Regulation of tumour immunity by CD25⁺ T cells

Introduction

The impact of the immune system on tumour growth in humans is poorly understood. Several studies have, however, identified T cells, in particular CD8⁺ cytotoxic T lymphocytes (CTLs), that recognize antigenic peptides expressed on autologous tumour cells.^1,2 These tumour antigens include tumour-specific antigens that arise through mutations in normal genes; shared tumour antigens such as cancer–testis antigens; differentiation antigens; and other self-antigens that may be over-expressed in malignant cells. In some cases, where malignancies arise as a result of virus infection, tumour cells may also express viral antigens. Although antigens of each category have been found to act as targets for T-cell responses in human patients, the ability of these T cells to promote tumour rejection remains unproven.³

Experimental evidence obtained using murine models indicates that either CD4⁺ or CD8⁺ T-cell responses to certain tumour-associated self-antigens can indeed result in tumour rejection.^4–7 Melanocyte differentiation antigens are the best-characterized self-antigens that act as targets for tumour (melanoma) rejection.⁸ Vigorous immune responses to melanocyte antigens have been shown to result, not only in tumour regression, but also in the development of autoimmune vitiligo. Vitiligo is a disease characterized by skin depigmentation following the destruction of non-malignant melanocytes, often by T cells.⁹ Overall, these findings indicate that the same effector T cells can be involved in both autoimmunity and tumour immunity. Thus, mechanisms of peripheral tolerance, which inhibit the activation of autoreactive T cells, also impinge upon the induction of T-cell responses to tumour-associated self-antigens.

Several mechanisms account for the lack of reactivity of T cells specific for self-antigens. Many self-reactive T cells are deleted by mechanisms of negative selection in the thymus.^10,11 Subsequently, some T cells that escape this process die following encounter with antigen in the periphery, whilst others survive but become functionally inactivated (anergic).¹² Other self-reactive T cells, residing in lymphoid tissues, may be oblivious to self-antigens expressed in peripheral sites¹³ particularly in the absence of any pro-inflammatory stimuli.¹⁴ When, however, these cells do meet self-antigens through, for example, infection or tissue injury, other pathways may also be called upon to limit the activation of such cells. An increasing body of evidence indicates that the immunomodulatory potential of a population of CD4⁺ CD25⁺ T cells, named CD25⁺ regulatory cells, represents one such pathway.^15,16

CD25⁺ Regulatory T Cells and Tumour Immunity

CD25⁺ regulatory T cells comprise 5–10% of peripheral CD4⁺ T cells in naïve mice, and have been shown in several in vivo murine models to prevent the induction of autoimmune disease^17–19 and inflammatory disease.²⁰ Since, as described above, tumour immunity is in part an autoimmune process, it was postulated that CD25⁺ regulatory cells would also inhibit the generation of immune responses to tumours.^21,22 Depletion of these cells, using CD25-specific monoclonal antibodies, has indeed been shown, in a variety of different mouse strains, to promote rejection of several transplantable, murine tumour cell lines including melanoma, fibrosarcoma, leukaemia and myeloma.^21–24 These studies imply that CD25⁺ regulatory cells normally inhibit the generation of effective anti-tumour immune responses in vivo (Fig. 1). Treatment of mice with CD25-specific antibodies may not however, represent the optimal means of inducing tumour immunity through regulatory cell depletion. The CD25 marker is not unique for regulatory T cells since following activation, this molecule, the interleukin-2 (IL-2) receptor α-chain, is also transiently expressed on conventional effector T cells. Antibodies specific for CD25 therefore deplete, not only regulatory cells, but also some effector cells that may be involved in tumour rejection. Although, in the murine models described above the beneficial effects of regulatory cell depletion outweighed the disadvantages of depleting some effector T cells, more specific targeting of the regulatory T-cell population should further improve the efficiency of tumour immunity (see below).

Several studies indicate that CD25⁺ regulatory cells inhibit the activation of both CD4⁺ and CD8⁺ T cells in vitro and that rejection of primary tumours in mice depleted of CD25⁺ regulatory cells is T-cell-dependent. In some cases (melanoma and a fibrosarcoma), both CD4⁺ and CD8⁺ T cells were found to be required for tumour rejection whilst only CD8⁺ T cells were found to be required for rejection of a murine myeloma.^21,23 It is possible that the role of CD4⁺ T cells is simply to behave as helper cells for the activation of CD8⁺ CTLs, particularly in the case where the tumour cells are poorly immunogenic. In tumour-inoculated mice depleted of CD25⁺ regulatory T cells, CD8⁺ T cells and CD4^– CD8^– cells exhibit tumour-specific and non-specific killing activity, respectively.²² The majority of such tumour non-specific CD4^– CD8^– effector cells are natural killer (NK) cells that kill a broad spectrum of tumour cell lines in vitro.²² Their cytotoxic action may be critical for tumour rejection since, at least in the case of melanoma, mice depleted of both CD25⁺ regulatory cells and NK1.1⁺ cells are unable to reject the tumours (A. Gallimore, unpublished data). NK cells can also be produced in vitro from normal spleen cells by simply culturing them after depleting CD25⁺ regulatory T cells.²² Depletion of these cells leads to spontaneous activation and proliferation of self-reactive CD4⁺ T cells recognizing self-peptides/major histocompatibility complex (MHC) on antigen-presenting cells. The large amount of IL-2 thus produced by CD4⁺ T cells as the result of this autologous mixed lymphocyte reaction leads to the development of NK cells as lymphokine-activated killer (LAK) cells from the CD4^– CD8^– population in the spleen. It is also possible that CD25⁺ regulatory cells directly inhibit NK-cell activation. NK cells, and perhaps other cells of the innate immune system, activated as a result of CD25⁺ regulatory cell depletion could further augment tumour-specific CD4⁺ and CD8⁺ T-cell responses through tumour destruction and by promoting inflammatory conditions at the site of tumour challenge.

Several studies indicate that rejection of primary tumours inoculated into mice depleted of CD25⁺ regulatory cells is accompanied by the ability to reject second tumours inoculated up to at least 3 months later.^21,23,24 As described above, in the case of melanoma, CD4⁺, CD8⁺ T cells and NK1.1⁺ cells are required for rejection of primary tumours, although CD4⁺ T cells alone mediate rejection of the same tumour when inoculated 2–3 months later.²³ These results indicate that at least in this particular experimental system, different types of effector cells are involved in rejection of primary and secondary tumours.

The studies described above indicate that removal of CD25⁺ regulatory cells can provoke tumour-specific immune responses. Two separate studies have shown that immune responses to melanocyte differentiation antigens are more easily detected in CD25⁺ regulatory cell-depleted, melanoma-immune mice than in normal, tumour-bearing mice.^23,24 A proportion of the former but not the latter group of mice, also develop signs of autoimmune vitiligo. It is not yet clear, however, whether or not specific responses to melanocyte self-antigens are responsible for tumour rejection. As discussed below, the suppression exerted by CD25⁺ regulatory cells is thought to be largely antigen non-specific, suggesting that depletion of the regulatory T cells could uncover responses to a spectrum of antigens expressed by the tumour cells. In addition, the depletion could, as described above, promote pro-inflammatory conditions at the site of tumour challenge, facilitating induction of immune responses to many types of tumour antigens.

The antigen-specificity of the regulatory T cells engaged in suppressing tumour immunity remains to be determined. Recent studies indicate that CD25⁺ regulatory cells, at least some of them, develop in the thymus following high-avidity interactions with self-peptide/MHC ligands expressed on cortical epithelial cells.^25,26 The cells also display various accessory molecules at high levels, being phenotypically similar to antigen-primed cells.²⁷ The high avidity of CD25⁺ regulatory cells to the selecting ligands due to the high affinity of T-cell receptors or high-level expression of accessory molecules, or both, is thought to render the cells highly responsive to further stimulation by self-antigens and thereby facilitate their rapid mobilization and effector function (suppression) at the site of tissue damage.²⁸ The physiological trigger(s) for activation of regulatory cells have not been identified in vivo although it is tempting, in the case of the tumour models described above, to speculate that the tumour cells may themselves be involved. Given that normal T-cell repertoire contains tumour-reactive T cells as well as self-reactive T cells, it should not be surprising that regulatory T cells also recognize a broad range of antigens including tumour-specific as well as tumour-associated antigens. The trigger for CD25⁺ regulatory cell activity may not reside exclusively (if at all) in specificity for tumour antigens but rather in the context in which any antigen is presented to the immune system. The tumour environment, which is thought to be immunosuppressive due to high local concentrations of anti-inflammatory cytokines such as transforming growth factor-β, may lend itself to the activation of CD25⁺ regulatory T cells, perhaps through preventing activation of dendritic cells (DCs) and therefore favouring activation of the regulatory cells over conventional effector cells. Indeed the results of several in vitro studies imply that immature but not mature DCs promote regulatory T-cell activation^29,30 (Fig. 2). It remains to be determined, however, whether such regulatory T cells are activated from the naturally present CD25⁺ regulatory T cells or have differentiated from naive T cells on immature DCs.

As mentioned above, in vitro data indicate that once activated, the suppression CD25⁺ regulatory T cells exert is antigen-non-specific,^31,32 and these cells could therefore inhibit immune responses to all tumour antigens regardless of whether these antigens are self-antigens or not. The precise mechanism(s) of action of CD25⁺ regulatory cells are unclear although experimental evidence exists to suggest that their inhibitory effects are mediated by the production of immunosuppressive cytokines such as transforming growth factor-β and IL-10^33,34 whilst other studies argue that the inhibitory effects of regulatory T cells are cell-contact dependent.^31,32,35 Neither mechanism has so far been investigated in the context of inhibition of tumour immunity.

T-cell accessory molecules play distinctive roles in activating CD25⁺ regulatory T cells or modifying their functions. For example, these cells constitutively express cytotoxic T-lymphocyte antigen-4 (CTLA-4; CD152). CTLA-4 is required for activation of the regulatory T cells, as blockade of CTLA-4 expressed on them abrogates suppression in vivo and in vitro.^20,36 CTLA-4 is also known to transduce negative signals in activated T cells, thereby leading to attenuation of T-cell responses. It is therefore likely that in vivo blockade of CTLA-4 may enhance tumour immunity by augmenting the activity of effector T cells as well. Glucocorticoid-induced tumour necrosis factor (TNF) receptor superfamily related gene (GITR), a member of the TNF receptor superfamily, is also predominantly expressed on CD25⁺ regulatory T cells.³⁷ In contrast to CTLA-4, ligation but not blockade of GITR molecules abrogates CD25⁺ T-cell-mediated suppression. GITR ligation also has co-stimulatory activity, enhancing effector activities of T cells. Thus, in contrast to anti-CD25 antibody treatment, which may eliminate not only regulatory T cells but also tumour effector T cells (see above), an advantage of anti-CTLA-4 antibody or anti-GITR antibody treatment is that it can attenuate T-cell regulation and enhance T-cell effector activity, synergistically promoting tumour immunity.

Concluding Remarks

CD4⁺ CD25⁺ T cells with phenotypic and functional properties similar to those described for murine regulatory cells have been identified in humans.^38–42 It will be very interesting therefore to discover whether or not these cells also prevent the generation of effective anti-tumour immune responses in cancer patients. The clinical benefits of treating patients with antibodies specific for CD25 must however, be approached with caution. Depletion of CD25⁺ cells could, for example, result in depletion of effector T cells important for controlling chronic infections with pathogens. Of greater concern is the possibility that depletion of regulatory cells could trigger the development of autoimmunity. So far, in the murine models described above, treatment with CD25-specific monoclonal antibodies, did not, in the majority of cases, result in the development of autoimmune disease. Development of autoimmunity following removal of regulatory cells may however, depend on several factors such as the individuals' genetic background, and/or infection with certain pathogens. Thus, as outlined above, many important questions regarding CD25⁺ regulatory cells remain to be answered before the immunotherapeutic potential of manipulating these cells in cancer treatment can be fully exploited.