Cancer Gene Therapy: The power of negative thinking

Promising new results in a well-established murine model of transplantable prostate cancer could lead to strategies that seek to exploit a fourth pathway for the gene therapy of cancer.

Cancer is fundamentally a genetic disorder associated with modification of genes that block apoptotic pathways, thereby causing abnormal necrotic cell death and expression of genes which enable unscheduled cell growth, particularly in nominal cancer stem cells, which have learned and inculcated these dark secrets.¹ Thus, much of the excitement about gene therapy for cancer at its beginning posited developing strategies to replace these altered genes (the first pathway) thereby returning a normal phenotype. While attractive, this notion was limited by the bilateral Achilles heels of targeting (getting the right genes into the right cells in the right places) and the irksome immune response to the more effective viral delivery systems. Perhaps the most intriguing current strategy of this ilk is to deliver antisense constructs to tumors to limit expression of antiapoptotic genes such as Bcl-2, which has indeed met with some success in the clinic.²

A second pathway is the use of genetic constructs to confer novel properties on a cell, often to elicit an effective immune response to the tumor using cytokines or antigens, an approach that our group and others have championed for quite some years^{3, 4, 5, 6, 7, 8, 9, 10} with a modicum of success. For example, our studies in the late 1980s (the first gene therapy studies in the United States) focused on the adoptive transfer of genetically modified antitumor T cells first designed solely to mark them,⁵ and subsequently to use them to deliver additional gene products such as TNFα. The importance of such attempts may have been primarily in elucidating the underlying immune mechanisms that are complex, highly evolved, remarkably redundant and synergistic. The application of GM-CSF as an immune adjuvant, for example, was first identified using such gene transfer strategies¹¹ and we can imagine future gene transfer studies using newly identified cytokines and costimulatory molecules to similarly help us define the biology of these factors. The third pathway of cancer gene therapy, and actually one of the first applied clinically, is to use nominally tumor-selective viruses that mediate oncolytic effects on the tumor.¹²

A fourth emergent pathway for cancer gene therapy focuses on strategies designed to modify the biology of the tumor-reactive inflammatory cells. Zhang et al¹³ new study, recently published in Cancer Research, is an excellent example of one such strategy to make T cells resistant to the altered and nominally disordered tumor microenvironment (see below). The authors adoptively transferred T cells which recognized the prostatic tumor, but that had been genetically modified to bind but not respond to TGFβ. The notion that TGFβ is important in cancer has a long history, with many tumors being able to both produce it and limit their own responsiveness to it by downregulating receptors. The immunologic consequences of local TGFβ elaboration by the tumor most likely involve both direct and indirect immunosuppression as well as the ability of TGFβ to promote tissue repair and wound healing. TGFβ limits NK and T-cell production of IFNγ, thereby checking the so-called TH1 response. Limiting this response, in turn, leads to inhibition of tumor growth and enhancment of the cytolytic response to the tumor. TGFβ is delivered not only by the tumor but likely also by the resplendent and renascent T-suppressor cell, found abundantly within various tumors.^{14, 15} This same group first tested the strategy of blocking TGFβ effects by engineering bone marrow cells to express a dominant-negative receptor in all emergent T cells. However, this strategy could be dangerous because it could lead to autoimmune T cells as well as antitumor T cells arising.¹⁶ By contrast, the new strategy of introducing TGFβ into tumor-specific T cells is more likely to be clinically applicable.

There is strong evidence from murine models and inflammatory infiltrate, which indicates that early events in tumorigenesis are associated with dendritic cell (DC) and T-cell infiltrates. It is clear then that cancers in adults arise most often as the end stage of chronic inflammation^{17, 18} Cancer in adults is thus an immunologic disorder, which generates multiple neoepitopes and common tumor antigens.

These findings fueled burgeoning academic and biotechnology industries that were designed to develop antigen and cytokine therapies, some as gene therapies, based on the very effective therapies in murine models. The average epithelial neoplasm has over 10 000 mutations as a consequence of emergent genomic instability, enabling a cornucopia of antigens for exploration. One example of such a cancer is prostate cancer, the fifth highest cause of cancer death in the USA. Prostate cancers^{13, 19} develop resistance to TGFβ during their development²⁰ from a variety of mechanisms including downregulation of the receptor itself or associated signaling pathways during their inflammatory genesis. This increased resistance protects them from the TGFβ elaborated in the tumor microenvironment. Not so for the susceptible T cell. Alternatives to genetically modifying the T cells to protect them from tumor-elaborated TGFβ are to apply reagents that modify the tumor TGFβ receptors themselves or to antagonize the local release by antisense constructs or antibodies.^{21, 22} Although intriguing, the global role of TGFβ and its ablation, being associated with lethal inflammation in murine models as well as its critical role in tissue remodeling and healing, limit enthusiasm for this strategy. Modification of the T cell confers specificity (through the cognate T-cell receptors specific for cancer antigen/MHC complexes) and limits the biologic effects of these cells nominally to the tumor itself. So it appears that TGFβ in the tumor microenvironment might indeed modify the ability of T cells to mediate their effector function.²³

The tumor microenvironment is replete with the ancient cytokines TGFβ and HMGB1, which are important in wound repair and inflammation. Intriguingly, TGFβ is highly conserved, with only a single amino-acid difference between man and mouse. This molecule modifies inflammation in important ways.^{24, 25} Similarly, the nuclear transcriptional regulatory factor, high-mobility group B1 (HMGB1), appears to promote inflammation when released from necrotic cells²⁶ and to drive the initial events in eliciting immune reactivity. HMGB1 differs only in two amino acids when comparing mouse and man. Like TGFβ, HMGB1 is probably in an evolutionary box that it cannot get out of. So both these cytokines are critically important for tissue homeostasis and the response to injury or damage.

To sensibly co-ordinate and direct cancer therapies, we need to recognize the critical role of these cytokines in normal tissues and the response to injury, while (Figure 1) promoting tumor eradication (1), eliciting an immune response (2), and/or promoting immune effector function (3). Future strategies will need to remain aware of these three coordinate approaches when confronting the problem of cancer. Nascent gene therapies should similarly pay heed to these notions, applying all three approaches coordinately when possible. ▪

Figure 1

Cancer therapeutic triangle. The last 50 years of evolving cancer therapies have primarily focused on means of either locally extirpating cancer (1) by surgical excision or radiation therapy. Adjuvant, therapeutic or, increasingly, neoadjuvant strategies to elicit apoptotic (non-necrotic) cell death most frequently uses chemotherapeutic agents, most commonly platinum-based regimens, often with taxanes or fluoro-pyrimidines. An alternative strategy that uses direct application of oncolytic viruses has been dealt with extensively in this journal and others.^{12, 27, 28, 29, 30, 31, 32} One of the major issues with this approach is that TGFβ³³ and HMGB1^{34, 35, 36, 37} often mediate the therapeutic response: paradoxically, these molecules can also suppress immune reactivity and promote reparative cell growth with associated angiogenesis and stromagenesis. Gene replacement or antisense strategies are similarly designed to enhance apoptotic death of tumors. Immune approaches that have been explored in earnest for 20 years typically try to elicit a novel immune response (2) by targeting the afferent limb using peptide or dendritic cell (DC) vaccines, systemic delivery of interferon α or use of adenoviral, canary pox or other viral vectors to deliver antigens or cytokines. The emergent and perhaps most interesting approaches predicate an already established immune response. These strategies are designed to promote survival and function of the (3) efferent limb of the immune response, which is deficient in the setting of the tumor microenvironment, perhaps as a consequence of HMGB1 and TGFβ local elaboration. Interleukin 2 administration likely overcomes some of the effects of suppressor T cells, which can be supplemented by antibodies to important immunological molecules regulating the immune response including antibodies to CTLA4, CD137, PD1 or the IL2R. Proteosome inhibitors appear to promote apoptotic death in the tumor targets and may limit elaboration of immunosuppressive factors. Making the adoptive transfer of T cells bullet-proof by making them TGB-β insensitive, as described, would represent one of the versions of enhancing effector pathways as a third emergent pathway in cancer gene therapy.