Cancer—avoiding immune detection

Introduction

We have come to understand the many diseases we call cancer by their genetic and epigenetic instability and the relationship the growing tumor has with its surrounding cellular microenvironment and with the host’s extensive immune system. Genetic instability in the tumor leads to the development of tumor-specific neoantigenic peptides and, on occasion, ectopic overexpression of genes not normally expressed in the tissue in which the tumor has developed. These non-self-antigenic peptides have an opportunity for presentation by major histocompatibility (MHC) class I molecules expressed on the tumor cells and for cross-presentation by MHC class II molecules expressed by classic antigen-presenting cells. The peptide-loaded MHC molecules present the antigenic epitopes to the T-cell population of the adaptive immune system and to natural killer (NK) cells of the innate immune system. The level of antitumor response by both innate and adaptive compartments of the immune system depends on numerous interactions between the tumor cells and various immune cells and with inflammatory cells of the tumor microenvironment (TME). Antitumor immunity occurs at multiple levels within the host, both locally at the tumor site and peripherally in the immune system. The patient’s antitumor response must react to a changing tumor burden and to therapeutic interventions requiring constant efforts of response renewal.

The intimate relationship between the tumor and the immune system occurs over the entire course of disease, from initiation through the progressive stages of growth, local invasion, and metastasis to distant sites. For tumors to survive and progress along this path requires that the tumor develop mechanisms that actively inhibit antitumor cell immune responses. ^, Much like the complexities of the antitumor immune response, there are multiple, equally complex mechanisms that function to enable the tumor cells to resist immune destruction. ^, Though much has been learned regarding the development of immune resistance to avoid immune tumor cell destruction, developing effective immune therapies that overcome this resistance remains a tremendous challenge. This chapter will address the intimate relationship between the tumor as it progresses and interacts with the patient’s immune system with a focus on the mechanisms that enable the tumor to evade immune destruction.

The presence of tumor immune suppression

The concept of the three “Es” of cancer—elimination, equilibrium, and escape—was first introduced in 2004 as a way of describing the relationship between the initiation and development of cancer and the host’s immune system. In the elimination phase, potentially malignant cells arising during the early initiation stage of cancer development are detected as nonself and eliminated as a result of their MHC class I presentation of neoantigens. They are recognized as foreign by the NK cells of the host’s innate immune system and by cytotoxic T lymphocytes (CTLs) of the adaptive immune response. This stage has been equated with the historic concept of immune surveillance. If the tumor cells escape elimination, they may enter the second stage, termed the equilibrium phase . In the equilibrium phase, surviving tumor cells, not eliminated by the immune system, are instead held in check by the immune system without being destroyed but without progression.

Most cancer clinicians recognize a second clinical component of this equilibrium phase. For example, it is not uncommon for aggressively treated newly diagnosed cancer to enter a period of complete response (CR) only to recur many years later. Clearly, these CR patients have undetected micrometastases after treatment and remain in a state of equilibrium or at least controlled slow progression. Clinically, we refer to this phase of cancer as disease-free survival (DFS) or in some cases stable disease (SD). During this phase of DFS or SD, the relationship between the tumor and the host’s immune system is one of controlled equilibrium. Finally, should the tumor progress to the third escape stage, its capacity to be immunogenic has been significantly altered and on balance it is no longer completely susceptible to immune cell destruction by the host.

As the tumor progresses through these three stages, it is said to undergo immunoediting. Immunoediting has been described as the complex process(s) by which the host immune system can either inhibit tumor growth or be suppressed/edited in ways that promote tumor progression. There are multiple targets and intracellular pathways that provide mechanisms by which tumor cells can escape both the innate and adaptive immune systems. These escape mechanisms will be described individually but almost certainly occur in a combinatory fashion for successful tumor escape ( Box 8.1 ).

Downregulation of or actual loss of MHC class I and II expression

As presented in Chapter 5 , the MHC class I molecules are the mechanism by which nonself tumor neoantigenic peptides are presented on the surface of tumor cells to the CD8 ⁺ T cells of the adaptive immune system. Though MHC class I molecules are essential for the presentation of foreign antigens such as infectious viral peptides or cancer antigens, conditions in which the expression of the MHC molecules is structurally defective or even absent are not uncommon in the genetically altered cancer cell. Importantly, defective or absent tumor cell MHC molecules does not interfere with tumor cell survival. It is recognized that tumor cell surface MHC class I neoantigen presentation is generally not a very efficient method of tumor antigen T-cell stimulation. In fact, cross-antigen presentation by MHC class II molecules expressed on antigen-presenting cells (APCs) is generally the more effective mechanism for stimulating the CD8 ⁺ T cells and other T cells of the immune system. Tumor antigen cross-presentation requires the antigenic peptides produced in the tumor cell to be released and taken up by APCs such as the dendritic cells (DCs) and presented to T cells via the MHC class II molecules rather than MHC class I.

Nevertheless, any impairment in tumor cells to the normal expression of genes involved in the processing of potential tumor antigenic proteins and in their presentation to the T-cell receptor (TCR) of the CD8 ⁺ T cells will have a significant negative effect on the host’s antitumor response. For example, let us begin with the dysregulation of the MHC chromosome 6p region’s human leukocyte antigen (HLA) genes. The class I region of the polymorphic HLA locus includes, among others, the highly polymorphic HLA-A , HLA-B , and HLA-C genes ( Fig. 8.1 ). These genes generate the 45 kDa classic α-chain (362–366 amino acids) of the class I molecule. The MHC class I β ₂ microglobulin chain of 12 kDa is coded for by a gene on chromosome 15, a site different from the HLA region located on chromosome 6p21.3. Two of the three class I α domains, α ₁ and α ₂ , create the MHC domain containing the neoantigen peptide-binding cleft. The class II MHC molecule has a similar overall physical structure but is the product of two HLA genes containing one α ₁ and one β ₁ domain that fold to generate the groove that holds the antigenic peptide (for an in-depth review, see Monos and Winchester ).

It is easy to understand that within the genomically unstable tumor cell there is ample opportunity for somatic mutations to occur in the HLA genes, resulting in aberrant expression or actual silencing of HLA class I and II molecules. An example of HLA-I gene dysregulation and dysregulation of genes involved in antigen processing and presentation can be found in studies at the University of Cologne reported by Thelen and colleagues. They determined the expression level of 24 genes known to be involved in antigen presentation. The genes were selected based on their known downregulation in at least a portion of patients for one or more of the tumor types selected.7, They found that the mean expression levels for 7 of 24 genes were increased greater than twofold in the tumor microenvironment (TME) of tumor samples compared with normal tissues. They selected a twofold change as a cutoff in order to determine evidence for impaired gene expression. They found TME downregulation to occur in at least one patient of the following genes β2M , HLA-A , HLA-B , HLA-C , ERAP1 , ERAP2 , PDIA2 , NLRC5 , UBB , UBC , and LMP10 . This downregulation was reported in >10% of analyzed tumor samples and occurred in more than 3 (>12.5%) of the included genes in 45 of 142 patients (32%). No alterations in expression level were found in 23% of patients. They noted that expression patterns of the 24 genes across tumor types were quite heterogeneous. Though this represents an early effort to show HLA-I gene dysregulation as an important aspect of immunoediting during cancer progression, it clearly emphasizes both the complexity and the critical relationship of HLA expression to immunotherapy.

The loss of both copies of MHC class I heavy chain genes or of β2M will, of course, eliminate the expression of a major component of the antigen presentation molecule and has been observed in cancers such as melanoma with the loss of HLA-A2 expression. Many cancers have also been found to have a loss of only one copy of the heavy chain or of only β2M (heterozygous). ^, MHC class I heavy chains are codominantly expressed but the genetic alteration at the level of the heavy chain gene may be heterozygous. Montesion and colleagues surveyed 59 different cancer types and observed that in 17% there was a significant loss of MHC I expression secondary to a loss of heterozygosity ^, (for an in-depth review, see Dhatchinamoorthy et al. ).

There are a few facts to keep in mind regarding the function of MHC class I molecules. First, as part of their normal functions, all cells generate essential proteins and, as a result, present short peptides derived from those proteins on their cell surface as part of the peptide–MHC (pMHC) complex. When the peptide is nonself as in cancer, it is recognized as such by the immune system—innate or adaptive—for the cell to be destroyed. In all cells, the expression of the MHC complex on the cell surface is governed by interferons, especially interferon gamma (INF-γ) and the IFN signaling pathway. Stimulation with INF-γ increases the expression of MHC class I and therefore of detectable MHC class I molecules on the surface of tumor cells. The loss of MHC in cancer or the suppression of MHC expression correlates significantly with the progression of the cancer, indicating the significance of this mechanism in tumor resistance and poor prognosis ( Fig. 8.2 ).

MHC low expression metastases are also known to respond poorly to immunotherapy. Many genes coding for proteins intimately involved as factors in the ubiquitin–proteasome production of the neoantigenic peptides can also be subject to genetic alteration in the tumor cell. Such genetic alterations in the proteins comprising the processing machinery of the proteasome can alter the neoantigenic peptide, causing a loss of its antigenic epitope or simply a loss of the peptide altogether. Such alteration can also occur as the potential antigenic peptides leave the proteasome and enter the endoplasmic reticulum where they are loaded on the MHC class I molecules (see also Chapter 5 , Fig. 5.2 ).

Many of the genes involved in the preparation of immunogenic peptides have similar transcription and translation governance, and many of the genes for these processing factors are actually chromosomally located in association with the HLA heavy chain genes on chromosome 6p. Thus genetic alterations affecting these genes may manifest in low expression of MHC molecules. Studies show that within the tumor itself there exists significant tumor cell heterogeneity in terms of MHC expression and therefore the robustness of T-cell antigen presentation. This heterogeneity in levels of pMHC cell expression is also found in relationship to the various metastases of a given tumor both as to the metastatic site (lung, liver, brain, etc.) and within the actual individual metastases. Low or absence of detectable MHC class I has been reported to varying degrees in patients for many of the most common human malignancies including breast, prostate, colorectal, melanoma, head and neck, and non-small cell lung cancer (NSCLC), for example. It also appears that the level of expression may vary over the course of disease and in association with therapy. Finally, a loss of MHC class I has been reported to be associated with resistance to anti-programmed cell death protein-1 (PD-1) immunotherapy. ^,

It is easy to understand how in a genomically unstable tumor cell there exists ample opportunity (see Chapter 5 ) for genetic variants to cause defects that prevent optimal MHC molecule formation and optimal tumor antigen generation suitable for pMHC class I presentation. For example, any loss of the key molecules such as transporter for antigen processing (TAP), tapasin, or endoplasmic reticulum aminopeptidases (ERAP1 and ERAP2) will cause a loss of pMHC expression and presentation on the cell surface. Tumor cells that sustain homozygous deletion or frameshift variants of the HLA region of chromosome 6p will result in low expression or complete loss of MHC class I antigenic peptide presentation. Some alterations are unique to a specific cancer, whereas others may be found in numerous patients and in a variety of cancer types ( Box 8.2 ).

A number of clinical correlations with low or absent MHC class I are important to note. For example, the presence of low MHC class I expression in breast cancer is associated with tumors that contain fewer tumor-infiltrating lymphocytes (TILs) than the tumors demonstrating a high degree of MHC class I expression. Second, though the low expression of pMHC molecules or even a total absence prevents activation of the T cells and triggering of adaptive immune response against the tumor, it does not have a negative effect on cancer cell viability or tumor cell proliferation. The low expression or loss of MHC, as expected, does strongly indicate a worse cancer prognosis. Third, and perhaps even more important, is the association of low or absent MHC class I expression with development of resistance to anti-PD-1 and anti-PD-L1 immunotherapy. A similar negative correlation is associated with MHC class I loss in other epithelial tumors such as melanoma, cancers of the colon and rectum, and head and neck cancer ^, (for an in-depth review, see Gettinger et al. ).

MHC class I molecules are more densely presented on the cells of the hematologic lineage. MHC class II molecules, in contrast, are predominantly found on the surface of antigen-presenting cells such as dendritic cells (DCs), B cells, and classic macrophages. Their major role is the presentation of antigenic peptides such as from viral infections and cancer cells produced in other cells from which these peptides are released. The released antigenic peptides are internalized by the APCs for further processing and for loading onto the MHC class II molecules for surface presentation. The antigenic peptide–MHC class II molecule is primarily interactive with the CD4 ⁺ T cells. CD4 ⁺ T cells function to support and enhance the CD8 ⁺ CTL of the adaptive immune response and to generate effective memory T cells.

In addition to a direct interaction of the antigenic peptide bearing tumor-specific MHC class II molecule (tsMHC class II) with the TCR of CD4 ⁺ T cells, tsMHC class II molecules have been implicated in the stimulation directly or indirectly of other T-cell subsets. Secretion of several known activating cytokines, such as IFN-γ, can be triggered by tsMHC II interaction with CD4 ⁺ T helper (Th) cells. On the other hand, the stimulation of CD4 ⁺ T cells may result in activation of the subclass of regulatory T cells (Tregs), which function to suppress the immune response favoring cancer progression.

Some tumor cells during the cancer’s progression have been shown by immunofluorescence to actually present both MHC class II and MHC class I molecules. Tumor-specific MHC class II expression may increase the immune response to the tumor. In fact, this finding has been associated with improved prognosis in response to immunotherapy. The binding groove of the MHC class II molecules can bind longer peptide chains than can be held in the groove of the MHC class I molecule and also accept peptide chains having more branched side chains that are more immunostimulatory when interacting with the TCR.

The MHC class II molecule is heterodimeric in contrast to MHC class I, and studies have demonstrated that expression of the class I and class II molecules is independently regulated in cancer. This independent regulation may in part account for observed variations in immunotherapy responses. For example, MHC class I molecular assembly and expression are not normally controlled except in T cells. Its expression, however, can be induced and enhanced by IFN-γ and by activation of the NF-κB signaling pathway. The expression of MHC class II, in contrast, is tightly controlled by the “master transcription control factor,” a protein (class II major histocompatibility complex transactivator) encoded by the CIITA gene located on 6p.

There are a number of intracellular alterations in signaling pathways that can contribute to the suppression of MHC class II expression. For example, JAK/STAT cell signaling is necessary for upregulation of MHC class II molecule expression and, as such, mutations occurring in JAK result in suppressed IFN signaling and suppressed expression of MHC class II molecules. Also, with activation of RAS/MAPK signaling in breast cancer, there is suppression of MHC class II expression. Clearly, from clinical studies the extent to which MHC class II expression is suppressed is a strong biomarker for a poor response to immunotherapy. ^,

In summary, increasing evidence from clinical studies of solid tumors suggests the following: (1) the presence of a significant increase in tsMHC class II expression is associated with an improved disease-free survival (DFS) and overall survival (OS) perhaps as a result of increased antigenic recognition by the adaptive immune system, (2) significant surface expression of tsMHC class II is required for CD4 ⁺ T-cell (including T-cell subsets Th1 and Treg) activation, and (3) the increased presence of tsMHC class II correlates with improved response to anti-PD-1 and anti-PD-L1 immunotherapy. The apparent significant role that the neoantigen-loaded MHC class II molecule plays in the immune control of the tumor’s progression highlights why it is extremely important when anticipating the use of immunotherapy to determine whether the tumor is showing a low or possible absence of tsMHC class II expression or an intermediate or high level of expression, as well as upon which cell types the expression is occurring. This information is increasingly predictive of likely response.

Immune checkpoints and checkpoint ligands

Tumors, as noted, not only escape recognition and destruction by the host’s immune system but can actively block normal T cell–directed antitumor activity. One such method is by modulating immune inhibitory signals termed immune checkpoints . The immune checkpoints and their biologic pathways represent one of the most exciting anticancer therapeutic opportunities since their introduction in 2011 and, as such, one of the most intensely studied. In normal health, these immune checkpoints operate to prevent immunity to self (autoimmune diseases) and to self-proteins when the host is responding to infectious pathogens.

Cytotoxic T lymphocyte–associated antigen 4 (CTLA-4) was the first immune checkpoint receptor to be studied clinically. CTLA-4 functions primarily to depress the early T-cell response to tumor neoantigens by countering the TCR costimulatory receptor CD28. ^, Antigen recognition by the TCRs of T cells and the subsequent activation of CD8 ⁺ CTLs increases the expression of the CTLA-4 receptor on the T cells. TCR activation also suppresses the numbers of CD4 ⁺ T helper cells and their activity, along with an increase in numbers of CD4 ⁺ Tregs, which have an immunosuppressive effect. Taken together, these multiple actions have a suppressive effect on the antitumor immune response and favor tumor progression.

Another important checkpoint inhibitor is the programmed cell death protein 1 (PD-1). PD-1 receptor protein is expressed on T cells with the function of limiting T-cell activity in peripheral tissues at the time of pathogenic inflammation and of cancer initiation and development. In cancer, the ligand to PD-1 (PD-L1) is expressed in tumor cells and other cells of the TME. A high or enhanced expression of the ligand PD-L1 by the tumor causes increased stimulation of the T-cell PD-1 receptor, resulting in suppression of these T cells and their functions; in doing so, it promotes tumor growth. PD-1 activation generates suppression of two important antitumor activities. It triggers the induction of apoptosis to antigen-specific T cells and blocks the apoptosis of Tregs. PD-1 is a member of the immunoglobulin gene superfamily and is expressed as a transmembrane protein with a C-end cytoplasmic tail and an N terminus IgV-like domain. PD-1 is found to be highly expressed on TCR-activated tumor-specific T lymphocytes and is also found expressed to some extent on NK cells, B cells, dendritic cells, macrophages, and monocytes.

PD-1, in conditions where the immune T cells are challenged with a pathogen and inflammation, functions in the antigen-reactive T cell to prevent ineffective or harmful immune responses to self-tissues, helping to maintain a condition of immune self-tolerance. When expressed on immune cells that are being activated by cancer antigens, PD-1 has a very different role. When faced with cancer, PD-1 acts to interfere with the antitumor response by downregulating the T-cell responses, causing a decrease in T-cell proliferation and clonal expansion, a decrease in secretion of cytokines, and an increase in T-cell demise ( Fig. 8.3 ). Other described checkpoints include YIM-3, LAG-3, and VISTA. Targeting the immune checkpoints with blocking humanized monoclonal antibodies has emerged as a strategy to generate beneficial antitumor immune responses (for an in-depth review, see Han et al. ).

The ligand that binds to the PD-1 receptor, PD-L1, is expressed by cancer cells and, as such, plays the role of enabling the tumor cells to escape immune attack. PD-L1 is a small (33 kDa) B7 transmembrane glycoprotein with extracellular domains of Ig- and IgC. PD-L1 is also found to be expressed by DCs, macrophages, and B cells, as well as some epithelial cells, during inflammation. IFN-γ is known to increase the expression of PD-L1 in most cancers, and the pathway involved in PD-L1 expression also appears involved in several additional protumor stimulatory actions such as tumor cell proliferation, epithelial-to-mesenchymal transition, and expansion of the tumor stem cell population in some tumors. A presentation of all of the intrinsic cell signaling pathways involved in interacting to provide influence/control over the PD-L1/PD-1 axis in cancer is beyond the scope of this chapter but it has been well reviewed by others. Of increasing interest is how the genetic variants that define and alter the tumor microenvironment (TME) over time also manifest themselves in the up- and downregulation of various microRNAs that directly affect these signaling pathways. Undoubtedly, this will be a critical map to develop to optimally guide the integration of future immunotherapies.