|Home | About | Journals | Submit | Contact Us | Français|
In real estate and in cancer, location and neighborhood are of prime importance. There is considerable evidence that the signals derived from the tumor microenvironment have a major impact on the growth of tumors and determine their clinical behavior1. In some instances, these signals have been shown to dominate over the genotype and oncogene activation in tumor cells1. Infiltration by cells of myeloid origin is a common feature of most human tumors, and has been extensively studied in the context of solid tumors2. In myeloma, bone marrow is the dominant site of tumor growth. Nearly ten years ago, some groups demonstrated that tumor lesions in myeloma were extensively infiltrated with cells with dendritic morphology expressing myeloid markers3–5. However the initial finding that these cells were infected with Kaposi sarcoma herpesvirus (KSHV) soon became controversial. Nonetheless, it has become increasingly evident that the interaction between tumor cells and other cells in the bone marrow microenvironment have a major effect on the biology of tumor cells6. While the initial studies utilized bone marrow “stroma” as a loosely defined collection of diverse cell types and non-cellular elements, it is likely that the tumor cells mediate specific interactions with defined cell types, necessitating the need to focus on distinct cell types in the tumor microenvironment. For example, interactions between myeloma tumor cells and osteoclasts promote the survival of myeloma tumor cells in culture.
What about extramedullary plasmacytomas, wherein the tumor cells grow predominantly outside the marrow? The conventional wisdom has been that these cells may be “independent” of signals derived from the microenvironment. In this issue, Peng et al show that plasmacytoma lesions are also infiltrated with CD68+ myeloid cells with a dendritic morphology7. These studies therefore extend the findings from medullary disease to be a general feature of plasma cell tumors. These findings also raise several other questions. Why and how are these cells recruited to tumors? Is their recruitment essential to the growth and survival of these tumors? If the latter is true, it would argue that the extramedullary lesions are not truly independent of stromal influence.
Recent studies have shown that dendritic cells with both myeloid or plasmacytoid phenotype may enhance the clonogenicity and survival of myeloma tumor cells8–10. These interactions may also help protect myeloma cells from cell death induced by certain anti-tumor therapies9. The close proximity between myeloid cells and tumor cells observed on immunohistochemical studies suggests that cell-cell contact between these cell types may be critical for their effects on tumor cells10. It is notable that myeloid cells are often found to contaminate preparations of myeloma plasma cells isolated with magnetic beads, which also argues for close interaction between these cells as a common occurrence11. It is likely that the cross talk between tumor and myeloid cells is bidirectional, and that tumor cells may also have major effects on the properties of infiltrating myeloid cells2.
Two aspects of the biology of tumor associated myeloid cells may be particularly relevant in myeloma. First is their plasticity12. For example, it has been demonstrated that such cells may contribute to tumor associated angiogenesis13,14. As osteoclasts are typically derived from myeloid progenitors, it is also possible that the infiltrating myeloid cells may also serve as direct precursors for myeloma associated osteoclasts. Another implication of this plasticity is that these cells may serve as a two edged sword. For example, while under certain conditions these cells may promote tumor growth, under other conditions, they may be important in mediating innate immune surveillance15.
Another interesting property of myeloid cells is their capacity to mediate cell fusion events16. Indeed, cell fusion events underlie the formation of osteoclasts from myeloid progenitors under physiologic settings. The potential ability of tumors to usurp these mechanisms may allow them to fundamentally and potentially genetically alter their microenvironment. Improved understanding of the biology, heterogeneity, and functional properties and plasticity of myeloid cells that infiltrate myeloma lesions therefore may have major implications for understanding pathogenesis of these tumors. It is likely that single markers will not be sufficient to define the phenotypic and functional heterogeneity of these cells. Development of new tools / markers that allow specific multiparameter identification of these cells in tissue sections is therefore needed.
If tumors indeed need to actively recruit myeloid cells, can this property be exploited therapeutically? It is increasingly recognized that selective targeting of anticancer agents to tumor cells may enhance their therapeutic potential. Some examples already in the clinic are radiolabeled antibodies and immunotoxin conjugates used to treat hematologic malignancies. The property of tumor cells to attract myeloid cells has also led to consideration of their use as cellular vectors for targeting tumors17. The capacity of myeloid cells to be recruited to and survive in hypoxic regions of tumors make them particularly appealing as cell based vectors. As another example of this approach, Peng et al demonstrate that myeloid cells may serve as efficient vectors for the delivery of measles virus to myeloma cells in xenograft models7. This approach could in principle be extended to other anti-myeloma agents or biologicals as well, so long as they are not directly toxic to myeloid cells. In other words, the very property of myeloma cells to attract infiltration by myeloid cells to facilitate their survival may provide the basis for novel approaches to myeloma therapeutics. The very cell types that may help myeloma cells survive may be their underbelly !!
The author reports no conflict of interest. MVD is supported by funds from NIH.