Cytotoxic T Lymphocytes 

The main interest to molecular biology of cytotoxic T lymphocytes is probably mechanistic. These cells are able to destroy other cells. How do they achieve this? The 40-odd year-old history of this field started with immunology and defense, and the cellular phenomenology that was possible at the time. It now culminates with the realization that the long-sought mechanisms of cytotoxicity convey in fact molecular signals of cell death. Thus, fundamental research on cytotoxic T cells now goes beyond immunology, becoming part of the booming research effort on programmed cell death.

It was shown in 1960 that white peripheral blood cells from alloimmune dogs were able to lyse in vitro target cells bearing the corresponding alloantigens (1). We now know that many different cell types have this ability to destroy other cells, in particular lymphoid cells, and among these T (i.e., thymus-derived) lymphocytes. There are two varieties: CD4+ and CD8+ lymphocytes. When activated, both varieties can be cytotoxic through a Fas-based mechanism, and in addition CD8+ cells can be cytotoxic via a perforin-granzyme-based mechanism. These two mechanisms will be considered in more detail below. A main role of cytotoxic T cells is to lyse syngeneic cells presenting surface determinants originating from viruses or other intracellular microorganisms. These, when within cells, are out of reach of other immune system effectors. Destruction of infected cells destroys the microorganism or exposes it to other effector mechanisms. Cytotoxic T cells thus fight intracellular pathogens. Interestingly, cytotoxic T cells can also destroy normal, uninfected cells, such as Fas-bearing activated lymphocytes, thus in part ensuring down-regulation of immune responses. Neither the perforin- nor the Fas-based mechanisms in isolation, nor both together, seem to account fully for graft rejection.

 When an effector cell approaches a target cell, it “recognizes” it through an array of diverse molecules at the surface of the effector cell (TCR/CD3/peptide, CD4 or CD8, LFA-1, etc.) and of the target cell (Class I or Class II MHC, ICAM-1, etc.). How, then, does the cytotoxic cell destroy the target cell? First indications were obtained at a phenomenological level. Engagement of the above-mentioned molecules leads to effector cell activation, including an increase in the concentration of free cytoplasmic calcium (2) within a matter of minutes, immediately followed by reorientation within the effector cell of the secretory apparatus towards an area facing the target cell (3). This suggested the possible participation of secretory phenomena in at least one mechanism of cytotoxicity. Many cytotoxic T cells indeed contain peculiar cytoplasmic granules. Their content can be released by degranulation in the extracellular fluid on target cell recognition.

 Among the phenomena then observed in the target cell is the early disintegration of nuclear DNA (4, 5) into fragments of around 180 bp, and multiples thereof, corresponding to the size of nucleosome-shielded DNA, which is due to the action of one or several endonucleases activated early in the sequence of events leading to target cell death. Other events in this sequence include an influx of calcium, condensation of the cytoplasm, fragmentation of the cytoplasm and nucleus, and only very late secondary membrane disruption. This phenomenology of cell death is very similar, if not identical, for example in developmental circumstances. In all cases, the same program seems to govern the course of events in a cell that dies, including, as we now know, a similar cascade of molecular events and the same “apoptotic” morphological traits (6). What may be different is merely the signaling of this death program. The nature of the signals produced by cytotoxic cells has been largely unraveled by studies at the molecular level.

1. The Perforin/Granzyme B Mechanism of Cytotoxicity

 A first approach to identify molecules involved in cytotoxicity has been to look for effector cell molecules themselves endowed with cytotoxic activity. This approach led to the detection (7, 8,(  characterization, and cloning of perforin, a 60-kDa protein present in granules in the cytoplasm of many cytotoxic T and NK (natural killer) cells. At high concentrations, it can act on membranes as a calcium-dependent channel-former; most interestingly, it shows significant homology to the terminal components of the complement system cascade. A formal demonstration that perforin was involved in a mechanism of cytotoxicity was provided by several groups (9), all showing that cytotoxicity mediated by T cells was greatly impaired in mice made perforin-deficient through gene targeting.

Other molecules present in granules of cytotoxic T cells were identified by subtractive cloning. Prominent among molecules isolated this way were a number of serine proteinases, such as CTLA-1/CCP1/Granzyme B (10-12) and H Factor/CTLA-3/Granzyme A (10, 12, 13). Granzyme B was demonstrated through gene targeting to be required in a mechanism of cytotoxicity (14).

 A current view on the cooperation between perforin and Granzyme B within the same mechanism of cytotoxicity is as follows. On recognition of the target cell, perforin- and granzyme B–containing granules in the effector cell reorient towards the target cell and are triggered to exocytose. Granzyme B may somehow enter the target cell and migrate into a target cell compartment where it would be innocuous, but from which it can be released by perforin at sublytic concentrations (15-17). Intracellularly released Granzyme B would then activate the standard programmed cell death cascade, through direct cleavage and therefore activation (18) of some of the ICE-family cysteine proteases called caspases. However, while target cell death mediated by Granzyme B (or at least by purified granules) may involve caspase activation for phenomena such as DNA fragmentation, it may, interestingly, use a noncaspase pathway for cell membrane disruption (19).

 2. The Fas-Based Mechanism of Cytotoxicity

 Through a somatic cell genetic approach it was found (20) that another mechanism of T cell–mediated cytotoxicity required a protein called Fas/APO-1/CD95 (21, 22) at the target cell surface. This in turn led to the cloning of the Fas ligand that is expressed at the effector cell surface (23). Fas belongs to a family of receptors (including Fas, TNF-R1, DR3, DR4, DR5), most or all of which are very efficient at transducing a signal interpreted as a cell death signal, while Fas ligand belongs to a family of corresponding ligands (including the Fas ligand, the TNFs, TRAIL, etc.). Involvement of Fas in cytotoxicity may be a reflection in vitro of the main roles of the Fas system in vivo, such as down-regulation of the immune response (24), protection against the immune system (25, 26), or potential major physiopathological effects (27).

Target cell Fas and effector cell Fas ligand define molecularly the Fas-based mechanism of cytotoxicity. In this mechanism, when an effector cell encounters a target cell, engagement of the T cell receptor of the effector cell by target cell major histocompatibility complex (MHC) leads to expression of the Fas ligand at the effector cell surface. Fas-ligand expression can be induced on CD8+ and on Th0 and Th1 CD4+ T cells (28-30) following specific recognition of antigen through the T cell receptor. This induction seems to involve the activation of tyrosine kinases, requires RNA and protein synthesis and the presence of calcium, and is inhibited by cyclosporin A (31-34). Other molecules may be involved, such as Myc and Max, 9-cis-retinoic acid, and its receptors, the orphan receptor for steroids, Nur77, and ALG-3.

Once expressed, effector cell Fas ligand engages target cell Fas, which leads to target cell death. In more detail, engagement of Fas leads in a matter of seconds to protein recruitment via the Fas “death domain,” the cytoplasmic segment of Fas that is necessary and sufficient to transduce a death signal (35, 36) . First, FADD/MORT-1 (37, 38) directly binds the death domain of Fas via its own C-terminal death domain. FADD/MORT-1 also includes an N-terminal death effector domain through which it associates with another molecule, FLICE/MACH (39, 40). FLICE/MACH includes two death effector domains at its N terminus and, strikingly, a caspase-homology domain at its C terminus (39, 40). Thus, the Fas-FADD-FLICE complex provides a remarkably direct link from a membrane signal to caspase activation, which is required for programmed cell death.

3. Concluding Remarks

 The perforin-based and the Fas-based pathways account for most, and perhaps all, of T cell–mediated cytotoxicity (41-44), at least as assessed in a 4-hr assay in vitro. Other molecules, such as TNF-R1 or the TRAIL receptors, may play a role in longer assays. The Fas pathway is used by both CD4+ (mostly Th1) and CD8+ effector cells, while the latter also use the perforin/granzyme pathway. Thus, in CD8+ T cells, two signals may stem from the T cell receptor/CD3 complex on antigen-specific recognition, one of them leading to granule exocytosis and the other one leading to transcription of the Fas ligand gene. Neither the exact nature of these distinct signals (45), nor whether they can be triggered simultaneously in a given cytotoxic cell is known as yet.

More generally, both mechanisms of T cell–mediated cytotoxicity seem to act by signaling the caspase activation step of the evolutionarily conserved programmed cell death cascade within the target cell. In this sense, the emergence in evolution of T cell–mediated cytotoxicity has not required the invention of new mechanisms of killing, but merely of new ways of signaling a preexisting programmed cell death cascade.

References

1. A. Govaerts (1960) J. Immunol. 85, 516–522. 

2. M. Poenie, R. Y. Tsien, and A.-M. Schmitt-Verhulst (1987) EMBO J. 6, 2223–2232. 

3. A. Kupfer, G. Dennert, and S. J. Singer (1985) J. Mol. Cell. Immunol. 2, 37–49. 

4. J. H. Russell, V. Masakovski, T. Rucinsky, and G. Phillips (1982) J. Immunol. 128, 2087–2094. 

5. J. J. Cohen et al. (1985) Adv. Exp. Med. Biol. 184, 493–508. 

6. J. F. R. Kerr, A. H. Wyllie, and A. R. Currie (1972) Br. J. Cancer 26, 239–257. 

7. P. A. Henkart (1985) Annu. Rev. Immunol. 3, 31–58. 

8. E. R. Podack (1985) Immunol. Today 6, 21–27. 

9. D. Kägi et al. (1994) Nature 369, 31–37. 

10. J.-F. Brunet et al. (1986) Nature 322, 268–271. 

11. C. G. Lobe et al. (1986) Science 232, 858–861. 

12. D. Masson and J. Tschopp (1987) Cell 49, 679–685. 

13. H. K. Gershenfeld and I. L. Weissman (1986) Science 232, 854–858. 

14. J. W. Heusel et al. (1994) Cell 76, 977–987. 

15. C. J. Froelich et al. (1996) J. Biol. Chem. 271, 29073–29079. 

16. D. A. Jans et al. (1996) J. Biol. Chem. 271, 30781–30789. 

17. L. Shi et al. (1997) J Exp. Med. 185, 855–866. 

18. A. J. Darmon, D. W. Nicholson, and R. C. Bleackley (1995) Nature 377, 446–448. 

19. A. Sarin et al. (1997) Immunity 6, 209–215. 

20. E. Rouvier, M.-F. Luciani, and P. Golstein (1993) J. exp. Med. 177, 195–200. 

21. S. Yonehara, A. Ishii, and M. Yonehara (1989) J. Exp. Med. 169, 1747–1756. 

22. B. C. Trauth et al. (1989) Science 245, 301–305. 

23. T. Suda, T. Takahashi, P. Golstein, and S. Nagata (1993) Cell 75, 1169–1178. 

24. J. H. Russell, B. Rush, C. Weaver, and R. Wang (1993) Proc. Nat. Acad. Sci. USA 90, 4409–4413.

25. D. Bellgrau et al. (1995) Nature 377, 630–632. 

26. T. S. Griffith et al. (1995) Science 270, 1189–1192. 

27. L. E. French and J. Tschopp (1997) Nat. Med. 3, 387–388. 

28. T. Suda et al. (1995) J. Immunol. 154, 3806–3813.

29. F. Ramsdell et al. (1994) Internat. Immunol. 6, 1545–1553. 

30. L. L. Carter and R. W. Dutton (1995) J. Immunol. 155, 1028–1031. 

31. A. Anel et al. (1994) Eur. J. Immunol. 24, 2469–2476. 

32. A. Anel et al. (1995) Eur. J. Immunol. 25, 3381–3387. 

33. F. Vignaux et al. (1995) J. Exp. Med. 181, 781–786. 

34. M.-F. Luciani and P. Golstein (1994) Roy. Soc. Phil. Trans. B 345, 303–309. 

35. N. Itoh and S. Nagata (1993) J. Biol. Chem. 268, 10932–10937. 

36. M. P. Boldin et al. (1995) J. Biol. Chem. 270, 387–391. 

37. M. P. Boldin et al. (1995) J. Biol. Chem. 270, 7795–7798. 

38. A. M. Chinnaiyan, K. O''Rourke, M. Tewari, and V. M. Dixit (1995) Cell 81, 505–512. 

39. M. P. Boldin, T. M. Goncharov, Y. V. Goltsev, and D. Wallach (1996) Cell 85, 803–815. 

40. M. Muzio et al. (1996) Cell 85, 817–827. 

41. D. Kägi et al. (1994) Science 265, 528–530. 

42. B. Lowin, M. Hahne, C. Mattmann, and J. Tschopp (1994) Nature 370, 650–652. 

43. H. Kojima et al. (1994) Immunity 1, 357–364. 

44. C. M. Walsh et al. (1994) Proc. Natl. Acad. Sci. USA 91, 10854–10858. 

45. M. T. Esser, B. Krishnamurthy, and V. L. Braciale (1996) J. Exp. Med. 183, 1697–1706.