TGF-betas as key mediators of tumor-stromal interactions: opportunities for prevention

ED03-02

TGF-βs as candidate master regulators of stroma/parenchymal interactions and environmental sensing within tissues. It is increasingly well-appreciated that normal tissue homeostasis is maintained through the integrated activity of many different cellular components, all of which have the potential to impact on the carcinogenic process if the integrative control mechanisms go awry. TGF-βs are highly pleiotropic polypeptide growth factors that may occupy a nodal position in the molecular networks that mediate this functional integration . The TGF-βs arose relatively late in metazoan evolution as organismal complexity was increasing , and it is plausible that they evolved to coordinate and fine tune higher order cell-cell interactions and organization. They have a number of properties that would particularly fit them for this role. Expression of TGF-β ligands and receptors is essentially ubiquitous, and nearly every cell type shows some biological response to TGF-β stimulation. TGF-βs are secreted into the extracellular milieu as biologically latent molecules where they must be activated in order to elicit a biological response . Typically the latent TGF-β complexes are bound to extracellular matrix components, where they are poised to act as local environmental sensors. Activation of TGF-β is an acute response to many forms of stress or injury that serves to restore normal homeostasis. However chronic activation of TGF-β is associated with the pathogenesis of many diseases, including fibroproliferative disorders and cancer .
 Biological effects of TGF-βs on stromal cells. TGF-βs have effects on essentially all major cellular compartments of the stroma, including the vasculature, immune cells and fibroblasts . In the vascular compartment, TGF-βs promote blood vessel maturation through direct effects on endothelial cells, and by increasing pericyte and smooth muscle cell coverage. However, they can also impact indirectly on angiogenesis by regulating the local production of pro-angiogenic and anti-angiogenic factors. In the immune compartment, TGF-βs can initiate an acute inflammatory response by promoting leukocyte recruitment, but their overall effect on immune function is generally suppressive. Thus TGF-βs inhibit the generation or effector function of cytotoxic T-cells, NK cells, dendritic cells and macrophage/ monocytes, and they promote the generation and activity of immunosuppressive regulatory T-cells. Indeed, the TGF-β1 null mouse dies of a multifocal inflammatory response, suggesting that tonic inhibition of the immune system by TGF-β1 is critical to prevent self-reactivity and autoimmune disease. Finally, TGF-β modulates the phenotype of stromal fibroblasts, promoting differentiation to the myofibroblast phenotype, with particularly important effects on extracellular matrix and cytokine production. Clearly, these three activities, namely promotion of angiogenesis, immune suppression and modulation of the composition of the acellular microenvironment, if engaged chronically would tend to promote tumor progression.
 Evidence that activity of the TGF-β pathway is critical for normal homeostatic interactions between stroma and epithelia. TGF-βs are potent inhibitors of epithelial cell proliferation, and initial studies on the role of TGF-β in tumorigenesis focused on loss of TGF-β response in the epithelial compartment. Unexpectedly however, recent work has suggested that stromal responses to TGF-β are critical for the maintenance of homeostasis in the overlying epithelium . Conditional knockout of the type II TGF-β receptor specifically in fibroblasts was shown to promote dysplasia or frank carcinogenesis in the epithelia of the prostate and stomach. Similarly, conditional knockout of the downstream TGF-β signaling component Smad4 in T-cells lead to development of colorectal carcinomas. In both cases, loss of TGF-β response in the stromal compartment resulted locally in the inappropriate secretion of growth factors or cytokines that were tumor promoting for the particular epithelium. The data suggest that in normal homeostasis, tonic low level TGF-β activity is critical for maintaining an anti-carcinogenic phenotype in the stroma. Indeed, mice and humans with germline mutations or polymorphisms that reduce TGF-β pathway function tend to be tumor-prone, but since these lesions have a direct impact on the epithelial cells as well, it is difficult to dissect out the relative contribution of the impaired stroma to this phenotype.
 The TGF-β-mediated parenchymal-stromal dialog becomes distorted in tumor progression. The mouse studies demonstrate that an intact TGF-β response in the stroma is important for normal epithelial homeostasis, but to date there is no clinical evidence that somatic inactivation of the TGF-β pathway occurs in the stromal compartment during carcinogenesis. However, the TGF-β-mediated dialog between tumor and stroma frequently gets distorted in a different way during cancer progression. Many advanced human tumors show elevated TGF-β expression, which correlates with metastasis and poor prognosis . Although TGF-β clearly can have direct pro-progression effects on the tumor parenchyma, preclinical studies suggest that the more important effect of elevated TGF-β is to generate a more permissive tumor stroma, primarily through suppression of effective anti-tumor immune surveillance and promotion of angiogenesis, as discussed in more detail below. Thus it appears that normal parenchymal/stromal interactions involving TGF-β are very delicately balanced, and that chronic resetting of the system away from its equilibrium position either by reduced or enhanced TGF-β pathway activation will result in the generation of a tumor-promoting stroma.
 Targeting TGF-β to restore an anti-carcinogenic stroma: prospects for tertiary prevention. If overexpression of TGF-β by advanced tumors does indeed generate a tumor-promoting stroma, then TGF-β antagonism should reverse this process and prevent further tumor progression. Clearly the delicate balance between the undesirable effects of too much and too little TGF-β, and the general complexity of TGF-β biology make the design of strategies for TGF-β pathway blockade in vivo somewhat tricky. However, a number of preclinical studies now suggest that this approach may be feasible . In one of the earliest studies of this type, our lab generated a transgenic mouse that chronically overexpressed an antibody-like TGF-β antagonist . We found that these mice were relatively protected from the development of metastatic disease, both in the context of spontaneously arising metastatic tumors driven by overexpression of the Her2/Neu oncogene, and when metastatic cells were injected intravenously. Unexpectedly, we saw no acceleration of primary tumorigenesis despite the fact that TGF-β functions as a tumor suppressor in the Neu-initiated mammary epithelium, and we saw little evidence for the autoimmune manifestations that would have been predicted from the phenotype of the TGF-β1 knockout mouse. A number of other studies, using a variety of strategies to antagonize TGF-β have come to similar conclusions (reviewed in ).
 To address the mechanisms underlying this metastasis suppression, we have used the 4T1 transplantable model of metastatic mammary cancer in syngeneic Balb/c mice. By depleting select populations of immune cells, we can show that the efficacy of an anti-TGF-beta antibody is dependent on both the innate and the adaptive arms of the immune system, suggesting that TGF-β antagonism is acting to restore effective anti-tumor immune responses. We can further show that TGF-β antagonism has additional effects on the tumor cell itself that make the tumor more visible to the immune system and more susceptible to immune cell killing. Finally, we also found a decrease in microvessel density in tumors and metastases treated with anti-TGF-β, and at the RNA level we saw alterations in expression of extracellular matrix genes that have previously been implicated in metastasis susceptibility. Individually most of the effects of TGF-β antagonism were relatively small in magnitude and local to the tumor site. Thus, consistent with the notion that TGF-β occupies a nodal position in mediating mediating tumor-stroma interactions, TGF-β antagonism has effects on virtually every cell type in the tumor, and aggregately these effects combine to modify tumor/stromal interactions in such a way that further tumor progression is impeded.
 Clinical prospects. Encouraged by these and similar preclinical results, a number of pharmaceutical and biotechnology companies are developing TGF-β antagonists for use in a cancer setting . A fully human monoclonal antibody and a small molecule TGF-β receptor kinase antagonist are now in Phase I trials in cancer patients. It will be interesting to see whether this approach, which allows simultaneous targeting of so many cells in the cancer microenvironment, will offer advantages over more narrowly targeted approaches.
 (1)
 Bierie B, Moses HL. Nat Rev Cancer 2006;6(7):506-20.
 (2)
 Putnam NH, et al. Science 2007;317(5834):86-94.
 (3)
 ten Dijke P, Arthur HM. Nat Rev Mol Cell Biol 2007;8(11):857-69.
 (4)
 Blobe GC, Schiemann WP, Lodish HF. N Engl J Med 2000;342(18):1350-8.
 (5)
 Bhowmick NA, et al. Science 2004;303(5659):848-51.
 (6)
 Bhowmick NA, Moses HL. Curr Opin Genet Dev 2005;15(1):97-101.
 (7)
 Kim BG, et al. Nature 2006;441(7096):1015-9.
 (8)
 Gold LI. Crit Rev Oncog 1999;10(4):303-60.
 (9)
 Akhurst RJ. Curr Opin Investig Drugs 2006;7(6):513-21.
 (10)
 Yang YA, et al. J Clin Invest 2002;109(12):1607-15.