How Tumours Escape Mass Destruction

TJ Stewart1 and SI Abrams2

Laboratory of Tumor Immunology and Biology, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA

It is now well established that the immune system can effector mechanisms, resulted in a significantly elevated control neoplastic development and growth in a process incidence and rate of spontaneous, chemically induced termed immunosurveillance. A link between host immuno- or autochthonous tumour formation (Smyth et al., 2000, surveillance and neoplastic progression is revealed in cases 2006; Takeda et al., 2001; Pollard, 2004; Dunn et al., where the immune response becomes compromised due to 2006; Bui and Schreiber, 2007; Lin and Pollard, 2007; genetic or other pathological conditions, resulting in a Street et al., 2007). Similarly, immunocompromised substantially increased incidence and rate ofspontaneous patients, particularly transplant recipients, appear tumour formation in both preclinical animal models and to be more susceptible to certain types of neoplasms patients. It has also been demonstrated in tumour-bearing (Penn, 1999; Pawelec et al., 2002). Thus, the generation hosts that the tumorigenic process itselfcan promote a and regulation of appropriate pro-inflammatory inter- state ofimmunosuppression that, in turn, facilitates actions can indeed control and promote the eradication neoplastic progression. The ability ofneoplastic popula- of susceptible neoplastic populations. tions to induce a hostile microenvironment through both However, tumour cells within the mass or lesion can also cell contact-dependent and -independent immunosuppres- employ mechanisms that circumvent or usurp these immune sive networks is a significant barrier to effective cell- reactions to enhance their own growth. These malignant mediated immunity and immunotherapy. Thus, a competent populations may do so in a number of ways, chiefly through immune system is integral for the control of neoplastic the production of diverse tumour-derived factors (TDFs) disease, and dissecting the plethora oftumour escape that function to (a) facilitate tumour growth in an autocrine mechanisms that disrupt this essential host defense fashion, (b) recruit host stromal cells, namely fibroblasts capability is integral for the development of effective and tumour-associated macrophages, which further nurture immunotherapeutic paradigms. the tumorigenic process, (c) recruit and engage multiple host Oncogene (2008) 27, 5894–5903; doi:10.1038/onc.2008.268 immune suppressive cell populations that downregulate innate or adaptive immune responses locally or systemically Keywords: tumour immunity; tumour escape; tumour and/or (d) suppress innate or adaptive immune responses microenvironment; immunosuppression; immuno- directly. However, when a productive antitumour immune surveillance response does develop, it can also serve as a biological selective pressure that promotes the emergence of more aggressive tumour escape variants in a process termed cancer immunoediting (Dunn et al., 2006; Smyth et al., The tumour microenvironment and immune suppression 2006; Bui and Schreiber, 2007). Therefore, although the goal of an antitumour response is to mediate tumour cell Compelling data support the view that the immune destruction, this very response may also have unintended system is important for the control of neoplastic tumour-promoting consequences due to immunoediting or development and growth (Dunn et al., 2006; Smyth the induction of an inflammatory response that may et al., 2006; Bui and Schreiber, 2007). If innate or stimulate tumour progression (Dunn et al., 2006; Smyth adaptive immunity becomes impaired or suppressed, et al., 2006; Bui and Schreiber, 2007; Stewart et al., 2007). tumour development can occur. In mouse models, it was In addition to the TDF-mediated mechanisms of demonstrated that the loss of immune function, tumour escape that drive a highly immune suppressive especially affecting interferon-g (IFN-g) production, microenvironment, neoplastic cells may escape the T lymphocytes and natural killer cells or their immune response by avoiding or reducing the efficiency of immune recognition. In the case of tumour-associated antigens (TAA), which represent self-Ag and are shared Correspondence: Dr SI Abrams, Department of Immunology, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263, by normal host cells, the efficiency of Ag recognition, USA. however, may not be affected. Instead, the naive E-mail: [email protected] peripheral T-cell pool will likely not contain high- 1Current address: Cancer Immunology Research Program, The Peter affinity T cells that will recognize self-Ag and any MacCallum Cancer Institute, Level 2, Smorgon Family Building, St Andrews Place, East Melbourne, Victoria 3002, Australia lower affinity T cells that will be present, may be 2Current address: Department of Immunology, Roswell Park Cancer tolerized or rendered anergic. TAA-specific, but func- Institute, Elm and Carlton Streets, Buffalo, NY 14263, USA tionally unresponsive, T-cell populations are evident in How tumours escape mass destruction TJ Stewart and SI Abrams 5895 tumour-infiltrating lymphocytes obtained from patients escape). Therefore, DCs have been shown to play a with malignant melanoma (Dudley and Rosenberg, major role in dictating whether T-cell priming or 2003). Following appropriate ex vivo stimulation, tolerance occurs (Steinman et al., 2003). For full T-cell however, these tumour-infiltrating lymphocytes can activation, the DC must have the capacity to process recover functional competence and mediate tumour and present TAA to both CD4 þ and CD8 þ T cells and regression in at least subsets of patients receiving simultaneously supply the relevant costimulatory autologous adoptive tumour-infiltrating lymphocyte signals. Such signals include cell-surface expression of therapy, particularly under joint conditions of non- a family of B7 molecules, as well as the production of myeloablative chemotherapy (Dudley et al., 2002; pro-inflammatory cytokines, namely interleukin (IL)-12, Dudley and Rosenberg, 2003). This clinical scenario which enable productive T-cell activation (Pardoll, 2003; indicates that the immune response can play an integral Steinman et al., 2003). However, inefficient T-cell role in the neoplastic process and, if appropriately activation may occur because of the absence of manipulated, can mediate significant antitumour activ- inflammatory mediators and an immunosuppressive ity. Moreover, these observations reveal that the tumour or tolerogenic host–tumour microenvironment consist- microenvironment can adversely affect the competence ing, in part, of immature DC that fail to express the of tumour-reactive lymphocytes and, in conjunction requisite pro-inflammatory and costimulatory charac- with depressed pro-inflammatory signals, facilitate teristics that are needed to induce a favourable tumour escape. Thus, it has become a challenging task antitumour response. to induce productive antitumour immune responses In the Ag-specific CD4 þ T-cell compartment, pro- against ‘self-TAA’ (Pardoll, 2003). gressive tumour growth induces various forms of It has also been demonstrated that clonally unresponsiveness and coincides with reduced therapeu- expanded tumour cells are genetically unstable and tic efficacy (Lee et al., 1999; Willimsky and Blanken- can readily acquire new mutations or downregulate stein, 2005). Therefore, in vivo interactions between tumour suppressor genes through epigenetic mechan- CD4 þ T cells and their Ag-bearing tumours lead to a isms (Hanahan and Weinberg, 2000; Vogelstein and functionally heterogeneous population of T cells com- Kinzler, 2004). Indeed, these genetic and epigenetic prising anergic, naive (unactivated) and inhibitory alterations underlie the mechanisms for aggressive (regulatory T cells, Treg) phenotypes (Lee et al., 1999; tumorigenic behaviour, including the nature and types Willimsky and Blankenstein, 2005; Zhou et al., 2006). of TDF that contribute to tumour immune escape. CD8 þ T cells can also exhibit tolerogenic characteristics During the evolution of the tumorigenic process, there is in tumour-bearing hosts. These characteristics can a dynamic cross talk with host cells, which can include manifest in a number of ways, such as through those of the immune system. Therefore, productive ignorance (Ochsenbein et al., 1999) or tolerance to tumour development may depend, in part, on the TAA (Overwijk et al., 2003), or through a phenomenon balance of positive and negative signals generated termed split anergy (Otten and Germain, 1991) by which during these host–tumour interactions. Ag-reactive CD8 þ T cells could be detected in the draining lymph nodes, but were deficient in one or more The phenomenon of immune tolerance effector functions. Overall, the balance of signals It is well recognized that neoplastic cells are antigenic, favouring immune tolerance over immune activation but often weakly to non-immunogenic; that is, they may explain, in part, tumour progression in the presence express Ag against which an immune response could of tumour-reactive T cells, which may have been be directed, but lack the capacity to create a ignored, rendered anergic or even deleted due to pro-inflammatory environment necessary to sustain mechanisms of tolerance. efficient T-cell activation and clonal expansion following So, what are some other key factors in the tumour Ag recognition. An important mechanism of tolerance microenvironment that influence the outcome of tumour induction is insufficient or inappropriate Ag processing regression versus tumour escape? As part of the dynamic and presentation of TAA by antigen-presenting cells dialogue within the tumour microenvironment, tumour (APCs) (Cuenca et al., 2003). Ordinarily APCs, cells may escape immune recognition or attack through mainly dendritic cells (DCs), monocytes/macrophages two general paradigms: (a) tumour cell contact-inde- and activated B cells, express or upregulate relevant pendent mechanisms that involve the secretion of TDFs, major histocompatibility complex (MHC) and costimu- which inhibit immune cells either directly or indirectly latory molecules under the appropriate inflammatory through engagement of third-party immune suppressive conditions. These cells are then capable of inducing cell sub-populations; and (b) tumour cell contact- immune activation by presenting TAA through either dependent mechanisms that involve direct interactions exogenous or endogenous (that is, cross-priming) between the tumour cell and the immune cell. pathways to CD4 þ or CD8 þ T cells, respectively. DCs are mainly responsible for the initiation or potentiation of an immune response; however, within Tumour cell contact-independent mechanisms a highly immunosuppressive tumour microenvironment, DC may fail to induce effector functions in the Ag- Tumour-derived factors specific T cells that they interact with (see also section Numerous TDFs with pro-apoptotic or immune below on contact-independent mechanisms of tumour suppressive properties have been characterized

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5896 (Figure 1), most notably, transforming growth factor-b the tumour microenvironment. This, in turn, impairs (TGF-b), vascular endothelial factor (VEGF), IL-10 DC maturation and activation, resulting in populations and prostaglandins (that is, PGE2). Although such of DC that are unable to support antitumour immunity. factors have diverse and multi-functional consequences, Moreover, such TDFs can induce endogenous STAT3 they may act on lymphocytes directly (for example, activation in affected DC populations, which alters their TGF-b, IL-10) or indirectly (for example, VEGF, PGE2) expression of costimulatory molecules and pro-inﬂam- through host APC populations, such as DCs and matory cytokine characteristics. Aberrant STAT3 signal- monocytes/macrophages, to render them non-stimula- ling may therefore have profound consequences on both tory or tolerogenic for adequate induction of host innate and adaptive immunity, and strategies that inhibit immune responses. STAT3 signalling in neoplastic cells or aberrant APC A pivotal mechanism accounting for the produc- populations may reinstate APC function and the tion of such tumour-derived immune suppressive engagement of antitumour activity. This has been factors involves STAT3 activation (Wang et al., 2004; demonstrated by using dominant-negative or antisense Kortylewski et al., 2005; Yu et al., 2007) (Figure 1). oligonucleotide approaches, as well as STAT3 antago- Constitutive STAT3 activity in neoplastic populations nists that impede constitutive tumour- or immune drives the production of cytokines, such as IL-10 and system-associated STAT3 activity (Wang et al., 2004; VEGF, that inhibit pro-inﬂammatory reactions within Kortylewski et al., 2005; Yu et al., 2007).

Figure 1 Cell contact-independent mechanisms of tumour escape. Tumour escape through cell contact-independent mechanisms occurs largely through the secretion of diverse tumour-derived factors (TDFs) with multi-functional consequences. Here, a number of major well-characterized TDFs are illustrated to reveal the complexity of interactions and the biological consequences. TDF may act directly on the tumour cell population itself and/or other host cell types to further advance the malignant process. Among the plethora of TDF that can promote tumour escape, suppress innate or adaptive immunity and facilitate tumour progression include vascular endothelial growth factor (VEGF), interleukin-10 (IL-10), reactive oxygen species (ROS), indoleamine-2,3-dioxgenase (IDO), prostaglandins (for example, PGE2) and transforming growth factor (TGF)-b. Moreover, various myeloid or lymphoid populations are also engaged and exploited in creating an immunosuppressive or tolerogenic network that further compromises the effectiveness of innate or adaptive immunity. These include tumour-associated macrophages (TAMs), regulatory T cells (Treg), natural killer T cells (NKTs), immature dendritic cells (iDCs), IDO-producing plasmacytoid DC (pDC) and myeloid-derived suppressor cells (MDSCs).

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5897 Tumour-derived TGF-b has pleiotropic consequences immunity in a unique capacity, they may also interface (Figure 1). It can mediate tumour escape and progression to maximize immune dysfunction (Stewart et al., 2007) by impacting the biological behaviour of neoplastic cells (Figure 1). directly and/or indirectly by affecting immune CD4 þ CD25 þ T cells, which also express the FoxP3, or stromal elements. Among its numerous activities, CTLA-4 and GITR markers (‘Treg cells’), are best TGF-b can promote the proliferation of host stromal defined by their ability to suppress effector T-cell cells, such as fibroblasts and tumour-associated macro- function in both animal tumour models and patients phages(Pollard,2004),which,inturn,maysecrete bearing a variety of tumour types (Sakaguchi, 2004; diverse angiogenic factors, such as VEGF (Kim et al., Curiel, 2007). Treg cells occur naturally and act to inhibit 2004; Reckamp et al., 2006). Tumour-associated macro- autoimmune responses (Shevach, 2000) but can also phages are recruited to the site of tumour growth through suppress the generation of tumour-specific T-cell res- a host of TDF, including colony-stimulating factor-1 ponses (Gallimore and Sakaguchi, 2002), possibly (CSF-1) and members of the monocyte chemoattractant through similar mechanisms. Increased numbers of Treg protein family (for example, MCP-1/CCL2). TGF-b cells have been found in the peripheral circulation of also induces stromal cells to produce PGE2 through the patients with a range of cancer types (Liyanage et al., upregulation of cyclooxygenase-2 activity (Reckamp 2002; Ichihara et al., 2003; Ormandy et al., 2005). et al., 2006), which favours tumour cell survival, In studies of ovarian carcinoma, these cells have been proliferation and apoptotic resistance. shown to accumulate at the tumour site through Tumour- or stromal-derived TGF-b abrogates T-cell CCL22-induced migration (Woo et al., 2001; Curiel differentiation, proliferation and effector function, in et al., 2004). Strategies that deplete Treg cells have also part, through the loss of IL-12 production and cytotoxic been shown to improve responses to therapy (Ercolini T lymphocyte lytic capability (Thomas and Massague, et al., 2005; Litzinger et al., 2007). Treg cells can suppress 2005; Li et al., 2006). The production of IL-4 and IL-10 by effector T cells and thus prevent the development of TGF-b-stimulated stromal cells can also tilt the immune antitumour immunity, although the mechanisms by response towards an inappropriate type-2 phenotype, which these cells mediate immunosuppression remain which is unable to sustain effective cell-mediated immu- elusive. nity. Increased production of PGE2 by stromal cells Tumour growth has also been shown to be accom- following stimulation by tumour-derived TGF-b will also panied by the expansion of a unique heterogeneous impair DC function and their ability to appropriately population of inhibitory myeloid cells, termed MDSC activate an antitumour T-cell response, in part, by (Figure 1), which in mice commonly co-express the decreasing the expression levels of MHC class I and II CD11b and Gr-1 differentiation markers (Kusmartsev molecules on these APC (Harizi et al., 2001). Tumour cells and Gabrilovich, 2006; Nagaraj and Gabrilovich, 2007). can also promote their own survival through the produc- Although more extensively studied in animal models, tion of these same TDFs, namely PGE2,TGF-b, VEGF, the generation of similar inhibitory myeloid populations IL-4 and IL-10 (Figure 1). Such TDFs, in turn, can has also been observed in various human cancers upregulate the expression of additional angiogenic activa- including carcinoma of the head and neck, non-small tors, such as matrix metalloproteinase-2 and matrix cell lung carcinoma, renal carcinoma, melanoma and metalloproteinase-9 (Vakkila and Lotze, 2004), as well as adenocarcinomas of the colon, breast and pancreas anti-apoptotic genes that encode c-FLIP, Bcl-xL and Mcl-1 (Almand et al., 2001; Schmielau and Finn, 2001; Zea (Espana et al., 2004; Jung et al., 2004). Increased resistance et al., 2005; Filipazzi et al., 2007; Ochoa et al., 2007). to apoptosis coupled with increased angiogenic activity Many tumour types have been shown to promote the and invasiveness through VEGF and matrix metallopro- expansion of MDSC in vivo through the secretion of one teinase production are key acquisitions that favour tumour or more TDFs such as granulocyte macrophage-CSF, escape and progression. VEGF, IL-10, IL-6, CSF-1 and PGE2 (Serafini et al., 2004; Rodriguez et al., 2005; Kusmartsev and Gabrilo- vich, 2006; Nagaraj and Gabrilovich, 2007; Ochoa et al., Immune suppressive subsets that contribute to tumour 2007; Sinha et al., 2007). escape A number of animal studies have implicated these As a result of these TDF-induced effects, lymphoid or CD11b þ Gr-1 þ cells in the suppression of both CD4 þ myeloid populations may also develop into potent and CD8 þ T cell-mediated immunity through diverse immune suppressive subsets that contribute to tumour pathways (Figure 1). Some of the mechanisms char- escape (Figure 1). In both mice and humans, a number acterized include (a) the secretion of TGF-b, which of immune suppressive subsets have been identified, downregulates CTL induction (Terabe et al., 2003); including Treg, myeloid-derived suppressor cells (b) the production of inhibitory enzymes, namely arginase (MDSCs) and indoleamine 2,3-dioxygenase (IDO)- 1, which leads to arginine depletion, and inducible nitric producing DC populations. Also, subsets of CD1- oxide synthase that results in the generation of reactive restricted natural killer T cells have been identified in oxygen species including nitric oxide, which ultimately mouse tumour models, which downregulate host im- alter T-cell signalling, activation and eventually their munosurveillance through IL-13 production and recruit- survival (Bronte and Zanovello, 2005; Kusmartsev and ment of MDSC (Terabe et al., 2000, 2003). Although Gabrilovich, 2006; Nagaraj and Gabrilovich, 2007; each cell type suppresses systemic and/or adaptive Ochoa et al., 2007); and (c) the promotion of tumour

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5898

Figure 2 Cell contact-dependent mechanisms of tumour escape. In addition to the arsenal of cell contact-independent mechanisms, tumour cells have acquired the ability to affect expression of a number of cell-surface molecules required for efﬁcient host T cell– tumour interactions and tumour cell destruction. Here, tumour escape may result from the loss or downregulation of key receptor/ ligand interactions important for cellular adhesion, immune recognition and activation; acquisition of apoptotic resistance due to the alterations in extrinsic or intrinsic death signalling pathways; or aberrant expression of cell-surface ligands that either downregulate T-cell activity, such as PD-L1/B7-H1 or B7-H4, or mediate T-cell death, such as TRAIL or FasL. PD-1, programmed death receptor-1; TRAIL, TNF-related apoptosis-inducing ligand.

angiogenesis through matrix metalloproteinase-9 produc- Finally, in addition to MDSC-mediated arginine tion, which regulates in part the bioavailability of VEGF depletion as a mechanism of immunosuppression, (Yang et al., 2004). tryptophan depletion in the tumour microenvironment Although these MDSCs are readily apparent in the may elicit a similar adverse functional outcome on T-cell bone marrow and peripheral lymphoid organs of mice, immunity (Zamanakou et al., 2007). This can be they are also found to infiltrate tumour tissue, consti- achieved through the action of the enzyme IDO that tuting about 5% of the total mass (Yang et al., 2004). In may be expressed by either neoplastic cells or myeloid fact, strategies to deplete or alter the function of MDSC populations, such as DC or monocytes/macrophages in vivo can improve antitumour responses alone or in that infiltrate the tumour lesion (Figure 1). In animal combination with other therapies (Terabe et al., 2003; models, IDO overexpression in immunogenic tumours Rodriguez et al., 2005; Suzuki et al., 2005; Fricke et al., enhances tumour aggressiveness, whereas, inhibitors of 2007). In human cancer patients, different phenotypic IDO augment CTL activity and reduce tumour growth subsets of ‘MDSC-like’ cells have been identified, in vivo. Therefore, it appears evident that tumours including those that express CD11b and CD15 (Zea develop a parasitic relationship with its host to usurp et al., 2005; Ochoa et al., 2007) or others that lack control of both myeloid and lymphoid compartments to CD11b, but express CD33 and CD34 (Almand et al., further perpetuate neoplastic growth and progression. 2001; Fricke et al., 2007). Furthermore, it remains to be fully understood whether such MDSC-like populations in humans influence tumour progression. Nonetheless, it Tumour cell contact-dependent mechanisms has been reported that in patients receiving therapies that modulate ‘MDSC-like’ populations, immune status Tumour cells may also escape destruction through three or function can be elevated (Mirza et al., 2006; Fricke general cell contact-dependent scenarios (Figure 2): et al., 2007). (1) as a consequence of the loss or downregulation of

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5899 cell-surface molecules critical for adhesion or immune provides a positive signal important for lymphocyte recognition and activation (for example, MHC/Ag activation and proliferation, CTLA-4/B7 transduces an expression or its intracellular components); (2) through opposing negative signal, leading to the subsequent aberrant expression and engagement of receptor–ligand downregulation of the T cell-mediated immune res- interactions that adversely affect lymphocyte survival or ponse. Therefore, in neoplasia, a therapeutically their effector functions (for example, Fas ligand (FasL), relevant T-cell response may be disengaged in the face B7-H1, B7-H4); and (3) through acquisition of genetic of progressive tumour growth, which may alter the or epigenetic alterations that increase apoptotic resis- balance from induction of tumour immunity to one that tance to death receptor-mediated pathways. favours tumour escape (Figure 2). This possibility is supported by several studies that revealed that CTLA-4 blockade alone or in combination with other therapies Alterations in receptor–ligand interactions enhanced tumour rejection efficiency in vivo (Leach The downregulation of key cell-surface molecules impor- et al., 1996; van Elsas et al., 1999; Phan et al., 2003; tant for efficient effector–target interactions is a common Demaria et al., 2005). tactic acquired by neoplastic cells to evade immune Compared with the restricted expression of CTLA-4, recognition and destruction (Whiteside, 2006) (Figure 2). programmed death receptor-1 (PD-1) is expressed on a A frequent abnormality observed in many tumour cells broad range of immune cells, including mature T and B involves alterations in the expression of MHC molecules cells, thymocytes and myeloid cells (Blank and Mack- or in the components of the Ag-processing pathways ensen, 2007). The interaction of PD-1 with either of its (Ferris et al., 2006; Lopez-Albaitero et al., 2006; Chang two ligands, PD-L1 (also known as B7-H1) or PD-L2, and Ferrone, 2007). Tumour cells can develop mutations has been described to negatively regulate the prolifera- that result in the misprocessing or mispresenting of TAA tion and cytokine production of T cells (Freeman et al., so that an appropriate Ag-presenting complex (MHC-b2 2000; Latchman et al., 2001). The inhibitory effects of microglobulin/peptide) does not form, or cannot be PD-1 were initially observed when PD-1-deficient mice recognized, on the tumour cell surface. Downregulation developed autoimmune diseases (Nishimura et al., of, or mutations in, the transporter associated with Ag- 2001). PD-L1 is strongly expressed on a variety of processing proteins (TAP1 and TAP2) and components of tumours (Iwai et al., 2002) and inversely correlates with the immunoproteasome (LMP2 and LMP7) prevents patient prognosis (Ohigashi et al., 2005; Thompson the normal processing of TAA. Generally, tumours et al., 2006). In animal models, PD-L1/B7-H1 blockade are not efficient at directly presenting TAA to the immune has also been shown to enhance therapeutic efficacy system for immune activation or immune-mediated (Hirano et al., 2005). destruction. Another recently described member of the B7 family, In addition to altered MHC/Ag expression, tumour termed B7-H4, is inducibly expressed by myeloid cells fail to function as effective APCs because of the and lymphoid populations (Sica et al., 2003; Flies and lack of expression of costimulatory molecules that are Chen, 2007). As with PD-L1 (B7-H1), the engagement of necessary for proficient T-cell activation. These costi- B7-H4 abrogates T-cell immunity. Here, the mechanism mulatory molecules include members of the B7 family, of suppression is thought to involve cell cycle arrest and which when downregulated lead to T-cell unresponsive- decreased IL-2 production. Interestingly, several studies ness despite circumstances of MHC-restricted Ag now demonstrate the expression of B7-H4 in certain presentation, as previously discussed. In both animal cancer types, including ovarian, renal and mammary models and patients with advanced cancer, impairment carcinomas (Krambeck et al., 2006; Kryczek et al., 2006; in T-cell signalling and lytic function has also been Sadun et al., 2007), which implicates its involvement as a demonstrated (Koneru et al., 2005; Whiteside, 2006). In negative regulator of adaptive immunity. This notion is tumour-infiltrating lymphocytes, there is a decrease in supported by the observation that B7-H4 blockade can CD3z chain expression, as well as the tyrosine kinases augment tumour regression in vivo (Kryczek et al., p56lck and p59fyn, which all play a role in the proximal 2006). Taken collectively, these studies strengthen the TCR signalling events that lead to optimal T-cell roles of PD-L1 (B7-H1)- and B7-H4-dependent path- activation (Koneru et al., 2005). Although the precise ways in cell contact-dependent mechanisms of tumour mechanisms causing alterations in T-cell signalling evasion (Figure 2). remain to be fully understood, arginine depletion It has been demonstrated, at least in some patients, through MDSC may contribute to this immune defect that the number of circulating T cells, particularly (Serafini et al., 2004; Rodriguez et al., 2005; Kusmartsev CD8 þ T cells, is decreased (Kuss et al., 2004) because of and Gabrilovich, 2006; Nagaraj and Gabrilovich, 2007; tumour-induced mechanisms of apoptosis (Hoffmann Ochoa et al., 2007; Sinha et al., 2007). Therefore, et al., 2002). In a mechanism analogous to that tumour-induced alterations in TCR signalling can involved in the maintenance of T-cell tolerance to impair adaptive immunity (Figure 2). normal tissue Ag, it has been proposed that under Although the upregulation and engagement of certain circumstances in vivo, Fas-bearing activated CTLA-4 is not necessarily a tumour escape mechanism, T cells undergo programmed cell death due to their tumour cells can exploit this physiological mechanism engagement with FasL expressed or released by (that is, for their own advantage (Abrams, 2004; Korman et al., soluble FasL) certain tumour types. Tumour-derived 2006). In contrast to CD28/B7 engagement, which microvesicles that contain FasL have also been

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5900 demonstrated in the sera of cancer patients (Andreola revealed that antitumour CTL responses in vivo can et al., 2002). Therefore, tumours may escape T-cell concurrently mediate tumour regression and tumour attack by inducing apoptosis in the effector cell progression through the selection and outgrowth of population either at the site of tumour formation or in residual Faslo variants that possess enhanced malignant circulating T cells, by shedding FasL-containing micro- potential. In fact, such Faslo variants were not only more vesicles (Figure 2). Another death receptor ligand, aggressive in vivo, but also significantly more refractory termed tumour necrosis factor-related apoptosis-indu- to CTL-adoptive immunotherapy (Liu et al., 2005b, cing ligand, has also been found to be expressed on 2006). In addition, in both mouse and human solid human melanomas (Bron et al., 2004) and has been tumour models, it was demonstrated that biological associated with the apoptosis of tumour-infiltrating selection against Fas-responsive cells within parental or lymphocytes (Giovarelli et al., 1999; Bosque et al., primary tumour cell lines, using surrogate sources of Fas 2005). It is thought that tumour cells have usurped these engagement, generated Faslo sub-populations with physiological death pathways as a means of eliminating enhanced malignant ability (Liu et al., 2003; Liu and infiltrating effector cells to maintain a state of ‘privilege’ Abrams, 2003a). Thus, Fas-mediated cytotoxicity can within the tumour microenvironment. unintentionally enforce an immune selective pressure, resulting in the survival, escape and outgrowth of Faslo Apoptotic resistance variants, which demonstrate more aggressive behaviour. In terms of molecular events occurring within the Further characterization of Faslo variants led to the neoplastic population, apoptotic resistance is now identification of ICAM-1 downregulation as an addi- considered an important hallmark of tumour progres- tional molecular determinant important for enhanced sion (Hanahan and Weinberg, 2000). The loss of malignant behaviour (Liu et al., 2005a). Immune sensitivity to cell death may reflect resistance to intrinsic selection of Faslo variants may thus be a novel (mitochondrial) or extrinsic (death receptor) forms of mechanism of immune escape during the effector/target apoptotic induction. In the case of increased resistance interaction within the local tumour microenvironment. to intrinsic cell death, this may occur through the Subsequently, microarray studies revealed the modulation of pro- and anti-apoptotic members of the differential expression of the two genes affecting Fas Bcl-2 family (Figure 2). In the case of increased responsiveness under pro-inflammatory conditions (that resistance to extrinsic cell death, such as Fas, this may is, following IFN-g sensitization): (a) interferon con- occur through downmodulation of the receptor itself, sensus sequence-binding protein, also known as inter- defects in caspase family members or dysregulation of feron regulatory factor-8 (IRF-8); and (b) caspase-1 of the signalling pathway due to the overexpression of key the Fas signalling pathway (Liu and Abrams, 2003b; anti-apoptotic proteins, such as FLICE inhibitory Yang et al., 2007). Incidentally, IRF-8 expression was proteins (FLIP), inhibitors of apoptosis or survivin, originally discovered as an IFN-g-inducible transcrip- which also affect caspase activation. The potential tion factor essential for regulating normal myelopoiesis importance of Fas loss of function in tumour escape (Holtschke et al., 1996). In these solid tumour studies and tumour progression reflects two important con- (Liu and Abrams, 2003b; Yang et al., 2007), IRF-8 siderations in cancer biology and tumour immunology: expression was directly associated with apoptotic (a) Fas downregulation has been noted in the progres- sensitivity, but inversely related to malignant phenotype. sion of a range of human malignancies (Keane et al., Using RNA interference strategies, it was then demon- 1996; Krammer et al., 1998; von Reyher et al., 1998; strated that IRF-8 expression in solid tumours was a Owen-Schaub et al., 2000; Worth et al., 2002; Liu et al., causal determinant for their response to cytotoxicity, 2003; Gordon et al., 2005); and (b) if Fas-mediated including Fas-mediated apoptosis, g-irradiation and cytotoxicity is an important host defense mechanism host antitumour immunosurveillance mechanisms during the effector phase of the immune reaction (Greeneltch et al., 2007). Although IRF-8 was originally and if it becomes compromised, this could lead to the identified in the regulation of normal and neoplastic emergence of Fas-resistant (Faslo)tumourescapevariants myeloid development, these findings revealed a new with potentially enhanced malignant capabilities. Faslo functional role for IRF-8 in non-haematopoietic malig- variants have been shown to exhibit enhanced tumour nancies. Thus, IRF-8 downregulation in solid neoplasms growth and reduced sensitivity to CTL-based immu- may provide a previously unrecognized molecular notherapy, as shown in mouse models of experimental determinant for tumour escape, as well as a potential lung metastasis (Liu et al., 2005b, 2006). Ultimately, target for therapeutic modulation. tumour cells that fail to die, through conventional oncological treatments (that is, radiotherapy, chemotherapy) or experimental immunotherapies, have a clear and Conclusions distinctive survival advantage to persist, accumulate additional genetic or epigenetic modifications and As the control of tumour development, growth or progress to a more malignantly proficient phenotype. progression involves dynamic and chronic interactions Several studies have examined whether an antitumour with the host immune system, alterations in the CD8 þ CTL response actually ‘selects’ such neoplastic molecular events of both neoplastic tissue and key sub-populations with enhanced resistance to Fas- elements of the immune system will collectively influence induced death (Liu et al., 2005b, 2006). These studies malignant proficiency. The mechanisms of tumour

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5901 escape are broad and complex, and reflect both cell factors, as opposed to any single factor alone, interface contact-dependent and -independent interactions. It is to impact the outcome between tumour immunity and also likely that new mechanisms of tumour escape will tumour escape. Accordingly, each preclinical or clinical be identified over time, as more knowledge and insights situation needs to be assessed carefully to best under- are gained in cancer biology. Nevertheless, the net effect stand and predict the potential physiological benefit of of these interactions is a highly immunosuppressive or immune intervention, and whether the therapeutic hostile environment that favours tumour cell survival regimen has the capacity to overcome such defined and proliferation, and the disengagement or inactivation mechanisms of tumour self-protection or escape. of critical host defense mechanisms. These tumour escape mechanisms therefore represent potentially signi- Acknowledgements ficant challenges to successful cancer immunotherapy. It is important to emphasize that because neoplastic This research was supported (in part) by the Intramural progression is a dynamic, chronic and complex process, Research Program of the NIH, National Cancer Institute, it is likely that varying and distinct combinations of Center for Cancer Research.

References

Abrams SI. (2004). Role of anti-CTLA-4 therapies in the treatment of immunity in patients after clonal repopulation with antitumor cancer. Curr Opin Mol Ther 6: 71–77. lymphocytes. Science 298: 850–854. Almand B, Clark JI, Nikitina E, van Beynen J, English NR, Knight SC Dunn GP, Koebel CM, Schreiber RD. (2006). Interferons, immunity et al. (2001). Increased production of immature myeloid cells in and cancer immunoediting. Nat Rev Immunol 6: 836–848. cancer patients: a mechanism of immunosuppression in cancer. Ercolini AM, Ladle BH, Manning EA, Pfannenstiel LW, J Immunol 166: 678–689. Armstrong TD, Machiels JP et al. (2005). Recruitment of latent Andreola G, Rivoltini L, Castelli C, Huber V, Perego P, Deho P et al. pools of high-avidity CD8(+) T cells to the antitumor immune (2002). Induction of lymphocyte apoptosis by tumor cell secretion of response. J Exp Med 201: 1591–1602. FasL-bearing microvesicles. J Exp Med 195: 1303–1316. Espana L, Fernandez Y, Rubio N, Torregrosa A, Blanco J, Sierra A. Blank C, Mackensen A. (2007). Contribution of the PD-L1/PD-1 (2004). Overexpression of Bcl-xL in human breast cancer cells pathway to T-cell exhaustion: an update on implications for chronic enhances organ-selective lymph node metastasis. Breast Cancer Res infections and tumor evasion. Cancer Immunol Immunother 56: Treat 87: 33–44. 739–745. Ferris RL, Whiteside TL, Ferrone S. (2006). Immune escape associated Bosque A, Pardo J, Martinez-Lorenzo MJ, Lasierra P, Larrad L, with functional defects in antigen-processing machinery in head and Marzo I et al. (2005). Human CD8+ T cell blasts are more sensitive neck cancer. Clin Cancer Res 12: 3890–3895. than CD4+ T cell blasts to regulation by APO2L/TRAIL. Eur J Filipazzi P, Valenti R, Huber V, Pilla L, Canese P, Iero M et al. (2007). Immunol 35: 1812–1821. Identification of a new subset of myeloid suppressor cells in Bron LP, Scolyer RA, Thompson JF, Hersey P. (2004). Histological peripheral blood of melanoma patients with modulation by a expression of tumour necrosis factor-related apoptosis-inducing granulocyte-macrophage colony-stimulation factor-based antitumor ligand (TRAIL) in human primary melanoma. Pathology 36: vaccine. J Clin Oncol 25: 2546–2553. 561–565. Flies DB, Chen L. (2007). The new B7s: playing a pivotal role in tumor Bronte V, Zanovello P. (2005). Regulation of immune responses by immunity. J Immunother (1997) 30: 251–260. L-arginine metabolism. Nat Rev Immunol 5: 641–654. Freeman GJ, Long AJ, Iwai Y, Bourque K, Chernova T, Nishimura H Bui JD, Schreiber RD. (2007). Cancer immunosurveillance, immunoe- et al. (2000). Engagement of the PD-1 immunoinhibitory receptor diting and inflammation: independent or interdependent processes? by a novel B7 family member leads to negative regulation of Curr Opin Immunol 19: 203–208. lymphocyte activation. J Exp Med 192: 1027–1034. Chang CC, Ferrone S. (2007). Immune selective pressure and HLA Fricke I, Mirza N, Dupont J, Lockhart C, Jackson A, Lee JH et al. class I antigen defects in malignant lesions. Cancer Immunol (2007). Vascular endothelial growth factor-trap overcomes defects Immunother 56: 227–236. in dendritic cell differentiation but does not improve antigen-specific Cuenca A, Cheng F, Wang H, Brayer J, Horna P, Gu L et al. (2003). immune responses. Clin Cancer Res 13: 4840–4848. Extra-lymphatic solid tumor growth is not immunologically ignored Gallimore A, Sakaguchi S. (2002). Regulation of tumour immunity by and results in early induction of antigen-specific T-cell anergy: CD25+ T cells. Immunology 107: 5–9. dominant role of cross-tolerance to tumor antigens. Cancer Res 63: Giovarelli M, Musiani P, Garotta G, Ebner R, Di Carlo E, Kim Y 9007–9015. et al. (1999). A ‘stealth effect’: adenocarcinoma cells engineered to Curiel TJ. (2007). Tregs and rethinking cancer immunotherapy. J Clin express TRAIL elude tumor-specific and allogeneic T cell reactions. Invest 117: 1167–1174. J Immunol 163: 4886–4893. Curiel TJ, Coukos G, Zou L, Alvarez X, Cheng P, Mottram P et al. Gordon N, Arndt CA, Hawkins DS, Doherty DK, Inwards CY, (2004). Specific recruitment of regulatory T cells in ovarian Munsell MF et al. (2005). Fas expression in lung metastasis from carcinoma fosters immune privilege and predicts reduced survival. osteosarcoma patients. J Pediatr Hematol Oncol 27: 611–615. Nat Med 10: 942–949. Greeneltch K, Schneider M, Steinberg SM, Liewehr DJ, Stewart TJ, Demaria S, Kawashima N, Yang AM, Devitt ML, Babb JS, Allison JP Liu K et al. (2007). Host immunosurveillance controls tumor et al. (2005). Immune-mediated inhibition of metastases after growth via IRF-8-dependent mechanisms. Cancer Res 67: treatment with local radiation and CTLA-4 blockade in a mouse 10406–10416. model of breast cancer. Clin Cancer Res 11: 728–734. Hanahan D, Weinberg RA. (2000). The hallmarks of cancer. Cell 100: Dudley ME, Rosenberg SA. (2003). Adoptive-cell-transfer 57–70. therapy for the treatment of patients with cancer. Nat Rev Cancer Harizi H, Juzan M, Grosset C, Rashedi M, Gualde N. (2001). 3: 666–675. Dendritic cells issued in vitro from bone marrow produce PGE(2) Dudley ME, Wunderlich JR, Robbins PF, Yang JC, Hwu P, that contributes to the immunomodulation induced by antigen- Schwartzentruber DJ et al. (2002). Cancer regression and auto- presenting cells. Cell Immunol 209: 19–28.

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5902 Hirano F, Kaneko K, Tamura H, Dong H, Wang S, Ichikawa M Litzinger MT, Fernando R, Curiel TJ, Grosenbach DW, Schlom J, et al. (2005). Blockade of B7-H1 and PD-1 by monoclonal Palena C. (2007). The IL-2 immunotoxin denileukin diftitox reduces antibodies potentiates cancer therapeutic immunity. Cancer Res regulatory T cells and enhances vaccine-mediated T-cell immunity. 65: 1089–1096. Blood 110: 3192–3201. Hoffmann TK, Dworacki G, Tsukihiro T, Meidenbauer N, Gooding Liu K, Abrams SI. (2003a). Alterations in Fas expression are W, Johnson JT et al. (2002). Spontaneous apoptosis of circulating characteristic of, but not solely responsible for, enhanced metastatic T lymphocytes in patients with head and neck cancer and its clinical competence. J Immunol 170: 5973–5980. importance. Clin Cancer Res 8: 2553–2562. Liu K, Abrams SI. (2003b). Coordinate regulation of IFN consensus Holtschke T, Lohler J, Kanno Y, Fehr T, Giese N, Rosenbauer F et al. sequence-binding protein and caspase-1 in the sensitization (1996). Immunodeficiency and chronic myelogenous leukemia-like of human colon carcinoma cells to Fas-mediated apoptosis by syndrome in mice with a targeted mutation of the ICSBP gene. Cell IFN-gamma. J Immunol 170: 6329–6337. 87: 307–317. Liu K, Caldwell SA, Abrams SI. (2005a). Cooperative disengagement Ichihara F, Kono K, Takahashi A, Kawaida H, Sugai H, Fujii H. of Fas and intercellular adhesion molecule-1 function in neoplastic (2003). Increased populations of regulatory T cells in peripheral cells confers enhanced colonization efficiency. Cancer Res 65: blood and tumor-infiltrating lymphocytes in patients with gastric 1045–1054. and esophageal cancers. Clin Cancer Res 9: 4404–4408. Liu K, Caldwell SA, Abrams SI. (2005b). Immune selection and Iwai Y, Ishida M, Tanaka Y, Okazaki T, Honjo T, Minato N. (2002). emergence of aggressive tumor variants as negative consequences of Involvement of PD-L1 on tumor cells in the escape from host Fas-mediated cytotoxicity and altered IFN-gamma-regulated gene immune system and tumor immunotherapy by PD-L1 blockade. expression. Cancer Res 65: 4376–4388. Proc Natl Acad Sci USA 99: 12293–12297. Liu K, Caldwell SA, Greeneltch KM, Yang D, Abrams SI. (2006). Jung YJ, Kim JY, Park JH. (2004). TGF-beta1 inhibits Fas-mediated CTL adoptive immunotherapy concurrently mediates tumor regres- apoptosis by regulating surface Fas and cFLIPL expression in sion and tumor escape. J Immunol 176: 3374–3382. human leukaemia/lymphoma cells. Int J Mol Med 13: 99–104. Liu K, McDuffie E, Abrams SI. (2003). Exposure of human primary Keane MM, Ettenberg SA, Lowrey GA, Russell EK, Lipkowitz S. colon carcinoma cells to anti-Fas interactions influences the (1996). Fas expression and function in normal and malignant breast emergence of pre-existing Fas-resistant metastatic subpopulations. cell lines. Cancer Res 56: 4791–4798. J Immunol 171: 4164–4174. Kim R, Emi M, Tanabe K, Uchida Y, Toge T. (2004). The role of Fas Liyanage UK, Moore TT, Joo HG, Tanaka Y, Herrmann V, ligand and transforming growth factor beta in tumor progression: Doherty G et al. (2002). Prevalence of regulatory T cells is molecular mechanisms of immune privilege via Fas-mediated apop- increased in peripheral blood and tumor microenvironment of tosis and potential targets for cancer therapy. Cancer 100: 2281–2291. patients with pancreas or breast adenocarcinoma. J Immunol 169: Koneru M, Schaer D, Monu N, Ayala A, Frey AB. (2005). Defective 2756–2761. proximal TCR signaling inhibits CD8+ tumor-infiltrating lympho- Lopez-Albaitero A, Nayak JV, Ogino T, Machandia A, Gooding W, cyte lytic function. J Immunol 174: 1830–1840. DeLeo AB et al. (2006). Role of antigen-processing machinery in the Korman AJ, Peggs KS, Allison JP. (2006). Checkpoint blockade in in vitro resistance of squamous cell carcinoma of the head and neck cancer immunotherapy. Adv Immunol 90: 297–339. cells to recognition by CTL. J Immunol 176: 3402–3409. Kortylewski M, Kujawski M, Wang T, Wei S, Zhang S, Pilon-Thomas Mirza N, Fishman M, Fricke I, Dunn M, Neuger AM, Frost TJ et al. S et al. (2005). Inhibiting Stat3 signaling in the hematopoietic (2006). All-trans-retinoic acid improves differentiation of myeloid system elicits multicomponent antitumor immunity. Nat Med 11: cells and immune response in cancer patients. Cancer Res 66: 1314–1321. 9299–9307. Krambeck AE, Thompson RH, Dong H, Lohse CM, Park ES, Kuntz Nagaraj S, Gabrilovich DI. (2007). Myeloid-derived suppressor cells. SM et al. (2006). B7-H4 expression in renal cell carcinoma and Adv Exp Med Biol 601: 213–223. tumor vasculature: associations with cancer progression and Nishimura H, Okazaki T, Tanaka Y, Nakatani K, Hara M, survival. Proc Natl Acad Sci USA 103: 10391–10396. Matsumori A et al. (2001). Autoimmune dilated cardiomyopathy Krammer PH, Galle PR, Moller P, Debatin KM. (1998). CD95(APO- in PD-1 receptor-deficient mice. Science 291: 319–322. 1/Fas)-mediated apoptosis in normal and malignant liver, colon, Ochoa AC, Zea AH, Hernandez C, Rodriguez PC. (2007). Arginase, and hematopoietic cells. Adv Cancer Res 75: 251–273. prostaglandins, and myeloid-derived suppressor cells in renal cell Kryczek I, Zou L, Rodriguez P, Zhu G, Wei S, Mottram P et al. carcinoma. Clin Cancer Res 13: 721s–726s. (2006). B7-H4 expression identifies a novel suppressive macrophage Ochsenbein AF, Klenerman P, Karrer U, Ludewig B, Pericin M, population in human ovarian carcinoma. J Exp Med 203: 871–881. Hengartner H et al. (1999). Immune surveillance against a solid Kusmartsev S, Gabrilovich DI. (2006). Role of immature myeloid cells tumor fails because of immunological ignorance. Proc Natl Acad Sci in mechanisms of immune evasion in cancer. Cancer Immunol USA 96: 2233–2238. Immunother 55: 237–245. Ohigashi Y, Sho M, Yamada Y, Tsurui Y, Hamada K, Ikeda N et al. Kuss I, Hathaway B, Ferris RL, Gooding W, Whiteside TL. (2004). (2005). Clinical significance of programmed death-1 ligand-1 and Decreased absolute counts of T lymphocyte subsets and their programmed death-1 ligand-2 expression in human esophageal relation to disease in squamous cell carcinoma of the head and neck. cancer. Clin Cancer Res 11: 2947–2953. Clin Cancer Res 10: 3755–3762. Ormandy LA, Hillemann T, Wedemeyer H, Manns MP, Greten TF, Latchman Y, Wood CR, Chernova T, Chaudhary D, Borde M, Korangy F. (2005). Increased populations of regulatory T cells in Chernova I et al. (2001). PD-L2 is a second ligand for PD-1 and peripheral blood of patients with hepatocellular carcinoma. Cancer inhibits T cell activation. Nat Immunol 2: 261–268. Res 65: 2457–2464. Leach DR, Krummel MF, Allison JP. (1996). Enhancement of Otten GR, Germain RN. (1991). Split anergy in a CD8+ T cell: antitumor immunity by CTLA-4 blockade. Science 271: 1734–1736. receptor-dependent cytolysis in the absence of interleukin-2 Lee PP, Yee C, Savage PA, Fong L, Brockstedt D, Weber JS et al. production. Science 251: 1228–1231. (1999). Characterization of circulating T cells specific for tumor- Overwijk WW, Theoret MR, Finkelstein SE, Surman DR, de Jong LA, associated antigens in melanoma patients. Nat Med 5: 677–685. Vyth-Dreese FA et al. (2003). Tumor regression and autoimmunity Li MO, Wan YY, Sanjabi S, Robertson AK, Flavell RA. (2006). after reversal of a functionally tolerant state of self-reactive CD8+ Transforming growth factor-beta regulation of immune responses. T cells. J Exp Med 198: 569–580. Annu Rev Immunol 24: 99–146. Owen-Schaub L, Chan H, Cusack JC, Roth J, Hill LL. (2000). Fas and Lin EY, Pollard JW. (2007). Tumor-associated macrophages Fas ligand interactions in malignant disease. Int J Oncol 17: 5–12. press the angiogenic switch in breast cancer. Cancer Res 67: Pardoll D. (2003). Does the immune system see tumors as foreign or 5064–5066. self? Annu Rev Immunol 21: 807–839.

Oncogene How tumours escape mass destruction TJ Stewart and SI Abrams 5903 Pawelec G, Ouyang Q, Colonna-Romano G, Candore G, Lio D, Terabe M, Matsui S, Noben-Trauth N, Chen H, Watson C, Caruso C. (2002). Is human immunosenescence clinically relevant? Donaldson DD et al. (2000). NKT cell-mediated repression of Looking for ‘immunological risk phenotypes’. Trends Immunol 23: tumor immunosurveillance by IL-13 and the IL-4R-STAT6 path- 330–332. way. Nat Immunol 1: 515–520. Penn I. (1999). Posttransplant malignancies. Transplant Proc 31: Terabe M, Matsui S, Park JM, Mamura M, Noben-Trauth N, 1260–1262. Donaldson DD et al. (2003). Transforming growth factor-beta Phan GQ, Yang JC, Sherry RM, Hwu P, Topalian SL, Schwartzen- production and myeloid cells are an effector mechanism through truber DJ et al. (2003). Cancer regression and autoimmunity which CD1d-restricted T cells block cytotoxic T lymphocyte- induced by cytotoxic T lymphocyte-associated antigen 4 blockade in mediated tumor immunosurveillance: abrogation prevents tumor patients with metastatic melanoma. Proc Natl Acad Sci USA 100: recurrence. J Exp Med 198: 1741–1752. 8372–8377. Thomas DA, Massague J. (2005). TGF-beta directly targets cytotoxic Pollard JW. (2004). Tumour-educated macrophages promote tumour T cell functions during tumor evasion of immune surveillance. progression and metastasis. Nat Rev Cancer 4: 71–78. Cancer Cell 8: 369–380. Reckamp KL, Krysan K, Morrow JD, Milne GL, Newman RA, Thompson RH, Kuntz SM, Leibovich BC, Dong H, Lohse CM, Tucker C et al. (2006). A phase I trial to determine the optimal Webster WS et al. (2006). Tumor B7-H1 is associated with poor biological dose of celecoxib when combined with erlotinib in prognosis in renal cell carcinoma patients with long-term follow-up. advanced non-small cell lung cancer. Clin Cancer Res 12: Cancer Res 66: 3381–3385. 3381–3388. Vakkila J, Lotze MT. (2004). Inflammation and necrosis promote Rodriguez PC, Hernandez CP, Quiceno D, Dubinett SM, Zabaleta J, tumour growth. Nat Rev Immunol 4: 641–648. Ochoa JB et al. (2005). Arginase I in myeloid suppressor cells is van Elsas A, Hurwitz AA, Allison JP. (1999). Combination induced by COX-2 in lung carcinoma. J Exp Med 202: 931–939. immunotherapy of B16 melanoma using anti-cytotoxic T lympho- Sadun RE, Sachsman SM, Chen X, Christenson KW, Morris WZ, Hu cyte-associated antigen 4 (CTLA-4) and granulocyte/macrophage P et al. (2007). Immune signatures of murine and human cancers colony-stimulating factor (GM-CSF)-producing vaccines induces reveal unique mechanisms of tumor escape and new targets for rejection of subcutaneous and metastatic tumors accompanied by cancer immunotherapy. Clin Cancer Res 13: 4016–4025. autoimmune depigmentation. J Exp Med 190: 355–366. Sakaguchi S. (2004). Naturally arising CD4+ regulatory t cells for Vogelstein B, Kinzler KW. (2004). Cancer genes and the pathways they immunologic self-tolerance and negative control of immune control. Nat Med 10: 789–799. responses. Annu Rev Immunol 22: 531–562. von Reyher U, Strater J, Kittstein W, Gschwendt M, Schmielau J, Finn OJ. (2001). Activated granulocytes and granulocyte- Krammer PH, Moller P. (1998). Colon carcinoma cells use different derived hydrogen peroxide are the underlying mechanism of mechanisms to escape CD95-mediated apoptosis. Cancer Res 58: suppression of t-cell function in advanced cancer patients. Cancer 526–534. Res 61: 4756–4760. Wang T, Niu G, Kortylewski M, Burdelya L, Shain K, Zhang S et al. Serafini P, Carbley R, Noonan KA, Tan G, Bronte V, Borrello I. (2004). Regulation of the innate and adaptive immune responses by (2004). High-dose granulocyte-macrophage colony-stimulating fac- Stat-3 signaling in tumor cells. Nat Med 10: 48–54. tor-producing vaccines impair the immune response through the Whiteside TL. (2006). Immune suppression in cancer: effects on recruitment of myeloid suppressor cells. Cancer Res 64: 6337–6343. immune cells, mechanisms and future therapeutic intervention. Shevach EM. (2000). Regulatory T cells in autoimmmunity*. Annu Rev Semin Cancer Biol 16: 3–15. Immunol 18: 423–449. Willimsky G, Blankenstein T. (2005). Sporadic immunogenic tumours Sica GL, Choi IH, Zhu G, Tamada K, Wang SD, Tamura H et al. avoid destruction by inducing T-cell tolerance. Nature 437: 141–146. (2003). B7-H4, a molecule of the B7 family, negatively regulates Woo EY, Chu CS, Goletz TJ, Schlienger K, Yeh H, Coukos G et al. T cell immunity. Immunity 18: 849–861. (2001). Regulatory CD4(+)CD25(+) T cells in tumors from Sinha P, Clements VK, Fulton AM, Ostrand-Rosenberg S. (2007). patients with early-stage non-small cell lung cancer and late-stage Prostaglandin E2 promotes tumor progression by inducing myeloid- ovarian cancer. Cancer Res 61: 4766–4772. derived suppressor cells. Cancer Res 67: 4507–4513. Worth LL, Lafleur EA, Jia SF, Kleinerman ES. (2002). Fas expression Smyth MJ, Dunn GP, Schreiber RD. (2006). Cancer immunosurveil- inversely correlates with metastatic potential in osteosarcoma cells. lance and immunoediting: the roles of immunity in suppressing Oncol Rep 9: 823–827. tumor development and shaping tumor immunogenicity. Adv Yang D, Thangaraju M, Greeneltch K, Browning DD, Schoenlein PV, Immunol 90: 1–50. Tamura T et al. (2007). Repression of IFN regulatory factor 8 by Smyth MJ, Thia KY, Street SE, MacGregor D, Godfrey DI, Trapani DNA methylation is a molecular determinant of apoptotic JA. (2000). Perforin-mediated cytotoxicity is critical for surveillance resistance and metastatic phenotype in metastatic tumor cells. of spontaneous lymphoma. J Exp Med 192: 755–760. Cancer Res 67: 3301–3309. Steinman RM, Hawiger D, Nussenzweig MC. (2003). Tolerogenic Yang L, DeBusk LM, Fukuda K, Fingleton B, Green-Jarvis B, Shyr Y dendritic cells. Annu Rev Immunol 21: 685–711. et al. (2004). Expansion of myeloid immune suppressor Stewart TJ, Greeneltch KM, Lutsiak ME, Abrams SI. (2007). Gr+CD11b+ cells in tumor-bearing host directly promotes tumor Immunological responses can have both pro- and antitumour angiogenesis. Cancer Cell 6: 409–421. effects: implications for immunotherapy. Exp Rev Mol Med 9: 1–20. Yu H, Kortylewski M, Pardoll D. (2007). Crosstalk between cancer Street SE, Zerafa N, Iezzi M, Westwood JA, Stagg J, Musiani P et al. and immune cells: role of STAT3 in the tumour microenvironment. (2007). Host perforin reduces tumor number but does not increase Nat Rev Immunol 7: 41–51. survival in oncogene-driven mammary adenocarcinoma. Cancer Res Zamanakou M, Germenis AE, Karanikas V. (2007). Tumor immune 67: 5454–5460. escape mediated by indoleamine 2,3-dioxygenase. Immunol Lett 111: Suzuki E, Kapoor V, Jassar AS, Kaiser LR, Albelda SM. (2005). 69–75. Gemcitabine selectively eliminates splenic Gr-1+/CD11b+ myeloid Zea AH, Rodriguez PC, Atkins MB, Hernandez C, Signoretti S, suppressor cells in tumor-bearing animals and enhances antitumor Zabaleta J et al. (2005). Arginase-producing myeloid suppressor immune activity. Clin Cancer Res 11: 6713–6721. cells in renal cell carcinoma patients: a mechanism of tumor evasion. Takeda K, Hayakawa Y, Smyth MJ, Kayagaki N, Yamaguchi N, Cancer Res 65: 3044–3048. Kakuta S et al. (2001). Involvement of tumor necrosis factor-related Zhou G, Drake CG, Levitsky HI. (2006). Amplification of tumor- apoptosis-inducing ligand in surveillance of tumor metastasis by specific regulatory T cells following therapeutic cancer vaccines. liver natural killer cells. Nat Med 7: 94–100. Blood 107: 628–636.

Oncogene