- Email: [email protected]
CD25+ regulatory T cells and tumor immunity
Immunology Letters 85 (2003) 141 /143 www.elsevier.com/locate/
CD25 regulatory T cells and tumor immunity Emma Jones a, Michaela Dahm-Vicker a, Denise Golgher b, Awen Gallimore a, a
b
Nuffield Department of Medicine, John Radcliffe Hospital, Oxford, UK. The Cancer Sciences Division, University of Southampton, School of Medicine, Southampton, UK
Abstract Tumor cells express a range of antigens including self-antigens (those whose expression is shared by normal host tissue) and nonself antigens (such as those that arise as a result of mutations in normal cellular genes or in the case of some tumors, viral antigens). Immune responses to both types of antigen have been identified in human patients with cancer and in murine tumor models. In both cases, these responses are typically weak and generally fail to result in tumor rejection. Accumulating evidence indicates that a population of T cells, namely CD25 regulatory cells, is at least partly responsible for the poor immunogenicity of tumor cells. This evidence is discussed in the context of a murine model of melanoma. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Immunogenicity; Melanoma; Mutations; Tumor
1. Anti-tumor immunity and autoimmunity Immune responses to tumor-associated self-antigens are limited by mechanisms of immunological tolerance. Many self-reactive T cells are deleted in the thymus and those that persist in the periphery either exhibit too low affinity for their corresponding MHC/peptide complexes to impact significantly upon host tissue/tumor cells or are functionally silent [1 /3]. Under certain, but as yet poorly defined, conditions however, self-reactive T cells can be triggered to proliferate and target host tissue thereby indicating that the immune system does have the capacity to mount sufficiently high-affinity T cell responses against self-antigens to cause autoimmune disease. Conditions that provoke autoimmunity should also facilitate development of immune responses to selfantigens expressed on tumor cells.
2. CD25 Regulatory T cells and tumor immunity Recent in vivo studies, performed in murine models, have shown that CD25 T cells, which comprise 5/10% Corresponding author. Address: University of Wales College of Medicine, Heath Park, Cardiff, CF14 4XX Wales, UK. Tel.: /442920-745249; fax: /44-2920-744905 E-mail address: [email protected] (A. Gallimore).
of peripheral CD4 T cells in naive mice, prevent the induction of a variety of autoimmune diseases [4,5]. Since, as described above, tumor immunity can be an autoimmune process, the hypothesis that CD25 cells inhibit the generation of immune responses to tumors was tested. Depletion of CD25 regulatory T cells using monoclonal antibodies (mAb) specific for CD25, was indeed shown, in a variety of different mouse strains, to promote rejection of several transplantable, murine tumor cell-lines [6 /9]. Rejection of tumors in mice depleted of CD25 regulatory cells is T cell dependent although the requirement for CD8 T cells, CD4 T cells or both varies according to the tumor and probably the mouse strain used. In the case of the melanoma cell-line B16F10, both CD4 and CD8 T cells are required for tumor rejection in CD25-specific mAb-treated B6 mice [9]. Although CD25 regulatory T cells have been studied mainly in the context of autoimmunity, in vitro data has shown that, once activated, the cells may suppress both CD4 and CD8 T cells in an antigennon-specific fashion [4,5]. Thus, the T cells involved in tumor rejection may recognize both self- and non-self antigens. Shimizu et al. found that spleen cells from B6 mice depleted of CD25 regulatory T cells and inoculated with B16F10 exhibited tumor-non-specific killing activity characteristic of NK cells [6]. In order to determine
0165-2478/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 7 8 ( 0 2 ) 0 0 2 4 0 - 7
142
E. Jones et al. / Immunology Letters 85 (2003) 141 /143
Fig. 1. NK1.1 cells are required for rejection of B16F10 in vivo following depletion of CD25 cells. In each graph a single line represents an individual mouse. (a) Mice were injected i.p. with 1 mg isotype control mAbs (anti-rat b-gal) on days 3 and 1 prior to subcutaneous s.c. inoculation with 2 /104 B16F10 cells. As indicated, tumor growth was observed in 5 out of 5 mice. (b) Mice were injected i.p. with 1 mg NK1.1-specific mAbs [15] on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumor growth was observed in 5 out of 5 mice. (c) Mice were injected i.p. with 1 mg CD25-specific mAbs [16] on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumour growth was observed in 2 out of 5 mice. (d) Mice were injected i.p. with 1 mg NK1.1-specific and CD25-specific mAbs on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumor growth was observed in 5 out of 5 mice.
whether or not NK cells are essential for the tumor rejection observed in the absence of regulatory cells, we treated mice with mAbs specific for CD25 and NK1.1 prior to tumor inoculation. The results, shown in Fig. 1, indicate that NK1.1 cells are critical for rejection of B16F10 in mice depleted of CD25 regulatory T cells since in the absence of NK1.1 cells, all of the antiCD25-mAb-treated mice developed tumors. Experiments performed in vitro indicate that CD25 T cells influence NK cell activity rather indirectly since removal of the regulatory cell population from spleen cell cultures induces proliferation of CD4 T cells in response to self-MHC/peptide complexes and the large amount of IL-2 produced by these CD4 T cells subsequently promotes NK cell activity [6]. Despite these findings, it would be interesting to explore the possibility that CD25 regulatory cells directly inhibit NK cell activation in vivo.
inflammatory cytokines such as transforming growth factor-b (TGF-b) and interleukin-10 (IL-10) by the tumor cells [10,11]. We determined whether or not TGF-b and/or IL-10 were involved in promoting growth of the melanoma cell-line B16F10 in B6 mice. Mice were treated either with mAbs specific for TGF-b or the IL10 receptor (IL-10R) and subsequently inoculated with B16F10 cells. The results of the experiment, shown in Fig. 2a/d, indicate that, when compared to mice treated with control mAbs (a) and mice treated with CD25specific mAbs (b), inhibition of IL-10 binding had no
3. Inflammation, immunosuppression and tumor immunity The poor immunogenicity of tumor cells is exemplified by the general failure of these cells to induce T cell responses to foreign antigens that may have arisen naturally as a result of mutations or that are expressed as a result of experimental manipulation. Possible reasons for this lack of immunogenicity include a failure of the cells to activate neighbouring antigen presenting cells (APCs) such as dendritic cells (DCs) that may have picked up and processed antigens from dying tumor cells. The failure of tumor cells to activate APCs may arise, not only from lack of tumor-elicited pro-inflammatory signals but also from the expression of anti-
Fig. 2. Tumors can be rejected in vivo following neutralization of TGFb but not IL-10. Mice were either injected i.p. with 1 mg of isotype control mAbs (a), CD25-specific mAbs [16] (b), IL-10R-specific mAbs [17] (c) or TGF-b-specific mAbs [18] on days 3 and 1 prior to inoculation with 105 B16F10 cells. Each line represents an individual mouse and the numbers in parenthesis indicate the number of tumorfree mice.
E. Jones et al. / Immunology Letters 85 (2003) 141 /143
effect on tumor growth (c) whilst neutralization of TGFb resulted in slower growth and in some mice, complete rejection of the tumor cells (d). Several cell-types could be responsible for production of TGF-b in vivo including the tumor cells or infiltrating lymphocytes including CD25 regulatory T cells. Indeed, the results of some studies imply that CD25 regulatory cells exert their immunosuppressive effects through production of TGFb [12]. The source of TGF-b in the experimental model described here is currently under investigation.
143
their activation. Indeed, danger signals may promote conventional T cell activation, not only through activation of DCs, but also through inhibiting CD25 regulatory T cell activity. Thus, as for activation of conventional T cells, activation of CD25 regulatory T cells could be dependent upon the context in which antigens are presented rather than the ability of the cells to discriminate between self and non-self antigens.
Acknowledgements 4. What activates CD25
regulatory T cells?
The physiological trigger for the activation of CD25 regulatory cells in the periphery has not been identified. T cell receptor repertoire analyses have revealed no significant differences in Va or Vb gene usage between regulatory cells and other conventional T cells suggesting that these cells recognize a diversity of antigens in much the same way as conventional T cells [13]. CD25 regulatory T cells, unlike conventional T cells are, however, thought to recognize self-antigens with high affinity, thus CD25 regulatory cells may respond equally well to both foreign antigens and self-antigens. The ability to discriminate between self and foreign antigens may therefore not be the trigger for regulatory T cell activation. The trigger could relate rather to the immunosuppressive nature of the tumor environment. In a situation such as a tumor site, the immunosuppressive cytokines secreted by the tumor may prevent adequate activation of APCs that have picked up and processed tumor antigens, and favor stimulation of CD25 regulatory T cells over conventional T cells. As mentioned above, these CD25 regulatory T cells could potentially recognize both self- and non-self antigens expressed by the tumor cells. Due to the antigen non-specific nature of their immunosuppressive effects, all effector T cell responses to tumor-derived antigens would be inhibited. It has been proposed that conventional T cells are activated in response to ‘dangerous’ stimuli (e.g. heat shock proteins) rather than the ability to discriminate between self- and non-self antigens [14]. According to the hypothetical situation described above for the activation of CD25 regulatory T cells, these cells require no danger signals, but rather the opposite, for
This work was supported by The Wellcome Trust (grant no. GR056527MA).
References [1] C. Bouneaud, P. Kourilsky, P. Bousso, Immunity 13 (2000) 829 / 840. [2] J.W. Kappler, N. Roehm, P. Marrack, Cell 49 (1987) 273 /280. [3] F. Ramsdell, B.J. Fowlkes, Science 248 (1990) 1342 /1348. [4] E.M. Shevach, Annu. Rev. Immunol. 18 (2000) 423 /449. [5] S. Sakaguchi, Cell 101 (2000) 455 /458. [6] J. Shimizu, S. Yamazaki, S. Sakaguchi, J. Immunol. 163 (1999) 5211 /5218. [7] S. Onizuka, I. Tawara, J. Shimizu, S. Sakaguchi, T. Fujita, E. Nakayama, Cancer Res. 59 (1999) 3128 /3133. [8] Jones, E., Dahm-Vicker, M, Simon, A.K., Green, A., Powrie, F., Cerundolo, V., Gallimore, A., 2002. Cancer Immunity [serial online] 2002; 2:1. Available from http://www.cancerimmunity.org/ v2p1/020201.htm [9] R.P. Sutmuller, L.M. van Duivenvoorde, A. van Elsas, T.N. Schumacher, M.E. Wildenberg, J.P. Allison, R.E. Toes, R. Offringa, C.J. Melief, J. Exp. Med. 194 (2001) 823 /832. [10] H. Maeda, H. Kuwahara, Y. Ichimura, M. Ohtsuki, S. Kurakata, A. Shiraishi, J. Immunol. 155 (1995) 4926 /4932. [11] H. Maeda, A. Shiraishi, J. Immunol. 156 (1996) 73 /78. [12] F. Powrie, J. Carlino, M.W. Leach, S. Mauze, R.L. Coffman, J. Exp. Med. 183 (1996) 2669 /2674. [13] T. Takahashi, Y. Kuniyasu, M. Toda, N. Sakaguchi, M. Itoh, M. Iwata, J. Shimizu, S. Sakaguchi, Int. Immunol. 10 (1998) 1969 / 1980. [14] P. Matzinger, Annu. Rev. Immunol. 12 (1994) 991 /1045. [15] G.C. Koo, J.R. Peppard, Hybridoma 3 (1984) 301 /303. [16] J.W. Lowenthal, R.H. Zubler, M. Nabholz, H.R. MacDonald, Nature 315 (1985) 669 /672. [17] C. Asseman, S. Mauze, M.W. Leach, R.L. Coffman, F. Powrie, J. Exp. Med. 190 (1999) 995 /1004. [18] J.R. Dasch, D.R. Pace, W. Waegell, D. Inenaga, L. Ellingsworth, J. Immunol. 142 (1989) 1536 /1541.
CD25 regulatory T cells and tumor immunity Emma Jones a, Michaela Dahm-Vicker a, Denise Golgher b, Awen Gallimore a, a
b
Nuffield Department of Medicine, John Radcliffe Hospital, Oxford, UK. The Cancer Sciences Division, University of Southampton, School of Medicine, Southampton, UK
Abstract Tumor cells express a range of antigens including self-antigens (those whose expression is shared by normal host tissue) and nonself antigens (such as those that arise as a result of mutations in normal cellular genes or in the case of some tumors, viral antigens). Immune responses to both types of antigen have been identified in human patients with cancer and in murine tumor models. In both cases, these responses are typically weak and generally fail to result in tumor rejection. Accumulating evidence indicates that a population of T cells, namely CD25 regulatory cells, is at least partly responsible for the poor immunogenicity of tumor cells. This evidence is discussed in the context of a murine model of melanoma. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Immunogenicity; Melanoma; Mutations; Tumor
1. Anti-tumor immunity and autoimmunity Immune responses to tumor-associated self-antigens are limited by mechanisms of immunological tolerance. Many self-reactive T cells are deleted in the thymus and those that persist in the periphery either exhibit too low affinity for their corresponding MHC/peptide complexes to impact significantly upon host tissue/tumor cells or are functionally silent [1 /3]. Under certain, but as yet poorly defined, conditions however, self-reactive T cells can be triggered to proliferate and target host tissue thereby indicating that the immune system does have the capacity to mount sufficiently high-affinity T cell responses against self-antigens to cause autoimmune disease. Conditions that provoke autoimmunity should also facilitate development of immune responses to selfantigens expressed on tumor cells.
2. CD25 Regulatory T cells and tumor immunity Recent in vivo studies, performed in murine models, have shown that CD25 T cells, which comprise 5/10% Corresponding author. Address: University of Wales College of Medicine, Heath Park, Cardiff, CF14 4XX Wales, UK. Tel.: /442920-745249; fax: /44-2920-744905 E-mail address: [email protected] (A. Gallimore).
of peripheral CD4 T cells in naive mice, prevent the induction of a variety of autoimmune diseases [4,5]. Since, as described above, tumor immunity can be an autoimmune process, the hypothesis that CD25 cells inhibit the generation of immune responses to tumors was tested. Depletion of CD25 regulatory T cells using monoclonal antibodies (mAb) specific for CD25, was indeed shown, in a variety of different mouse strains, to promote rejection of several transplantable, murine tumor cell-lines [6 /9]. Rejection of tumors in mice depleted of CD25 regulatory cells is T cell dependent although the requirement for CD8 T cells, CD4 T cells or both varies according to the tumor and probably the mouse strain used. In the case of the melanoma cell-line B16F10, both CD4 and CD8 T cells are required for tumor rejection in CD25-specific mAb-treated B6 mice [9]. Although CD25 regulatory T cells have been studied mainly in the context of autoimmunity, in vitro data has shown that, once activated, the cells may suppress both CD4 and CD8 T cells in an antigennon-specific fashion [4,5]. Thus, the T cells involved in tumor rejection may recognize both self- and non-self antigens. Shimizu et al. found that spleen cells from B6 mice depleted of CD25 regulatory T cells and inoculated with B16F10 exhibited tumor-non-specific killing activity characteristic of NK cells [6]. In order to determine
0165-2478/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 2 4 7 8 ( 0 2 ) 0 0 2 4 0 - 7
142
E. Jones et al. / Immunology Letters 85 (2003) 141 /143
Fig. 1. NK1.1 cells are required for rejection of B16F10 in vivo following depletion of CD25 cells. In each graph a single line represents an individual mouse. (a) Mice were injected i.p. with 1 mg isotype control mAbs (anti-rat b-gal) on days 3 and 1 prior to subcutaneous s.c. inoculation with 2 /104 B16F10 cells. As indicated, tumor growth was observed in 5 out of 5 mice. (b) Mice were injected i.p. with 1 mg NK1.1-specific mAbs [15] on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumor growth was observed in 5 out of 5 mice. (c) Mice were injected i.p. with 1 mg CD25-specific mAbs [16] on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumour growth was observed in 2 out of 5 mice. (d) Mice were injected i.p. with 1 mg NK1.1-specific and CD25-specific mAbs on days 3 and 1 prior to s.c. inoculation with 2 /104 B16F10 cells. Tumor growth was observed in 5 out of 5 mice.
whether or not NK cells are essential for the tumor rejection observed in the absence of regulatory cells, we treated mice with mAbs specific for CD25 and NK1.1 prior to tumor inoculation. The results, shown in Fig. 1, indicate that NK1.1 cells are critical for rejection of B16F10 in mice depleted of CD25 regulatory T cells since in the absence of NK1.1 cells, all of the antiCD25-mAb-treated mice developed tumors. Experiments performed in vitro indicate that CD25 T cells influence NK cell activity rather indirectly since removal of the regulatory cell population from spleen cell cultures induces proliferation of CD4 T cells in response to self-MHC/peptide complexes and the large amount of IL-2 produced by these CD4 T cells subsequently promotes NK cell activity [6]. Despite these findings, it would be interesting to explore the possibility that CD25 regulatory cells directly inhibit NK cell activation in vivo.
inflammatory cytokines such as transforming growth factor-b (TGF-b) and interleukin-10 (IL-10) by the tumor cells [10,11]. We determined whether or not TGF-b and/or IL-10 were involved in promoting growth of the melanoma cell-line B16F10 in B6 mice. Mice were treated either with mAbs specific for TGF-b or the IL10 receptor (IL-10R) and subsequently inoculated with B16F10 cells. The results of the experiment, shown in Fig. 2a/d, indicate that, when compared to mice treated with control mAbs (a) and mice treated with CD25specific mAbs (b), inhibition of IL-10 binding had no
3. Inflammation, immunosuppression and tumor immunity The poor immunogenicity of tumor cells is exemplified by the general failure of these cells to induce T cell responses to foreign antigens that may have arisen naturally as a result of mutations or that are expressed as a result of experimental manipulation. Possible reasons for this lack of immunogenicity include a failure of the cells to activate neighbouring antigen presenting cells (APCs) such as dendritic cells (DCs) that may have picked up and processed antigens from dying tumor cells. The failure of tumor cells to activate APCs may arise, not only from lack of tumor-elicited pro-inflammatory signals but also from the expression of anti-
Fig. 2. Tumors can be rejected in vivo following neutralization of TGFb but not IL-10. Mice were either injected i.p. with 1 mg of isotype control mAbs (a), CD25-specific mAbs [16] (b), IL-10R-specific mAbs [17] (c) or TGF-b-specific mAbs [18] on days 3 and 1 prior to inoculation with 105 B16F10 cells. Each line represents an individual mouse and the numbers in parenthesis indicate the number of tumorfree mice.
E. Jones et al. / Immunology Letters 85 (2003) 141 /143
effect on tumor growth (c) whilst neutralization of TGFb resulted in slower growth and in some mice, complete rejection of the tumor cells (d). Several cell-types could be responsible for production of TGF-b in vivo including the tumor cells or infiltrating lymphocytes including CD25 regulatory T cells. Indeed, the results of some studies imply that CD25 regulatory cells exert their immunosuppressive effects through production of TGFb [12]. The source of TGF-b in the experimental model described here is currently under investigation.
143
their activation. Indeed, danger signals may promote conventional T cell activation, not only through activation of DCs, but also through inhibiting CD25 regulatory T cell activity. Thus, as for activation of conventional T cells, activation of CD25 regulatory T cells could be dependent upon the context in which antigens are presented rather than the ability of the cells to discriminate between self and non-self antigens.
Acknowledgements 4. What activates CD25
regulatory T cells?
The physiological trigger for the activation of CD25 regulatory cells in the periphery has not been identified. T cell receptor repertoire analyses have revealed no significant differences in Va or Vb gene usage between regulatory cells and other conventional T cells suggesting that these cells recognize a diversity of antigens in much the same way as conventional T cells [13]. CD25 regulatory T cells, unlike conventional T cells are, however, thought to recognize self-antigens with high affinity, thus CD25 regulatory cells may respond equally well to both foreign antigens and self-antigens. The ability to discriminate between self and foreign antigens may therefore not be the trigger for regulatory T cell activation. The trigger could relate rather to the immunosuppressive nature of the tumor environment. In a situation such as a tumor site, the immunosuppressive cytokines secreted by the tumor may prevent adequate activation of APCs that have picked up and processed tumor antigens, and favor stimulation of CD25 regulatory T cells over conventional T cells. As mentioned above, these CD25 regulatory T cells could potentially recognize both self- and non-self antigens expressed by the tumor cells. Due to the antigen non-specific nature of their immunosuppressive effects, all effector T cell responses to tumor-derived antigens would be inhibited. It has been proposed that conventional T cells are activated in response to ‘dangerous’ stimuli (e.g. heat shock proteins) rather than the ability to discriminate between self- and non-self antigens [14]. According to the hypothetical situation described above for the activation of CD25 regulatory T cells, these cells require no danger signals, but rather the opposite, for
This work was supported by The Wellcome Trust (grant no. GR056527MA).
References [1] C. Bouneaud, P. Kourilsky, P. Bousso, Immunity 13 (2000) 829 / 840. [2] J.W. Kappler, N. Roehm, P. Marrack, Cell 49 (1987) 273 /280. [3] F. Ramsdell, B.J. Fowlkes, Science 248 (1990) 1342 /1348. [4] E.M. Shevach, Annu. Rev. Immunol. 18 (2000) 423 /449. [5] S. Sakaguchi, Cell 101 (2000) 455 /458. [6] J. Shimizu, S. Yamazaki, S. Sakaguchi, J. Immunol. 163 (1999) 5211 /5218. [7] S. Onizuka, I. Tawara, J. Shimizu, S. Sakaguchi, T. Fujita, E. Nakayama, Cancer Res. 59 (1999) 3128 /3133. [8] Jones, E., Dahm-Vicker, M, Simon, A.K., Green, A., Powrie, F., Cerundolo, V., Gallimore, A., 2002. Cancer Immunity [serial online] 2002; 2:1. Available from http://www.cancerimmunity.org/ v2p1/020201.htm [9] R.P. Sutmuller, L.M. van Duivenvoorde, A. van Elsas, T.N. Schumacher, M.E. Wildenberg, J.P. Allison, R.E. Toes, R. Offringa, C.J. Melief, J. Exp. Med. 194 (2001) 823 /832. [10] H. Maeda, H. Kuwahara, Y. Ichimura, M. Ohtsuki, S. Kurakata, A. Shiraishi, J. Immunol. 155 (1995) 4926 /4932. [11] H. Maeda, A. Shiraishi, J. Immunol. 156 (1996) 73 /78. [12] F. Powrie, J. Carlino, M.W. Leach, S. Mauze, R.L. Coffman, J. Exp. Med. 183 (1996) 2669 /2674. [13] T. Takahashi, Y. Kuniyasu, M. Toda, N. Sakaguchi, M. Itoh, M. Iwata, J. Shimizu, S. Sakaguchi, Int. Immunol. 10 (1998) 1969 / 1980. [14] P. Matzinger, Annu. Rev. Immunol. 12 (1994) 991 /1045. [15] G.C. Koo, J.R. Peppard, Hybridoma 3 (1984) 301 /303. [16] J.W. Lowenthal, R.H. Zubler, M. Nabholz, H.R. MacDonald, Nature 315 (1985) 669 /672. [17] C. Asseman, S. Mauze, M.W. Leach, R.L. Coffman, F. Powrie, J. Exp. Med. 190 (1999) 995 /1004. [18] J.R. Dasch, D.R. Pace, W. Waegell, D. Inenaga, L. Ellingsworth, J. Immunol. 142 (1989) 1536 /1541.