“The road to metastasis is paved with tumor-microenvironment interactions”, claimed Dr. Isaac Witz from Tel Aviv University, Israel, in his keynote speech at the first Tianjin Forum on Tumor Microenvironment (http://TFTM.nankai.edu. cn), an international conference held at Nankai University in Tianjin, China, on July 2–4, 2010. About 300 cancer researchers and students attended the conference. Thirty lecturers from North America, Europe and Asia discussed their findings on how the interactions between cancer cells and host normal cells in the microenvironment may determine whether the cancer cells develop into a malignant tumor, metastasize, or remain dormant, even die. Dr. Witz briefly reviewed in a historic perspective how cancer researchers came to realize the importance of the interactions between cancer cells and their microenvironment in the processes of cancer development and metastasis. He pointed out that there are two types of microenvironments that cancer cells reside in: the microenvironment of the primary, local tumor and the microenvironment of secondary, metastatic sites. He emphasized that the interactions between tumor cells and their microenvironment are dynamic and bi-directional and that each interaction partner shapes the phenotype of the other partner. The relationship between cancer cells and their microenvironment is often likened to “seed and soil”. The hypothesis was first put forward by Stephen Paget (Lancet 1:571, 1889). “What is it that decides what organs shall suffer a case of disseminated cancer?” Paget wondered. “When a plant goes to seed, its seeds are carried in all directions,” he wrote. “But they can only live and grow if they fall on congenial soil.” It may be useful to point out the conceptual differences between Paget’s perception and the modern paradigm. According to Paget’s hypothesis the microenvironment either supports metastasis by supplying growth promoting factors or alternatively inhibits metastasis by growth inhibitors. The contemporary paradigm, however, assigns to the tumor microenvironment an inductive, adaptive and selective function: The tumor is directed into one or several possible molecular evolution pathways by signals originating in native or modified microenvironmental factors. Many of these pathways may lead to metastasis. Dr. Robert Kerbel from University of Toronto, another keynote speaker, discussed circumstances by which certain cytotoxic drugs can rapidly alter the tumor microenvironment in such a way as to decrease their efficacy. Specifically, he summarized results showing that bolus injections of certain cytotoxic chemotherapy agents at maximum tolerated doses such as paclitaxel or cyclophosphamide, or microtubule inhibiting ‘vascular disrupting agents’ can cause a rapid mobilization of various cell populations from the bone marrow compartment, which then migrate to and colonize the drug treated tumors. These retained cell populations include endothelial progenitor cells, but likely other cell types as well including monocytes, macrophages, and mesenchymal derived stem cells. These cell populations can promote tumor repopulation, at least in part, by stimulating tumor angiogenesis and vasculogenesis. This acute reactive host response appears to be driven by systemic induction of multiple chemokines and cytokines such as granulocyte colonystimulating factor (G-CSF) and stromal derived factor-1 (SDF-1). The mobilization/tumor colonization response by such bone marrow-derived cells (BMDC) can be blocked, at least in part, by several pharmacologic maneuvers. These include administration of vascular endothelial cell growth factor (VEGF) pathway targeting antibodies (which, incidentally, provides potential explanation about how administration of drugs such as bevacizumab may enhance the efficacy of chemotherapy, i.e., by suppressing such host BMDC responses and thus extending the duration of the tumor response), or by anti-SDF-1 antibodies.