., 2006). In binding to its ligands, CTLA-4 sequesters CD80 and CD86 and may actively remove the ligands from the surface of APCs further limiting CD28 engagement with the ligands (Qureshi et al., 2011). Aside from its inhibitory effects, CTLA-4 is a downstream transcriptional target of FOXP3 and its expression on regulatory T cells (T regs) enhances the suppressive activity of this subset of T cells (Fontenot et al., 2003; Hori et al., 2003; Wing et al., 2008; Peggs et al., 2009).
Recognition of antigens presented by tumor cells is critical for an effective T cell response. Tumors arise from normal cells in which genetic alterations lead to uncontrolled proliferation and transformation into malignant cells (Hanahan and Weinberg, 2000). As such antigens presented by non-viral tumors are of self-proteins or variations thereof. Identification of tumor antigens can be accomplished in several ways. They can be identified using serological analysis of recombinant tumor cDNA expression library (Kawakami et al., 1994) or synthetic peptide libraries (Townsend et al., 1986) by T cell clones. Alternatively, peptides can be isolated from MHC molecules and analyzed by mass spectrometry (Hunt et al., 1992). Over 170 antigenic peptides derived from 60 human antigens that are recognized by T cells have been identified (Renkvist et al., 2001). These antigens are generally classified into four groups: 1) cancer/testis or germline antigens, 2) tissue-specific differentiation proteins, 3) mutated self-proteins, and 4) those derived from post-translational modifications (PTMs) of proteins such as phosphopeptides (Boon and van der Bruggen, 1996; Zarling et al., 2000; Zarling et al., 2006; Mohammed et al., 2008; Depontieu et al., 2009; Zarling et al., 2014). Despite the fact that many tumor antigens are of unmutated self-proteins, CD8 (van der Bruggen et al., 1991; Cox et al., 1994; Van den Eynde et al., 1995) and CD4 (Topalian et al., 1994; Chaux et al., 1999; Pieper et al., 1999) T cell responses against antigens in all classes can be generated, sometimes spontaneously in humans (Bakker et al., 1994; Kawakami et al., 1994; Robbins et al., 1996).
The amount of tumor antigen present can modulate immune surveillance. Suboptimal activation of T cells may occur due to the lack of the necessary costimulatory molecules on tumor cells or because of induction of APCs that are ineffective at stimulating T cells (Gabrilovich et al., 1996; Menetrier-Caux et al., 1998; Munn et al., 2004; Hargadon et al., 2006). In situations where antigen is limited, highly avid CD8 T cells, which require only small amounts of antigen for activation, are needed to generate productive immune responses. Otherwise, those tumors go unnoticed by the immune system (Dudley et al., 1999; Zeh et al., 1999). Several studies in mice (Urban et al., 1986; Shankaran et al., 2001) and in humans (Slingluff et al., 2000; Yamshchikov et al., 2005) have shown that an ongoing immune response can place selective pressure that leads to the emergence of antigenic loss variants of tumors that escape immune recognition. This suggests that some tumor antigens are dispensable and not functionally important for tumor formation and/or metastasis. It also illuminates the need to identify and target functionally relevant tumor antigens for cancer therapy (Hirohashi et al., 2009). Antigens derived from proteins that regulate the malignant properties of tumors are necessary because such antigens are needed to maintain the malignant phenotype of tumors. Additionally targeting such antigens may counteract the emergence of antigenic loss variants of tumors as downregulation of the proteins would not be beneficial to the tumor. Systematic evaluation and careful selection of tumor antigens that are 1) differentially expressed in tumors and minimally expressed in normal cells, 2) functionally relevant to malignancy, and 3) antigenically distinct and immunogenic should provide a collection of peptides that will be more efficacious in clinical trials.
In the process of differentiation, T cells acquire certain properties and attributes that dictate their functional capacity. In the context of effector function, CD4 T cells can differentiate into several different subsets (Murphy and Stockinger, 2010; Zhu, 2018). TH1 that produce interferon gamma (IFNγ) and tumor necrosis factor (TNF) to activate macrophages and recruit leukocytes. TH2 subsets that secrete IL-4 and IL-13 and stimulate mast cells and eosinophils. Other subsets such as T regs hinder effective T cell response (Kim and Cantor, 2014). The differentiation of CD4 T cells into various subsets is heavily influenced by the cytokines present in the microenvironment. Polarization of CD4 T cells into the proinflammatory TH1 phenotype is mediated by IL-12 and IFNγ (Hsieh et al., 1993; Perez et al., 1995) while TH2 polarization is regulated by IL-2 and IL-4 (Le Gros et al., 1990; Swain et al., 1990). TH17 CD4 T cell differentiation is mediated by transforming growth factor beta (TGFβ), IL-6, IL-21 and/or IL-23 (Veldhoen et al., 2006; Zhou et al., 2007). T regs comprise natural and inducible T regs cells that require TGFβ and IL-2 for differentiation (Chen et al., 2003; Davidson et al., 2007; Takeuchi and Nishikawa, 2016). Both natural and inducible T regs can suppress effector function of CD4 and CD8 T cells through direct cellular interactions and through the release of cytokines and growth factors such as IL-10, IL-35, and TGFβ that negatively regulate effector function. For example, IL-10 has been shown to suppress TH17 effector cytokine production, while TGFβ dampens the effector response by regulating TH1 and TH17 differentiation and subsequent T cell tolerance (Li et al., 2007; Chaudhry et al., 2011). These cytokines have also been shown to block antigen presentation and interactions between conventional T cells and APCs in the TME. CD8 T cells differentiate into cytotoxic cells with the ability to directly kill tumor cell targets by producing cytokines and producing lytic granules including granzymes and perforin (Breart et al., 2008; Schietinger et al., 2013). Some subsets of T cells differentiate into circulatory effectors that move through tissues, lymphatics, and blood while others acquire f
