As discussed earlier in this chapter, NK cells are likely derived in the bone marrow from a com- mon precursor with T cells. NK cells have clearly distinctive characteristics, both morphologi- cally and biochemically. While T cells carrying the T-cell receptor recognize and respond to MHC antigens bearing peptide epitopes, NK cells are inhibited in their activity and replication by such MHC complexes. This inhibition is mediated by receptors of somewhat limited diver- sity, that recognize distinct MHC class I antigen epitopes and are shared by a large number of MHC alleles (Seebach and Waneck, 1997; Raulet, 1999). Therefore many have argued that NK cells are the basis of host immunosurveillance, since the NK cell can play a major role in the host immunity to many neoplasms as well as certain virus infections. In these instances, the downregulation of the MHC class I and class II antigen would eliminate the inhibitory effect of these antigens on NK-cell activity toward such cells, thus allowing for the destruction of these cells. Studies have strongly suggested that such destruction occurs through the perforin pathway (Figure 19.19) (van den Broek et al., 1995). However, NK cells do produce cytokines including TNF-α and interferon-γ, the latter also involved in the activation of NK cells in the presence of other cytokines such as interleukin-2 (Naume and Espevik, 1994).
As might be expected, neoplastic cells with the greatest downregulation of MHC anti- gens—namely, metastatic lesions well into the stage of progression—are more susceptible to the effects of NK cells. This has been shown in genetic systems in mice that are deficient in NK cells (Talmadge et al., 1980), where metastases of transplanted neoplasms normally sensitive to natural killer activity were much more extensive than in wild-type mice. Other studies showed similar effects in in vitro and in vivo systems, even with human cells (Hanna, 1985). Thus, in a heterogeneous population of neoplastic cells in vivo, it would be those cells that are least differ- entiated, exhibiting the greatest downregulation of MHC antigens, that would be susceptible to NK cell–mediated lysis (cf. Uchida, 1986). However, animals with a large tumor burden exhibit impaired NK-cell cytotoxicity (Boom et al., 1988); in another model system, NK cells appar- ently play a significant role early in the growth of the neoplasm, fostering the generation of anti- neoplastic cytotoxic T cells (Kurosawa et al., 1995).
Macrophages, Dendritic Cells, and Antigen-Presenting Cells
Of the monocyte-macrophage-dendritic cell group, it is the dendritic cell that is most efficient at presenting antigen to naïve T cells and stimulating their proliferation in response to specific MHC-bound peptide epitopes. It is only in the last two decades that the dendritic cell has been distinguished from the monocyte found in the circulation and from the macrophage, a primarily phagocytic cell. While the latter two cell types presumably may act as antigen-presenting cells, they are 1% to 10% as effective as dendritic cells in antigen presentation. Dendritic cells are characterized morphologically by the presence of numerous membrane processes that can ex- tend for hundreds of micrometers from the cell. Such cells are far more effective at pinocytosis of soluble antigen than macrophages or monocytes, while the latter are more efficient at phago- cytosis of large particles such as cellular debris and bacteria (cf. Fong and Engleman, 2000). Dendritic cells also occur from a variety of hematopoietic lineages, including monocytes, granu- locytes, T cells, B cells, and others. They are found in a variety of tissues including the skin, intestinal mucosa, lymph nodes, and spleen. As expected, dendritic cells from different precur- sor lineages exhibit different populations of antigens and receptors on their surfaces (Björck,
1999). Thus, the principal role of the dendritic cell in the immunological host–tumor relation- ship is the processing of antigens and presentation to various T-cell populations, which would then interact with appropriate neoplastic cells provided they carry the MHC class I or class II protein antigens. However, there is substantial evidence that the less efficient antigen-presenting cells, particularly macrophages, have the ability to recognize neoplastic cells selectively and ei- ther destroy them or impede their growth (McBride, 1986; Fauve, 1993). Activation of macro- phages occurs through their interaction with cytokines, particularly interferon γ (cf. Killion and Fidler, 1998).
Macrophages occur within neoplasms in both animals and humans. Normann (1985), from a number of different published reports, suggested that intratumoral macrophage density was related to tumor growth, as seen diagrammatically in Figure 19.26. It appears that macrophage killing of neoplastic cells involves a process that can discriminate between neoplastic and nor- mal cells by a process independent of MHC antigens, tumor-specific antigens, the cell cycle, or the histogenetic phenotype of the transformed cell (cf. Whitworth et al., 1990). Neoplasms that are regressing, possibly through some degree of apoptosis, usually exhibit macrophages throughout the neoplastic tissue; however, in those neoplasms that are proliferating and continu- ing in the stage of progression, especially sarcomas, macrophages are confined to the periphery of the neoplasm (cf. Killion and Fidler, 1998). Normann (1985) reported that the induction time of neoplasms in rodents varies with their macrophage content; the more rapidly growing neo- plasms are those containing fewer infiltrating macrophages. However, the accumulation of mac- rophages in neoplasms does not necessarily correlate with the metastatic properties or the immunogenicity of the neoplasms. In some instances, the presence in the neoplasm of nonacti- vated (noncytotoxic) macrophages may actually enhance the growth of the neoplasm (cf. Killion and Fidler, 1998).
While the mechanism of this antineoplastic effect of macrophages is not understood, some evidence suggests that macrophage recognition of plasma membrane phosphatidylserine is at least correlated with an increased binding of macrophages to neoplastic cells. Since the mecha- nism for the specific recognition by activated macrophages of neoplastic cells is nonimmunolog- ical and requires cell-to-cell contact, it would appear that some membrane recognition
Figure 19.26 Diagrammatic representation of macrophage density within neoplasms during growth. Of the four phases indicated, phases I and II have been demonstrated for micrometastases and might be pre- sumed to exist for carcinoma in situ as well as early invasive neoplasms. Phases III and IV have been dem- onstrated for autochthonous and transplanted neoplasms and phase IV for metastases. (Adapted from Normann, 1985, with permission of the author and publisher.)
phenomenon of activated macrophages towards neoplastic cells is related to this immunological host–tumor relationship.
There have also been relationships between dendritic cells in human neoplasms and the prognosis of the lesion. Zeid and Muller (1993) demonstrated that well-differentiated squamous cell carcinomas of the lung contained a much higher density of a specific type of dendritic cell, the Langerhans cell, than did poorly differentiated lesions. Another interesting neoplasm in the human in which the immunocytic infiltrate is directly related to the prognosis is breast carci- noma displaying a considerable infiltration by lymphocytes and monocytes. Such neoplasms are termed medullary carcinomas (Figure 18.11). While the infiltrate contains a variety of T cells, both suppressor and cytotoxic, macrophages and NK cells are also seen in the intraneoplastic cell population (Naukkarinen and Syrjänen, 1990). In line with the relatively favorable progno- sis of such lesions, significant apoptosis of neoplastic cells was also demonstrated in medullary carcinomas of the breast (Grekou et al., 1996).
Dendritic Cells, Suppressors, and Ultraviolet Epidermal Carcinogenesis
As discussed in Chapter 13, ultraviolet radiation is a known carcinogen for both humans and animals. More than six decades ago, Roffo (1933) and Rusch et al. (1941) demonstrated that ultraviolet radiation, particularly UVB lying within the range of 275 to 315 nm, was carcino- genic for the epidermis of both mice and rats. Later studies by Kripke (1974) demonstrated the interesting phenomenon that in syngeneic mice, epidermoid carcinomas induced by UVB were quite immunogenic and could not be readily transplanted into nonirradiated syngeneic hosts. However, transplants into UV-irradiated hosts readily grew, even after relatively short exposure to UV radiation (Figure 19.27). Not shown in the figure is the fact that epidermal neoplasms developing after UV irradiation do not grow when transplanted into nonirradiated hosts (see above). This fact indicates that the host-immune response is quite effective against this particular neoplasm in this species. Other studies have also demonstrated that chronic UV irradiation en- hances carcinogenesis by benzo[a]pyrene-induced (Gensler, 1988) and N-methyl-N′-nitro-N-ni- trosoguanidine–induced skin cancers in similarly irradiated mice (Gensler, 1992). The latter experiment indicates that the chronic UV irradiation effect is not mediated through activation of the chemical carcinogen. Furthermore, UV radiation does not suppress the normal or inducible NK-cell activity in association with the appearance of the epidermal carcinomas in mice (Steer- enberg et al., 1997).
Cutaneous Immunity and Ultraviolet B Radiation
As expected, because of the strategic location of the skin as an interface with the environment, this organ has its own components of the immune system, as outlined diagrammatically in Fig- ure 19.28. Langerhans cells are virtually eliminated or markedly reduced from skin exposed to UVB radiation (Streilein et al., 1994; Thiers et al., 1984; Meunier, 1999). The inhibition of anti- gen-presenting activity of dendritic cells resulting from UV radiation in the skin appears to be dependent on UV-induced DNA damage in the cutaneous antigen-presenting cell population (Vink et al., 1997). Contact hypersensitivity to various antigens is lost from the area of skin that is subjected to the UV radiation (cf. Streilein et al., 1994). Such cells appear to have lost the ability to present antigens to T cells, resulting in the failure of the induction of Th1 responses but still allowing Th2 responses to occur (Simon et al., 1990; Cruz, 1992; cf. Meunier, 1999). A direct test of this conclusion is seen in Table 19.13. These studies then underline the importance of the role of Langerhans cells in the skin in presenting antigen, including tumor antigens, from
Figure 19.27 A. Time course for development of primary skin cancers and the susceptibility to the growth of transplanted skin cancers in C3H mice continuously exposed three times per week to UVB radi- ation. B. The percentage susceptibility of mice treated for 12 weeks (3 times per week) with UVB radiation to the growth of transplants of UV-induced neoplasms of the skin. (Adapted from Kripke, 1986, with per- mission of the author and publisher.)
epidermoid carcinomas to the immune system, with subsequent rejection of the transplant in the case of the UV-induced carcinomas.