Cancer has been primarily considered to be a consequence of disorders in the tumor cells themselves, resulting in unrestrained cell growth and invasion. Mounting evidence in the past few years indicated that cancers might be the result of complex disorders in the interaction between the corresponding cell type with its microenvironment. Disorders in the microenvironment are viewed as a major cause for the breakdown of tissue homeostasis and tumor development. Cancer can therefore be considered as a “disorganized tissue,” with the cancer stem cells at the top of the hierarchy of fairly heterogeneous tumor tissues. This concept, first demonstrated in human acute myeloid leukemia, has attracted much attention among basic and clinical researchers.

Cancer stem cells are perceived, in analogy to normal stem cells, as the real culprit in tumor development, progression, and relapse. Several studies have implicated that cancer stem cells are responsible for metastasis and for tumor recurrence after initial tumor control by chemotherapy. Similarly to their normal counterparts, cancer stem cells are resistant to conventional treatment because of their slow proliferation rate. They have preserved the ability to self-renew and to generate large populations of more differentiated and heterogeneous descendants. In contrast to the strictly regulated hierarchical organization in normal tissue homeostasis, genetic instability and phenotypic plasticity allow cancer cells to dynamically enter and exit from stem-cell states. Despite all the controversies concerning their identification, the concept of cancer stem cells has provided a better understanding of intra-tumoral heterogeneity, tumor dormancy, and their propensity to develop resistance and recurrence.

Thus far, cancer stem cells, or tumor-initiating cells, have been defined by certain surface markers, and by their ability to engraft and form tumors when a relative low number of cells are implanted into immune-deficient mouse models. Some studies have reported putative cancer stem cells in established cancer cell lines in vitro. The significance and biologic relevance of such results obtained in long-term cultured cell populations with inherent genetic and phenotypic instability have, however, been severely challenged.

The first evidence for the existence of cancer stem cells was derived from hematological malignancies. Based on cell-surface marker expressions characteristic for normal hematopoietic stem cells, the same constellation has been applied to enrich leukemia-initiating cells. A small subset of such slowly dividing cells derived from patients with acute myeloid leukemia was able to induce leukemia in xenotransplant models, which are still considered to be the ultimate proof for the concept of leukemia stem cells. Major problems have been encountered in attempting to translate this knowledge into the clinic, though, as the prospective identification of leukemia stem cells, their efficient separation, and the definition of their biologic properties still constitute major challenges. Moreover, the leukemia stem cells have recently been shown to be fairly heterogeneous. The clinical relevance of leukemia stem cell research, the role of cancer stem cells in development of refractoriness, and novel strategies to overcome this resistance is reviewed by Buss and Ho, pp. 2328.

Putative cancer stem cells have been identified in several other solid tumors based on the expression of some typical stem cell surface markers and their growth potential following isolation. Conceptually and in analogy to normal tissues, the number of cancer stem cells should be small within a given tumor cell population. Genetic modifications and interactions with the tumor microenvironment may affect the number of cancer stem cells that have accumulated. So far no markers, single or combined, could be defined unequivocally to specifically identify cancer stem cells in solid tumors and to localize them within their particular microenvironment.

Recognition of cancer stem cells is further complicated by recent observations demonstrating phenotypic plasticity in cancer cell populations both in culture and in tumors by which differentiated tumor cells may acquire stem cell traits and thus increase the stem cell pool. Characteristics of cancer stem cells and their phenotypic plasticity with, for example, acquisition of mesenchymal properties by epithelial cells, the so-called epithelial-mesenchymal transition (EMT), as well as the potential role of such transitions in cancer stem cell behavior are reviewed by Scheel and Weinberg, pp. 2310.

The niche is the environment in which stem cells reside and is responsible for maintaining the self-renewal capacity and an undifferentiated state, as exemplified by the bone marrow for hematopoietic stem cells. Identification of the mechanisms and signaling processes between tumor and stromal cells within the niche has been complicated by the fact that stem cells represent such a rare population and are difficult to identify in vivo. As the transcription factors Oct4, Nanog, and Sox2 are essential regulators for maintaining stemness of embryonic stem cells, it has been suggested that they might also play a role in maintenance of cancer stem cells.

As reviewed in detail by Cabarcas, Mathews and Farrer, pp. 2315, evidence is emerging that the factors responsible for maintaining the cancer stem cell also utilize the same cell-signaling pathways traditionally used to maintain processes such as inflammation, EMT, hypoxia, and angiogenesis.

Finally, Sun and Wang, pp. 2337, reported in an original paper that c-Met could serve as a novel marker for head neck squamous cell carcinoma cells (HNSCC) recovered from heterotransplants of fresh tumor specimens, and that the c-Met+ HNSCC population had cancer stem cell capacities and were highly chemoresistant and metastatic.

We are very grateful to all the authors for their excellent contributions.