OTHER THEMES PUBLISHED IN THIS IMMUNOLOGY IN THE CLINIC REVIEW SERIES
Metabolic Diseases, Host Responses, Allergies, Autoinflammatory Diseases, Type 1 diabetes and viruses.
OTHER THEMES PUBLISHED IN THIS IMMUNOLOGY IN THE CLINIC REVIEW SERIES
Metabolic Diseases, Host Responses, Allergies, Autoinflammatory Diseases, Type 1 diabetes and viruses.
Despite complex genomic and epigenetic abnormalities, many cancers are irrevocably dependent on an initiating oncogenic lesion whose restoration to a normal physiological activation can elicit a dramatic and sudden reversal of their neoplastic properties. This phenomenon of the reversal of tumorigenesis has been described as oncogene addiction. Oncogene addiction had been thought to occur largely through tumour cell-autonomous mechanisms such as proliferative arrest, apoptosis, differentiation and cellular senescence. However, the immune system plays an integral role in almost every aspect of tumorigenesis, including tumour initiation, prevention and progression as well as the response to therapeutics. Here we highlight more recent evidence suggesting that oncogene addiction may be integrally dependent upon host immune-mediated mechanisms, including specific immune effectors and cytokines that regulate tumour cell senescence and tumour-associated angiogenesis. Hence, the host immune system is essential to oncogene addiction.
Oncogene addiction is the phenomenon by which even highly complex tumour cells that are a consequence of multiple genetic and epigenetic changes become exquisitely dependent upon a single oncogene for their continued growth and survival [1,2]. Early studies illustrated that, in tumour cells, the in vitro suppression of an oncogene or the restoration of expression of a tumour suppressor could be sufficient to induce the sustained loss of their neoplastic features .
More recently, conditional transgenic mouse models have been used to explore the tumour-specific consequences of the suppression of oncogenes including MYC, RAS, BRAF and BCR-ABL[4–10]. The specific consequences of oncogene inactivation in a tumour are dependent upon cellular and genetic context and can include proliferative arrest, apoptosis , differentiation [5,6] and senescence  as well as the inhibition of angiogenesis [12,13]. Upon even brief inactivation of an oncogene, the diversity of these outcomes is evidenced by the induction of a permanent loss of the neoplastic phenotype in osteosarcoma and lymphoma [4,5], but not in epithelial tumours such as hepatocellular carcinoma and breast carcinoma [6,10]. In some cases, the inactivation of the oncogene fails to cause significant tumour regression such as in a murine model of MYC-induced lung adenocarcinoma . Thus, in many but not all cases, the inactivation of an oncogene that initiates tumorigenesis is sufficient to reverse tumorigenesis.
The clinical relevance of oncogene addiction was ensconced more firmly after the development of several effective targeted therapeutics [15,16]. The advent of potent agents such as imatinib for chronic myelogenous leukaemia and gastrointestinal stromal tumours , trastuzumab for the treatment of breast cancer  and PLX4032 for the treatment of melanoma , among other drugs , has galvanized interest in exploiting oncogene addiction for cancer therapy and understanding the underlying principles by which it works.
The mechanism of oncogene addiction has been largely presumed to be cell autonomous and to occur by processes intrinsic and exclusively dependent upon biological programmes within a tumour cell. Several mechanisms have been proposed for oncogene addiction, including the notion of abnormal tumour cell genetic circuitry , reversibility of tumorigenesis , oncogenic shock  and synthetic lethality . However, the host microenvironment is well established to play a critical role in how oncogenes initiate tumorigenesis [25–28], suggesting strongly that host factors might similarly play an important role in oncogene addiction.
The notion of an intimate relationship between tumour cells and host immune cells was first posited more than a century ago by Rudolf Virchow . The immune system is integral to almost every aspect of tumorigenesis, including tumour initiation [30,31], prevention  and progression . Tumours appear to undergo immune editing that is important to both their generation and therapeutic destruction [34,35]. Tumorigenesis is a consequence of interactions between incipient neoplastic cells and host stromal cells, including immune cells, endothelial cells and fibroblasts, as well as extracellular matrix components and secreted factors . The immune system plays a complex role in tumorigenesis , and immune effectors and their secreted factors have been implicated in the initiation of tumorigenesis [30,31], tumour growth, survival and metastastic dissemination as well as in immune surveillance and prevention of tumour growth .
Correspondingly, in mouse models and in human patients, various components of the immune system have been implicated in tumorigenesis. Immune effectors including macrophages, T and B cells have been shown to either have a role in promoting [37–39] or inhibiting [40–43] tumour growth, depending on the particular neoplastic context. Moreover, other immune cells such as natural killer (NK) cells  can inhibit metastasis, whereas CD4+ T cells  and macrophages  have been shown to promote metastasis. Similarly, in human patients, immune effectors appear to both promote and suppress tumorigenesis. Thus, pathological autoimmune stimulation or inflammation can be associated with increased tumorigenesis [29,47–49], whereas hosts that are immune compromised also may exhibit many magnitudes increased incidence of tumours . Similarly, the presence or absence of immune effectors, such as CD4+ T cells, in a particular tumour microenvironment can have either a favourable  or a non-favourable prognosis . Hence, immune cells and cytokines play a complex role in both the pathogenesis of tumorigenesis and the therapeutic response of tumours.
Finally, oncogene expression has been shown in some circumstances to influence the immune response significantly [52–56]. Activation of the RET oncogene in normal human thymocytes induces an inflammatory response leading to tumour tissue remodelling, angiogenesis and metastasis, all of which contribute to the maintenance of the transformed state of the tumour . Oncogenic RAS up-regulates expression of the cytokines interleukin (IL)-6  and IL-8  which, in turn, contributes to tumorigenesis. In a MYC-induced model of lymphoma a robust activation of macrophages is associated with tumour suppression . Furthermore, endogenous MYC levels have also been shown to maintain the angiogenic tumour microenvironment in certain tumour models . The dynamic conversation between oncogenes and the tumour microenvironment suggested that their interplay could also be fundamental to oncogene addiction (see Table 1).
|Oncogene||Tumour type||Involved immune compartment||Immune-mediated mechanism||References|
|MYC||T cell acute lymphoblastic lymphoma||Adaptive immunity (CD4+ T cells)||Induction of senescence and suppression of angiogenesis|||
|BCR-ABL||Pro-B cell acute lymphocytic leukaemia||Adaptive immunity (CD4+ T cells)||Induction of senescence and suppression of angiogenesis|||
|P53||Hepatocellular carcinoma||Innate immunity (neutrophils, macrophages, NK cells)||Tumour clearance|||
|Other oncogenic links to tumour immunity|
|MYC||B cell lymphoma; pancreatic islet cell tumour||Innate immunity (macrophages; mast cells)||Macrophages induce senescence; mast cells promote angiogenesis||[42,66]|
|RAS||Cervical cancer; renal cell carcinoma||Innate immunity (neutrophils)||Neutrophils recruited by IL-8, IL-6 secretion||[58,59]|
|MET||Papillary thyroid carcinoma||?Innate immunity/activation of proinflammatory programme||Innate cells recruitment via proinflammatory cytokines/chemokines||[52,57]|
|PML||Acute promyelocytic leukaemia; prostate carcinoma||Adaptive immunity (CD8+ T cells)||PML influences MHC class I antigen presentation||[53,54]|
|BRAF||Melanoma||Adaptive immunity (dendritic cells, CTLs)||BRAF inhibition up-regulates antigen presentation and decreases IL-10, IL-6||[55,56]|
The immune response has also been shown to be essential to the efficacy of therapeutics [61–63]. Experimental and clinical evidence illustrates that patient host immunity contributes to the response to anti-tumour therapy. Patients with impaired host immunity probably have decreased overall and progression-free survival in a variety of solid and haematological malignancies [64,65]. In colorectal carcinomas, the type, density and intratumoral location of the T cell infiltrate has proved a more robust predictor of patient outcome than the tumour–node–metastasis (TNM) or Duke's classification . More generally, the host immune status influences the efficacy of conventional chemoradiation therapies .
Similarly, in mouse models the immune system has been shown to be critical to therapeutic response. Mouse models of hepatocellular carcinoma, pancreatic tumour and B cell lymphoma have implicated innate immune members such as mast cells  and macrophages  as barriers to tumour growth and facilitators of tumour regression. In mouse models of colon and breast adenocarcinomas, chemotherapeutic agents and radiation therapies have been shown to elicit immunogenic apoptosis of cancer cells .
Multiple mechanisms of the immune contribution to the therapeutic response have been suggested, including both innate and adaptive immune effectors as well as specific cytokines [61–63]. Recently, it has been proposed that the restoration in tumour cells of the ‘eat me’ and ‘find me’ immune stimulatory signals could potentially be used therapeutically to treat cancer [67,68]. Hence, the promotion of both the adaptive and innate arms of host immunity may be highly useful towards the complete elimination of tumour cells [67,68].
Hence, the notion that immune effectors may be important for the both the genesis and therapy of tumours is based upon extensive previous findings. Less clear is whether oncogene inactivation specifically mediates tumour regression through immune-dependent mechanisms. Recently, CD4+ T cells have been implicated in the mechanism of tumour regression upon inactivation in mouse models of MYC- or BCR-ABL-induced haematopoietic tumorigenesis . Oncogene inactivation in MYC-induced tumours in severely immunodeficient mice resulted in significantly delayed kinetics of tumour regression and failed to eradicate tumour cells completely, leaving up to 1000-fold more minimal residual disease (MRD) than in wild-type hosts. Thus, oncogene addiction appears to comprise both cell-autonomous and non-cell-autonomous mechanisms (see Fig. 1a,b) .
CD4+ T cells, and not the canonical anti-tumour cytotoxic CD8+ T cells, emerged as the key immune effectors of sustained tumour regression upon MYC inactivation. CD4+ T cells trafficked to sites of tumour involvement as early as 4 days after MYC inactivation and persisted for up to 3 weeks. Importantly, other effectors are also recruited to the tumour site, suggesting their possible contribution . CD4+ T cells contributed to oncogene addiction by enforcing both the induction of cellular senescence and the suppression of angiogenesis , processes characterized previously as hallmarks of oncogene addiction (see Fig. 2). The mechanistic basis is not entirely clear, but CD4+ T cells express many cytokines thought to play a role in the regulation of one or both of these processes [71–74]. In particular the pleiotropic protein, thrombospondin-1 (TSP-1), was identified as a critical mediator of CD4+ T cell-mediated sustained tumour regression upon MYC inactivation. TSP-1 could potentially play a multi-faceted role in contributing to remodelling of the tumour microenvironment upon oncogene inactivation.
Produced by a panoply of cells, including activated CD4+ T cells [69,75], TSP-1 is a potent anti-angiogenic and immune modulatory cytokine that can induce apoptosis of endothelial cells and regulate T cell chemotaxis . Moreover, TSP-1 has been shown to activate latent transforming growth factor (TGF)-β. Notably, TGF-β can play a tumour suppressive role in the tumour microenvironment [78,79]. Also, TGF-β can contribute to both the restraint of tumour onset as well as oncogene addiction through the regulation of cellular senescence upon MYC activation and inactivation [42,80]. Thus, it is tempting to speculate that TSP-1 may contribute to oncogene addiction via an influence on TGF-β. However, other cytokines are also likely to play a role including eotaxin-1, IL-5, interferon (IFN)-γ and tumour necrosis factor (TNF)-α as well as the down-regulation of ‘pro-tumour’ cytokines such as vascular endothelial growth factor (VEGF), IL-1β and monocyte chemoattractant protein (MCP)-1 upon MYC inactivation . Whether these chemokines also contribute more generally to the phenomenon of oncogene addiction remains to be seen.
CD4+ T cells co-ordinate multiple components of both the innate and adaptive immune system . Therefore, the contribution of other immune effectors to the mechanisms of oncogene addiction is likely. These results are consistent with observations in other murine models of oncogene-induced hepatocellular carcinoma, pancreatic tumour and B cell lymphoma that have implicated innate immune members such as mast cells  and macrophages  as barriers to tumour growth and facilitators of tumour regression.
Notably, the restoration of the p53 tumour suppressor had been shown previously to induce tumour senescence, elicit chemokine expression and induce the activation and recruitment of innate immune cells that contribute to tumour clearance . Thus, the restoration of normal function of a single tumour suppressor or oncogene elicits oncogene addiction through changes in the tumour microenvironment dependent upon various host immune effectors.
The apparent requirement of an intact host immune system in mediating oncogene addiction underscores the potential role of immune effectors in mediating the efficacy of targeted therapeutics. The kinetics of tumour cell elimination, the degree of tumour elimination, the elimination of minimal residual disease (MRD) and the duration of a clinical response could all be dictated by the host immune status (Fig. 2). Oncogene inactivation appears to directly antagonize many of the hallmark features of tumorigenesis (Fig. 1b), while the immune system appears to play a fundamental role in contributing not only to how oncogene activation initiates these features, but equally importantly to the reversal of these features upon oncogene inactivation (Fig. 2). Specifically, the ability of the tumour to regulate self-renewal versus cellular senescence and the capacity of the host to regulate the angiogenic state may both be tightly coupled to the ability of CD4+ T cells to regulate other immune effectors and cytokines (Fig. 2). These mechanisms may also contribute to tumour dormancy , the notion that there can be a pause or latency in cancer progression. Future targeted therapeutic strategies could include targeting genes in the senescence pathway through the induction of p53 activity or modulating genes in the cell cycle machinery . Targeted therapeutic strategies that modulate the expression of genes that control angiogenesis are used currently in the clinic with limited success , and more effective strategies need to be designed and implemented.
Specific host immune effectors and chemokines may profoundly influence the quality of tumour regression elicited by targeted oncogene inactivation, radiation therapy and chemotherapy. Experimental strategies to identify and develop novel anti-neoplastic therapies through in vitro or in vivo model systems that fail to account for host immunity may severely underestimate potentially powerful treatments. Clinically, many anti-cancer therapies cause immunosuppression and lymphodepletion that may undermine their efficacy . The careful choice of a combination of targeted and immune therapy may therefore be more efficacious in mediating sustained tumour regression .
The authors would like to acknowledge current members of the Felsher laboratory for critical discussion and previous members who have contributed to characterizing various models of oncogene addiction. Within the Felsher laboratory, studies of the tumour microenvironment have been funded by the Burroughs Welcome Fund Career Award, the Damon Runyon Foundation Lilly Clinical Investigator Award, NIH RO1 grant number CA 089305, 105102, National Cancer Institute's In-vivo Cellular and Molecular Imaging Center grant number CA 114747, Integrative Cancer Biology Program grant number CA 112973, NIH/NCI PO1 grant number CA 034233, the Leukaemia and Lymphoma Society Translational Research grant number R6223-07 (D.W.F.), the Stanford Graduate Fellowship (K.R.), the Stanford Medical Scholars Research Fellowship (P.B.) and the Howard Hughes Medical Institute Research Training Fellowship (P.B.).
The authors declare no competing financial interests.