The Problem of Cancer Dormancy

Abstract

The hiatus observed in the progression of cancer after diagnosis and treatment in a large proportion of patients has led to the notion that a state of cancer dormancy must exist during tumor progression. However, research on this stage of cancer has been limited due to the lack of appropriate models and clinical correlates. Fortunately, the last decade has seen the development of new cancer dormancy models, whole animal and intravital imaging techniques and the molecular characterization of minimal residual disease. These studies enabled researchers to reveal intriguing mechanisms and molecular determinants that define tumor dormancy. It is imperative to understand the basic mechanisms of dormancy, as this will accelerate the development of new markers of progression and novel therapeutic opportunities to induce dormancy and/or eradicate dormant disease. This issue of Cell Cycle includes a “Spotlight on Cancer Dormancy” highlighting major contributions to the field of cancer dormancy from basic and clinical studies. We anticipate that this will initiate a forum of discussion on the problem of cancer dormancy and stimulate investigators to study this rather unexplored but undeniably relevant clinical stage of cancer progression.

A STAGE OF DORMANCY IN CANCER PROGRESSION

It is an indisputable clinical fact that cancer cells, which disseminate from the primary tumor before its removal, can persist in a dormant state for a very long time, before initiating overt metastases.^1,2 Because in the past the existence of such cells could only be deduced, and not actually proven, it was impossible to attribute a specific mechanism to the phenomenon of dormancy.^2-5 Currently, genetic analyses of the primary tumors can predict with relative certainty whether spread has occurred.^6,7 Moreover, with the advent of new techniques of detection, isolation and characterization of very small number of cells from patients,⁸ as well as generation of experimental models of dormancy^9-11 (reviewd in this issue, refs. ^12-20), new insights are beginning to emerge. It appears that dormant cancer cells can persist by completely withdrawing from the cell cycle, or by continuing to slowly cycle and die at an equivalent rate. It is believed that the former is the result of loss of pro-proliferative signals while the latter is due to an inability to mount an angiogenic response.^11,21-23 Also, it appears that the disseminated cells that don’t resume proliferation but persist, even in face of chemotherapy,²⁴ might have developed adaptive mechanisms to survive under stress.²⁵ Some of these properties resemble those of stem cells.²⁶

It is my opinion that identifying mechanisms that keep disseminated cells in a persistent dormant state will allow interventions that might prevent their reemergence as overt metastases, while overcoming their survival response to stress might lead to their elimination. Both scenarios would be of benefit to cancer patients.

In this introductory article I provide a focused synopsis of the salient findings described in the reviews in the “Spotlight on Cancer Dormancy” in this issue of Cell Cycle, and put them in context of the important, and mostly unanswered questions in the field.

CLUES ABOUT DORMANCY FROM THE CLINIC AND MOUSE MODELS: IS THERE A ROLE FOR TUMOR STEM CELLS?

Three reviews in this issue (Marches et al.,¹³ Laufs et al.¹⁶ and Klein and Hötzel¹⁸) provide tantalizing information from minimal residual disease (MRD) studies regarding how dormancy may manifest itself in the clinic.^1-3,8,27 Klein and Hötzel¹⁸ suggest that almost all tumors undergo what appears to be an obligatory dormancy stage that can last in some cases ~5 years. They show that very few genetic alterations are found in a large proportion of early disseminated tumor cells suggesting that either genetic alterations not detected by CGH are taking place or that the lack of growth of these initial departing tumor cells may be driven initially more by epigenetic than genetic alterations. Laufs et al.,¹⁶ describe the use of staining for the urokinase receptor (uPAR) along with cytokeratin-18 in bone marrow disseminated tumor cells as a way to stage cancer patients. Their findings showing that uPAR staining is an accurate predictor of poor prognosis and recurrence in gastric cancer^28,29 correlated well with mechanistic studies from the Ossowski lab (see Ranganathan et al. and Laufs et al. refs. ¹⁶ and ¹⁹) where it was found that uPAR overexpression and signaling^21,30 was important to prevent dormancy of malignant carcinoma cells.³¹ Thus, these studies^1,28 support further exploration of strategies such as humanized monoclonal antibodies or small molecules that may restrict disease recurrence by blocking uPAR function.

Marches et al.,¹³ describe the detection of circulating tumor cells (CTCs) in blood >20 years after patients have been deemed cured.²⁷ They hypothesize that the existence of tumor stem cells may explain this sub-clinical condition of cancer. They propose that rather than the existence of a tumor stem cell population remaining alive in circulation it is possible that tumor stem cells that have disseminated and formed micrometastases might shed cells into circulation, which may explain their findings. This hypothesis is in agreement with the studies reviewed by Felsher²⁰ on Myc inactivation in a transgenic liver tumor mouse model.^26,32 It is proposed that two possibilities might explain the transition between tumor cell growth and quiescence due to Myc inactivation: either Myc transformation occurred in a liver stem cell that upon oncogene inactivation differentiated into a normal appearing and functioning cell, but that still retains stem cell properties that allow them to resume growth after Myc-reactivation or, alternatively Myc may activate a stem cell program in cells that had already committed to a specific lineage.^26,32 It is possible that many of the questions regarding tumor cell dormancy will be answered if it can be demonstrated that disseminated tumor stem cells can regain quiescence. Given that the primary tumor may generate communal signals that “recruit” the tumor stem cells, it is possible that upon dissemination and lodging in distant organs tumor stem cells that are unable to reproduce their expansion niche may enter a program of dormancy, a natural state for normal stem cells.^26,33,34 Alternatively, the bone marrow, where disseminated tumor cells can be found, may serve as a survival reservoir for tumor stem cells. Thus, it will be important to determine the functional relationship between dormant tumor cells and their surrounding bone marrow or target organ stroma (e.g., extracellular matrix, fibroblasts, inflammatory cells).^2,35 Further, the possibility that a tumor stem cell that carries mutations that can give origin to a primary tumor would reprogram into a low proliferation or growth arrest mode, questions the dominance of oncogenic signaling over cellular programs that can induce quiescence.

An interesting possibility is that tumor cell dormancy may represent an adaptive response to a non-propitious microenvironment.⁹ Tumor cells that are not fit to survive the dissemination may be rapidly cleared by apoptosis.³⁶ Tumor cells that have all the tools to resume growth will do so.³⁷ A third scenario is a phenotype where the genetic and/or epigenetic alterations in tumor cells do not favor immediate growth but give a selective advantage through the adaptation to stress (e.g., hypoxia, hyperoxia, low nutrient availability, chemotherapy).^9,25 This phenotype could be dependent on programs described by Ranganathan et al.,¹⁹ where pathways that allow cells adapt to stress (e.g., p38^SAPK, endoplasmic reticulum stress) would halt tumor cell growth until the conditions are once more favorable.²⁵ This behavior could similarly apply to a model of tumor stem cell dormancy, as stem cells are known to withdraw from cell cycle and to resist drug treatments due to the expression of MDR pumps.^26,33,34 Progress made in the tumor stem cell research field will very likely provide insight into these possibilities.

Most of the mouse models of cancer attempt recapitulating carcinogenesis and primary tumor development. Comparatively, much less effort has been dedicated to model the progression of MRD and metastasis development and these models rarely include chemotherapy. Only recently some mouse models have taken advantage of inducible expression systems for oncogenes such as Myc and Her2/neu or malignancy promoting factors such as TGFβ.^38-40 While some of these studies have addressed the problem of metastasis they remain mostly focused on primary tumor development. Nevertheless, such transgenic mouse systems may be useful to model the progression of the disease including surgery of the primary tumor, chemotherapeutic treatment, MRD, dormancy and recurrence of disseminated disease.

TWO POTENTIAL MECHANISMS FOR TUMOR DORMANCY: WHAT CAN WE LEARN FROM THEM?

The general consensus that emerges from these reviews is that two stages might explain tumor dormancy: one that involves the induction of growth arrest and prolonged survival of disseminated single or small groups of tumor cells^21,41 and a second that occurs in a small tumor mass where a constant balance between apoptosis and proliferation keeps the lesion small and undiagnosed.^42,43 As these are general outputs of tumor cell behavior it is anticipated, and reflected in the reviews presented here, that different mechanisms might regulate these outcomes.

The existence of cancer dormancy due to a growth arrest of tumor cells is supported by evidence that within tissues where primary tumors are developing (see White et al. and Ranganathan et al., refs. ¹⁵ and ¹⁹, this issue) or that harbor disseminated cells (Townson and Chambers¹² and Ranganathan et al.,¹⁹ this issue) and have a functional vasculature, tumor cells are found to be in a nonproliferative mode. Dormant tumor cells have been found to be in a G₀/G₁ arrest which was linked to reduced ERK mitogenic signaling,^21,31 and to be associated with negative staining for proliferation markers (e.g., PCNA or Ki67).^1,22 In models of breast and head and neck cancer it appears that a failure of tumor cells to coordinate integrin (e.g., α5β1.β1) and growth factor signaling (e.g., EGFR, Her2/neu) may result in the deactivation of mitogenic signaling^21,39,44,45 and activation of p38^SAPK stress pathways,⁴⁶ that can lead to a state of tumor dormancy.

Naumov et al.¹⁷ and Indraccolo et al.¹⁴ describe how a failure to activate the angiogenic “switch” can maintain the tumor cell population in a constant balance between apoptosis and proliferation^23,47 and cause dormancy. Indraccolo et al.¹⁴ provide novel insight as to the requirement for high levels of angiogenic factors to interrupt dormancy and subsequently low or moderate levels to maintain a functional vasculature in the now growing tumors. Marches et al.¹³ describe how in a model of B-cell lymphoma, immunizing toward the BCL1 Ig idiotype can induce tumor dormancy.¹¹ Although in this model there is a balance between proliferation and cell death, it does not seem to depend on the angiogenic capacity of the tumor cells but on B-cell receptor signaling and upregulation of p21^CIP/WAF1.^11,48 It will be interesting to learn how a tumor cell population that achieves a balance of apoptosis and proliferation for years and accumulates genetic modifications can avoid undergoing an angiogenic switch and escape dormancy and whether VEGFR and PDGFR inhibitors targeting the endothelium and pericytes, respectively,⁴⁹ can induce and maintain dormancy. Similarly it will be interesting to learn how this balance and the angiogenic switch is regulated in tumor stem cells, given that they could potentially enter a growth arrest and remain quiescent.^26,32-34 As highlighted by Marches et al., the precision of the events controlling the balance between proliferation and death seems extraordinary. Therefore, it is important to understand its underlying operation to be able to activate cell death pathways, which would allow eradicating the residual disease.²⁷

The notion of reprogramming malignant cells is not new.^20,26 Still, the existence of tumor dormancy, which could be defined as a reprogramming towards normalcy, challenges some well-established concepts in cancer progression. For example how is it possible that tumor cells that have accumulated genetic and epigenetic alterations that led to the formation of a primary tumor³⁷ somehow revert to a growth arrest mode after dissemination? This implies that common genetic alterations thought to be dominant (e.g., Ras mutations, ErbB2 amplification) may be overridden by signals that force the cells into a quiescent state. If this is true it suggests that the ability of tumor cells to fulfill the hallmarks of cancer³⁷ may be context dependent. This is an exciting possibility as it suggests that there is still plasticity in these malignant cells to convert them into a dormant phenotype.

POTENTIAL THERAPEUTIC APPROACHES

Understanding the mechanisms that induce dormancy and drug resistance of dormant cells would allow generating at least two strategies to improve treatment. Therapies aimed at inducing dormancy may help overcome conventional drug resistance that is based on the ability of drugs to induce cell killing. Thus, reprogramming malignant cells into a growth arrest would allow converting the disease that would be otherwise untreatable, into a chronic asymptomatic condition. The combination of mitogenic signaling inhibitors (e.g., Mek, EGFR, Met, IGFR inhibitors)^50-52 with activators of stress pathways (e.g., p38, JNK) may induce reprogramming of tumor cells into dormancy (see Ranganathan et al., ref. 19). A second option, which is highlighted in the reviews presented here, envisions inducing killing of dormant tumor cells during the dormancy stage^24,25 (see Townson and Chambers, Marches et al., and Ranganathan et al., refs. ^{12, 13} and¹⁹). However, in order to conclusively determine whether a patient has dormant disseminated disease, new markers might have to be developed (see also Marches et al., ref. 13). Studies in the MRD field have identified several markers to achieve this goal.¹ However, most of these markers inform about the recurrence of the disease (e.g., uPAR, ErbB2)^1,28,29 rather than the state of tumor cell dormancy. It is possible that ongoing research in experimental models combined with the molecular characterization of CTCs or tumor cells in bone marrow will very likely reveal markers of dormancy. To achieve this goal it will be important to enhance collaborations between basic research labs and clinical labs to test the markers of dormancy identified in experimental models in samples from bone marrow, lymph nodes and CTCs. These will be important for staging of cancer and for developing novel targets for therapies.

The data covered by these reviews describe new exciting and provocative hypotheses and in depth data to support them. Collectively they represent major advances in our understanding of the problems associated with cancer dormancy.