- Email: [email protected]
Immunotherapy: target the stroma to hit the tumor
Review
TRENDS in Molecular Medicine
Vol.11 No.5 May 2005
Immunotherapy: target the stroma to hit the tumor Thomas Kammertoens1, Thomas Schu¨ler2 and Thomas Blankenstein1,3 1
Institute of Immunology, Charite´ Campus Benjamin Franklin, 12200 Berlin, Germany German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany 3 Max-Delbru¨ck-Center for Molecular Medicine, 13092 Berlin, Germany 2
For decades it has been assumed that T cells reject tumors essentially by direct killing. However, solid tumors are composed of malignant cells and a variety of different nonmalignant cells, referred to as tumor stroma. Stromal cells, such as endothelial cells, fibroblasts and inflammatory cells, often support tumor growth. Here, we discuss new findings showing that the tumor stroma is an important target during T-cellmediated tumor rejection. Cytotoxic molecules and cytokines produced by T cells inhibit or destroy the stromal ‘infrastructure’, thereby withdrawing essential resources and leading to tumor infarction and subsequent T-cell-mediated elimination of residual tumor cells. These findings are important for the development of more effective and specific immunotherapies for cancer. Tumor rejection requires T cells Tumor rejection usually requires CD8C and CD4C T cells that recognize tumor-specific peptides on major histocomptability complex (MHC) class I and class II molecules, respectively (see Glossary). Whereas CD8C cytotoxic T lymphocytes (CTLs) usually are the main effector cells in most tumor models, CD4C T cells are necessary for the activation [1] and survival of CD8C effector T cells [2] and, in some models, as effector cells that can reject tumors in the absence of CD8C T cells [3]. CD8C T cells use cytokines, including interferon-g (IFN-g) and tumor necrosis factor-a (TNF-a), and cytotoxic molecules, such as perforin, for tumor elimination. The relative contribution of direct tumor-cell killing and cytokine action for tumor rejection depends upon the tumor model. However, IFN-g is essential for tumor rejection in almost all models analyzed to date [4–18]. CD8C T cells, when in vitro activated to produce interleukin-4 (IL-4), can also reject tumors [19,20]. CD4C T cells express IFN-g or IL-4, and are commonly termed Th1 and Th2 cells, respectively. Although these cytokines are often involved in immune regulation, they are discussed here by virtue of their effector function during tumor rejection. Cytotoxicity by CD4C T cells appears to be the exception rather than the rule. Both Th1 and Th2 cells can mediate tumor rejection [6,21–23]. It was assumed that CD4C T cells reject tumors by a delayed-type hypersensitivity-like reaction, during Corresponding author: Blankenstein, T. ([email protected]). Available online 8 April 2005
which they attract and activate innate effector cells, such as tumoricidal macrophages or natural killer (NK) cells [3]. However, firm evidence for the involvement of macrophages in this scenario is rare, because it is difficult to distinguish whether CD4C T cells induce the tumoricidal activity of macrophages or whether macrophages infiltrate tumors as a result of tumor-cell death caused by the CD4C T cells. T suppressor cells or regulatory T cells that might modulate anti-tumor immune responses will not be discussed. Because only small numbers of tumor cells, and virtually never large established tumors, can be rejected by vaccination-induced T cells, the adoptive transfer of effector T cells has often been used. In this situation, a variety of factors influence tumor rejection. The best chance for successful therapy exists if: the number of transferred T cells is high [24,25]; T cells can easily expand and their survival is supported [26]; T cells are of high avidity [27]; the tumor antigen is expressed in large Glossary Tumor stroma: Solid tumors consist not only of malignant cells but also nonmalignant cells, such as fibroblasts, endothelial cells and inflammatory cells, which form a tumor-promoting microenvironment that is termed the tumor stroma. Angiogenesis: Growing tumors require a blood supply to provide nutrients and oxygen. The formation of new blood vessels is termed angiogenesis. Antigen cross-presentation: Antigen cross-presentation describes the cellular uptake of exogenous protein antigens and their processing and presentation by MHC class I molecules. Cross-presentation by dendritic cells is required for the initiation of CD8C T-cell responses. Tumor-associated fibroblasts (TAFs): TAFs are connective-tissue cells of mesenchymal origin that are found in most solid tumors. Their role in tumorpromotion is poorly understood. However, their production of vascular endothelial growth factor (VEGF) contributes to angiogenesis. Tumor-associated macrophages (TAMs): TAMs are found in nearly all solid tumors. Depending on their activation status, macrophages can have tumoricidal activity or alternatively promote tumor growth (e.g. by providing angiogenic factors). Endothelial cells: Endothelial cells form the inner lining of blood vessels (the endothelium). Their function is to participate in the regulation of the exchange of gases, fluids, nutrients and cells between blood and the respective tissues. A population of cells in close association with the endothelium is pericytes, which are thin, contractile cells of mesenchymal origin. Cytotoxic T lymphocytes (CTLs): CD8C CTLs recognize antigenic peptides in the context of MHC class I. Upon activation, CTLs can directly kill target cells by cytotoxic molecules, such as perforin and ganzyme B, and secrete different effector cytokines (e.g. IFN-g and TNFa). T helper cells: CD4C T helper (TH) cells recognize specific peptides in the context of MHC class II molecules. Th cells can be subdivided into at least two different subsets. Th1 cells secrete IFN-g and primarily stimulate cellular immune responses. Th2 cells secrete IL-4 and IL-5 and mainly support humoral immune responses.
www.sciencedirect.com 1471-4914/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.molmed.2005.03.002
226
Review
TRENDS in Molecular Medicine
amounts and cannot be selected against [28]; tumorantigen-derived peptides can be efficiently cross-presented [29,30]; and tumor burden is low, such that the T cells can efficiently infiltrate the tumor tissue [31,32]. If any of these factors is suboptimal, the risk of failure of T-cell therapy increases. In this review, different mechanisms concerning how T cells or their products contribute to tumor rejection will be discussed. We will show that the stroma, in addition to the tumor cells themselves, is an important target for successful tumor rejection by T cells. It is important to note that the mechanisms by which T cells reject tumors are relevant for vaccine-induced or adoptively transferred T cells but probably cannot explain spontaneous immune responses against primary non-virus-associated tumors [33,34]. A better understanding of how T cells reject tumors and which cells they attack will be of crucial importance for the development of successful immunotherapies against cancer.
The tumor stroma Solid tumors are complex tissues that are maintained by dynamic interactions between tumor cells and the microenvironment [inflammatory cells, the vasculature, fibroblasts and the extracellular matrix (ECM)], commonly referred to as tumor stroma (Figure 1). The interactions between stroma and tumor cells have been reviewed [35–39]. Stromal cells provide growth factors, blood supply, mechanical support and metalloproteases and remove waste and dead cells. The emerging concept is that premalignant cells can be controlled by a normal stroma and, upon a change in the microenvironment (for example, by inflammatory stimuli), tissue remodeling and cancer progression is started. For example, the abrogation of transforming growth factor-b (TGF-b) signaling in fibroblasts led to stomach and prostate tumors [40], indicating that stromal fibroblasts with intact TGF-b signaling suppressed the oncogenic potential in adjacent epithelia. By contrast, tumors modulate their environment and keep stromal cells in an activated and tumor-promoting state. Of note is the necessity for even small tumors to induce angiogenesis: the sprouting of new blood vessels from existing blood vessels [41]. This process requires a complex cellular and molecular cooperation between different tissue and inflammatory cells, such as endothelial cells, fibroblasts and macrophages. Under inflammatory conditions, such as wound healing, or in growing tumors, endothelial cells are activated, start to divide and form new blood vessels. Frequently detected growth factors that activate endothelial cells are vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF). Tumor cells, tumor-associated macrophages (TAMs) and fibroblasts (TAFs) produce VEGF and are involved in tumor-induced angiogenesis [42–45]. Both TAMs and TAFs are abundantly detected in tumors
Vol.11 No.5 May 2005
ECM GF F G
T M BV P E TRENDS in Molecular Medicine
Figure 1. The cellular composition of solid tumors. Solid tumors are complex tissues that are maintained by dynamic interactions between cancer cells (blue) and the microenvironment. This microenvironment is referred to as tumor stroma. The tumor stroma is composed of different cell types, such as fibroblasts (F), pericytes (P), endothelial cells (E) and immune cells, including T cells (T), granulocytes (G) and macrophages (M). Stromal cells fulfill a variety of functions. Fibroblasts produce extracellular matrix (ECM) proteins for mechanical support. Endothelial cells that make up the inner lining of blood vessels (BV) ensure a blood supply for growing tumors by neoangiogenesis. Hematopoietic stromal cells can provide tumor-promoting growth factors (GFs) and macrophages remove waste and dead cells. Shown is a schematic representation of the cellular composition of tumor tissue that might, depending upon the tumor type and stage, vary considerably.
and often support tumor growth [43,44]. If macrophages were inhibited from infiltrating tumors, angiogenesis was impaired and tumors were rejected [46,47]. The deficiency of macrophages in cancer-prone (MMTVpy-midT) transgenic mice delayed the development of invasive mammary carcinomas [48]. TAFs are a morphologically homogeneous but functionally heterogeneous group of mesenchymal cells. Some fibroblasts are tissue-resident matrix synthesizing and/or matrix-degrading cells, whereas others are contractile cells (myofibroblasts), circulating precursor cells (fibrocytes) or blood-vessel-associated pericytes. Fibroblasts can produce several growth factors, cytokines, chemokines and matrix-degrading enzymes and, therefore, might modulate the local immune response [44]. In most systems analyzed, TAFs have pro-tumorigenic effects, such that the co-injection of fibroblasts facilitated the growth of low-tumorigenic cells [49]. The importance of VEGF-producing TAFs for angiogenesis and tumor growth was recently shown [45]. The inhibition of tumor angiogenesis as a therapeutic strategy is recognized and several angiogenesis inhibitors, such as anti-VEGF antibodies, have been developed [50]. Only recently it became clear that effective immune therapy primarily acts on the tumor stroma and similarly interferes with tumorinduced angiogenesis.
Figure 2. T cells target the stroma and hit the tumor by multiple mechanisms. (a) CD8C effector T cells recognize MHC-class-I–antigen-peptide complexes on cancer cells (blue) and kill them with perforin (pfp). Although this mechanism is often used, it is usually not sufficient for tumor rejection. (b) CD8C effector T cells recognize MHC-class-I– antigen-peptide complexes on cancer cells and produce IFN-g, which acts on tumor stroma (e.g. endothelial cells; yellow), resulting in the inhibition of tumor-induced www.sciencedirect.com
Review
TRENDS in Molecular Medicine
(a)
Vol.11 No.5 May 2005
227
(b) CD8
CD8
IFNγ TCR
TCR pfp
MHC I
MHC I Ag
Ag Cancer cell killed
(c)
CD4 (TH1)
Angiostasis
CD4 (TH2)
(d)
IL-4 IFNγ
TCR
TCR TCR
MHC II E
TAM
TAM
Ag
Ag
Angiostasis
Angiostasis
CD8
(e)
TAF
MHC II
(f)
CD8
TCR
TCR pfp
pfp MHC I MHC I TAM
non-BM cells
(CD11b+)
Ag Stroma cell killed
Ag Stroma cell killed TRENDS in Molecular Medicine
angiogenesis (angiostasis). (c) IFN-g-producing CD4C effector T cells (Th1), as do CD8C T cells, induce angiostasis in MHC class IIK tumors. Therefore, MHC-class-IIexpressing tumor stroma cells must have taken up, processed and presented tumor antigens by MHC class II molecules. These are presented to Th1 cells, which produce IFN-g that acts, directly or indirectly, on endothelial cells. Because of their abundance, we assume that tumor-associated macrophages (TAMs) are MHC class IIC cells, but this has yet to be proven. (d) IL-4-producing CD4C effector T cells (Th2), similar to Th1 cells, induce angiostasis, using tumor-associated fibroblasts (TAFs) as targets. (e) CD8C effector T cells recognize antigen on antigen cross-presenting tumor-associated (CD11bC) macrophages and kill them with perforin. (f) CD8C effector T cells recognize antigen on antigen cross-presenting non-bone-marrow-derived tumor stroma cells (non-BM cells), such as endothelial cells or fibroblasts, and kill them with perforin. Abbreviations: Ag, tumor antigen; TCR, T-cell receptor. www.sciencedirect.com
228
Review
TRENDS in Molecular Medicine
CD8C T cells recognize and kill cancer cells It is often assumed that the major mechanism of tumor rejection is the direct recognition and killing of tumor cells by CD8C effector T cells; therefore, they are termed CTLs. CTLs recognize MHC-class-I–peptide complexes on target cells and, with the help of perforin, lyse them (Figure 2a) [51]. Although cytotoxicity is often necessary for complete tumor rejection, it is usually not sufficient. This becomes obvious from studies showing that tumor rejection by CD8C effector T cells correlated better with IFN-g and TNF-a production than with cytotoxicity [52,53]. More importantly, a number of recent studies show that IFN-g or IFN-g-receptor gene-deficient (knockout; KO) mice cannot reject tumors in most systems, although these mice develop normal numbers of T cells and cytotoxic responses [4–18]. As demonstrated in perforin-KO mice or with adoptively transferred T cells from perforin-KO mice, perforin (or Fas-ligand, which is another cytotoxic molecule) was necessary for tumor rejection in some, but not all, models, but they generally required IFN-g [4]. IFN-g from CD8C T cells inhibits angiogenesis Schu¨ler et al. [13] analyzed the cells on which CD8C effector T cells recognize antigen during tumor rejection. They used CD8C OT-1 T cells, which were derived from T-cell-receptor-transgenic mice, that were specific for a peptide of the model antigen ovalbumin (OVA) presented by H2-Kb molecules. Following in vitro activation and adoptive transfer, OT-1 T cells rejected, in most of the mice, B16 melanoma cells that were transfected to express OVA. Bone-marrow reconstitution experiments demonstrated that bone-marrow-derived cells were not required to present OVA to achieve tumor rejection, indicating that that antigen recognition on tumor cells was sufficient for tumor rejection by the CD8C effector T cells. Interestingly, OT-1 T cells could not reject B16-OVA tumors in IFNg-receptor-KO mice, showing that IFN-g produced by the CD8C effector T cells acted on the host cells. In IFN-g-KO mice, OT-1 T cells rejected the tumor. Therefore, IFN-g expression by the transferred OT-1 T cells was sufficient for tumor rejection [13]. Because IFN-g expression by CD8C effector T cells is tightly regulated and requires antigen recognition, it is likely that IFN-g acted on stromal cells in the vicinity of the tumor cells. Qin et al. [12] investigated the cells within the tumor stroma on which IFN-g from CD8C T cells acted upon to induce tumor rejection. They showed that tumor rejection mediated by vaccination-induced CD8C T cells was always preceded by the inhibition of tumor-induced angiogenesis and that endothelial (CD31C) cells were arrested at the rim of the tumor. Angiostasis and tumor rejection was also observed in perforin-KO mice but not in IFN-g-KO mice. Only adoptively transferred CD8C effector T cells from wild-type, but not IFN-g-KO, mice inhibited angiogenesis in experimental lung metastases [12]. Together, these data suggest that CD8C effector T cells recognize antigen on tumor cells and express IFN-g that acts on tumor stroma, probably endothelial cells (Figure 2b). One cannot exclude, however, that antigen cross-presentation by non-bone-marrow-derived tumor stroma cells (e.g. endothelial cells) contributed to tumor rejection. It is www.sciencedirect.com
Vol.11 No.5 May 2005
interesting to note that CD8C effector T cells inhibited, but could not completely prevent, tumor growth in IFNg-receptor-KO mice [13]. This suggests that cytotoxicity slows down tumor growth but is not sufficient for its long-term inhibition. CD4C T cells recognize tumor antigen-presenting stromal cells and induce angiostasis through IFN-g It has been observed that, in some models, CD4C T cells are the main effector cells during tumor rejection, even if tumor cells do not express detectable levels of MHC class II molecules [3]. In a model of CD4C T-cell-mediated tumor rejection, it was shown that IFNg receptor was essential for tumor rejection [6]. Tumor rejection in immunized wild-type mice was equally efficient regardless of whether or not the tumor cells expressed IFNg-receptor. In fact, in most, but not all, systems, the upregulation of MHC molecules by IFN-g appears to have a minor role during tumor rejection [4,6,21,54–56]. CD4C T cells in both immunized IFNg-receptor-KO and wild-type mice produced IFN-g in a tumor-specific fashion, homed to the challenge site and, upon transfer into wild-type mice, rejected IFNg-receptor-negative tumors. Bone-marrowchimeric mice revealed that IFNg-receptor expression on non-bone-marrow-derived cells was necessary and sufficient for tumor rejection. Similar to tumor rejection by CD8C T cells, CD4C T cells inhibited tumor-induced angiogenesis by IFN-g. Because tumor cells did not express MHC class II molecules in this model, one can postulate that MHC class IIC tumor stroma cells presented tumor-derived antigens to CD4C T cells, thereby inducing IFN-g and inhibiting angiogenesis (Figure 2c). The MHC class IIC stromal cells that present antigen to the CD4C effector T cells remain elusive, but dendritic cells or, because of their abundance in most solid tumors, macrophages are potential candidates. It should be emphasized that the presented mechanisms should not be seen as working separately, but it is likely that several mechanisms cooperate during tumor rejection. CD4C T cells recognize stromal cells and induce angiostasis through IL-4 In adoptive T-cell-transfer experiments, IL-4-producing CD4C T cells were shown to eliminate B16 lung metastases [23]. Because B16 cells do not express MHC class II molecules, one has to assume that MHC class IIC stromal cells presented tumor-derived antigens to effector CD4C T cells to induce IL-4 production by the T cells. Tumors that are transfected to produce IL-4 are rejected in wildtype mice and are inhibited in the long term in T-cell-KO mice by a paracrine action on host cells [57,58]. It was assumed that tumor-reactive Th2 cells attract and activate eosinophils, which then kill tumor cells [23]. The role of eosinophils during IL-4-mediated tumor rejection is, however, controversial because IL-5-KO mice that lack eosinophils reject IL-4-producing tumors as efficiently as do wild-type control mice [59]. Recent data gave a better insight into the mechanism of IL-4-induced tumor rejection, if one assumes that the expression of IL-4 by the tumor mimics the local production by tumor-reactive Th2 cells that accumulate
Review
TRENDS in Molecular Medicine
at the tumor site. Using bone-marrow-chimeric mice, it was shown that IL-4-receptor (IL4R) expression on both hematopoietic and nonhematopoietic cells was sufficient for the rejection of IL-4-producing tumors [60]. TAFs were identified as one target of IL-4. In situations in which tumor cells did not produce IL-4 or host cells could not respond to IL-4, TAFs infiltrated tumors in an IL-4-independent fashion and were associated with blood vessels. In IL-4-secreting tumors grown in wild-type mice, TAFs were similarly present but more-pronounced ECM was detected, indicating that IL-4-responsive fibroblasts in the presence of IL-4 were differently activated. Blood vessels in these tumors were virtually absent and the tumors were subsequently rejected. These findings are compatible with the notion that IL-4 has profound effects on fibroblasts and suppresses angiogenesis [61]. The importance of fibroblasts as a target for IL-4 was underlined by the demonstration that IL-4-producing tumors could be rejected, even if TAFs were the only cells that could respond to IL-4 [60]. Based on these data, it is postulated that Th2 cells recognizing tumor-derived antigen on MHC class IIC tumor stroma cells produce IL-4 that acts on fibroblasts, thereby inhibiting their participation in tumor-induced angiogenesis (Figure 2d). Because mice with selective IL4R expression in bonemarrow-derived cells also rejected IL-4-producing tumors, it is interesting to note that bone-marrow-derived fibroblasts and myofibroblasts accumulated in tumors; therefore, it is possible that bone-marrow-derived fibroblasts supported tumor-induced angiogenesis [62]. CD8C T cells kill bone-marrow-derived stromal cells cross-presenting tumor antigen An interesting mechanism of how T cells contribute to tumor rejection was recently shown. In an adoptivetransfer model with tumor-specific CD8C effector T cells, tumor rejection required, in addition to tumor and nonbone-marrow-derived cells, bone-marrow-derived cells to cross-present tumor-derived antigen [63]. CD8C effector T cells lysed antigen cross-presenting CD11bC tumor stroma cells, typically monocytes and macrophages, probably by perforin (Figure 2e). Tumor rejection and cytolysis of CD11bC tumor stroma cells occurred only when the tumor expressed a large amount of antigen so that it could be cross-presented to the CD8C effector T cells. Tumors often express macrophage chemoattractants [43] and TAMs promote tumor growth, for example by contributing to angiogenesis [42]. The killing of TAMs by CD8C effector T cells is, therefore, an effective strategy to destroy the tumor infrastructure. CD8C T cells kill non-bone-marrow-derived stromal cells cross-presenting tumor antigen A mechanism suggested by the data of Spiotto et al. [63] is shown in Figure 2f. In their model, tumor cells express large amounts of antigen, which is cross-presented by bone-marrow-derived and non-bone-marrow-derived stromal cells. This enables adoptively transferred CD8C T cells to efficiently kill tumor cells directly and eliminate cross-presenting stromal cells, which is required to prevent the outgrowth of antigen loss variants [63]. The www.sciencedirect.com
Vol.11 No.5 May 2005
229
requirement for antigen cross-presentation by non-bonemarrow-derived cells, such as endothelial cells or fibroblasts, strongly suggests that they are targets of CD8C effector T cells. It is therefore likely that non-bonemarrow-derived TAFs or endothelial cells were killed, probably by perforin, which was required for rapid tumor regression (Figure 2f) [63]. Together, the above examples (Figure 2a–f) illustrate that T cells can use a variety of mechanisms to destroy tumors. Cyclophosphamide induces IFNg-receptor-dependent destruction of the tumor vasculature The above experiments indicated that IFN-g produced by T cells primarily inhibited the early phase of tumorinduced angiogenesis. However, IFN-g can also contribute to the destruction of an established tumor stroma [64]. In some models, a single injection of cyclophosphamide (Cy) induces immune-mediated rejection of Cy-resistant, established and vascularized tumors. Starting as early as 6 h after Cy-injection, the tumor-infiltrating T cells were inactivated, TAMs stopped producing IL-10 and started to produce IFN-g and, with the same kinetics, the tumor blood vessels were destroyed. Remarkably, blood vessel destruction and tumor rejection required IFN-g-receptor expression on cells of the tumor-bearing host, probably tumor stroma cells. The disappearance of (Meca32C or CD31C) endothelial cells preceded central tumor necrosis and the subsequent elimination of residual tumor cells by CD8C T cells. These data confirm that IFN-g targets, either directly or by secondary factors, endothelial cells. They also suggest that Cy, which is thought to eliminate suppressor T cells [65] and facilitate homeostatic T-cell proliferation [26], contributes to tumor rejection by modulating the tumor stroma. Concluding remarks Armed with cytokines and cytotoxic molecules, T cells use a multitude of strategies to reject tumors. They can recognize tumor antigens on tumor and multiple stroma cell types and destroy or modulate the function of multiple tumor stroma cell types, such as TAMs, fibroblasts and endothelial cells. Although phenotypically different effector T cells might target different tumor stroma cells, they act on the stroma in a redundant fashion to withdraw its blood supply. However, the tumor stroma can also act as a barrier [66] and prevent the access of T cells to the tumor [32]. Effective therapeutic modalities to modulate the stroma such that T-cell entry into tumors is improved have already been found, such as CpG oligonucleotides [31] or cyclophosphamide [64]. As our understanding of tumorstroma–immune-cell interactions increases, we can optimistically anticipate more precise and effective immunotherapies to be developed. Acknowledgements This review was supported by grants from the Deutsche Forschungsgemeinschaft and the Deutsche Krebshilfe, Dr. Mildred Scheel Stiftung.
References 1 Schoenberger, S.P. et al. (1998) T-cell help for cytotoxic T lymphocytes is mediated by CD40–CD40L interactions. Nature 393, 480–483
230
Review
TRENDS in Molecular Medicine
2 Bourgeois, C. et al. (2002) A role for CD40 expression on CD8C T cells in the generation of CD8C T cell memory. Science 297, 2060–2063 3 Greenberg, P.D. (1991) Adoptive T cell therapy of tumors: mechanisms operative in the recognition and elimination of tumor cells. Adv. Immunol. 49, 281–355 4 Blankenstein, T. and Qin, Z. (2003) The role of IFN-g in tumor transplantation immunity and inhibition of chemical carcinogenesis. Curr. Opin. Immunol. 15, 148–154 5 Winter, H. et al. (1999) Tumor regression after adoptive transfer of effector Tcells is independent of perforin or Fas ligand (APO-1L/CD95L). J. Immunol. 163, 4462–4472 6 Qin, Z. and Blankenstein, T. (2000) CD4C T cell-mediated tumor rejection involves inhibition of angiogenesis that is dependent on IFNg receptor expression by nonhematopoietic cells. Immunity 12, 677–686 7 Prevost-Blondel, A. et al. (2000) Differential requirement of perforin and IFN-g in CD8 T cell-mediated immune responses against B16.F10 melanoma cells expressing a viral antigen. Eur. J. Immunol. 30, 2507–2515 8 Peng, L. et al. (2000) T cell-mediated tumor rejection displays diverse dependence upon perforin and IFN-g mechanisms that cannot be predicted from in vitro T cell characteristics. J. Immunol. 165, 7116–7124 9 Winter, H. et al. (2001) Immunotherapy of melanoma: a dichotomy in the requirement for IFN-g in vaccine-induced antitumor immunity versus adoptive immunotherapy. J. Immunol. 166, 7370–7380 10 Wigginton, J.M. et al. (2001) IFN-g and Fas/FasL are required for the antitumor and antiangiogenic effects of IL-12/pulse IL-2 therapy. J. Clin. Invest. 108, 51–62 11 van Elsas, A. et al. (2001) Elucidating the autoimmune and antitumor effector mechanisms of a treatment based on cytotoxic T lymphocyte antigen-4 blockade in combination with a B16 melanoma vaccine: comparison of prophylaxis and therapy. J. Exp. Med. 194, 481–481 12 Qin, Z. et al. (2003) A critical requirement of IFNg-mediated angiostasis for tumor rejection by CD8C T cells. Cancer Res. 63, 4095–4100 13 Schu¨ler, T. and Blankenstein, T. (2003) CD8C effector T cells reject tumors by direct antigen recognition but indirect action on host cells. J. Immunol. 170, 4427–4431 14 Poehlein, C.H. et al. (2003) TNF plays an essential role in tumor regression after adoptive transfer of perforin/IFN-g double knockout effector T cells. J. Immunol. 170, 2004–2013 15 Kowalczyk, D.W. et al. (2003) Vaccine-induced CD8C T cells eliminate tumors by a two-staged attack. Cancer Gene Ther. 10, 870–878 16 Nanni, P. et al. (2004) Immunoprevention of mammary carcinoma in HER-2/neu transgenic mice is IFN-g and B cell dependent. J. Immunol. 173, 2288–2296 17 Hollenbough, J.A. et al. (2004) The rate of the CD8-dependent initial reduction in tumor volume is not limited by contact-dependent perforin, Fas ligand, or TNF-mediated cytolysis. J. Immunol. 173, 1738–1743 18 Wu, T.H. et al. (2004) Long-term suppression of tumor growth by TNF requires a Stat1- and IFN response factor 1-dependent IFNg pathway but not IL-12 or IL-18. J. Immunol. 172, 3243–3251 19 Dobrzanski, M.J. et al. (2000) Type 1 and type 2 CD8C effector T cell subpopulations promote long-term tumor immunity and protection to progressively growing tumor. J. Immunol. 164, 916–925 20 Dobrzanski, M.J. et al. (2001) Role of effector cell-derived IL-4, IL-5, and perforin in early and late stages of type 2 CD8 effector cellmediated tumor rejection. J. Immunol. 167, 424–434 21 Mumberg, D. et al. (1999) CD4C T cells eliminate MHC class IInegative cancer cells in vivo by indirect effects of IFN-g. Proc. Natl. Acad. Sci. U. S. A. 96, 8633–8638 22 Nishimura, T. et al. (1999) Distinct role of antigen-specific T helper type 1 (Th1) and Th2 cells in tumor eradication in vivo. J. Exp. Med. 190, 617–627 23 Mattes, J. et al. (2003) Immunotherapy of cytotoxic T cell-resistant tumors by T helper 2 cells: An eotaxin and STAT6-dependent process. J. Exp. Med. 197, 387–393 24 Nguyen, L.T. et al. (2002) Tumor growth enhances cross-presentation leading to limited T cell activation without tolerance. J. Exp. Med. 195, 423–435 www.sciencedirect.com
Vol.11 No.5 May 2005
25 Lyman, M.A. et al. (2004) A spontaneously arising pancreatic tumor does not promote the differentiation of naive CD8C T lymphocytes into effector CTL. J. Immunol. 172, 6558–6567 26 Dudley, M.E. and Rosenberg, S.A. (2003) Adoptive-cell-transfer therapy for the treatment of patients with cancer. Nat. Rev. Cancer 3, 666–675 27 Sadovnikova, E. and Stauss, H.J. (1996) Peptide-specific cytotoxic T lymphocytes restricted by nonself major histocompatibility complex class I molecules: reagents for tumor immunotherapy. Proc. Natl. Acad. Sci. U. S. A. 93, 13114–13118 28 Spiotto, M.T. et al. (2002) Increasing tumor antigen expression overcomes ‘ignorance’ to solid tumors via crosspresentation by bone marrow-derived stromal cells. Immunity 17, 737–747 29 Wolkers, M.C. et al. (2004) Antigen bias in T cell cross-priming. Science 304, 1314–1317 30 Norbury, C.C. et al. (2004) CD8C T cell cross-priming via transfer of proteasome substrates. Science 304, 1318–1321 31 Garbi, N. et al. (2004) CpG motifs as proinflammatory factors render autochthonous tumors permissive for infiltration and destruction. J. Immunol. 172, 5861–5869 32 Ganss, R. et al. (2004) Overcoming tumor-intrinsic resistance to immune effector function. Eur. J. Immunol. 34, 2635–2641 33 Qin, Z. and Blankenstein, Th. (2004) A cancer immunosurveillance controversy. Nat. Immunol. 5, 3–4 34 Blankenstein, Th. and Qin, Z. (2003) Chemical carcinogens as foreign bodies and some pitfalls regarding cancer immune surveillance. Adv. Cancer Res. 90, 179–207 35 Tlsty, T.D. and Hein, P.W. (2001) Know thy neighbor: stromal cells can contribute oncogenic signals. Curr. Opin. Genet. Dev. 11, 54–59 36 Bissell, M.J. and Radisky, D. (2001) Putting tumours in context. Nat. Rev. Cancer 1, 46–54 37 Coussens, L.M. and Werb, Z. (2002) Inflammation and cancer. Nature 420, 860–867 38 Mueller, M.M. and Fusenig, N.E. (2004) Friends or foes – bipolar effects of the tumour stroma in cancer. Nat. Rev. Cancer 4, 839–849 39 Philip, M. et al. (2004) Inflammation as a tumor promoter in cancer induction. Semin. Canc. Biol. 14, 433–439 40 Bhowmick, N.A. et al. (2004) TGF-b signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303, 848–851 41 Folkman, J. (2003) Fundamental concepts of the angiogenic process. Curr. Mol. Med. 3, 643–651 42 Polverini, P.J. and Leibovich, J. (1984) Induction of neovascularization in vivo and endothelial proliferation in vitro by tumor-associated macrophages. Lab. Invest. 51, 635–642 43 Mantovani, A. et al. (2002) Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol. 23, 549–555 44 Silzle, T. et al. (2004) The fibroblast: sentinel cell and local immune modulator in tumor tissue. Int. J. Cancer 108, 173–180 45 Dong, J. et al. (2004) VEGF-null cells require PDGFRa signalingmediated stromal fibroblast recruitment for tumorigenesis. EMBO J. 23, 2800–2810 46 Richter, G. et al. (1993) Interleukin 10 transfected into Chinese hamster ovary cells prevents tumor growth and macrophage infiltration. Cancer Res. 53, 4134–4137 47 Di Carlo, E. et al. (1998) Local release of interleukin-10 by transfected mouse adeno-carcinoma cells exhibits pro- and anti-inflammatory activity and results in a delayed tumor rejection. Eur. Cytokine Netw. 9, 61–68 48 Lin, E.Y. et al. (2001) Colony-stimulating factor 1 promotes progression of mammary tumors to malignancy. J. Exp. Med. 193, 727–739 49 Camps, J.L. et al. (1990) Fibroblast-mediated acceleration of human epithelial tumor growth in vivo. Proc. Natl. Acad. Sci. U. S. A. 87, 75–79 50 Kerbel, R. and Folkman, J. (2002) Clinical translation of angiogenesis inhibitors. Nat. Rev. Cancer 2, 727–739 51 van den Broek, M.F. and Hengartner, H. (2000) The role of perforin in infections and tumour surveillance. Exp. Physiol. 85, 681–685 52 Barth, R.J. et al. (1991) Interferon g and tumor necrosis factor have a role in tumor regressions mediated by murine CD8C tumor-infiltrating lymphocytes. J. Exp. Med. 173, 647–658 53 Becker, C. et al. (2001) Adoptive tumor therapy with T lymphocytes enriched through an IFN-g capture assay. Nat. Med. 7, 1159–1162
Review
TRENDS in Molecular Medicine
54 Dighe, A.S. et al. (1994) Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN g receptors. Immunity 1, 447–456 55 Pulaski, B.A. et al. (2002) Interferon-g-dependent phagocytic cells are a critical component of innate immunity against metastatic mammary carcinoma. Cancer Res. 62, 4406–4412 56 Segal, J.G. et al. (2002) The role of IFN-g in rejection of established tumors by IL-12: source of production and target. Cancer Res. 62, 4696–4703 57 Tepper, R.I. et al. (1989) Murine interleukin-4 displays potent antitumor activity in vivo. Cell 57, 503–512 58 Hock, H. et al. (1993) Mechanisms of rejection induced by tumor celltargeted gene transfer of interleukin 2, interleukin 4, interleukin 7, tumor necrosis factor, or interferon g. Proc. Natl. Acad. Sci. U. S. A. 90, 2774–2778 59 Noffz, G. et al. (1998) Neutrophils but not eosinophils are involved in growth suppression of IL-4-secreting tumors. J. Immunol. 160, 345–350
Vol.11 No.5 May 2005
231
60 Schuler, T. et al. (2003) Tumor rejection by modulation of tumor stromal fibroblasts. J. Exp. Med. 198, 1487–1493 61 Volpert, O.V. et al. (1998) Inhibition of angiogenesis by interleukin 4. J. Exp. Med. 188, 1039–1046 62 Direkze, N.C. et al. (2004) Bone marrow contribution to tumorassociated myofibroblasts and fibroblasts. Cancer Res. 64, 8492–8495 63 Spiotto, M.T. et al. (2004) Bystander elimination of antigen loss variants in established tumors. Nat. Med. 10, 294–298 64 Ibe, S. et al. (2001) Tumor rejection by disturbing tumor stroma cell interactions. J. Exp. Med. 194, 1549–1559 65 North, R.J. (1982) Cyclophosphamide-facilitated adoptive immunotherapy of an established tumor depends on elimination of tumorinduced suppressor T cells. J. Exp. Med. 155, 1063–1074 66 Singh, S. et al. (1992) Stroma is critical for preventing or permitting immunological destruction of antigenic cancer cells. J. Exp. Med. 175, 139–146
Endeavour the quarterly magazine for the history and philosophy of science You can access Endeavour online via ScienceDirect, where you’ll find a collection of beautifully illustrated articles on the history of science, book reviews and editorial comment.
Featuring Sverre Petterssen and the Contentious (and Momentous) Weather Forecasts for D-Day, 6 June 1944 by J.R. Fleming Food of Paradise: Tahitian breadfruit and the Autocritique of European Consumption by P. White and E.C. Spary Two Approaches to Etiology: The Debate Over Smoking and Lung Cancer in the 1950s by M. Parascandola Sicily, or sea of tranquility? Mapping and naming the moon by J. Vertesi The Prehistory of the Periodic Table by D. Rouvray Two portraits of Edmond Halley by P. Fara and coming soon Fighting the ‘microbe of sporting mania’: Australian science and Antarctic exploration in the early twentieth century by P. Roberts Learning from Education to Communicate Science as a Good Story by A. Negrete and C. Lartigue The Traffic and Display of Body Parts in the Early-19th Century by S. Alberti and S. Chaplin The Rise, Fall and Resurrection of Group Selection by M. Borrello Pomet’s great ‘‘Compleat History of Drugs’’ by S. Sherman Sherlock Holmes: scientific detective by L. Snyder The Future of Electricity in 1892 by G.J.N. Gooday The First Personal Computer by J. November Baloonmania: news in the air by M.G. Kim and much, much more . . . Locate Endeavour on ScienceDirect (http://www.sciencedirect.com) www.sciencedirect.com
TRENDS in Molecular Medicine
Vol.11 No.5 May 2005
Immunotherapy: target the stroma to hit the tumor Thomas Kammertoens1, Thomas Schu¨ler2 and Thomas Blankenstein1,3 1
Institute of Immunology, Charite´ Campus Benjamin Franklin, 12200 Berlin, Germany German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany 3 Max-Delbru¨ck-Center for Molecular Medicine, 13092 Berlin, Germany 2
For decades it has been assumed that T cells reject tumors essentially by direct killing. However, solid tumors are composed of malignant cells and a variety of different nonmalignant cells, referred to as tumor stroma. Stromal cells, such as endothelial cells, fibroblasts and inflammatory cells, often support tumor growth. Here, we discuss new findings showing that the tumor stroma is an important target during T-cellmediated tumor rejection. Cytotoxic molecules and cytokines produced by T cells inhibit or destroy the stromal ‘infrastructure’, thereby withdrawing essential resources and leading to tumor infarction and subsequent T-cell-mediated elimination of residual tumor cells. These findings are important for the development of more effective and specific immunotherapies for cancer. Tumor rejection requires T cells Tumor rejection usually requires CD8C and CD4C T cells that recognize tumor-specific peptides on major histocomptability complex (MHC) class I and class II molecules, respectively (see Glossary). Whereas CD8C cytotoxic T lymphocytes (CTLs) usually are the main effector cells in most tumor models, CD4C T cells are necessary for the activation [1] and survival of CD8C effector T cells [2] and, in some models, as effector cells that can reject tumors in the absence of CD8C T cells [3]. CD8C T cells use cytokines, including interferon-g (IFN-g) and tumor necrosis factor-a (TNF-a), and cytotoxic molecules, such as perforin, for tumor elimination. The relative contribution of direct tumor-cell killing and cytokine action for tumor rejection depends upon the tumor model. However, IFN-g is essential for tumor rejection in almost all models analyzed to date [4–18]. CD8C T cells, when in vitro activated to produce interleukin-4 (IL-4), can also reject tumors [19,20]. CD4C T cells express IFN-g or IL-4, and are commonly termed Th1 and Th2 cells, respectively. Although these cytokines are often involved in immune regulation, they are discussed here by virtue of their effector function during tumor rejection. Cytotoxicity by CD4C T cells appears to be the exception rather than the rule. Both Th1 and Th2 cells can mediate tumor rejection [6,21–23]. It was assumed that CD4C T cells reject tumors by a delayed-type hypersensitivity-like reaction, during Corresponding author: Blankenstein, T. ([email protected]). Available online 8 April 2005
which they attract and activate innate effector cells, such as tumoricidal macrophages or natural killer (NK) cells [3]. However, firm evidence for the involvement of macrophages in this scenario is rare, because it is difficult to distinguish whether CD4C T cells induce the tumoricidal activity of macrophages or whether macrophages infiltrate tumors as a result of tumor-cell death caused by the CD4C T cells. T suppressor cells or regulatory T cells that might modulate anti-tumor immune responses will not be discussed. Because only small numbers of tumor cells, and virtually never large established tumors, can be rejected by vaccination-induced T cells, the adoptive transfer of effector T cells has often been used. In this situation, a variety of factors influence tumor rejection. The best chance for successful therapy exists if: the number of transferred T cells is high [24,25]; T cells can easily expand and their survival is supported [26]; T cells are of high avidity [27]; the tumor antigen is expressed in large Glossary Tumor stroma: Solid tumors consist not only of malignant cells but also nonmalignant cells, such as fibroblasts, endothelial cells and inflammatory cells, which form a tumor-promoting microenvironment that is termed the tumor stroma. Angiogenesis: Growing tumors require a blood supply to provide nutrients and oxygen. The formation of new blood vessels is termed angiogenesis. Antigen cross-presentation: Antigen cross-presentation describes the cellular uptake of exogenous protein antigens and their processing and presentation by MHC class I molecules. Cross-presentation by dendritic cells is required for the initiation of CD8C T-cell responses. Tumor-associated fibroblasts (TAFs): TAFs are connective-tissue cells of mesenchymal origin that are found in most solid tumors. Their role in tumorpromotion is poorly understood. However, their production of vascular endothelial growth factor (VEGF) contributes to angiogenesis. Tumor-associated macrophages (TAMs): TAMs are found in nearly all solid tumors. Depending on their activation status, macrophages can have tumoricidal activity or alternatively promote tumor growth (e.g. by providing angiogenic factors). Endothelial cells: Endothelial cells form the inner lining of blood vessels (the endothelium). Their function is to participate in the regulation of the exchange of gases, fluids, nutrients and cells between blood and the respective tissues. A population of cells in close association with the endothelium is pericytes, which are thin, contractile cells of mesenchymal origin. Cytotoxic T lymphocytes (CTLs): CD8C CTLs recognize antigenic peptides in the context of MHC class I. Upon activation, CTLs can directly kill target cells by cytotoxic molecules, such as perforin and ganzyme B, and secrete different effector cytokines (e.g. IFN-g and TNFa). T helper cells: CD4C T helper (TH) cells recognize specific peptides in the context of MHC class II molecules. Th cells can be subdivided into at least two different subsets. Th1 cells secrete IFN-g and primarily stimulate cellular immune responses. Th2 cells secrete IL-4 and IL-5 and mainly support humoral immune responses.
www.sciencedirect.com 1471-4914/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.molmed.2005.03.002
226
Review
TRENDS in Molecular Medicine
amounts and cannot be selected against [28]; tumorantigen-derived peptides can be efficiently cross-presented [29,30]; and tumor burden is low, such that the T cells can efficiently infiltrate the tumor tissue [31,32]. If any of these factors is suboptimal, the risk of failure of T-cell therapy increases. In this review, different mechanisms concerning how T cells or their products contribute to tumor rejection will be discussed. We will show that the stroma, in addition to the tumor cells themselves, is an important target for successful tumor rejection by T cells. It is important to note that the mechanisms by which T cells reject tumors are relevant for vaccine-induced or adoptively transferred T cells but probably cannot explain spontaneous immune responses against primary non-virus-associated tumors [33,34]. A better understanding of how T cells reject tumors and which cells they attack will be of crucial importance for the development of successful immunotherapies against cancer.
The tumor stroma Solid tumors are complex tissues that are maintained by dynamic interactions between tumor cells and the microenvironment [inflammatory cells, the vasculature, fibroblasts and the extracellular matrix (ECM)], commonly referred to as tumor stroma (Figure 1). The interactions between stroma and tumor cells have been reviewed [35–39]. Stromal cells provide growth factors, blood supply, mechanical support and metalloproteases and remove waste and dead cells. The emerging concept is that premalignant cells can be controlled by a normal stroma and, upon a change in the microenvironment (for example, by inflammatory stimuli), tissue remodeling and cancer progression is started. For example, the abrogation of transforming growth factor-b (TGF-b) signaling in fibroblasts led to stomach and prostate tumors [40], indicating that stromal fibroblasts with intact TGF-b signaling suppressed the oncogenic potential in adjacent epithelia. By contrast, tumors modulate their environment and keep stromal cells in an activated and tumor-promoting state. Of note is the necessity for even small tumors to induce angiogenesis: the sprouting of new blood vessels from existing blood vessels [41]. This process requires a complex cellular and molecular cooperation between different tissue and inflammatory cells, such as endothelial cells, fibroblasts and macrophages. Under inflammatory conditions, such as wound healing, or in growing tumors, endothelial cells are activated, start to divide and form new blood vessels. Frequently detected growth factors that activate endothelial cells are vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF). Tumor cells, tumor-associated macrophages (TAMs) and fibroblasts (TAFs) produce VEGF and are involved in tumor-induced angiogenesis [42–45]. Both TAMs and TAFs are abundantly detected in tumors
Vol.11 No.5 May 2005
ECM GF F G
T M BV P E TRENDS in Molecular Medicine
Figure 1. The cellular composition of solid tumors. Solid tumors are complex tissues that are maintained by dynamic interactions between cancer cells (blue) and the microenvironment. This microenvironment is referred to as tumor stroma. The tumor stroma is composed of different cell types, such as fibroblasts (F), pericytes (P), endothelial cells (E) and immune cells, including T cells (T), granulocytes (G) and macrophages (M). Stromal cells fulfill a variety of functions. Fibroblasts produce extracellular matrix (ECM) proteins for mechanical support. Endothelial cells that make up the inner lining of blood vessels (BV) ensure a blood supply for growing tumors by neoangiogenesis. Hematopoietic stromal cells can provide tumor-promoting growth factors (GFs) and macrophages remove waste and dead cells. Shown is a schematic representation of the cellular composition of tumor tissue that might, depending upon the tumor type and stage, vary considerably.
and often support tumor growth [43,44]. If macrophages were inhibited from infiltrating tumors, angiogenesis was impaired and tumors were rejected [46,47]. The deficiency of macrophages in cancer-prone (MMTVpy-midT) transgenic mice delayed the development of invasive mammary carcinomas [48]. TAFs are a morphologically homogeneous but functionally heterogeneous group of mesenchymal cells. Some fibroblasts are tissue-resident matrix synthesizing and/or matrix-degrading cells, whereas others are contractile cells (myofibroblasts), circulating precursor cells (fibrocytes) or blood-vessel-associated pericytes. Fibroblasts can produce several growth factors, cytokines, chemokines and matrix-degrading enzymes and, therefore, might modulate the local immune response [44]. In most systems analyzed, TAFs have pro-tumorigenic effects, such that the co-injection of fibroblasts facilitated the growth of low-tumorigenic cells [49]. The importance of VEGF-producing TAFs for angiogenesis and tumor growth was recently shown [45]. The inhibition of tumor angiogenesis as a therapeutic strategy is recognized and several angiogenesis inhibitors, such as anti-VEGF antibodies, have been developed [50]. Only recently it became clear that effective immune therapy primarily acts on the tumor stroma and similarly interferes with tumorinduced angiogenesis.
Figure 2. T cells target the stroma and hit the tumor by multiple mechanisms. (a) CD8C effector T cells recognize MHC-class-I–antigen-peptide complexes on cancer cells (blue) and kill them with perforin (pfp). Although this mechanism is often used, it is usually not sufficient for tumor rejection. (b) CD8C effector T cells recognize MHC-class-I– antigen-peptide complexes on cancer cells and produce IFN-g, which acts on tumor stroma (e.g. endothelial cells; yellow), resulting in the inhibition of tumor-induced www.sciencedirect.com
Review
TRENDS in Molecular Medicine
(a)
Vol.11 No.5 May 2005
227
(b) CD8
CD8
IFNγ TCR
TCR pfp
MHC I
MHC I Ag
Ag Cancer cell killed
(c)
CD4 (TH1)
Angiostasis
CD4 (TH2)
(d)
IL-4 IFNγ
TCR
TCR TCR
MHC II E
TAM
TAM
Ag
Ag
Angiostasis
Angiostasis
CD8
(e)
TAF
MHC II
(f)
CD8
TCR
TCR pfp
pfp MHC I MHC I TAM
non-BM cells
(CD11b+)
Ag Stroma cell killed
Ag Stroma cell killed TRENDS in Molecular Medicine
angiogenesis (angiostasis). (c) IFN-g-producing CD4C effector T cells (Th1), as do CD8C T cells, induce angiostasis in MHC class IIK tumors. Therefore, MHC-class-IIexpressing tumor stroma cells must have taken up, processed and presented tumor antigens by MHC class II molecules. These are presented to Th1 cells, which produce IFN-g that acts, directly or indirectly, on endothelial cells. Because of their abundance, we assume that tumor-associated macrophages (TAMs) are MHC class IIC cells, but this has yet to be proven. (d) IL-4-producing CD4C effector T cells (Th2), similar to Th1 cells, induce angiostasis, using tumor-associated fibroblasts (TAFs) as targets. (e) CD8C effector T cells recognize antigen on antigen cross-presenting tumor-associated (CD11bC) macrophages and kill them with perforin. (f) CD8C effector T cells recognize antigen on antigen cross-presenting non-bone-marrow-derived tumor stroma cells (non-BM cells), such as endothelial cells or fibroblasts, and kill them with perforin. Abbreviations: Ag, tumor antigen; TCR, T-cell receptor. www.sciencedirect.com
228
Review
TRENDS in Molecular Medicine
CD8C T cells recognize and kill cancer cells It is often assumed that the major mechanism of tumor rejection is the direct recognition and killing of tumor cells by CD8C effector T cells; therefore, they are termed CTLs. CTLs recognize MHC-class-I–peptide complexes on target cells and, with the help of perforin, lyse them (Figure 2a) [51]. Although cytotoxicity is often necessary for complete tumor rejection, it is usually not sufficient. This becomes obvious from studies showing that tumor rejection by CD8C effector T cells correlated better with IFN-g and TNF-a production than with cytotoxicity [52,53]. More importantly, a number of recent studies show that IFN-g or IFN-g-receptor gene-deficient (knockout; KO) mice cannot reject tumors in most systems, although these mice develop normal numbers of T cells and cytotoxic responses [4–18]. As demonstrated in perforin-KO mice or with adoptively transferred T cells from perforin-KO mice, perforin (or Fas-ligand, which is another cytotoxic molecule) was necessary for tumor rejection in some, but not all, models, but they generally required IFN-g [4]. IFN-g from CD8C T cells inhibits angiogenesis Schu¨ler et al. [13] analyzed the cells on which CD8C effector T cells recognize antigen during tumor rejection. They used CD8C OT-1 T cells, which were derived from T-cell-receptor-transgenic mice, that were specific for a peptide of the model antigen ovalbumin (OVA) presented by H2-Kb molecules. Following in vitro activation and adoptive transfer, OT-1 T cells rejected, in most of the mice, B16 melanoma cells that were transfected to express OVA. Bone-marrow reconstitution experiments demonstrated that bone-marrow-derived cells were not required to present OVA to achieve tumor rejection, indicating that that antigen recognition on tumor cells was sufficient for tumor rejection by the CD8C effector T cells. Interestingly, OT-1 T cells could not reject B16-OVA tumors in IFNg-receptor-KO mice, showing that IFN-g produced by the CD8C effector T cells acted on the host cells. In IFN-g-KO mice, OT-1 T cells rejected the tumor. Therefore, IFN-g expression by the transferred OT-1 T cells was sufficient for tumor rejection [13]. Because IFN-g expression by CD8C effector T cells is tightly regulated and requires antigen recognition, it is likely that IFN-g acted on stromal cells in the vicinity of the tumor cells. Qin et al. [12] investigated the cells within the tumor stroma on which IFN-g from CD8C T cells acted upon to induce tumor rejection. They showed that tumor rejection mediated by vaccination-induced CD8C T cells was always preceded by the inhibition of tumor-induced angiogenesis and that endothelial (CD31C) cells were arrested at the rim of the tumor. Angiostasis and tumor rejection was also observed in perforin-KO mice but not in IFN-g-KO mice. Only adoptively transferred CD8C effector T cells from wild-type, but not IFN-g-KO, mice inhibited angiogenesis in experimental lung metastases [12]. Together, these data suggest that CD8C effector T cells recognize antigen on tumor cells and express IFN-g that acts on tumor stroma, probably endothelial cells (Figure 2b). One cannot exclude, however, that antigen cross-presentation by non-bone-marrow-derived tumor stroma cells (e.g. endothelial cells) contributed to tumor rejection. It is www.sciencedirect.com
Vol.11 No.5 May 2005
interesting to note that CD8C effector T cells inhibited, but could not completely prevent, tumor growth in IFNg-receptor-KO mice [13]. This suggests that cytotoxicity slows down tumor growth but is not sufficient for its long-term inhibition. CD4C T cells recognize tumor antigen-presenting stromal cells and induce angiostasis through IFN-g It has been observed that, in some models, CD4C T cells are the main effector cells during tumor rejection, even if tumor cells do not express detectable levels of MHC class II molecules [3]. In a model of CD4C T-cell-mediated tumor rejection, it was shown that IFNg receptor was essential for tumor rejection [6]. Tumor rejection in immunized wild-type mice was equally efficient regardless of whether or not the tumor cells expressed IFNg-receptor. In fact, in most, but not all, systems, the upregulation of MHC molecules by IFN-g appears to have a minor role during tumor rejection [4,6,21,54–56]. CD4C T cells in both immunized IFNg-receptor-KO and wild-type mice produced IFN-g in a tumor-specific fashion, homed to the challenge site and, upon transfer into wild-type mice, rejected IFNg-receptor-negative tumors. Bone-marrowchimeric mice revealed that IFNg-receptor expression on non-bone-marrow-derived cells was necessary and sufficient for tumor rejection. Similar to tumor rejection by CD8C T cells, CD4C T cells inhibited tumor-induced angiogenesis by IFN-g. Because tumor cells did not express MHC class II molecules in this model, one can postulate that MHC class IIC tumor stroma cells presented tumor-derived antigens to CD4C T cells, thereby inducing IFN-g and inhibiting angiogenesis (Figure 2c). The MHC class IIC stromal cells that present antigen to the CD4C effector T cells remain elusive, but dendritic cells or, because of their abundance in most solid tumors, macrophages are potential candidates. It should be emphasized that the presented mechanisms should not be seen as working separately, but it is likely that several mechanisms cooperate during tumor rejection. CD4C T cells recognize stromal cells and induce angiostasis through IL-4 In adoptive T-cell-transfer experiments, IL-4-producing CD4C T cells were shown to eliminate B16 lung metastases [23]. Because B16 cells do not express MHC class II molecules, one has to assume that MHC class IIC stromal cells presented tumor-derived antigens to effector CD4C T cells to induce IL-4 production by the T cells. Tumors that are transfected to produce IL-4 are rejected in wildtype mice and are inhibited in the long term in T-cell-KO mice by a paracrine action on host cells [57,58]. It was assumed that tumor-reactive Th2 cells attract and activate eosinophils, which then kill tumor cells [23]. The role of eosinophils during IL-4-mediated tumor rejection is, however, controversial because IL-5-KO mice that lack eosinophils reject IL-4-producing tumors as efficiently as do wild-type control mice [59]. Recent data gave a better insight into the mechanism of IL-4-induced tumor rejection, if one assumes that the expression of IL-4 by the tumor mimics the local production by tumor-reactive Th2 cells that accumulate
Review
TRENDS in Molecular Medicine
at the tumor site. Using bone-marrow-chimeric mice, it was shown that IL-4-receptor (IL4R) expression on both hematopoietic and nonhematopoietic cells was sufficient for the rejection of IL-4-producing tumors [60]. TAFs were identified as one target of IL-4. In situations in which tumor cells did not produce IL-4 or host cells could not respond to IL-4, TAFs infiltrated tumors in an IL-4-independent fashion and were associated with blood vessels. In IL-4-secreting tumors grown in wild-type mice, TAFs were similarly present but more-pronounced ECM was detected, indicating that IL-4-responsive fibroblasts in the presence of IL-4 were differently activated. Blood vessels in these tumors were virtually absent and the tumors were subsequently rejected. These findings are compatible with the notion that IL-4 has profound effects on fibroblasts and suppresses angiogenesis [61]. The importance of fibroblasts as a target for IL-4 was underlined by the demonstration that IL-4-producing tumors could be rejected, even if TAFs were the only cells that could respond to IL-4 [60]. Based on these data, it is postulated that Th2 cells recognizing tumor-derived antigen on MHC class IIC tumor stroma cells produce IL-4 that acts on fibroblasts, thereby inhibiting their participation in tumor-induced angiogenesis (Figure 2d). Because mice with selective IL4R expression in bonemarrow-derived cells also rejected IL-4-producing tumors, it is interesting to note that bone-marrow-derived fibroblasts and myofibroblasts accumulated in tumors; therefore, it is possible that bone-marrow-derived fibroblasts supported tumor-induced angiogenesis [62]. CD8C T cells kill bone-marrow-derived stromal cells cross-presenting tumor antigen An interesting mechanism of how T cells contribute to tumor rejection was recently shown. In an adoptivetransfer model with tumor-specific CD8C effector T cells, tumor rejection required, in addition to tumor and nonbone-marrow-derived cells, bone-marrow-derived cells to cross-present tumor-derived antigen [63]. CD8C effector T cells lysed antigen cross-presenting CD11bC tumor stroma cells, typically monocytes and macrophages, probably by perforin (Figure 2e). Tumor rejection and cytolysis of CD11bC tumor stroma cells occurred only when the tumor expressed a large amount of antigen so that it could be cross-presented to the CD8C effector T cells. Tumors often express macrophage chemoattractants [43] and TAMs promote tumor growth, for example by contributing to angiogenesis [42]. The killing of TAMs by CD8C effector T cells is, therefore, an effective strategy to destroy the tumor infrastructure. CD8C T cells kill non-bone-marrow-derived stromal cells cross-presenting tumor antigen A mechanism suggested by the data of Spiotto et al. [63] is shown in Figure 2f. In their model, tumor cells express large amounts of antigen, which is cross-presented by bone-marrow-derived and non-bone-marrow-derived stromal cells. This enables adoptively transferred CD8C T cells to efficiently kill tumor cells directly and eliminate cross-presenting stromal cells, which is required to prevent the outgrowth of antigen loss variants [63]. The www.sciencedirect.com
Vol.11 No.5 May 2005
229
requirement for antigen cross-presentation by non-bonemarrow-derived cells, such as endothelial cells or fibroblasts, strongly suggests that they are targets of CD8C effector T cells. It is therefore likely that non-bonemarrow-derived TAFs or endothelial cells were killed, probably by perforin, which was required for rapid tumor regression (Figure 2f) [63]. Together, the above examples (Figure 2a–f) illustrate that T cells can use a variety of mechanisms to destroy tumors. Cyclophosphamide induces IFNg-receptor-dependent destruction of the tumor vasculature The above experiments indicated that IFN-g produced by T cells primarily inhibited the early phase of tumorinduced angiogenesis. However, IFN-g can also contribute to the destruction of an established tumor stroma [64]. In some models, a single injection of cyclophosphamide (Cy) induces immune-mediated rejection of Cy-resistant, established and vascularized tumors. Starting as early as 6 h after Cy-injection, the tumor-infiltrating T cells were inactivated, TAMs stopped producing IL-10 and started to produce IFN-g and, with the same kinetics, the tumor blood vessels were destroyed. Remarkably, blood vessel destruction and tumor rejection required IFN-g-receptor expression on cells of the tumor-bearing host, probably tumor stroma cells. The disappearance of (Meca32C or CD31C) endothelial cells preceded central tumor necrosis and the subsequent elimination of residual tumor cells by CD8C T cells. These data confirm that IFN-g targets, either directly or by secondary factors, endothelial cells. They also suggest that Cy, which is thought to eliminate suppressor T cells [65] and facilitate homeostatic T-cell proliferation [26], contributes to tumor rejection by modulating the tumor stroma. Concluding remarks Armed with cytokines and cytotoxic molecules, T cells use a multitude of strategies to reject tumors. They can recognize tumor antigens on tumor and multiple stroma cell types and destroy or modulate the function of multiple tumor stroma cell types, such as TAMs, fibroblasts and endothelial cells. Although phenotypically different effector T cells might target different tumor stroma cells, they act on the stroma in a redundant fashion to withdraw its blood supply. However, the tumor stroma can also act as a barrier [66] and prevent the access of T cells to the tumor [32]. Effective therapeutic modalities to modulate the stroma such that T-cell entry into tumors is improved have already been found, such as CpG oligonucleotides [31] or cyclophosphamide [64]. As our understanding of tumorstroma–immune-cell interactions increases, we can optimistically anticipate more precise and effective immunotherapies to be developed. Acknowledgements This review was supported by grants from the Deutsche Forschungsgemeinschaft and the Deutsche Krebshilfe, Dr. Mildred Scheel Stiftung.
References 1 Schoenberger, S.P. et al. (1998) T-cell help for cytotoxic T lymphocytes is mediated by CD40–CD40L interactions. Nature 393, 480–483
230
Review
TRENDS in Molecular Medicine
2 Bourgeois, C. et al. (2002) A role for CD40 expression on CD8C T cells in the generation of CD8C T cell memory. Science 297, 2060–2063 3 Greenberg, P.D. (1991) Adoptive T cell therapy of tumors: mechanisms operative in the recognition and elimination of tumor cells. Adv. Immunol. 49, 281–355 4 Blankenstein, T. and Qin, Z. (2003) The role of IFN-g in tumor transplantation immunity and inhibition of chemical carcinogenesis. Curr. Opin. Immunol. 15, 148–154 5 Winter, H. et al. (1999) Tumor regression after adoptive transfer of effector Tcells is independent of perforin or Fas ligand (APO-1L/CD95L). J. Immunol. 163, 4462–4472 6 Qin, Z. and Blankenstein, T. (2000) CD4C T cell-mediated tumor rejection involves inhibition of angiogenesis that is dependent on IFNg receptor expression by nonhematopoietic cells. Immunity 12, 677–686 7 Prevost-Blondel, A. et al. (2000) Differential requirement of perforin and IFN-g in CD8 T cell-mediated immune responses against B16.F10 melanoma cells expressing a viral antigen. Eur. J. Immunol. 30, 2507–2515 8 Peng, L. et al. (2000) T cell-mediated tumor rejection displays diverse dependence upon perforin and IFN-g mechanisms that cannot be predicted from in vitro T cell characteristics. J. Immunol. 165, 7116–7124 9 Winter, H. et al. (2001) Immunotherapy of melanoma: a dichotomy in the requirement for IFN-g in vaccine-induced antitumor immunity versus adoptive immunotherapy. J. Immunol. 166, 7370–7380 10 Wigginton, J.M. et al. (2001) IFN-g and Fas/FasL are required for the antitumor and antiangiogenic effects of IL-12/pulse IL-2 therapy. J. Clin. Invest. 108, 51–62 11 van Elsas, A. et al. (2001) Elucidating the autoimmune and antitumor effector mechanisms of a treatment based on cytotoxic T lymphocyte antigen-4 blockade in combination with a B16 melanoma vaccine: comparison of prophylaxis and therapy. J. Exp. Med. 194, 481–481 12 Qin, Z. et al. (2003) A critical requirement of IFNg-mediated angiostasis for tumor rejection by CD8C T cells. Cancer Res. 63, 4095–4100 13 Schu¨ler, T. and Blankenstein, T. (2003) CD8C effector T cells reject tumors by direct antigen recognition but indirect action on host cells. J. Immunol. 170, 4427–4431 14 Poehlein, C.H. et al. (2003) TNF plays an essential role in tumor regression after adoptive transfer of perforin/IFN-g double knockout effector T cells. J. Immunol. 170, 2004–2013 15 Kowalczyk, D.W. et al. (2003) Vaccine-induced CD8C T cells eliminate tumors by a two-staged attack. Cancer Gene Ther. 10, 870–878 16 Nanni, P. et al. (2004) Immunoprevention of mammary carcinoma in HER-2/neu transgenic mice is IFN-g and B cell dependent. J. Immunol. 173, 2288–2296 17 Hollenbough, J.A. et al. (2004) The rate of the CD8-dependent initial reduction in tumor volume is not limited by contact-dependent perforin, Fas ligand, or TNF-mediated cytolysis. J. Immunol. 173, 1738–1743 18 Wu, T.H. et al. (2004) Long-term suppression of tumor growth by TNF requires a Stat1- and IFN response factor 1-dependent IFNg pathway but not IL-12 or IL-18. J. Immunol. 172, 3243–3251 19 Dobrzanski, M.J. et al. (2000) Type 1 and type 2 CD8C effector T cell subpopulations promote long-term tumor immunity and protection to progressively growing tumor. J. Immunol. 164, 916–925 20 Dobrzanski, M.J. et al. (2001) Role of effector cell-derived IL-4, IL-5, and perforin in early and late stages of type 2 CD8 effector cellmediated tumor rejection. J. Immunol. 167, 424–434 21 Mumberg, D. et al. (1999) CD4C T cells eliminate MHC class IInegative cancer cells in vivo by indirect effects of IFN-g. Proc. Natl. Acad. Sci. U. S. A. 96, 8633–8638 22 Nishimura, T. et al. (1999) Distinct role of antigen-specific T helper type 1 (Th1) and Th2 cells in tumor eradication in vivo. J. Exp. Med. 190, 617–627 23 Mattes, J. et al. (2003) Immunotherapy of cytotoxic T cell-resistant tumors by T helper 2 cells: An eotaxin and STAT6-dependent process. J. Exp. Med. 197, 387–393 24 Nguyen, L.T. et al. (2002) Tumor growth enhances cross-presentation leading to limited T cell activation without tolerance. J. Exp. Med. 195, 423–435 www.sciencedirect.com
Vol.11 No.5 May 2005
25 Lyman, M.A. et al. (2004) A spontaneously arising pancreatic tumor does not promote the differentiation of naive CD8C T lymphocytes into effector CTL. J. Immunol. 172, 6558–6567 26 Dudley, M.E. and Rosenberg, S.A. (2003) Adoptive-cell-transfer therapy for the treatment of patients with cancer. Nat. Rev. Cancer 3, 666–675 27 Sadovnikova, E. and Stauss, H.J. (1996) Peptide-specific cytotoxic T lymphocytes restricted by nonself major histocompatibility complex class I molecules: reagents for tumor immunotherapy. Proc. Natl. Acad. Sci. U. S. A. 93, 13114–13118 28 Spiotto, M.T. et al. (2002) Increasing tumor antigen expression overcomes ‘ignorance’ to solid tumors via crosspresentation by bone marrow-derived stromal cells. Immunity 17, 737–747 29 Wolkers, M.C. et al. (2004) Antigen bias in T cell cross-priming. Science 304, 1314–1317 30 Norbury, C.C. et al. (2004) CD8C T cell cross-priming via transfer of proteasome substrates. Science 304, 1318–1321 31 Garbi, N. et al. (2004) CpG motifs as proinflammatory factors render autochthonous tumors permissive for infiltration and destruction. J. Immunol. 172, 5861–5869 32 Ganss, R. et al. (2004) Overcoming tumor-intrinsic resistance to immune effector function. Eur. J. Immunol. 34, 2635–2641 33 Qin, Z. and Blankenstein, Th. (2004) A cancer immunosurveillance controversy. Nat. Immunol. 5, 3–4 34 Blankenstein, Th. and Qin, Z. (2003) Chemical carcinogens as foreign bodies and some pitfalls regarding cancer immune surveillance. Adv. Cancer Res. 90, 179–207 35 Tlsty, T.D. and Hein, P.W. (2001) Know thy neighbor: stromal cells can contribute oncogenic signals. Curr. Opin. Genet. Dev. 11, 54–59 36 Bissell, M.J. and Radisky, D. (2001) Putting tumours in context. Nat. Rev. Cancer 1, 46–54 37 Coussens, L.M. and Werb, Z. (2002) Inflammation and cancer. Nature 420, 860–867 38 Mueller, M.M. and Fusenig, N.E. (2004) Friends or foes – bipolar effects of the tumour stroma in cancer. Nat. Rev. Cancer 4, 839–849 39 Philip, M. et al. (2004) Inflammation as a tumor promoter in cancer induction. Semin. Canc. Biol. 14, 433–439 40 Bhowmick, N.A. et al. (2004) TGF-b signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303, 848–851 41 Folkman, J. (2003) Fundamental concepts of the angiogenic process. Curr. Mol. Med. 3, 643–651 42 Polverini, P.J. and Leibovich, J. (1984) Induction of neovascularization in vivo and endothelial proliferation in vitro by tumor-associated macrophages. Lab. Invest. 51, 635–642 43 Mantovani, A. et al. (2002) Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol. 23, 549–555 44 Silzle, T. et al. (2004) The fibroblast: sentinel cell and local immune modulator in tumor tissue. Int. J. Cancer 108, 173–180 45 Dong, J. et al. (2004) VEGF-null cells require PDGFRa signalingmediated stromal fibroblast recruitment for tumorigenesis. EMBO J. 23, 2800–2810 46 Richter, G. et al. (1993) Interleukin 10 transfected into Chinese hamster ovary cells prevents tumor growth and macrophage infiltration. Cancer Res. 53, 4134–4137 47 Di Carlo, E. et al. (1998) Local release of interleukin-10 by transfected mouse adeno-carcinoma cells exhibits pro- and anti-inflammatory activity and results in a delayed tumor rejection. Eur. Cytokine Netw. 9, 61–68 48 Lin, E.Y. et al. (2001) Colony-stimulating factor 1 promotes progression of mammary tumors to malignancy. J. Exp. Med. 193, 727–739 49 Camps, J.L. et al. (1990) Fibroblast-mediated acceleration of human epithelial tumor growth in vivo. Proc. Natl. Acad. Sci. U. S. A. 87, 75–79 50 Kerbel, R. and Folkman, J. (2002) Clinical translation of angiogenesis inhibitors. Nat. Rev. Cancer 2, 727–739 51 van den Broek, M.F. and Hengartner, H. (2000) The role of perforin in infections and tumour surveillance. Exp. Physiol. 85, 681–685 52 Barth, R.J. et al. (1991) Interferon g and tumor necrosis factor have a role in tumor regressions mediated by murine CD8C tumor-infiltrating lymphocytes. J. Exp. Med. 173, 647–658 53 Becker, C. et al. (2001) Adoptive tumor therapy with T lymphocytes enriched through an IFN-g capture assay. Nat. Med. 7, 1159–1162
Review
TRENDS in Molecular Medicine
54 Dighe, A.S. et al. (1994) Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN g receptors. Immunity 1, 447–456 55 Pulaski, B.A. et al. (2002) Interferon-g-dependent phagocytic cells are a critical component of innate immunity against metastatic mammary carcinoma. Cancer Res. 62, 4406–4412 56 Segal, J.G. et al. (2002) The role of IFN-g in rejection of established tumors by IL-12: source of production and target. Cancer Res. 62, 4696–4703 57 Tepper, R.I. et al. (1989) Murine interleukin-4 displays potent antitumor activity in vivo. Cell 57, 503–512 58 Hock, H. et al. (1993) Mechanisms of rejection induced by tumor celltargeted gene transfer of interleukin 2, interleukin 4, interleukin 7, tumor necrosis factor, or interferon g. Proc. Natl. Acad. Sci. U. S. A. 90, 2774–2778 59 Noffz, G. et al. (1998) Neutrophils but not eosinophils are involved in growth suppression of IL-4-secreting tumors. J. Immunol. 160, 345–350
Vol.11 No.5 May 2005
231
60 Schuler, T. et al. (2003) Tumor rejection by modulation of tumor stromal fibroblasts. J. Exp. Med. 198, 1487–1493 61 Volpert, O.V. et al. (1998) Inhibition of angiogenesis by interleukin 4. J. Exp. Med. 188, 1039–1046 62 Direkze, N.C. et al. (2004) Bone marrow contribution to tumorassociated myofibroblasts and fibroblasts. Cancer Res. 64, 8492–8495 63 Spiotto, M.T. et al. (2004) Bystander elimination of antigen loss variants in established tumors. Nat. Med. 10, 294–298 64 Ibe, S. et al. (2001) Tumor rejection by disturbing tumor stroma cell interactions. J. Exp. Med. 194, 1549–1559 65 North, R.J. (1982) Cyclophosphamide-facilitated adoptive immunotherapy of an established tumor depends on elimination of tumorinduced suppressor T cells. J. Exp. Med. 155, 1063–1074 66 Singh, S. et al. (1992) Stroma is critical for preventing or permitting immunological destruction of antigenic cancer cells. J. Exp. Med. 175, 139–146
Endeavour the quarterly magazine for the history and philosophy of science You can access Endeavour online via ScienceDirect, where you’ll find a collection of beautifully illustrated articles on the history of science, book reviews and editorial comment.
Featuring Sverre Petterssen and the Contentious (and Momentous) Weather Forecasts for D-Day, 6 June 1944 by J.R. Fleming Food of Paradise: Tahitian breadfruit and the Autocritique of European Consumption by P. White and E.C. Spary Two Approaches to Etiology: The Debate Over Smoking and Lung Cancer in the 1950s by M. Parascandola Sicily, or sea of tranquility? Mapping and naming the moon by J. Vertesi The Prehistory of the Periodic Table by D. Rouvray Two portraits of Edmond Halley by P. Fara and coming soon Fighting the ‘microbe of sporting mania’: Australian science and Antarctic exploration in the early twentieth century by P. Roberts Learning from Education to Communicate Science as a Good Story by A. Negrete and C. Lartigue The Traffic and Display of Body Parts in the Early-19th Century by S. Alberti and S. Chaplin The Rise, Fall and Resurrection of Group Selection by M. Borrello Pomet’s great ‘‘Compleat History of Drugs’’ by S. Sherman Sherlock Holmes: scientific detective by L. Snyder The Future of Electricity in 1892 by G.J.N. Gooday The First Personal Computer by J. November Baloonmania: news in the air by M.G. Kim and much, much more . . . Locate Endeavour on ScienceDirect (http://www.sciencedirect.com) www.sciencedirect.com