The development of cancer represents a failure of immune surveillance, because the immune system has the capacity to recognize tumor-associated antigens and develop specific T cell responses to those antigens. The ability to intervene and enhance the immune system to achieve a beneficial antitumor response remains an area of intense clinical research. Considerable progress has been made in expanding our knowledge of the targets for an immune response and about the full repertoire of cellular and humoral constituents involved in the generation of an effective antitumor response. Tumor cells display a variety of mechanisms by which they evade immune detection and destruction and render the immune response ineffective. With a more complete understanding of these escape mechanisms, clinical investigators are devising strategies to enhance the development of a robust immune response in the tumor-bearing host (active tumor immunity) or, alternatively, by the adoptive transfer of activated effector cells or tumor-specific antibodies into the tumor-bearing host (passive tumor immunity). Several strategies hold promise for substantial therapeutic benefit. Finally, antibodies that recognize tumor-associated antigens can aid in the pathologic diagnosis of cancer and facilitate the staging of cancer in vivo and the detection of recurrent cancer.
Overview of the Immune Response
The cellular and humoral arms of the immune response feature lymphocytes and antibodies, respectively, but extend beyond those elements. The cellular response consists of multiple components, including different subsets of lymphocytes (helper T cells, cytotoxic T cells, B cells, and natural killer [NK] cells) and antigen-presenting cells (APCs) that include blood monocytes, tissue macrophages, and dendritic cells. Different components of the immune system communicate with one another by direct cell-to-cell interaction via various adhesion molecules and receptors on the cell surface or through secreted chemicals (usually proteins) called cytokines that circulate and bind to specific receptors on effector cells. Cytokines can be either immune stimulatory or immune inhibitory; examples of cytokines include interleukins and interferons.
The cellular and soluble constituents of the immune response work in concert to recognize foreign (nonself) antigens (e.g., proteins, glycoproteins, and glycolipids) that are expressed by or secreted from tumor cells. The afferent limb of the immune response (immune recognition) subsequently triggers a highly specific response by the effector limb, in which other immune cells and their soluble products attack tumor cells bearing the same antigens recognized by the afferent limb. The success or failure of the afferent limb in recognizing tumor cells and of the effector limb in attacking them is influenced by a variety of mitigating factors.
The identification of the cellular constituents of the immune response and knowledge of their functions have been aided by the development of monoclonal antibodies created by immunization of mice against human immune cells. Each monoclonal antibody recognizes a single glycoprotein antigen that reflects the expression of a unique cell surface receptor. By convention, these receptors and the cells that express them have been assigned a cluster designation (CD) number (e.g., CD3, CD4, CD8).
Host Immune Response to Cancer
The existence of a host immune response to cancer is supported by the following observations:
1. Histologic analysis of excised human and animal tumors has demonstrated varying degrees of immune cellular infiltration (lymphocytes and APCs), which suggests the recruitment of these cells in response to neoplastic proliferation. When analyzed, the activity of these tumor-infiltrating lymphocytes is often found to be specific for the autologous tumor, with little to no activity against unrelated tumor targets. Accompanying this mononuclear infiltration of tumors is the elaboration of various cytokines that are associated with an ongoing immune response. T cells that exhibit specific reactivity against autologous tumor have been isolated from patients with melanoma, breast cancer, ovarian cancer, and colorectal cancer.1
2. Long-term remissions are induced in small numbers of patients who receive some form of immunologic therapy for their advanced cancers.
3. There are documented (albeit infrequent) reports of spontaneous remissions in patients with melanoma and renal cell carcinoma that are believed to be immune mediated.
The possible existence of immune surveillance mechanisms that prevent the development of cancer is supported by the finding that immunodeficient individuals and patients undergoing long-term treatment with immunosuppressive drugs are at greater risk for cancer than the general population. Experimental evidence that directly links immune mechanisms to the defense against cancer comes from classic experiments in which immunized mice rejected chemically induced syngeneic tumors.2 Naive syngeneic mice (i.e., mice not previously exposed to tumor) were protected against tumor growth by immunization with killed tumor cells administered before challenge with viable tumor cells [see Figure 1]. The specificity of this immune response was demonstrated by the lack of protective effect of prior immunization with killed cells from an unrelated tumor. The important role of lymphoid cells in this immune response was demonstrated by the protective effect of spleen cells from a tumor-bearing mouse administered to a naive host (adoptive transfer of immune cells) before tumor challenge with live tumor cells. In these experiments, the cyto-toxic T cell population (CD8+ T cells) within the spleens of tumor-bearing mice demonstrated unique protective activity upon transfer. Transfer of helper T cells (CD4+ T cells), B cells, or serum-derived antibodies did not confer protection in this model. The results of these classic experiments, in which syn-geneic tumors could be rejected in naive hosts receiving prior immunization or by adoptive transfer of immune cells, support the existence of tumor-associated antigens in animal models and form the basis of much of the experimental work in human tumor immunology.
Figure 1 Immune-mediated rejection of transplanted tumors. Whereas the original host and naive syngeneic mice immunized with killed cells from the original tumor do not experience tumor growth after transplantation of tumor cells, nonimmunized naive syngeneic mice and mice immunized with killed cells from an unrelated tumor do experience tumor growth.
Targets for Immune Response to Cancer
The ability to immunize naive animals against cancer led to the concept of tumor-specific or tumor-associated antigens that serve as the target of an inducible effector response. The development of strategies to enhance this immune response against cancer has depended upon the further identification and characterization of these tumor-related antigens.3 The work of many investigators over the past decade has confirmed the existence of at least five classes of potential tumor-associated antigens that are recognized by and stimulate T cells: oncoviral proteins, tumor-associated antigens, mutated or overexpressed oncogene or tumor-suppressor gene antigens, differentiation (or lineage-specific) antigens, and abnormal posttranslational modification of self-proteins.
Oncoviral proteins
Oncoviral proteins represent cellular antigens encoded by the genomes of oncogenic viruses. Examples of oncoviral proteins include the human papillomavirus E6 (HPV-E6) and HPV-E7 antigens, found in cervical carcinoma; and the Ep-stein-Barr virus (EBV) EBNA-1 antigen, found in Burkitt lym-phoma and nasopharyngeal carcinoma.
Tumor-associated antigens
Tumor-associated antigens (also referred to as tumor-testis antigens) are proteins that are normally expressed during the course of embryonic development and in the human adult testis; they become abnormally expressed by the cancer cells of adult individuals. Examples of tumor-associated antigens include the MAGE-1 and MAGE-3 proteins, which are expressed by a variety of tumor cells, including melanoma, glioma, and breast carcinoma.
Mutated or overexpressed oncogene or tumor-suppressor gene antigens
Mutated or overexpressed oncogene or tumor-suppressor gene antigens represent the protein products of mutated or overexpressed cellular oncogenes or tumor-suppressor genes found in a variety of tumor cells. Examples of this class of tumor-associated antigens are p21ras (expressed in a number of carcinomas), the p210 product of the bcr-abl translocation found in chronic myelogenous leukemia, the cyclin-dependent kinase 4 (CdK4) and |-catenin proteins found in melanoma, the HER-2/neu protein found in breast carcinoma and other cancers, the caspase-8 protein expressed in certain squamous cell carcinomas, and the p53 tumor-suppressor gene product found in multiple tumors.
Differentiation antigens
Differentiation (or lineage-specific) antigens are proteins that are normally expressed in a tissue-specific fashion by normal cells but are coexpressed by tumor cells derived from the normal host tissue. Examples of differentiation antigens are tyrosinase, GP100, MART-1 antigens coexpressed by normal melanocytes and melanoma cells, and cell membrane immunoglobulin in a specific B cell clone.
Abnormal posttranslational modification of self-proteins
Self-proteins that have undergone abnormal posttranslation-al modification represent mutated forms of normal protein products; these modified self-proteins give rise to unique tumor-associated carbohydrate epitopes. An example of this class of tumor antigens is the MUC-1 antigen (featuring underglyco-sylated mucin), which is expressed by breast and pancreatic carcinomas.
In addition to the antigens that may be the targets of a T cell-directed immune response, there are other antigens that may be recognized by T cells as well as by antibodies resulting from deliberate immunization against these antigens. These targets of an antibody-mediated immune response include (1) tissue-specific differentiation antigens that, like the lineage-specific antigens, represent protein products shared between tumor cells and the tissues from which they are derived (e.g., CD20 and the surface immunoglobulin [Ig] idiotype, which are expressed by B cell lymphomas; and the prostate-specific antigen [PSA], expressed by prostate carcinoma cells); (2) oncofetal antigens that, like the tumor-associated antigens, represent self-proteins normally expressed during embryonic development but that are found on tumor cells (e.g., the carcinoembry-onic antigen [CEA], expressed by multiple carcinomas, and a-fetoprotein [AFP], expressed by hepatocellular carcinoma and germ cell tumors); and (3) altered or overexpressed glycolipid and glycoprotein antigens. These include the gangliosides GM2 and GD2 found on melanomas and neuroblastomas; and the mucin antigens CA125, CA19-9, and MUC-1, which are predominantly expressed by ovarian, pancreatic, and breast carcinomas, respectively.
Effector Mechanisms in the Anticancer Immune Response
The immune response against cancer includes both the cellular response and the humoral response.
Cellular immune response
The cellular effector response involves activity by five major cellular constituents. (1) CD8+ cytotoxic T cells recognize tumor-associated antigens that are presented in association with major histocompatibility (MHC) class I molecules [see Figure 2]. (2) CD4+ helper T cells recognize tumor-associated antigens that are presented in association with MHC class II molecules leading to cy-tokine release (help signals) for the generation and activation of cytotoxic T cells. (3) NK cells may kill tumor cells in a non-MHC-dependent fashion. Rather than binding directly to tumor-associated antigens, NK cells may be targeted to antibody-coated tumor cells via recognition and attachment to the Fc portion of the antibody. (4) Mononuclear phagocytes (macrophages), like NK cells, may be targeted to antibody-coated tumor cells, where their cytotoxic activity may depend upon the release of destructive proteases, cytokines (e.g., tumor necrosis factor [TNF]), and reactive oxidative intermediates (O2~). (5) Dendritic cells (DCs) are important APCs that can present antigen to both CD4+ and CD8+ T cells and are able to stimulate a naive T cell response (i.e., stimulation of a T cell response to an antigen to which the T cell has not been previously exposed).
The recognition of tumor target cells or tumor APCs depends on a direct binding interaction mediated by several receptor species [see Figure 2].4 The specificity of binding is conferred by the T cell receptor (TCR), in association with CD8 or CD4 surface glycoproteins, which recognizes antigen as presented by either class I (recognized by CD8+ T cells) or class II (recognized by CD4+ T cells) MHC antigens. Antigen recognition and concomitant T cell activation also depend on the binding of pairs of receptor and costimulatory molecules, including leukocyte function-associated antigen-1/intercellular adhesion molecule-1, CD28/B7, and CD2/leukocyte function-associated antigen-3. Helper T cell activation results in secretion of interleukin-2 (IL-2), which provides a necessary help signal to cytotoxic T cells. T cell binding to antigen on a target cell that does not express the requisite accessory molecules can lead to T cell apoptosis or anergy, thus preventing an immune response.
Figure 2 T cell interaction with a dendritic antigen-presenting cell (APC). Helper T cells recognize MHC class II-associated antigens. Cytotoxic T cells recognize MHC class I-associated antigens. Release of interleukin-2 (IL-2) by the helper T cell provides a so-called help signal to the cytotoxic T cells. (ICAM—intercellular adhesion molecule; LFA—leukocyte function-associated antigen)
After effective membrane binding between cytotoxic T cells and tumor target cells, actual tumor cell cytolysis is accomplished by one of two general mechanisms5: (1) a perforin-de-pendent mechanism, in which killer cell enzymes (granzymes) are released from cytotoxic T cells and gain access to target cells via perforin pores, endocytosis, or granzyme receptor-mediated uptake, leading to target cell membrane damage with subsequent necrosis and apoptosis; or (2) a newly recognized mechanism of target cell cytolysis that depends on the recognition of the membrane protein Fas (expressed by certain tumor cells) by the Fas ligand expressed by killer T cells. The binding of Fas and its ligand leads to the generation of proapoptotic death signals to the target cell nucleus, resulting in DNA fragmentation and apoptosis.
Humoral immune response
The recognition of tumor-associated antigens by CD4+ helper T cells may also elicit a B cell humoral response leading to the development of antibodies that recognize tumor-associated antigens. Unfortunately, despite this mechanism of antitumor immunity, the results of animal experiments suggest that naturally occurring antibodies play little role in effective antitumor response. However, the administration of manufactured antibodies targeted to tumor-associated antigens may be quite useful in eliciting an effective antitumor response. Specifically, these antibodies may facilitate antibody-dependent cellular cy-totoxicity (ADCC) by NK cells and macrophages, or they may be employed to kill tumor cells by direct mechanisms (e.g., induction of apoptosis or fixation of complement) or indirect mechanisms. In the latter case, antibodies may be conjugated (chemically linked) with cellular toxins (e.g., ricin or diphtheria toxin) or gamma radiation-emitting radionuclides; binding of the antibodies to tumor cells results in the death of those cells from in-ternalization of the toxin or exposure to ionizing radiation.
Evading Antitumor Immune Responses
Despite the fact that the immune system has the potential to recognize tumor-associated antigens and to marshall an effective cytotoxic response, this response often fails to prevent the local growth and distant spread of cancers. A better understanding of the mechanisms by which tumor cells can evade a host response is emerging. Tumor cells may evade detection by multiple mechanisms, all of which are probably operational to some extent.6 Tumor cells may express only low levels of tumor-associated antigens in cryptic sites that are covered by gly-cocalyx molecules, or they may undergo antigenic modulation by shedding tumor antigens. Tumor cells have been shown to express low levels or nondetectable amounts of MHC molecules or appropriate costimulatory molecules (e.g., B7) that are required for an effective immune response. In other circumstances, tumor cells that can stimulate an effective immune response may have already been eliminated, leaving only those cells that can evade an immune response. Finally, tumor cells have been found to secrete various soluble factors (e.g., transforming growth factor-| [TGF-|], IL-10, and Fas ligand) that have been found to be immunosuppressive.
Enhancement of Active Host Immune Responses
Two major approaches have been undertaken to counter the mechanisms by which tumor cells evade immune recognition: (1) more effective stimulation of a host immune response in tumor-bearing patients (active immunity) and (2) the adoptive transfer of cellular or humoral immunity to the tumor-bearing host (passive immunity).
