The many faces of tumor dormancy

The issue of tumor dormancy becomes increasingly acute as we progressively refine our diagnostic techniques, thereby enabling the detection of relatively small tumor cell deposits that previously would have escaped notice. This issue encompasses both dormant primary tumors and dormant micrometastatic or metastatic growths. In each of these cases, tumor growth may be so minimal that doubling times are measured over periods of years or decades, resulting in overall schedules of tumor progression that exceed the average lifespan of the individuals carrying them. Moreover, it is unclear whether the barriers to active proliferation that constrain such growths are imposed by the tissue microenvironments in which they find themselves or represent cell-autonomous defects that preclude the tumor cells from growing actively. Given (i) the multi-step nature of pathogenesis of adult tumors in virtually all tissues; (ii) the long lag times between the individual steps of these multi-step processes; and (iii) the long lives of most humans, it is unsurprising that autopsies performed on individuals in their 7th, 8th, and 9th decades of life reveal multiple, previously undetected incipient tumors that were never able to become clinically apparent. Their growth testifies to the fact that multi-step tumor pathogenesis proceeds at a slow rate in many tissues and in most of us, and that we may be blessed with reaching an old age without any of these complex programs achieving completion and becoming clinically apparent, i.e., it is plausible that the bodies of most aged humans are riddled with preneoplastic growths that are en route to becoming full-fledged malignancies but whose agenda of tumor formation is cut short by death from other causes. In many of these cases, we can imagine that studying the multi-step pathogenesis of tumors which commonly arise in the relevant tissues and become clinically apparent will give a clear explanation of the "dormancy" of these preneoplastic growths: they have had enough time to complete some of the early steps of the program leading to full-blown tumors, but not enough time to complete all of these steps. Study of such dormant growths will not, on its own, create new conceptual paradigms. Instead, it will simply illuminate the early chapters of the more complex stories of how clinically detectable tumors arise. Another set of dormant growth in anatomical sites of primary tumor formation may eventually prove to be more interesting and conceptually novel: the cells in these tumors may have acquired a series of initial changes (for example, somatic mutations) that have led them down a path that is, ultimately, a dead end. Thus, the set of initially sustained somatic changes may, because of their nature, not pave the way for other changes to follow that, in aggregate, will culminate in the formation of a fully malignant tumor. Accordingly, these initiating changes may, for reasons still unknown, perturb tissue histomorphology to yield hyperplastic or dysplastic tissues, but the subsequent steps that would be required to convert these into bona fide neoplastic cells may not be likely to occur, even within an essentially unlimited period of time. Multiple sources of tumor dormancy. As an alternative, these initiating steps may affect a cell of origin whose differentiation program is, for various complex reasons, intrinsically incompatible with spawning a neoplastic growth, whatever the initiating somatic alterations (for example, mutations) are that happen to strike this cell. By either mechanism, the preneoplastic cell nests are stuck in a state from which they cannot emerge and, as such, appear to be "dormant". In one sense, this term is most unfortunate, since it implies that such growths can, under the proper conditions, be awakened, or at least re-awakened, when in fact these cells may be irreversibly blocked from re-entry into active growth. Another type of dormant growth is to be found in the metastatic derivatives of a primary tumor. The invasion-metastasis cascade, as it is often called, is responsible for physically disseminating primary cancer cells to a diverse array of sites throughout the body. While some might propose a tropism of cells that causes them to migrate preferentially to specific target organs throughout the body, it seems more likely that cancer cells are spread via hematogenous dissemination to a vast array of capillary beds and become physically trapped in them. Some of these cells may then extravasate, entering the parenchyma of the surrounding tissues. It is likely that the same cell-motility and cell-invasion programs, such as the epithelial-mesenchymal transition (EMT), which initially enabled such cells to enter into the circulation and survive transport to distant sites, also empower such cells to extravasate. However, these programs of invasion and metastasis are unlikely to facilitate the adaptation of already disseminated cells to their newfound homes. Thus, at the moment of arrival, it is unlikely that the vast majority of recently disseminated cells are adapted to these alien microenvironments, which are likely to differ substantially from those within the tissues in which the primary tumors arose. The subsequent ability of such cells to found a micrometastasis and, in the longer term, a macroscopic metastasis—the latter process being termed "colonization"—would seem to depend on adaptive solutions that these cells discover long after they have been seeded in distant tissue microenvironments. In such microenvironments, Darwinian selective pressure will place a strong premium on the rare cells that do devise successful solutions to the adaptation problem. This suggests that many dormant tumors, largely micrometastases, are simply clusters of cells that are attempting to survive and expand without having developed the requisite means of doing so, for example, by learning to adapt to the spectrum of growth factors, cytokines, and extracellular matrix components that together constitute and define each distinct tissue microenvironment in the body. In the event that the cells within a micrometastasis do indeed succeed, on rare occasions, in solving the problem of adaptation, they are now equipped to launch the final, lethal step of tumor progression. These cells are capable of metastasis, since they descend from cells that brought them to their site of initial metastasis in the first place. In addition, these cells have now solved the "adaptation problem". Together, these two attributes should enable this metastasis, which is likely to grow rapidly soon after, adaptation is achieved, to serve as the center (nidus) from which the synchronous seeding of hundreds, perhaps thousands of secondary metastases emanates; each of these secondarily seeded cells should be able to begin to proliferate robustly shortly after its arrival in a new home, since such cells will, as said, have solved the adaptation problem. This scenario raises many questions, one of which concerns the mechanisms by which initially disseminated micrometastatic cancer cells solve the adaptation problem and thereby acquire the ability to colonize. In general, if micrometastatic deposits contain cells that are out of the active growth-and-division cycle and thus are truly dormant, it would seem unlikely that they will ever be able to assemble the traits required for adaptation and colonization; instead such acquisition is likely to require active cell cycling and the continual reshuffling of tumor cell genomes. Hence, colonization must arise in micrometastases that have, over extended periods of time, contained cells that have been in the active growth-and-division cycle. Since such micrometastases remain by definition very small, it seems highly likely that their active growth and division is counterbalanced by an equivalent amount of cell death each generation, yielding no net change in overall cell number. Such proliferation and compensating cell death may occur, for example, under conditions where tumor-associated neovascularization has been ineffective and thus unable to sustain micrometastatic deposits above a certain very small size. It seems highly likely that the various types of dormant cancer cells described above exist in various patients. Yet another type of dormancy is more speculative, but seems highly plausible. Its existence relates to the commonly held presumption that metastasis begins only when primary cancer cells have evolved to a state of high-grade malignancy. But in fact, this notion rests on very little experimental evidence. An alternative notion is that metastatic dissemination begins relatively early in the multi-step development of the primary tumor, long before malignant progression has reached its endpoint in this tumor. Cancer cells disseminating from such early tumors may enter into the general circulation because of defective, highly permeable, tumor-associated vasculature, and may end up being trapped subsequently in a variety of distantly located capillary beds. Importantly, these cells may not yet have acquired the full array of traits that are needed in order to enable autonomous, tumorigenic growth to proceed. This would suggest that premalignant cells are dispersed and deposited in significant numbers in a variety of tissue sites. Since such cells are incapable of truly autonomous tumorigenic growth, they are likely to remain as small, even single-cell deposits, unable to proliferate in their sites of dissemination, yet surviving for extended periods of time for reasons that are unclear at present. In general, micrometastatic deposits have been viewed as the potential precursors of macroscopic metastases, but in this case, it seems plausible that these cells are predestined to be dead ends, unable to ever acquire the additional cell-associated phenotypes that are needed to generate macroscopic metastases. This notion, if validated, would cast new light on the process of metastasis, in that it would suggest an interesting notion: that the qualities of the disseminated cells (as indicated by their genotypes and their phenotypes) may vary during the course of the formation of a primary tumor, reflecting and paralleling the progressively changing tumor cells within the primary tumor; as such a primary tumor evolves, it may shed cells with increasingly higher degrees of neoplastic growth. Accordingly, it is only the cancer cells that are shed relatively late in primary tumor progression that have the possibility of eventually spawning macroscopic metastases. Taken together, these arguments suggest a variety of mechanistic reasons why cancer cells remain dormant, and in addition indicate that dormant micrometastatic cells may not threaten the lives of those carrying them, if only because many of these cells have virtually no possibility of ever evolving into aggressive growths.