The role of the immune system in the pathogenesis of cancer

The Role of the Immune System in the Pathogenesis of Cancer Janet W. Appelbaum

T

HE IMMUNE system is a highly integrated complex network of specialized cells and organs that has evolved to defend the host against foreign invaders (nonself). In many ways, cancer can be thought of as a foreign invader. The malignant transformation leading to cancer results in a heritable change in the DNA of the cell, and theoretically, this change should allow immune recognition of the cell as “nonself’ ’ promoting eradication by immunocytes. Although it has long been a dream of clinicians to use the immune system intelligently to treat cancer, effective immunotherapy remains elusive. Nonetheless, there is strong evidence for a role of the immune system in cancer. Documented, albeit rare, cases of spontaneous remission of renal cell cancer, lymphoma, and melanoma have been credited to immune effector mechanisms. The association of certain malignancies with congenital or acquired immunodeficiency diseases and the bimodal distribution of cancer in the very young and the very old suggests that an immature or debilitated immune system predisposes to malignancy. Recent advances in molecular biology, hybridoma technology, and genetic engineering have led to an increased appreciation and clarification of immune mechanisms and have continued the interest in modulating the immune response to promote tumor recognition and eradication. ’ This article will review the components of the immune response, the malignancies associated with immunocompromise, tumor antigens, immune effector mechanisms, and evasion of specific tumor immunity. OVERVIEW OF THE IMMUNE RESPONSE Definition

of Innate Versus Adaptive

Immunity

A broad overview of the immune response is difficult to present in a few paragraphs because of the intricacies and interrelationships that characterize the immune system. For a more detailed discussion, the reader is referred elsewhere.2-4 The immune response can be divided functionally into innate and adaptive components. Innate immunity is relatively nonspecific and it has the primary objective of prevention of infection. The innate sysSeminarS

in Oncology

Nursing,

Vol 8, No 1 (February),

1992:

pp 51-62

tern includes the physical barriers of the skin and mucous membranes and the chemical barriers found in the cytolytic agents located in the respiratory, genitourinary, and gastrointestinal tracts. The cellular components of the innate system consist of monocytes, macrophages, eosinophils, polymorphonuclear neutrophils (PMNs), and morphologically appearing large granular lymphocytes (LGLs) called natural killer (NK) cells. These leukocytes do not differentiate into long-lived memory cells and without the capacity for immunologic memory (anamnesis), resistance is not improved with repeated exposure to infectious agents. An invading organism must first transgress the skin or mucous membranes and then is confronted with soluble proteins such as lysozyme, acute-phase proteins, and complement which results in either direct cytolysis or promotes phagocytosis. Antiviral activity is initiated by NK cells activated by interferons. The value of an intact innate immune system is clearly appreciated in the setting of treatment-induced or disease-induced granulocytopenia or mucosal damage due to drugs or dehydration commonly seen in cancer patients. If the innate immune system is unable to stop the invader, the adaptive immune system comes into play. Monocytes , macrophages , thymus-derived lymphocytes (T cells), and bone marrow-derived lymphocytes (B cells) are the critical cellular components of the adaptive immune system. In contrast to the relative nonspecificity of the innate component, the adaptive component is highly antigen specific and has the capacity for anamnesis due to the differentiation of a subset of activated T and B cells into long-lived memory cells. Resistance or acquired immunity is evident on subsequent exposure to the same invading organism.

From the NeoRx Corporation, Seattle, WA. Janet W. Appelbaum, ARNP, MS: Medical Research Associate, NeoRx Corporation. Address reprint requests to Janet W. Appelbaum. ARNP, MS, NeoRx Corporation, 410 West Harrison, Seattle, WA 98119. Copyright 0 1992 by W.B. Saunders Company 0749-2081/92/0801-0007$5.00/O

51

52

Recognition of Antigen Antigens are substances capable of evoking an immune response. Proteins are the most ubiquitous, diverse, and well-characterized antigens. Not all antigens are equally immunogenic. Immunogenicity is not solely related to the structure of the antigen but depends as well on the organism being immunized. The route or mode of immunization, the dose of antigen, the degree of foreignness of the antigen, and its molecular size and complexity all contribute to immunogenicity.5 For a full-blown immune response to develop, an antigen must be able to be ingested by a macrophage, the primary antigen-presenting cell (APC), and subsequently must be displayed by the macrophage for recognition by the B and/or T lymphocyte. During the development of an immune response, T cells usually do not bind directly with intact protein but require accessory cells (APCs) to process and present antigen for recognition. Although the process is not fully clarified, the APC has the uncanny ability to internalize soluble antigen by pinocytosis, process or degrade it into a small number of its constituent peptides, bind these peptides to its own major histocompatibility complex (MHC), and express the antigen-MHC on its cell surface.6 B cells and other cells, especially those found in the skin and spleen (eg, Langerhans’ cells, follicular dendritic, and interdigitating cells) can function as APCS.~ T cells recognize a specific antigen-MHC by using their own membrane-bound T-cell receptor (TCR) complex. TCR is a member of the immunoglobulin superfamily and is composed of constant and variable domains homologous to immunoglobulin. The two variable regions form a pocket or cleft of antigen binding and the constant region anchors the TCR to the cell membrane. The constant region extends through the membrane into the cytoplasm. When the TCR binds to its specific antigen-MHC complex, intracellular activation signals are thought to be transmitted by the constant region of the TCR. Signals transmitted through the receptor are necessary for initiating effector functions such as the secretion of interleukins or cytokines .8 Helper lymphocytes, designated T4 or more recently CD4 + , bind to processed antigen in the context of class II MHC antigens and become activated with the addition of macrophage-secreted interleukin 1 (IL- 1). Once activated, T helper cells

JANET

W. APPELBAUM

secrete B-cell growth factor (BCGF) and B-cell differentiation factor (BCDF) which in turn activate B cells and lead to the production of antibody and IL-2. IL-2 induces a positive feedback effect on CD4+ cells which results in the growth of IL-2 receptor bearing cells and activates cytotoxic (T8 or CD8 + ) T cells (CTLs). CDS + cells recognize processed antigen in the context of class I MHC antigen. Activation of CTLs culminates in the cytotoxic events that lead to target cell death (Fig 1). Unlike T cells, B cells recognize antigen in solution or on cell surfaces by antigen specific immunoglobulin which is both expressed on the B-cell surface and secreted by plasma cells. The antigen-binding amino-terminal or hypervariable region accounts for antibody heterogeneity and allows for target specificity. The constant domain or FC (crystallizable fragment) receptor mediates host responses such as complement fixation by interacting with FC receptor-bearing host cells.’ Immunoglobulin is composed of two identical light chain (kappa or lambda) and two identical heavy chain polypeptides held together by sulfide and disulfide bonds. Variations in the heavy chain carbohydrate content and different (multimeric) configurations make up the five immunoglobulin classes (IgG, IgM, IgA, IgD, and IgE). Although both T and B cells are responsible for recognizing invading organisms, B cells are most important for host resistance to extracellular pathogens and T cells are generally important for host resistance to intracellular pathogens.* Both T-cell antigen and B-cell antigen binding require that the antigen and immunocyte be in close proximity and that multiple covalent bonds are formed between the hypervariable binding region and the antigen. lo Although B cells do not require antigen processing and presentation for initial recognition, antibody proliferation usually requires T-helper secreted lymphokines that have been activated by being exposed to processed antigen in the context of class II MHC.7 When either T or B cells bind the specific antigen, clonal proliferation of that population of cells is initiated. There may be multiple antigenic determinants or binding sites per foreign protein and therefore clonal proliferation to a single substance usually involves many clones. The rapid expansion of antigen specific T and B cell clones from memory cells after antigen re-exposure is the essential basis for adaptive immunity. However, if this expansion were unchecked, it could prove harmful to

IMMUNE

SYSTEM

53

AND CANCER

Fig 1. Overview of the adaptive immune response. (Adapted and reprinted with permission.‘3)

the host. Thus, T cell and antibody exert feedback control over their own clonal expansion by binding antigen and eliminating the stimulus for further proliferation. ‘i Further, suppressor T cells that down-regulate (reduce) T-helper and B-cell proliferation exist to help control the immune response. l2 Antibody production is also regulated by antibody directed against the hypervariable region of the immunoglobulin molecule. It has been estimated that an individual has a repertoire of lo7 B-cell clones each preprogrammed to produce antigen-specific antibody if that particular antigen is encountered. l3 When clonal expansion occurs, it serves as an immunogenic stimulus that leads to the production of anti-idiotypic antibodies that can suppress proliferation of the expanding clone. According to Goodman, i3 Niels Jeme proposed the Network Theory of anti-idiotypes stating that ev-

ery antibody has an internal image anti-idiotype. Although difficult to prove experimentally (beyond anti-anti-idiotype), his theory is accepted and thought to be critical to regulation of the immune response. DEVELOPMENT OF CANCER IN IMMUNODEFICIENCY STATES

The fact that cancer is more common in individuals with preexisting immunodeficiency provides compelling evidence for the role of the immune system in controlling cancer. Impaired immune responsiveness can be due to a congenital lack of specific immune compartments or due to acquired immune deficits secondary to drugs, infection, cancer, or autoimmunity.’ The increased incidence of cancer associated with immunodeficiency

54

JANET

is most marked with leukemias and lymphomas whereas the increase is less significant with the common solid tumors seen in the general population. The precise incidence of cancer in patients with congenital or primary immunodeficiency disease (Table 1) such as X-linked lymphoproliferative syndrome, Wiskott-Aldrich Syndrome, ataxiatelangiectasia, or severe combined immunodeficiency disease (SCID) is unknown because severe forms of these immunodeficiencies are rapidly fatal and milder clinical presentations may go undiagnosed. l4 Cancers associated with primary immunodeficiency disorders are leukemias (ALL, AML), Hodgkin’s disease (HD), non-Hodgkin’s lymphoma (NHL), and epithelial cancers such as gastric, liver, and ovarian cancer. The immune defects in congenital disorders are complex and include both humoral and cellular immune dysfunction and most often combinations of both.15 The fact that patients with SCID might not show a greater incidence of some cancers has been attributed to the retention of NK cell function despite dramatic cell-mediated and humoral immune deficits. l6 Cancer in Patients With Autoimmune Disease Autoimmune disorders such as rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and SjGgren’s syndrome (SS) are not simply the result of an immune response against an autoantigen. Disease initiation is believed to be multifactorial and to include genetic susceptibility accompanied by repeated viral or bacterial infection in a host with impaired immunoregulatory mechanisms . 17,‘* Abnormal regulation may be a result of faulty T cell “thymic education” where any cell recognizing a self-MHC is aborted and/or interference with suppressor mechanisms resulting in abnormal target recognition, cell proliferation, and immunosuppression. The actual incidence of can-

Immune

Syndrome

X-linked immunodeficiency syndrome Wiskott-Aldrich syndrome Ataxia-telangiectasia

Adapted

and reprinted

Impaired Complex,

cer in patients with autoimmune disease is unclear due to a number of factors including a lack of adequate controlled studies, the confounding effect of mutagenic and immunosuppressive drugs used in the treatment of autoimmune disease, and because the diagnosis of either cancer or autoimmune disease may facilitate the diagnosis of the &her condition.” Sela and Shoenfeld” conclude that patients with RA have 2 to 3 times greater risk of developing cancer and that this risk is further increased with the addition of immunosuppressive therapy. On the other hand, they conclude that malignancy and SLE in humans are not related (it appears to be related in murine SLE). SS has the clearest association with malignancy. In the large National Institutes of Health (NIH) series of 136 patients, 7 developed lymphoma, which is 44 times the expected prevalence in the female agematched controls from the general population. l9 As with congenital immunodeficiencies, the types of malignancies associated with autoimmune disease are lymphomas (HD, NHL), leukemias and multiple myeloma. It is becoming increasingly important to understand the nature of immunoregulatory function. Dysregulation as opposed to simple immunosuppression may be especially significant in the pathogenesis of cancers of the immune system (leukemias, lymphomas) and has been attributed to the combined effects of repetitive stimulation and lack of proper suppression. l4 Cancer in the Organ Transplant Population The development of specific immunosuppressive agents such as cyclosporin, FK-506 and Rapamycin (Wyeth-Ayerst Pharmaceuticals, Princeton, NJ), as well as increased sophistication in locating suitable organ donors and providing supportive care to transplant recipients, has led to an increased use of organ transplant for a variety of nonmalignant conditions. To be successful, an allogeneic organ transplant requires a balance be-

Defect

B cell responses multicompartmental

Malignant

to EBV antigens defects

Complex, multicompartmental defects; defective DNA repair after gamma irradiation with permission.‘4

W. APPELBAUM

Tumors

Risk Overall

NHL

35%

NHL, AML HD

15% to 37%

ALL, NHL, HD, nerve, ovarian, skin, stomach centers

12%

IMMUNE

SYSTEM

55

AND CANCER

tween adequate immunosuppression to maintain the graft and prevention of potentially lethal infection.20 Immunosuppression is directed toward T cells as they are largely responsible for graft rejection. In the acute graft period, the goal is to prevent the sensitization of T cells and their subsequent recognition and rejection of the graft as nonself. After the acute phase, the host can become unresponsive (tolerant) to the continued presence of donor MHC antigens by the development of donor-specific suppressor T cells.20 The data on cancer incidence among organ transplant patients represent a cohort of chronically immunosuppressed patients for whom there are no matched controls. However, the approximate 6% incidence represents a lOO-fold increased risk compared with an age-corrected general population.21 Most allogeneic organ grafts require chronic immunosuppression. Transplant-associated malignancies include anogenital carcinoma, lymphoma, Kaposi’s sarcoma, and squamous cell carcinoma of the skin. I4 Anogenital cancers are associated with human papilloma virus (HPV) and it is assumed that immunosuppression impairs immunosurveillance of viral antigens or that immunosuppression activates latent virally infected cells. l4 Transplant-associated lymphomas are largely of B-cell origin and are commonly extranodal (as opposed to the more common nodal presentation among nontransplant-related lymphoma) . There is an unusually high (33%) involvement of the central nervous system. i4 Nonlethal basal cell skin cancer is the most common histopathology for skin cancer in the nontransplant population. Virulent (60% mortality) squamous cell carcinoma of the skin is much more prevalent among the transplant population. l4 Finally, Kaposi’s sarcoma has been reported as 400 to 500 times increased in the transplant population compared with the general population. In the non-HIV-infected nontransplant population, Kaposi’s sarcoma affects men more than women and usually runs a noncritical course. However, the mortality rate among transplant patients is 25% and the typical male-to-female ratio is decreased from 9: 1 to 2: 1 in the transplant population. I4 How can these specific malignancies and their unique presentations be explained? One explanation is that chronic immunosuppression impairs immunosurveillance. Based on the information obtained from patients lacking T-cell- or B-cellmediated immunity, those cancers likely to have a

viral etiology (such as anogenital cancer and Epstein-Barr virus [EBV]-associated lymphoma) may indeed be related to defective immunosurveillance as a result of diminished T-, B-, and NK-cell antiviral function. Another explanation is that immunosuppressants may themselves be carcinogenic. However, although the data suggest that alkylating agents and radiation are carcinogenic, there is no evidence that cyclosporin and Rapamycin exert a direct carcinogenic effect. Thus, the increased incidence of the specific malignancies seen in transplant patients is likely related to the chronic allostimulation from the graft that provides an increased opportunity for random somatic mutation in the responding cells. These potentially malignant mutations are able to expand because of concomitant immunosuppressive agents administered to the patient. Cancer in Patients Infected With Human Immunodeficiency Virus Neoplasms associated with the acquired immunodeficiency syndrome (AIDS) include Kaposi’s sarcoma (KS), NHL (largely B-cell lymphomas, commonly extranodal and aggressive), squamous cell carcinoma of the head and neck, and rectal carcinomas. l4 These cancers arise in the setting of dramatic immunosuppression resulting from infection with the human immunodeficiency virus (HIV) which is compounded by immune suppression from its therapy. The CD4 + T lymphocyte and the macrophage are targeted by HIV inhibiting the initiation of the cell-mediated immune response, normal B cell growth and proliferation, and antigen processing and presentation. In general, impaired cellular immunity disarms the most powerful immune effector mechanisms against both viral and nonviral antigens. The progression and regression of KS has been correlated with the initiation and withdrawal of immunosuppressants in both animals and humans.21 It has been postulated that factors common to the immunodeficiency present in classic KS, occurring in elderly Mediterranean Jewish men, iatrogenic KS present in organ transplant patients, and epidemic KS (patients with AIDS) may contribute to the pathogenesis of both KS and possibly NHL. The search for these common factors continues and is currently focused on the over-expression of oncogene, possibly the c-myc oncogene.22 It is likely that multiple factors are involved in the biology of the various AIDS-associated neo-

56

JANET

plasms. Activation of latent herpes virus or HPV has been implicated in the development of squamous cell cancers. It is known that EBV can immortalize B cells and that unrestrained B-cell proliferation can increase the opportunity for somatic mutation. However, only about 30% of AIDSlymphomas contain the EBV genome.22 It is plausible that other DNA and/or RNA viruses may be involved in the etiology of AIDS-related NHL.22 Second Malignancies in Cancer Patients The largest population of cancer patients followed for secondary malignancies are those diagnosed with Hodgkin’s disease (HD). Tucker et a123 recently updated the 15year follow-up of 1,507 patients treated for HD at a single institution. Patients were found to have a 17.6% cumulative risk (sixfold excess risk) of second malignancy. Secondary neoplasms are commonly aggressive tumors that respond poorly to conventional chemoradiotherapy. The types of secondary neoplasms observed were consistent with those noted in other immunocompromised populations. The vast majority of de novo cancer patients are not known to have predisposing congenital or acquired alterations in immunity and secondary malignancies largely have been attributed to mutagenic therapy, particularly alkylating agents. 24 The observation of second malignancy in HD patients is complicated by the fact that immunologic abnormalities have been identified that persist for long disease-free intervals and are thought to appear concomitant with HD. These abnormalities are T-cell related and include depressed T-cell response to antigen, increased sensitivity to suppressor mechanisms, and decreased production of IL2. Alterations in immune function are exacerbated by therapy especially radiotherapy which has been shown to particularly depress CD4+ T lymphocytes. Decreased CD4+/CD8+ ratio persists in disease-free survivors.25 Despite the identification of immune alterations in patients with HD, alkylating agents are currently held responsible for the increased risk of leukemia.25 TUMOR ANTIGENS

Animal Models Tumor immunology is the study of the antigenic properties of malignant cells and the host’s im-

W. APPELBAUM

mune response to them. Animal models have been critical research tools in defining the basic principles of tumor immunology.26 The availability of inbred strains of mice in the early 1960s allowed tests of tumor antigenicity without the complication of histocompatibility.26A series of animal studies were performed investigating the antigenicity of tumors induced by chemicals, viruses, radiation, physical irritation, or those occurring spontaneously. The basic design of these studies was to induce a tumor in the primary host, excise it, and reimplant it in the primary host.26,27 Immunogenicity was measured as a function of tumor transplant acceptance (low immunogenicity) or rejection (high immunogenicity). These studies showed that tumors could be recognized and rejected and that the recognition was highly specific. Tumors induced by ultraviolet (UV) light were highly immunogenic and tumors induced by chemicals (usually polycyclic hydrocarbons) such as 3-methylcholanthrene (MCA) and dimethylbenzanthrene (DMBA) exhibited varying degrees of immunogenicity. Unfortunately, spontaneous tumors generally did not evoke any immunogenic response. The antigens in these types of animal studies are called tumor-specific transplantation antigens (TSTAs). TSTAs were noted to be extraordinarily heterogeneous . Subsequent studies sought to determine if protective immunity is present when tumor is excised from the primary host and that host is later challenged with a new secondary tumor.27 Chemical and UV light-induced tumors exhibited little or no cross-protective immunity, thus transplanted tumors grew and progressed. However, immunization of the host with killed cells from the primary tumor, a form of adoptive immunotherapy, did elicit subsequent transplant rejection.27 This rejection was specific to the primary tumor. For example, killed cells from an MCA-induced tumor would immunize against that MCA tumor but not another MCA-induced tumor.27 There does not appear to be any common TSTAs shared by chemically induced tumors even for that specific chemical. Viral Antigens A virus can incorporate itself into the DNA of the host cell by direct insertion or indirectly through reverse transcriptase resulting in the expression of virally encoded antigens, usually proteins or glycoproteins, on tumor cells. Animal

IMMUNE

SYSTEM

AND CANCER

models of virally induced tumors demonstrated that such tumors exhibit varying degrees of immunogenicity .27 More importantly, common antigens are expressed on all tumors induced by a specific viral system (eg, SV40, MuLV, and polyoma virus). In contrast to the highly heterogeneous antigen expression of chemically induced tumors, cross-protective immunity could be demonstrated with virally induced tumors because of common antigen expression. Thus, an SV40-induced tumor excised from the primary host would result in the transplant rejection of any subsequent SV40induced tumor. The gene that codes for the SV40 TSTA, called the SV40 T antigen has been identified, is present on all SV40-transformed cells and is required to produce the malignant phenotype.26 SV40 T antigen proximity to class I MHC molecules facilitates recognition by CTLs. The discovery of common viral antigens raised hopes of future antitumor vaccines for those tumors suspected of viral etiology such as HPVassociated cervical carcinoma and EBV-associated lymphoma. Immunization with live virus has been shown to prevent the development of primary and transplanted animal tumors in some cases.26 Unfortunately, obstacles inherent in vaccine design, the fact that viruses can mutate frequently, and the ability of certain viruses to prevent surface expression of class I antigens has not yet allowed the discovery of common viral antigens to be put to therapeutic advantage. ‘* Human Tumor Antigens Questions regarding the nature of antigen in human tumors obviously could not be answered by transplanting tumors in humans, although some studies have involved skin testing with killed tumor cells.29 Cultured tumor cell lines have been researched extensively as an antigen source but a number of limitations to this approach have been encountered. These include potential expression of neoantigens on cultured cells not present in vivo, altered antigenicity as a function of cell culture conditions, viral contamination of cultures, and alteration of cell lines with passage over time.29 Despite these and other limitations, human tumor antigens have been discovered and several have been well-characterized. Tumor-specific antigens (TSAs) ideally represent the best target for induction of an immune response against tumor cells in humans. Unfortu-

57

nately, isolation and characterization of unique human TSAs has not been accomplished. The general inability to detect TSAs on human tumors may be due to technical limitations of assay sensitivity as well as the fact that human tumors are weakly antigenic. Recent advances have allowed scientists to expand populations of reactive T- and B-cell clones from tumor-draining lymph nodes from a variety of solid tumors, leukemias, and lymphomas.’ These cells, while tumor reactive, have yet to be shown to be tumor-specific. Several tumorassociated antigens (TAAs) have been well characterized by using monoclonal antibodies. TAAs are not ideal immune targets for therapeutic purposes because of their concurrent expression on normal cells; however, there are potentially exploitable features of TAAs that make them interesting. There may be quantitative rather than qualitative differences of TAA expression on tumors as compared with normal cells, as has been shown in melanoma and certain solid tumor-associated glycoprotein antigens. 27 Oncofetal antigens are the most familiar TAAs to the oncology clinician. Normally expressed during embryonic and fetal development, oncofetal antigen, such as carcinoembryonic antigen (CEA), has been found in the serum of patients with a wide variety of adenocarcinomas including colon, lung, breast, and pancreas where CEA is shed from the tumor into the serum. CEA does not evoke an immune response. One explanation is that CEA is remembered as a “self” protein from early development.27 The increase and decrease of serum CEA has been useful as a tumor marker or indicator of progressive disease or response to therapy; however, elevation of CEA is also associated with non-neoplastic conditions such as emphysema and pancreatitis. Alpha-fetoprotein (AFP) normally is synthesized and secreted during fetal development by the yolk sac and the liver. Serum levels can be elevated in patients with testicular cancer and hepatocellular carcinoma. Like CEA, AFP is nonimmunogenic and may be found in the serum of patients with non-neoplastic conditions such as cirrhosis.27 Melanoma-associated antigens (MAA) have been studied extensively due to the relative ease of obtaining fresh tumor samples. Over 40 different MAAs have been described.29 Excised melanomas frequently have adjacent lymphocyte infiltration indicating that MAAs are immunogenic in vivo.

58

Immune Effector Mechanisms There are both T-cell-mediated and B-cellmediated mechanisms for neutralizing or eradicating foreign antigen. These mechanisms are highly integrated and interdependent. The term cellmediated immunity refers to primarily T-celldriven effector functions and the term humoral immunity refers to B-cell-driven functions. B-Cell EfSector Functions Antibody is effective in eradicating bacterial and viral antigens by forming antibody-antigen complexes that are cleared rapidly by the reticuloendothelial system (RES), by blocking viral binding sites on target cells, by opsonization which facilitates phagocytosis, and by activating the complement cascade that leads to complement-induced lysis and cell death by necrosis. Serological analyses, most commonly by direct or indirect immunofluorescence, have shown antibody to specific tumor in both animal and human tumor-bearing subjects. However, the presence or absence of antibody has not been correlated with tumor inhibition or growth.29 For antibody to have an antitumor effect, the tumor must express an antigen recognizable to B cells. In addition, the antigen also must be recognized by CD4+ T cells that are capable of secreting various lymphokines, including IL-4, IL-5, and IL-6, which stimulate B-cell proliferation and differentiation. ‘* Certain antibody subclasses, particularly IgG 1 and IgG3, are potent mediators of antibody-dependent cellular cytotoxicity (ADCC).30 Host NK cells, killer (K) cells, and macrophages that have FC-gamma receptors (FQR) adhere to target cells that are opsonized (coated) by ADCC antibodies. Cell death occurs by apoptosis, a process in which the cell rounds up and forms vesicles that fuse to the cell membrane. Cellular “bubbles” are then eliminated by phagocytosis.31 In early clinical trials with polyclonal antiserum and subsequently monoclonal antibodies, antibody was administered as native or unmodified molecules (ie, not conjugated to a cytotoxic moiety such as chemotherapy or isotope). ADCC was probably the mechanism responsible for the rare reports of tumor responses in these early trials. The majority of clinical trials to date have used murine monoclonal antibodies due to their relative ease of production. It is

JANET

W. APPELBAUM

thought that murine receptors will be suboptimal in mediating human antibody effector mechanisms. 32 Because murine monoclonals by themselves have resulted at best in only minor clinical responses, interest in the ability to incur ADCC responses using humanized antibodies has developed and is now possible as a result of advances in chimeric and humanized hybridoma technology where human FC receptors are available. The anti-idiotype regulatory network may play a unique and complex role in B-cell and T-cell effector mechanisms by stimulating or inhibiting antigen specific responses. Miller et a133reported sustained complete remission in a lymphoma patient treated with unmodified anti-idiotype antibody directed against his own B-cell lymphoma. However, most responses to anti-idiotype antibodies have been partial and temporary34; furthermore, making a unique antibody for each patient is an expensive and lengthy procedure. Therefore, current trials are investigating shared antiidiotypes35 (ie, anti-idiotypes common to lymphomas of several different individuals and not uniquely cloned for each patient), as well as the addition of either cytotoxic agents or radioisotopes conjugated to the antibody to achieve targetspecific cell kill. T-Cell-Mediated Effector Functions Intact cell-mediated immunity protects against fungal, viral, and protozoa1 infections. In addition to recognizing and responding to specific antigen, T cells interact with each other and other lymphocytes by secreting antigen nonspecific, hormonelike polypeptides called lymphokines, cytokines, or interleukins . It has been demonstrated that CTLs are the T-cell subset responsible for the recognition and rejection of tumor transplant in animals. This same subset is also responsible for graft rejection in organ transplantation and graft-versus-host disease (GVHD) in marrow transplant patients. Cytolysis by CTLs requires close proximity between effector and target cell. Membrane-bound cytoplasmic granules are responsible for the lytic events that lead to cell death by apoptosis. The granules contain enzymes such as aryl sulfatase, acid phosphatase, and B-glucuronidase.31 Interestingly, CTLs are immune to their own effector mechanisms and can kill more than one target cell of that specific clone. 3’

IMMUNE

SYSTEM

Animal models have been developed to determine the specific lymphocyte subsets involved in tumor lysis. One approach involves selective depletion of a cell line (eg, thymectomized, lethally irradiated mice reconstituted with T-depleted marrow or animals rendered humorally deficient by the chronic administration of high doses of antibody against immunoglobulin from birth). l2 Observations of tumor immunogenicity, growth rate, or eradication can then be made and compared with controls. These models have shown that although CD8 + cells are required for tumor lysis, CD4 + cells play an important role as well. Macrophages Activated macrophages have the ability to bind to transformed cells and lyse them. The nature of the macrophage receptor and its recognition structure is not known. The macrophage can induce target cell lysis by the release of tumor necrosis factor (TNF). TNF can kill tumor cells by directly binding to specific TNF cell-surface receptors or by nonspecific uptake by pinocytosis. The toxicity is thought to be due to the production of free radicals within the tumor cell. Normal cells can produce the enzyme superoxide dismutase which neutralizes free radicals; tumor cells seem unable to produce as much of this enzyme.27 Some tumor cell lines are known to be TNF-resistant by an as yet undefined mechanism. TNF also may act directly on tumor vessels resulting in thrombosis and ischemia and ultimately in necrosis. Although TNF has potent antitumor effects locally, clinical trials using TNF systemically have so far been disappointing largely due to the considerable toxicities of the agent. Interferon gamma (EN-y) is the primary activator of macrophages.6 In addition to TNF, macrophages also release IL- 1, potent proteases, and inhibitors of tumor-cell mitochondrial respiration and DNA synthesis.36 T-cell-secreted IFN-?I has been shown to up-regulate (increase) MHC antigen expression thus increasing the efficacy of the cellmediated antitumor response.27 Natural

59

AND CANCER

Killer Cells

NK cells represent 5% to 8% of the total white blood cell count. NK cells possess FCyR which enables them to participate in antibody-dependent cellular toxicity. NK cells are a distinct lymphocyte lineage as illustrated by patients with SCID

who have no B or T cells but do have NK cells. NK cells possess neither surface immunoglobulin nor T cell receptors. The structure responsible for antigen recognition has not been identified. Recognition of antigen does not require processing or presentation by APCs and is not MHC restricted,6 therefore NK cells are rapidly mobilized. NK cells are activated by multiple lymphokines including interferon alfa (IFN-a), interferon beta, and IL-2. Activation can be accomplished in vitro or in vivo resulting in lymphokine-activated killer (LAK) cells. Clinical trials with IL-2 alone or with LAK cells have shown clear responses, particularly in renal cell carcinoma and melanoma.37 However, because LAK cells are not totally target-specific, considerable toxicities have been seen. NK cells can lyse a wide variety of target cells but appear to have a predilection for virally infected cells and certain tumors. Some experiments indicate there may be several subsets of NK cells.3’ One such subset is the killer cell. Killer cells appear to differ in their mode of recognition but use a similar cell-killing process (apoptosis). EVASION OF IMMUNE EFFECTOR MECHANISMS

It is generally believed that somatic mutations occur constantly and at a high rate. Because animal studies showed that CTLs are responsible for recognition and rejection of TSTA-bearing tumors, it was postulated that the immune system evolved not only to protect the host from foreign pathogens but to recognize and eliminate mutant cells such as cancer cells. ” It was proposed early that immune surveillance was responsible for eradicating cancer cells before they become clinically apparent and that those that become apparent represent a defect in this protective mechanism. The theory of immune surveillance as proposed by Sir MacFarlane Bumet38 in 1971 has been variably accepted and highly criticized over the ensuing years. Criticism has been leveled due to the lack of clinical correlation between T-cell depletion (athymic mice, patients with the DiGeorge syndrome) and an increase in spontaneous malignancy. The fact that both nude mice and patients with the DiGeorge syndrome are fully competent with regard to macrophage function and NK cells may explain why spontaneous malignancies are not more common.16 A role for NK cell immune surveillance in terms

60

of virally induced tumors is currently accepted. ” There is an increased incidence of EBV-induced lymphoma in organ transplant patients who are T-cell depleted to reduce graft rejection and who have been shown to have low NK cell activity.16 The concept of cancer cells “sneaking through” or avoiding eradication by avoiding detection was proposed by Old et a1.39The thought is that most human tumors are weakly antigenic and low levels of tumor antigen at tumor inception fail to stimulate an immune response. After tumor growth and progression, the tumor load overwhelms and exceeds the capacity of the immune system to respond effectively. Animal experiments have not fully corroborated this concept. The phenomenon of “sneaking through” has been cited as support for the theory of immune surveillance as there is an increased incidence of cancer in immunosuppressed patients. However, as was discussed earlier, instead of a wide variety of cancers, the majority of tumors that occur in immunosuppressed individuals are leukemias , lymphomas , and tumors of epithelial origin. Leukemias and lymphomas are thought to result from an immunoregulatory abnormality rather than defective immune surveillance. l4 On the other hand, epithelial neoplasms are strongly antigenic and may reflect a failure in immune surveillance particularly in those patients maintained on chronic immunosuppressants. I4 Several ways in which tumors might evade immune detection not involving immune surveillance have been described. ” The importance of understanding tumor escape mechanisms is clearly essential to developing ways to thwart these “tactics.” Antigen-TCR binding is necessary for cytolysis. It has been suggested that molecules bound to the tumor cell surface, such as sialic acid, may mask tumor antigens and/or prevent lymphocyte binding.27 The enzyme neuramidase has been shown to decrease cell membrane sialic acid and increase lymphocyte binding. ” Tumors also may induce increased fibrin formation which can effectively mask antigen and/or inhibit binding. Serum levels of CEA and AFP represent antigen shed from the tumor. Many other shed antigens have been found in the serum of a variety of human tumors. Antibody complexing with shed antigen can induce tolerance, which causes a normally reactive T and/or B cell to become insensitive to activation signals. This insensitivity may be due to a “soaking up” of the reactive antibody clone by the shed antigen or by saturating all the low im-

JANET

W. APPELBAUM

munogenic antigen receptors and shielding the highly immunogenic receptors. ’ ’ In addition to shedding, surface antigen can be lost by internalization through endocytosis and by redistribution along the cell membrane,’ ’ referred to as antigenic modulation. Antigenic modulation has been shown to be both specific and reversible and to occur on both normal and transformed cells. Modulation is a distinct, adaptive phenotypic change and not the emergence of a resistant cell subset. “*27 The displacement of antigen along the cell membrane results in cells that are resistant to cytolysis despite the fact that antigens are not completely lost from the cell surface. ’ ’ Antigenic modulation may pose serious hurdles for monoclonal antibody-based therapy approaches that rely on target cell antigens for efficacy. It has been reported that an anti-common acute lymphocytic leukemia antigen (CALLA) antibody administered in very small doses led to antigenic modulation in a very short time ( 18 hours). ’ ’ However, not all antigens modulate and cells with antigens that do modulate also have stable cell surface antigens. Immunocyte-to-target-cell ratio is known to be critical to an effective immune response. Unfortunately, there are many factors contributing to cancer patients having a decrease in immunocyte subsets such as granulocytes, T and B cells, and NK cells. Immunosuppression most commonly results from chemotherapy and/or radiation used to treat cancer or in certain cases may be secondary to congenital or acquired immunodeficiency that exists before cancer. Current experiments are investigating whether it is possible to isolate T-cell clones that are specific for tumor-associated or tumor-specific antigens, expand these clones in vitro, and give them back to the patient. This form of tumor-specific adoptive immunotherapy is clearly effective in animal models and, as understanding of human tumor immunology expands, may prove effective in humans.40 SUMMARY

The observation that malignant cells express antigens that may be recognized by immunocytes and that immune effector mechanisms have the capability of destroying tumor cells has increased our appreciation of the biology of cancer and its relationship to immune function as well as offered new options for therapeutic intervention. Clinical trials are in progress to evaluate several different ap-

IMMUNE

SYSTEM

61

AND CANCER

proaches to modifying the host’s immune response against tumor. One approach is to administer agents that have direct activity against the malignancy. For example, antibody conjugates bring cytotoxic molecules of chemotherapy, radioisotopes, or toxins directly to the tumor.4 A second approach is to administer agents that modulate the host’s own antitumor response such as IFN-o and IFNy.4 Adoptive cellular immunotherapy aimed at isolating and expanding the host’s own tumor-specific lymphocytes and inducing activation and proliferation with lymphokines such as IL-2 has shown encouraging results. ’ Even though clinical data are still quite premature, it is reasonable to assume that

in the future immunomodulation including the stimulation of immune effector mechanisms to eradicate tumor, the reconstitution of immune deficiency in diseases such as AIDS, the suppression of immune function to avoid graft rejection and GVHD,4’ and the isolation and insertion of genes encoding tumor antigens into recombinant vectors to immunize the host to the tumor antigen will be commonly and successfully employed.’ ACKNOWLEDGMENTS The author thanks Gerald Sonnefeld, PhD, for his review of the manuscript; Phillip Greenberg, MD, and Fred Appelbaum, MD, for their advice and encouragement; and Pauline E. Sherman for preparation of the manuscript.

REFERENCES 1. Greenberg PD: Mechanisms of tumor immunology, in Stites DP, Terr AI (ed): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 580-587 2. Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991 3. Roitt I, Brostoff J, Male D (eds): Immunology (ed 2). New York, NY, Harper & Row, 1989 4. Abbas AK, Lichtman AH, Pober JS (eds): Cellular and Molecular Immunology. Philadelphia, PA, Saunders, 1991 5. Male D, Champion B, Cooke A: Antigenicity of proteins, in Male D, Champion B, Cooke A (eds): Advanced Immunology. Philadelphia, PA, Lippincott, 1987, pp 8.1-8.9 6. Lamer L: Cells of the immune response: Lymphocytes and monoculear phagocytes, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 61-72 7. Lydyard P, Grossi C: Cells involved in the immune response, in Roitt I, Brostoff J, Male D (eds): Immunology. Philadelphia, PA, Saunders, 1991, pp 2.2-2.17 8. Groverman J, Pames J: The T cell receptor, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 73-77 9. Turner M: Molecules which recognize antigen, in Roitt I, Brostoff J, Male D (eds): Immunology. Philadelphia, PA, Saunders, 1991, pp 5.1-5.11 10. Steward M: Antigen recognition, in Roitt I, Brostoff J, Male D (eds): Immunology. Philadelphia, PA, Saunders, 1991, pp 7.1-7.10 11. Ozer H. Tumor immunity and escape mechanisms in humans, in Mihich E (ed): Immunological Approaches to Cancer Therapeutics. New York, NY, Wiley, 1982, pp 39-61 12. Greenberg PD: Adoptive T cell therapy of tumors: Mechanisms operative in the recognition and elimination of tumor cells, in Dixon FJ (ed): Advances in Immunology, Vol 49. San Diego, CA, Academic Press, 1991, 281-355 13. Goodman JW: The immune response, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 34-44 14. Ziegler JL: Cancer in the immunocompromised host, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 588-598 15. Antmann AJ: T cell immunodeficiency disorders, in

Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 335-340 16. Lopez C: Natural killer cell functions. Clinical Immunol Today 13:1-12, 1985 17. Steinberg AD: Mechanisms of disordered immune regulation in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 432-437 18. Golub ES: Autoimmunity, in Golub ES (ed): Immunology A Synthesis. Sunderland, MA, Sinauck Associates, 1987, pp 48 l-496 19. Sela D, Shoenfeld Y: Cancer in autoimmune diseases. Semin Arthritis Rheum 18:77-87, 1988 20. Welsh K, Male D: Transplantation and rejection, in Roitt I, Brostoff J, Male D (eds): Immunology (ed 2). New York, NY, Harper & Row, 1989 pp 24.1-24.10 2 1. Groopman JE, Broder S: Cancer in AIDS and other immunodeficiency states, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer Principles and Practices of Oncology, Vol 2 (ed 3). Philadelphia, PA, Lippincott, 1989, pp 1953-1970 22. Kaplan LD, McGrath MS: AIDS-Associated nonHodgkin’s lymphoma in DeVita VT, Hellman S, Rosenberg SA (eds): AIDS Etiology, Diagnosis, Treatment and Prevention, Vol 4 (ed 2). Philadelphia, PA, Lippincott, 1991, pp l-l 1 23. Tucker MA, Coleman CN, Cox RX, et al: Risk of second cancers after treatment for Hodgkin’s disease. N Engl J Med 318:76-81, 1988 24. Coleman CN, Tucker MA: Secondary cancers, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer Principles and Practices of Oncology, Vol 2 (ed 3). Philadelphia, PA, Lippincott, 1989, pp 2181-2188 25. Hellman S, Jaffe ES, DeVita VT: Hodgkin’s disease, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer Principles and Practices of Oncology, Vol 2 (ed 3). Philadelphia, PA, Lippincott, 1989, pp 2696-1734 26. Kripke ML: Principles of tumor immunology. Immunol Allergy Clin North Am 10:595-606, 1990 27. Abbas AK, Lichtman AH, Pober JS: Immunity to tnmors, in Wonsiewicz MJ (ed): Cellular and Molecular Immunology. Philadelphia, PA, Saunders, 1991, pp 336-352 28. Anderson M, Piiabo S, Nilsson T, et al: Impaired intracellular transport of Class I MHC antigens as a possible means

62 for adenoviruses to evade immune surveillance. Cell 43:215222, 1985 29. Brown JM, Rosenberg SA: Serologic analysis of human solid tumor antigens, in Mihich E (ed): Immunological Approaches to Cancer Therapeutics. New York, NY, Wiley, 1982, pp l-37 30. Ortaldo JR, Woodhouse C, Morgan AC, et al: Analysis of effector cells in human antibody-dependent cellular cytotoxicity with murine monoclonal antibodies. .I Immunol 183:35663572, 1987 31. Male D, Champion B, Cooke A (eds): Cytotoxic lymphocytes, in Advanced Immunology. Philadelphia, PA, Lippincott, 1987, pp 7.1-8.0 32. Kipps T, Parham P, Punt J, et al: Importance of immunoglobulin isotype in human antibody-dependent, cellmediated cytotoxicity directed by murine monoclonal antibodies. J Exp Med 161:1-17, 1985 33. Miller RA, Maloney DG, Wamke RK, et al: Treatment of B cell lymphoma with monoclonal anti-idiotype antibody. N Engl J Med 306:517-522, 1983 34. Meeker TC, Lowder J, Maloney DG, et al: A clinical trial of anti-idiotypc therapy for B cell malignancy. Blood 65:1349-1363, 1985

JANET

W. APPELBAUM

35. Miller RA, Hart S, Samoszuk M, et al: Shared idiotypes expressed by human B-cell lymphomas. N Engl J Med 321851-857, 1989 36. Oppenheim JJ, Ruscetti FW, Fattynek C: Cytokines, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 78-100 37. Hawkins MJ: IL-2ILAK: Current status and possible future directions, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer Principles & Practice of Oncology Updates, Vol 3. Philadelphia, PA, Lippincott, 1989, pp l-14 38. Bumet FM: Immunological surveillance in neoplasia. Transplant Rev 7:3-25, 1971 39. Old LJ, Boyse EA, Clarke DA, et al: Antigenic properties of chemically induced tumors. Ann NY Acad Sci 101:80106, 1961 40. Rosenberg SA: Adoptive cellular therapy in patients with advanced cancer, in DeVita VT, Hellman S, Rosenberg SA (eds): Biologic Therapy of Cancer Updates, Vol 1. Philadelphia, PA, Lippincott, 1991, pp l-15 41. Jaffe HS, Sherwin SA: Immunomodulators, in Stites DP, Terr AI (eds): Basic and Clinical Immunology. Norwalk, CT, Appleton & Lange, 1991, pp 73-77