One of the major mechanisms by which the immune response deals with foreign or abnormal cells is to damage or destroy them. Such immunologic cytotoxicity may lead to complete loss of viability of the target cells (cytolysis) or an inhibition of the ability of the cells to continue growing (cytostasis). Immunologic cytotoxicity can be manifested against a wide variety of target cells. These include malignant cells, notmal cells from individuals unrelated to the responding host, and normal cells of the host that are infected with viruses or other microorganisms. In addition, the immune system can cause direct cytotoxic effects on some microorganisms, including bacteria, parasites, and fungi. Immunologic cytotoxicity is a principal mechanism by which the immune response copes with and often eliminates foreign materials or abnormal cells. Cytotoxic reactions are frequently observed as a major component of an immune response that develops following exposure to foreign cells or microorganisms. In addition, there is increasing recent evidence that cytotoxic reactions represent a major mechanism for natural immunity and resistance to such materials. In most instances, cytotoxicity by immune components involves the recognition of particular structures on the target cells. Additionally, the targets need to be susceptible to attack by the immune components. Some cells are quite resistant to immunologic cytotoxicity. This appears to represent a major mechanism by which they can escape control by the immune system.
There are a variety of mechanisms for immunologic cytotoxicity. The two main categories are antibodyand cell-mediated cytotoxicity. Within cell-mediated cytotoxicity, multiple effector cell types and mechanisms can be involved. This review will focus on cytotoxicity by lymphocytes. Furthermore, because my own expertise is largely restricted to natural killer (NK) cells and related effector cells, and since the characteristics and functions of cytotoxic T-lymphocytes (CTLs) have been extensively reviewed elsewhere,1 this article will emphasize information related to NK cells and discuss the similarities and differences in comparison with CTL.
TYPES OF LYMPHOCYTIC EFFECTOR CELLS AND THEIR CHARACTERISTICS
Immune T cells are the best known effectors of cellmediated cytotoxicity. Upon immunization with cells that differ from the host in their major histocompatibility complex (MHC) antigens and are therefore recognized as foreign, cytotoxic T cells are generated which have a potent ability to lyse target cells bearing the foreign MHC antigens. In addition, cytotoxic T cells can be generated against a variety of other foreign antigenic structures on the surface of cells, including antigens associated with tumors of the host and host cells infected with various microorganisms, particularly viruses. In general, cytotoxicity against cells with these other antigens is only displayed when the cells share the same MHC antigens with the responding host.
GENERAL CHARACTERISTICS OF CYTOTOXIC T CELLS AND NATURAL KILLER CELLS
A summary of some of the general characteristics of cytotoxic T cells is given in the Table. CTLs are typical small lymphocytes, although they may be larger and blastic in appearance during the early phases of induction in vitro or in vivo. They lack readily detectable azurophilic granules in their cytoplasm; occasionally, small numbers of such granules have been described, especially upon electron microscopic examination.2 In contrast, some clones of CTLs, maintained in culture in the presence of interleukin 2 (IL-2), have been found to have prominent azurophilic granules in their cytoplasm, very similar to those observed in large granular lymphocytes. Most CTLs have a characteristic cell surface phenotype, expressing T cell lineage markers such as CD3 in humans. Characteristically, CTLs are within the CD4 ~/CD8 + subpopulation with a phenotype of suppressor T cells, yet it seems likely that CTLs represent only a portion of T cells with these markers. However, to date there have been no cell surface antigens identified which are entirely selective for the CTL.
In regard to functional characteristics, CTLs are generally not detectable in the lymphoid organs of normal individuals. They must be induced by in vitro stimulation with the appropriate antigens or mitogens, or by in vivo immunization. In most instances, the development of CTL activity takes considerable time, with primary responses, even to alloantigens or other strong antigens, becoming detectable only after about 1 week. A characteristic feature of CTLs, as well as of other immune T cells, is to manifest memory, with more rapid induction of cytolytic activity in cells which have been previously exposed to the antigen.
Natural killer cells were discovered about 14 years ago3,4 during studies of cell-mediated cytotoxicity. Although investigators expected to find specific cytotoxic activity of tumor-bearing individuals against autologous tumor cells or against allogeneic tumors of similar or the same histologic type, appreciable cytotoxic activity was observed with lymphocytes from normal individuals.
Until recently, the cells responsible for NK activity could be defined only in a negative way, ie, by distinguishing them from typical T cells, B cells, or mononuclear phagocytes (MNPs). However, it is now possible to isolate highly enriched populations and show that the NK activity is closely associated with a subpopulation of lymphocytes, morphologically identified as large granular lymphocytes (LGL) (Table). LGLs comprise about 5% of peripheral blood lymphocytes and 1% to 3% of all mononuclear cells.5 LGLs, which contain azurophilic cytoplasmic granules, can be isolated by discontinuous density grathent centrifugation on Percoli.® LGLs are nonphagocytic, nonadherent cells that lack surface immunoglobulin or receptors for the third component of complement but display cell surface receptors for the Fc fragment of IgG.5 This latter quality allows them to bind antibodycoated target cells and mediate the phenomenon termed antibody-dependent cellular cytotoxicity (ADCC), a function previously attributed to the K (killer) cell.6 Although the same cells, ie, NK/K cells, seem able to mediate both forms of cytotoxity, NK activity is due to NK receptors discrete from the Fc receptors of K cells that interact with target cell-bound antibody.6,7
Some cell surface antigens, particularly those detectable by monoclonal antibodies, have been found on virtually all NK cells and therefore help to characterize the phenotype of these effector cells. Thus, most human NK cells react with the following monoclonal antibodies: B73.1, 3G8, Leu ll.8 The CD16 Fc receptors for IgG on LGLs react with 3G8 and Leu 11, also strongly expressed on polymorphonuclear neutrophils (PMNs).
Rabbit antisera to the glycolipid asialo GM, react well with NK cells and also with monocytes (MNPs) and granulocytes (PMNs).9 Monoclonal antibody OKTlO reacts with NKs and most thymocytes and activated lymphocytes.9 Monoclonal antibody OKMl reacts with NKs, MNPs, PMNs and platelets,9 while monoclonal antibody NKHl or Leu 19 which is quite selective for NK cells, is also expressed on a small percentage of T cells. Removal of cells bearing any of these markers, either by treatment with antibody plus complement or by negative selection immunoaffinity procedures, results in depletion of most or all detectable NK activity.
NK cells can also be characterized by a lack of expression of certain cell surface markers. For example, human NK cells have no detectable surface reactivity with monoclonal antibodies to pan-T-cell antigens such as CD5 or CD3, or to the CD4 T-helper antigen, nor do they express surface antigens detected by a number of monocyte-specific reagents such as MO2 and Leu ML.7-9
REGULATION OF CYTOTOXIC ACTIVITY OF EFFECTOR CELLS
In vivo treatment of allosensitized mice with interferon (IFN) has been reported to augment the reactivity of CTL. l0 However, studies in my laboratory have failed to confirm these results (unpublished observation) and the activity of human CTL, after in vitro sensitization in mixed lymphocyte cultures, was not boosted by IFN.11 In contrast, the NK-like activity which develops in such cultures was augmented by IFN.1 - Thus it appears that the human CTL-mediating alloimmune cytotoxicity are resistant to augmentation by IFN.
In contrast to the largely negative results with CTL, all three types of IFN (a, ß, *y), have been shown to augment the activity of NK cells potently.7,11·12 In vivo administration to mice or rats of a variety of IFNinducers, or of IFN itself, led to rapid boosting of NK activity.13 Similarly, incubation in vitro of lymphoid cells or of purified LGL with IFN induced considerable augmentation of NK activity.7,11-12
Although as discussed above, IFN does not appear to affect the activity of CTL, it can influence the level of activity generated from CTL precursors. Thus, the addition of IFN at the time of initiation of human mixed lymphocyte cultures resulted in the generation of considerably higher levels of CTL activity than in cultures without IFN. However, the mechanism responsible for this more efficient generation of CTL has not been determined.
Since the discovery by Morgan et al14 that human T cells could be maintained in continuous culture in the presence of T cell growth factor (TCGF), an extensive amount of research has been performed on the role of this factor in the growth and regulation of T cells. TCGF, now know as interleukin 2 (IL-2), is a key intermediary in the lymphoproliferative responses of T cells and the generation of CTL. 15,16 The response of T cells to IL-2 depends on the activation of the cells to express cell surface receptors for IL-2, and the human IL-2 receptor on T cells has been found to react with a monoclonal antibody, anti-Tac.17 The anti-Tac antibody has been found to inhibit the binding of radiolabeled IL-2 to T cell lines or to activated T cells expressing Tac, thereby strongly inhibiting the proliferative responses to IL-2.17
Lymphocytes sensitized in mixed lymphocyte cultures (MLC) represent primed cells that undergo rapid proliferation in response to the specific alloantigens during a secondary culture period, referred to as the primed lymphocyte test. By the addition of IL-2 at the end of the MLC, cultured T cells (MLC-CTL) could be generated which retained specific proliferative reactivity in primed lymphocyte tests.18 In most cases, the MLC-CTL also retained strong cytotoxic activity against the priming antigens, raising the possibility of growing large quantities of CTL, with reactivity to well-defined alloantigens. 19
In addition to the ability of IL-2 to promote the proliferation of T cells, IL-2 has been found to stimulate strongly the proliferation and cytotoxic activity of NK cells.20'21 Some intriguing evidence has even suggested that endogenous production of IL-2 may be responsible, or even required, for maintenance of spontaneous human NK activity. Preincubation of human LGL with monoclonal antibodies to IL-2 for about 20 hours resulted in a complete loss of NK activity, which could be reversed by addition of IL-2.21
The IL-2 boosting of NK activity appears to result from a direct interaction of the Iymphokine with LGL.22 This boosting was observed with Tac-negative, highly purified preparations of LGL and was not abrogated by monoclonal anti-Tac antibodies. Thus, it appears that the Tac antigen is not involved in this immunoregulatory effect of IL-2 on NK cells, whereas it is required for IL-2-induced proliferation of LGL as well as T cells. These results suggest that LGL may express receptors for IL-2 which are distinct from Tac and which mediate the boosting of cytotoxic reactivity.22
Recently, lymphokine-activated killer (LAK) cells have been described23 that share many of the characteristics of the previously described culture-activated NK-like cells. LAK cells have been activated after a short period of culture in vivo with highly purified IL-2 and display cytotoxic activity against a variety of autologous, allogeneic, and xenogeneic tumors. These cells were initially thought to lack markers typical of fresh NK cells, to be devoid of cytolytic activity prior to culture, but to develop T cell markers upon activation. 24 However, more recent studies in several laboratories have indicated that most LAK activity developing from blood or splenic lymphocytes is attributable to IL-2 -stimulated NK cells, and that in fact, most LAK cells and their progenitors have a phenotype characteristic of NK cells but not T cells.25
INTERACTION BETWEEN EFFECTOR AND TARGET CELLS WHICH LEAD TO CYTOTOXICITY
As a result of efforts to dissect the mechanisms involved in the interactions between cytotoxic effector cells and target cells, it has been possible to define a sequence of events which appear to be required for the lytic process. These steps were first described for CTL, 1 and subsequently a similar sequence has been shown to be involved in NK activity. 7 The main stages can be identified as: 1) recognition of target cells by effector cells; 2) binding of effectors to targets; 3) activation of lytic machinery of effector cells; 4) lytic effects on target cells, often referred to as the "lytic hit"; and 5) effector cell-independent dissolution of the targets. In the following discussion, some of the main features of each of these stages will be summarized, with particular emphasis on similarities and differences between CTL and NK cells.
Recognition of Target Cells by Effector Cells
The recognition event appears to be dependent on two types of structures, the cell surface recognition receptors on the effector cells, and the antigens or other structures on the target cells which need to be recognized.
Recognition Structures on Effector Cells. Much attention has been devoted to the nature of the T cell receptor, and there have been some recent exciting advances in our understanding of the biochemical and molecular biologic features of this receptor. In contrast, there has as yet been little documentation of the nature of the recognition receptors on NK cells. A series of studies were performed to determine whether NK cells have structures directly analogous to the T cell receptor. Most investigators now agree that highly purified NK cells show no evidence of T cell receptor expression, indicating that NK activity is dependent on a different type of receptor.26
Target Structures Recognized by Effector Cells. In most instances, the recognition of target cells by CTL is restricted by the major histocompatibility complex (MHC), and the target structures themselves seem to be closely associated with the MHC.1 It has been proposed that MHC proteins on target cells are critical molecular mediators by which the CTL determinants transmit signals required for the subsequent lysis of the target cells.1
In contrast to the extensive studies oh the target structures recognized by CTL, little definitive information exists about the nature of the target structures recognized by NK cells.
Post-Recognition Steps Leading to Lysis of Target Cells
After the requisite recognition of target cells by effectors and their binding together to form conjugates, a complex series of events are initiated which lead to lysis of the target cells. Many of the early biochemical changes in the effector cells, including a calcium-dependent step, have been found to be similar for CTL and NK cells.1,27 However, the actual mechanism underlying the lethal hir has been difficult to identify. There have been some suggestions that both CTL activity and NK activity are dependent on the generation of reactive oxygen intermediates, but subsequent studies have essentially ruled out this mechanism.28
A recent major advance in our understanding of the possible mechanism involved in lysis by NK cells has come from studies with cytoplasmic granules isolated from LGL.29 Observations of human LGL-target cell conjugates by electron microscopy indicated the secretion of granules in close proximity to the target cell membrane. To evaluate the possible role of such granules in the lytic mechanism, they were isolated from rat LGL Ieukemias with NK activity. Granules containing ß-glucuronidase and other lysosomal enzymes were purified on Percoli grathents and shown to have potent calcium-dependent lytic activity against tumor target cells and a wide variety of normal cells, particularly sheep erythrocytes.29 A potent protein cytolysin could be solubilized from the LGL granules. It was possible to isolate such highly cytolytic granules from normal rat and human LGL as well as from the rat leukemic LGL, whereas lysosomal granules isolated from liver, granulocytes, and a variety of other cell types had no detectable cytolytic activity.
The granule-induced lytic activity was found to be associated with the formation of ring-like structures on the surface of the target cells, detectable by electron microscopy.29 The LGL granules and the solubilized granule cytolysin could also be shown to lyse liposomes. The affected liposomes have pore-like lesions which were demonstrated by electron microscopy to be cylindrical structures, 15 nm in diameter, and quite similar to those previously observed on target cells lysed by NK cells or by K cells.29 The appearance and other features of the lesions were very similar to, but about twice the diameter of, the previously described channels produced by the membrane attack complex of complement. It is also of note that the ringlike structures associated with cytotoxicity by LGL are very similar to those described on target cells affected by CTL.30
Taken together, the available evidence fits well into the following sequence: NK cells, and perhaps also CTL, upon recognition of and binding to target cells are triggered to release the granule-associated cytolysin, which can insert into the target cell membrane and lead to the death of the targets.
Another somewhat similar pathway for the mechanism of cytotoxicity by NK cells has been described involving the release of a soluble cytotoxic factor, termed NK cytotoxic factor (NKCF), from NK cells upon stimulation by NK-susceptible target cells or lectins.28 NKCF has been shown to bind selectively to NK-susceptible target cells and then cause their lysis, but with kinetics considerably slower than those observed with intact NK cells or the LGL granule cytolysis. Although the difference in kinetics stands out as a major divergent point, most other characteristics related to NKCF have been very similar to those associated with NK activity. For example, a variety of inhibitors which were previously demonstrated to inhibit NK activity were also shown to inhibit the activity of NKCF. Initial attempts at purification have indicated that NKCF has apparent molecular weights of 18,000 and 36,000. The possible relationship between NKCF and the granule cytolysin is currently unclear.
Recently, studies have been performed to determine the possible relationship of NKCF to other known cytotoxic molecules. It has been possible to demonstrate that NKCF is distinct from tumor necrosis factor, lymphotoxin, or leukoregulin31 and thus appears to be a quite novel cytotoxic molecule. In contrast, mouse NC activity has been found to be dependent on the release of tumor necrosis factor upon interaction of NC cells with their target cells.32
Many of the features of CTL and NK cells and their lytic activity appear to be very similar. The main aspect of divergence appears to be related to their recognition receptors. It is possible that they are derived from two distinct lineages of effector cells, which simply share some properties. However, an alternative possibility is that NK cells represent an alternative thymic-independent pathway of T cell differentiation, and that they have basically the same lytic machinery as CTL but operate witli different recognition receptors for non-MHC-associated target structures. 33 Further detailed studies on the ontogeny and pathways of differentiation of NK cells and possible similarities in their lytic mechanisms may help to discriminate between these possibilities.
1. Berfce G: Cytolytic T- lymphocytes: How do they function.' Immunol Rev 1983; 72:5-42.
2. Zagury D: Direct analysis of individual killer T celb. Susceptibility of target cells to lysis and secretion of hydrolytic enzymes by CTL, in Clark WR. Goldstein P (eds): Mechanisms of Cell-Mediated Oitotoxirirx New York. Plenum Press, 1982; pp 149-156.
3. Rosenberg RB, Herberman RB, Levine PH, et al: Lymphocyte cytotoxicity reactions to leukemia-associated antiKens in identical twins. Int J Cancer 1972; 9:648.
4. Herberman RB. Ntmn ME, Lavrin DH: Natural cytotoxic reactivity of mouse lymphoid cells against syngeneic and allogeneic tumors. I. Distribution of reactivity and specificity, Int J Cancer 1975; 16:216.
5. Timonen T. Ortaldo JR, Herberman RB: Analysis by a single cell cytotoxicity assay of natural killer (NK) cell frequencies among human large granular lymphocytes and of the effects of interferon on their activity. J Immunol 1982; 128:2514-2521.
6. Kay HD, Bonnard GD, Wïst WH, et al: A functional comparison of human Fcreceptot-beanng lymphocytes active in natural cytotoxicity and antibody-dependent cellular cytotoxicity. I Immunol 1977; 118:2058-2066.
7. Herberman RB (ed): NK Cells and Omer Natural Effector Cells. New YoA. Academic Press, 1982.
8. Perussia D. Acuto D. Terhorst C et al: Human NK celb analyzed by B73.I, a monoclonal antibody blocking Fc receptor function. IL Studies of B73. 1 antibodyantigen interaction on the lymphocyte membrane. J Immunol 1983; 130:2142.
9. Ortaldo JR, Shanow SO, Timonen T, et al: Determination of surface antigens on highly purified human NK celts by flow cytometry with monoclonal antibodies. J Immunol 1981; 127:2401-2409.
10. Lindahl P, Leary P, Gresser 1: Enhancement by inteferon of the specific cytotoxicity of seratized lymphocytes. Proc Nod Acad Set USA 1972; 69:721-725.
11. Herberman RB, Ortaldo JR, Djeu JY, et al: Role of interferon in regulation of cytotoxicity by natural killer cells and macrophages. Ann NY Acad Sa 1980; 350:63.
12. Reynolds CW, Timonen T, Holde HT, et al: Natural killer (NK) cell activity in the rat: Analysis of effector cell morphology and effects of interferon on NK cell function in the athymic (nude) rat. Eur) Immunol 1982; 12:577-582.
13. OehlerjR, Lindsay LR. Nunn ME, et al: Natural cell-mediated cytotoxicity in rats. II. In vivo augmentation of NK-cell activity. JmJ Cancer 1978; 21:210-220.
14. Morgan DA, Ruscetti FW, Gallo RC: Selective in vitro growth of T lymphocytes from normal human bone marrows. Science 1976; 193:1007.
15. Bonnard GD, Yasaka K, Jacobson D: Ligand-activated T cell growth factor-induced proliferation. Absorption of T cell gnwth (acuir by activated T cells. ) Immunol 197°; 12 3:2704-2708.
16. Farrar WL, Johnson HM , Farrar JJ : Regulation of the production of immune interferon and cytotoxic T lymphocytes by imerleukin 2. 1 Immunol 1981; 126:1120.
17. Robb RJ, Greene WC: Direct demonstration of the identity of T cell growth factor binding protein and the Tac antigen. ) Exp Med 1983; 158:1332.
18. Bonnard GD: Long-term cultures of immunocompetent T lymphocytes, in Sell KW, Miller WV (eds): TJ* Lymphocyte. New York, Alan R Liss. 1981. ? 45.
19. Sehende! DJ1 Wank R, Bonnard GD: Genetic specificity of primary and secondary proliferative and cytotoxic responses of human lymphocytes grown in conditioned culture. Stand J Immunol 1980; 11:99-107.
20. Kedar BL. Ikerjiri B, Sredni B, et al: Propagation of mouse cytotoxic clones with characteristics of natural killer (NK) cells. Cell Immunol 1982;69:305-329.
21. Domzig W, Stadler BM, Herberman RB: lnterleukin-2 dependent of human natural killer (NK) cell activity. ) Immunol 1983; 130:1970-1983.
22. Ortaldo JR, Mason AT, Gerard JP, et al; Effects of natural and recombinant IL-2 on regulation of IFN production and natural killer activity: Lack of involvement of the Tac antigen for these immunoregulatory effects. J Immunol 1984; 133:779-783.
23. Grimm EA, Mazumder HZ, Zhang S, et al: Lymphokine-activated killer cell phenomenon, lysis of natural killer cell resistant fresh solid tumor cells by interleukin 2activated autologous human peripheral blood lymphocytes. / Exp Med 1982; 15:823-830.
24. Grimm EA1 Ramsey KM, Mazumder A, et al: Lymphokine-activated killer cell phenomenon. II. Precursor phenotype is serologically distinct from peripheral Tlymphocytes, memory cytotoxic thymus-derived lymphocytes and natural killer cells. J Exp Med 1983; 157:884-897.
25. HenSetman RB: Adoptive therapy of cancer with interleukin 2 UL-2) activated killer cells. Cancer Bulletin 1986; in press.
26. Young HA, Ortaldo JR, Herberman RB, et al: Analysis of T cell receptors in highly purified rat and human large granular lymphisytes (LGL): Lack of functional 1. 3 kbchain m-RNA. J Immunol 1986; 136:2701-2704.
27. Clark WR, Golstein P (eds): Mechanisms of CeU-Mediaied Cytuü>xkity. New York, Plenum Press. 1982.
28. Herberman RB, Callewaert D (eds): Mechanisms of Ouxinicu-v bi NK Oils. New York, Academic Press, 1985.
29. Henkart PA, Millard PJ, Reynolds CW, et al: Cytolytic activity of punfied cytoplasmic granules from cytotoxic rat LGL tumors. J Exp Med 1984; 160:75-93.
30. Dennett G. Rxlack ER: Cytolysis by H-2 specific T killer cells. ) Exp Med 1983: 157:1483.
31. Herberman RB, Reynolds CW, Ortaldo JR: Mechanisms erf cytotoxicity by natural killer (NK) cells. Annu Rev Immunol 1986; 4:651-680.
32. Ortaldo JR, Mason LH, Mathieson BL, et al: Mediation of mouse natural cytotoxic activity by tumor necrosis factor. Nature 1986; 321:700-702.
33. Grossman Z, Herberman RB: Natural killer cells and their relationship to T cells: Hypothesis on the role of T-cell receptor gene rearrangement on the coutse of adaptive differentiation. Cancer Res 1986; 46:2651-2658.
GENERAL CHARACTERISTICS OF CYTOTOXIC T CELLS AND NATURAL KILLER CELLS