Since the transplantation of UV-induced epidermal carcinomas in mice could be accomplished virtually anywhere in the organism, it was clear that the effects of UVB irradiation were not simply local, as suggested by the failure of contact hypersensitivity only in regions of UV radia- tion. It now appears that there are several mechanisms for such generalized immunosuppression by UV radiation. The first of these is the fact that suppressor T cells that are specific for the UV radiation–induced neoplasms (Trial and McIntyre, 1990) have been induced in the organism as a result of UV radiation. This may be because a different population of antigen-presenting cells in the skin, relatively nonsusceptible to UVB irradiation, can act as antigen-presenting cells for suppressor T lymphocytes that likely express the γδ T cell receptor (see below) (cf. Meunier,1999). In addition, there is evidence that immunosuppression by UV radiation of the skin is me- diated by the UV-induced formation of the abnormal cis isomer of urocanic acid from the prod- uct of the action of histidine ammonia-lyase in the skin, which is the trans isomer. A diagram of
Figure 19.28 Skin-associated lymphoid tissues (SALT) that initiate cutaneous immunity are located in the epidermis (E) and dermis (D) and are separated by the dermal-epidermal junction (J) and the basement membrane (BM). Dendritic cells in the dermis (DC) and Langerhans cells (LC) residing in the epidermis are the antigen-presenting cells of the skin. The dermis contains blood vessels [capillaries (CAP) and post- capillary venules (PCV)] and lymphatic channels. Afferent lymphatic ducts (AFF) carry lymph contents to lymph nodes (LN) containing lymphoreticular cells and postcapillary venules. Lymphocytes exit the LN via efferent lymphatics (EFF) into the thoracic duct (TD) and eventually into the bloodstream. (Adapted from Streilein et al., 1994, with permission of the authors and publisher.)
the reaction is seen in Figure 19.29 depicting primary metabolic products in the metabolism of histidine in skin and the effect of UV radiation. Evidence for this hypothesis was seen in animals having a mutant histidine-ammonia lyase gene with less than 10% of the wild-type levels of skin urocanic acid. Such animals did not exhibit the marked UV-induced immune suppression seen in wild-type animals (cf. Noonan and De Fabo, 1992). Other studies have also implicated TNF-α in the immunosuppressive effect of UV radiation (cf. Meunier, 1999).
γδ T Cells in the Immunological Host-Tumor Relationship
While more than 90% of T cells possess a T-cell receptor with α and β chains (Figures 19.10 and 19.11), about 1% to 10% of T cells in the peripheral blood and lymphoid organs express γ and δ
Figure 19.29 Initial reactions in histidine metabolism indicating the effect of ultraviolet radiation in the formation of cis-urocanic acid (UCA). (Adapted from Noonan and De Fabo, 1992, with permission of the authors and publisher.)
chains on their T-cell receptor. Their response to antigens is not restricted by classical MHC class I or class II antigen, but other MHC class I-related molecules such as CD1 may serve as antigen-presenting molecules for some γδ T cells (cf. Kabelitz, 1995). In a transgenic mouse system, immune resistance to acute T-cell leukemias depended on γδ T cells but was independ- ent of MHC antigen or TAP-2 peptide transporter expression (Penninger et al., 1995). There is also evidence that γδ T cells may recognize various stress-induced or heat-shock antigens within neoplastic cells (cf. Kabelitz, 1995; Groh et al., 1999). Some γδ T cells functioning as dendritic antigen-presenting cells may serve to activate suppressor T lymphocytes (cf. Meunier, 1999). However, in most instances the percentage of γδ T cells that infiltrate various human carcinomas is relatively low, although an accumulation of up to 30% of intrahepatic lymphocytes from can- cer patients may be γδ T cells (Seki et al., 1990). Thus, the actual role of γδ T cells in the immu- nological host-tumor relationship is not clear at this time, but there are clear suggestions that γδ T cells may contribute both to the host immune defense against neoplastic cells and the escape of neoplastic cells from the host immune system.
Upregulation of MHC Expression in Neoplastic Cells
As previously discussed (Table 19.10), several different factors were important in the down- regulation of MHC expression in both normal and neoplastic cells. This included several virus infections as well as the expression of certain oncogenes, the latter making such cells prone to lysis by NK cells (Versteeg et al., 1989). However, it is also apparent that several cytokines are capable of upregulating the expression of MHC antigens in both normal and neoplastic cells. Perhaps the best-studied of such MHC modulators are the interferons. As we shall see below, interferons have been used in the immunotherapy of neoplasia and are particularly important in activating natural killer cells as well as macrophages (see above). Interferons, especially inter- feron-γ, act as transcriptional activators of MHC class I genes by inducing transactivating nu- clear factors that bind specific interferon consensus sequences (cf. Tatake and Zeff, 1993). Interferon is capable of upregulating both class I and class II MHC antigens in normal and neoplastic cells, but there is significant diversity in this response in some melanomas (cf. Kappes and Strominger, 1988). In addition to inducing the expression of MHC antigens in some neoplasms that otherwise show essentially no expression, this cytokine can also induce the expression of such antigens in normal cells that express virtually no or very low levels of MHC class I antigens, as with cells in the central nervous system (Wong et al., 1984; Neumann et al., 1995).
In addition to interferons, tumor necrosis factor α may enhance the expression of both class I and class II MHC antigens, although reportedly, unlike interferon, this cytokine does not induce class I expression in class I–negative neoplasms (Singer and Maguire, 1990; Pfizenmaier et al., 1987).
A third example of the upregulation of MHC antigens that is presumably relevant to the immunological host–tumor relationship is the finding that expression of the hepatitis B virus X gene in both human and murine cell lines is associated with an upregulation of MHC antigen (Hu et al., 1990; cf. Rossner, 1992). The relevance of this to the development of hepatocellular carcinomas in patients with chronic hepatitis B virus infections is not clear at this time. How- ever, the fact that MHC class I and class II antigens may be upregulated or induced by cytok- ines and possibly viral products indicates that such agents are capable of overcoming the subversion of the immune response by neoplastic cells that do not normally express HLA anti- gens in vivo.
AUTOIMMUNITY AND THE HOST–TUMOR RELATIONSHIP
Although Paul Ehrlich (Ehrlich and Morgenroth, 1957) argued that the organism should never react to its own tissues, during the last century a number of diseases resulting from immune reactions within the organism to its own or “self” antigens have now been described. A discus- sion of the development of autoimmune disease is beyond the scope of this text; however, the interested student may consider the following references: Schwartz, 1993; Eisenbarth and Bellgrau, 1994; Mayes, 1999; Bach, 1995. Interestingly, Prehn and Prehn (1987) have presented arguments that neoplasia, at least in part, should itself be considered an autoimmune disease. However, here are considered only an artificially induced autoimmune condition leading to neo- plastic development and its potential application as well as some examples of autoimmunity stimulated in the host by antigens present in neoplasms. An artificial “autoimmune” disease has been induced in rodents by the production of “runt” disease resulting from a graft-versus-host (GVH) reaction in several mammalian species including the human. The basic requirements for a graft-versus-host reaction in vivo are (1) the graft must contain immunologically competent lymphocytes; (2) the host must be incapable of rejecting the graft either because of artificially induced immunological incompetence or because the host is tolerant to the engrafted cells; and (3) a degree of histoincompatibility must prevail between the graft and the host (cf. Seemayer et al., 1983).
The production of the GVH reaction resulting in runt disease is outlined in Figure 19.30; immunocompetent lymphoid and bone marrow tissues are removed from an adult animal that is genetically distinct from the recipient neonate. The cells from the adult survive within the neo- natal animal because the donor cells do not produce a rejection reaction in the host, owing to
Figure 19.30 Classic method for the production of “runt” disease or a graft-versus-host reaction. Immu- nocompetent cells from the spleen, lymph node, or bone marrow are removed from an adult animal and inoculated into a neonate. Since the neonate is immunologically deficient, it will not react to the donor cells, but the donor cells will react to the host tissues, resulting in a graft-versus-host reaction and runt disease with ultimate death.
neonatal tolerance. However, the cells from the adult donor do react to the host tissues, produc- ing both humoral and cell-mediated responses. Such a reaction of the donor cells damages the host tissues eventually to the point of death. During this process, the young animal does not grow—thus the term runt—and it may also exhibit other changes of the skin, tails, ears, and internal organs. A similar condition may be produced in the adult by making the recipient toler- ant to a small number of cells, insufficient to produce runt disease, and subsequently administer- ing a large number of immunocompetent cells from the same donor to the original recipient when this tolerant animal has grown to adulthood. This experimental model is analogous to sev- eral conditions found in the human under the general category of autoimmune diseases (Table
19.14), in which both humoral and cell-mediated immunity to the host’s own tissues are pro- duced by cellular populations within the host (cf. Theofilopoulos and Dixon, 1982).
One of the most interesting phenomena from such investigations was the description by Schwartz and André-Schwartz (1968) and Gleichmann et al. (1976) of the development of ma- lignant lymphomas in recipient mice undergoing a chronic GVH reaction. In this instance the reaction is produced by the injection of the parent’s spleen cells into its own neonatal offspring (F1). The same sequence of events occurs as is seen in Figure 19.30; however, the genotype of
the recipient is at least half identical to that of the donor. Therefore, the reaction in the host is
relatively mild, and most of the recipients survive with relatively little runting. Later in life, how- ever, many of these animals develop lymphomas. Gleichmann et al. (1976) have demonstrated that these lymphomas may develop from host cells, a mixture of host and donor cells, or donor cells only, although host-type lymphomas are seen most frequently. The initiation of lymphoma- genesis in F1 mice undergoing a chronic graft-versus-host reaction requires an immunological
reaction of donor lymphocytes, probably T cells, to incompatible H-2 gene products of the I
region on cells of the F1 recipient. In theory one would expect that the marked stimulation of the donor cells might ultimately lead to their neoplastic transformation. However, it is known that lymphoid cells of the host proliferate quite extensively in the GVH reaction, although the reason for this is not entirely understood. It would appear that the lymphomas resulting from these GVH reactions in genetically related animals are the result of chronic immunological stimula-
tion of lymphoid cells, although some evidence has indicated a role for oncogenic RNA viruses in lymphoma production. Goh and Klemperer (1977) have presented cytogenetic evidence for the leukemic transformation of grafted bone marrow cells in the human.
Although the model system depicted above (Figure 19.30) has no real counterpart in the human, with possible rare exceptions, the model does mimic autoimmune disease seen in the
Table 19.14 Idiopathic Autoimmune Diseases in the Human in Association with Neoplasms
human and other animals with the difference that in such diseases all cells involved are of host or self origin. In Table 19.14 is seen a brief listing of some of the major idiopathic autoimmune diseases in the human and their association with neoplasia. As noted, the relative risks involved are not extremely great, but all are significant. In the case of dermatomyositis, a disease present- ing as an inflammation of muscle and skin, the types of neoplasms found in patients with the disease parallel those observed in the general population but overall are at a higher percentage (Bernard and Bonnetblanc, 1993). In systemic lupus erythematosus, a disease affecting most tis- sues in the organism, the increased risk for various types of neoplasms in general was of the same order as that seen in dermatomyositis; however, a greater risk of breast cancer was seen in Caucasian women with lupus, while all women regardless of race also showed an increased risk of lung cancer (RR=3.1). In rheumatoid arthritis, a disease affecting joints, kidneys, and in later stages other tissues, lymphomas and leukemias were the most strikingly increased neoplasms (Mellemkjær et al., 1996). In systemic sclerosis or scleroderma, a disease affecting the skin and lungs, even higher incidences of specific carcinomas were noted, while lymphomas and leuke- mias were significantly associated with autoimmune thyroiditis.
While these autoimmune diseases are not considered neoplasms, they clearly, from our understanding of the immune process, have some of the characteristics of neoplasia. It is likely that clones of activated T and B cells act in a relatively autonomous manner to produce the reac- tions with self antigens, resulting in the disease pictures. Even more striking are some diseases of atypical lymphoid proliferation associated with autoimmune disease, such as “autoimmune disease–associated lymphadenopathy” and related diseases (cf. Koo et al., 1984) that may have the same biological basis as the idiopathic autoimmune diseases.
Already noted in Chapter 18 was the fact that neoplasms produce antigens to which the host mounts an immune response resulting in both humoral and cell-mediated reactions in target tissues. Similar “paraneoplastic” effects of antibodies and cell-mediated immunity induced by the production of antigens by neoplastic tissues include autoimmune hemolytic anemias occur- ring in association with a variety of carcinomas (Sokol et al., 1994) and chronic lymphocytic leukemia (cf. Kipps and Carson, 1993); autoimmune mucocutaneous disease associated with skin blistering and ulceration resulting from autoantibodies to keratin, also associated with lym- phomas (Anhalt et al., 1990); and vasculitis resulting from antibodies to antigens within the vas- cular system (Carsons, 1997). It is estimated that 5% of patients with vasculitis have a malignancy. There have been suggestions that these autoimmune phenomena seen in the host–tumor relationship could actually be utilized to induce a self immunity to the neoplasm provided one could select the appropriate antigen as the target for the host immune system (cf. Nanda and Sercarz, 1995; Pardoll, 1999). However, a variety of other techniques have been uti- lized in an attempt to take advantage of the peculiar immunobiology of the host–tumor relation- ship in attempts to prevent, cure, or eliminate the neoplastic disease in the host.