Tumor dormancy, oncogene addiction, cellular senescence, and self-renewal programs

Abstract

Cancers are frequently addicted to initiating oncogenes that elicit aberrant cellular proliferation, self-renewal, and apoptosis. Restoration of oncogenes to normal physiologic regulation can elicit dramatic reversal of the neoplastic phenotype, including reduced proliferation and increased apoptosis of tumor cells (Science 297(5578):63-64, 2002). In some cases, oncogene inactivation is associated with compete elimination of a tumor. However, in other cases, oncogene inactivation induces a conversion of tumor cells to a dormant state that is associated with cellular differentiation and/or loss of the ability to self-replicate. Importantly, this dormant state is reversible, with tumor cells regaining the ability to self-renew upon oncogene reactivation. Thus, understanding the mechanism of oncogene inactivation-induced dormancy may be crucial for predicting therapeutic outcome of targeted therapy. One important mechanistic insight into tumor dormancy is that oncogene addiction might involve regulation of a decision between self-renewal and cellular senescence. Recent evidence suggests that this decision is regulated by multiple mechanisms that include tumor cell-intrinsic, cell-autonomous mechanisms and host-dependent, tumor cell-non-autonomous programs (Mol Cell 4(2):199-207, 1999; Science 297(5578):102-104, 2002; Nature 431(7012):1112-1117, 2004; Proc Natl Acad Sci U S A 104(32):13028-13033, 2007). In particular, the tumor microenvironment, which is known to be critical during tumor initiation (Cancer Cell 7(5):411-423, 2005; J Clin Invest 121(6):2436-2446, 2011), prevention (Nature 410(6832):1107-1111, 2001), and progression (Cytokine Growth Factor Rev 21(1):3-10, 2010), also appears to dictate when oncogene inactivation elicits the permanent loss of self-renewal through induction of cellular senescence (Nat Rev Clin Oncol 8(3):151-160, 2011; Science 313(5795):1960-1964, 2006; N Engl J Med 351(21):2159-21569, 2004). Thus, oncogene addiction may be best modeled as a consequence of the interplay amongst cell-autonomous and host-dependent programs that define when a therapy will result in tumor dormancy.

Figure 1. Oncogene addiction elicits tissue-specific effects

Oncogene inactivation has been observed to have different outcomes depending upon the tissue origin of tumors, including proliferative arrest, differentiation, apoptosis, and/or cellular senescence. The specific consequences of oncogene addiction have a dramatic impact on whether targeted therapies will result in tumor dormancy, as in the case of MYC-induced hepatocellular carcinoma, or tumor elimination, as shown for MYC-induced lymphoma.

Figure 2. Oncogene addiction comprises both cancer cell-autonomous and non-cell-autonomous mechanisms of tumor regression

Oncogene inactivation reverses many of the hallmarks of cancer to lead to tumor regression through reinstatement of cell-intrinsic programs such as proliferative arrest, apoptosis, and differentiation, as well as host-dependent phenomena including immune system activation and destruction of the tumor vasculature. Notably, cellular senescence is both a tumor cell-intrinsic and a host-regulated consequence of oncogene addiction; however, it is likely only in an immune intact host that targeted therapy will result in complete tumor elimination.

Figure 3. Tumor dormancy versus tumor elimination is regulated by an intact host immune system

Oncogene inactivation results in tumor regression regardless of host immune status due to proliferative arrest, apoptosis, and/or differentiation. In an intact host, the immune system is subsequently activated to facilitate elimination of the residual tumor cells via induction of cellular senescence and destruction of the tumor vasculature. However, in the absence of an immune system, tumors establish a state of dormancy, due to lack of inhibition of self-renewal and angiogenesis, followed by eventual escape from therapy and tumor recurrence.

Figure 4. Modeling and predicting when oncogene inactivation will result in tumor dormancy versus tumor elimination

A combinatorial, iterative approach can be utilized to model the consequences of oncogene addiction, thereby allowing for prediction of response to targeted therapy in patients. Continuing progress in molecular imaging and biomarker discovery will be crucial for the validation of mathematical models of oncogene addiction in primary animal models of cancer. In parallel, the advent of new molecular imaging tools in the clinic will allow for incorporation of discoveries in these pre-clinical models into assessment of human cancer patient response to therapy.