It seems that the abundance of special information about mechanisms of tumor progression and the role of various molecules in metastasis obtained with different model systems of human blastomas rather prevents than promotes understanding of carcinogenesis and especially the control of tumor growth. There are different approaches to systematizing such information. We think that searching for prototypes of physiological reactions among pathological processes can be a rather promising approach. This approach is now not very popular in the case of metastasis of malignant tumors, which often appears to be a cascade of molecular processes as if created by Nature purposefully to generalize malignancies. But the “physiological approach” allows us not only to remove this apparent uniqueness of processes associated with tumor progression but also to subordinate different mechanisms involved in metastasis; this approach can also reveal yet unknown aspects of this process and pathways to control tumor dissemination.

The recently proposed concept of metastatic niches (Psaila and Lyden 2009) can be very helpful in searching for physiological prototypes of metastasis (Perelmuter and Manskikh 2012). This concept allowed us to quite otherwise elucidate many problems associated with metastasis and explain experimental data that could not be interpreted earlier. Although the concept of metastatic niches still has many blank spots, its development (especially on searching for a probable physiological prototype of the metastatic niche) can be very promising for comprehension of such problem of oncology as organ-preferential localization of metastases. But it must be stipulated beforehand that in the present work, first, it is admitted that metastasis of all, or at least the majority, of carcinomas and melanomas can be described by the concept of metastatic niches (although the available experimental and clinical data concern only a limited range of studied tumors) and, second, mesenchymal tumors will be deliberately not considered because by now about them there are no data necessary for the theory of niches. A clear subordination of the metastasizing stages is also emphasized – in the present paper we shall speak only about processes preceding formation of micrometastases leaving aside the problem of formation of a macroscopic node of a secondary tumor.

Although there is no organ which would be absolutely protected against development of metastases of malignant tumors, such metastases are relatively often developed in a rather limited number of “typical” localizations: regional (with respectively to the primary tumor location); lymphatic nodes (lymphogenous metastases); lungs, liver, bone marrow, and brain (hematogenous metastases); peritoneum and pleura. Much less often hematogenous metastases are found in kidneys, gonads, spleen, subcutaneous fat tissue, and extremely seldom in the walls of the gastrointestinal tract, uterus, heart, and skeletal muscles. It should be noted that in the overwhelming majority of cases metastasis into atypical locations is associated with the generalization of the process affecting many organs and tissues. However, the spleen is an interesting exception. This organ is rarely damaged by macrometastases, except for the cases of generalized tumors (especially melanomas), but it has been shown earlier that micrometastases in the spleen occur rather often, whereas is muscles micrometastases are virtually not found.

There is no doubt that localization of metastases is partially associated with specific features of the lymph and venous blood outflow from the region of the primary tumor location. Just this determines the development of lymphogenous metastases into the regional lymph nodes and of hematogenous metastases of abdominal cavity organ tumors (stomach and pancreas carcinomas, colorectal cancer) into the liver. However, it is impossible to explain the localization of metastases only by specific features of the vascular system responsible for delivery of tumor cells to the site of metastasis. Thus, the bone marrow and liver are usual sites for hematogenous metastasis of kidney tumors, although these organs are not located on the pathway of venous outcome from the liver. Mechanisms responsible for differences in organ vulnerability are intensively discussed in the literature, but there is still no integral concept describing the causes of organ-preferential metastasis.

The presence in blood of circulating tumor cells not always leads to development of macro- and micrometastases in target organs (Alix-Panabieres et al. 2008) and experimental works have shown the absence of a direct and constant correlation between the ability of endothelial cells to constitutively express selectins halting tumor cells and the adhesion of these cells on the endothelium, on one hand, and the sites of preferential localization of metastases, on the other hand (Wong et al. 1997). Even more interesting is the discovery of a phenomenon of “inefficient metastasis” when immigrated tumor elements are present in the target organs but fail to produce metastases (Bidard et al. 2008). These findings clearly suggest that formation and localization of micrometastases are more likely determined not by the presence of tumor elements in the blood flow but rather by some specific features of target organs (including those arising under the influence of the primary tumor) that are responsible for occupation of a suitable organ by blastoma cells and formation from them of a micrometastasis. The same findings also show that the halting of tumor cells in the target organs without some additional conditions is yet insufficient for development in them of metastases. This so-called “seed and soil” hypothesis was proposed by Stephen Paget in the beginning of the last century (Psaila and Lyden 2009), but only recent data filled it with concrete content. Researchers of D. Lyden’s group have established that the development of micrometastases in target organs is preceded by the accumulation in them of cells immigrated from the bone marrow and creating a stromal microenvironment that is adequate for the tumor and determines the development of metastases (Kaplan et al. 2005; Peinado et al. 2008). To describe this process, the concept of “niche” was proposed, borrowed from hematology where it was used for description of microenvironment that regulates the proliferation, homing, and differentiation of stem cells (Wilson and Trumpp 2006).

Lyden’s concept suggests the formation and step-by-step changes in the site of a future metastasis of the following forms of microenvironment: a premetastatic niche with bone marrow precursor cells without tumor elements; a micrometastatic niche characterized by the presence of a cluster of immature bone marrow and tumor cells; a macrometastatic niche with angiogenesis added to the preceding processes (Wels et al. 2008). According to scheme of Peinado et al. (2011), the formation of a micrometastatic niche is determined by several processes:

Macrometastatic formation niche and clinical manifestation of metastases request activation of angiogenesis and migration of bone marrow endothelial precursors to metastatic niches (Kaplan et al. 2005). Besides angiogenesis, a lot of other factors have significance for macrometastatic niche development. Using mouse models of spontaneous breast cancer, Gao et al. (2012) has shown enhanced recruitment of bone marrow–derived CD11b⁺Gr1⁺ myeloid progenitor cells in the premetastatic lungs. Various protumorigenic activities were associated with Gr1⁺ myeloid cells, including expression of proangiogenic factor BV8, metastasis-promoting LOX and MMP9, contribution to TGF-β–mediated metastasis, and immune tolerance. Gene expression profiling revealed that the myeloid cells from metastatic lungs express versican, a large extracellular matrix chondroitin sulfate proteoglycan. Notably, versican in metastatic lungs was mainly contributed by the CD11b⁺Ly6C^high monocytic fraction of the myeloid cells and not the tumor cells or other stromal cells. Versican attenuated the Smad-mediated epithelial-mesenchymal transition (EMT) signaling pathway as determined by the reduction of p-Smad2 levels and suppression of transcription factor Snail. The suppression of Smad2 pathway–induced mesenchymal-epithelial transition (MET) increased cell proliferation. In addition, versican did not impact apoptosis, suggesting that versican-mediated stimulation of MET and enhanced proliferation may be the main mechanisms of increased tumor outgrowth and formation of focal macrometastases. The contribution of fibroblasts and tumor cells to the metastatic lung was about tenfold lower compared with CD11b⁺Ly6C^high cells. Furthermore, no significant increase in fibroblast numbers was observed in the metastatic lungs compared with controls. Analysis of a cohort of patients with breast cancer from whom distal metastases were available showed clusters of myeloid cells expressing versican in the metastatic lungs. Immunohistochemistry and RT-PCR analysis showing enhanced versican expression in metastatic lungs, brain, and liver in patients with breast cancer compared with control normal lungs. Gao et al. (2012) has supposed that selectively targeting tumor-elicited myeloid cells or versican-mediated proliferation pathways, perhaps in combination with conventional chemotherapeutics, may represent a potential therapeutic strategy for combating metastatic disease.

Due to introduction of BMDCs as a new messenger, the concept of metastatic niches allows us to remove the contradictions enumerated in the beginning of this section. This concept also opens great prospects for control of metastasis. However, the metastatic niche theory is still not a completed concept. In particular, it remains unclear what specific events trigger the formation of a metastatic niche, i.e. lead to accumulation of fibronectin and changes in the endothelium favorable for BMDC homing. The formation of a metastatic niche is usually described as a process depending on the primary tumor (Peinado et al. 2011). However, it is well known from experimental oncology that intravenous injection of cells of some tumors can induce development of metastases in internal organs, and in some cases with a rather specific (mono-organ) location. This indicates that metastatic niches can be formed due to processes independent of the development of the primary tumor node, but due to some other physiological or pathophysiological reaction. Note that the theory of metastatic niches describes the formation of metastases in general not considering the question why metastases are formed mainly in particular “typical” sites and why the location of metastases and the type of metastatic disease vary in different patients. The known factors synthesized by the tumor and regulating the development of the niche (LOX, MIP-2, VEGFA, TGFβ, TNFα) are not organ-preferential. Nevertheless, Kaplan et al. (2005) and Hiratsuka et al. (2008) have attempted to study this problem experimentally. Mice with grafted Lewis lung carcinomas were injected with conditioned medium from melanoma B16 cells characterized by generalized non-selective metastasis, and as a result metastatic niches and micrometastases developed in various organs and tissues, including such atypical locations as the oviducts (Psaila and Lyden 2009). However, the molecular mechanisms underlying the phenomenon observed by D. Lyden’s group are still unclear.

We think that the questions presented above might be answered taking into account the events preceding the formation of a premetastatic niche as it is and to suppose that these events should be based on a physiological or pathophysiological process more or less independent of the development of the primary tumor node. The overall in situ conditions that precede the formation of a premetastatic niche (recruiting BMDCs into the target organ) and determine the localization of a future metastasis is reasonable to term as a “preniche”. It is important to note that we think the preniche plays a key role in BMDC homing and also is essential for emigration of tumor cells from the blood flow.

Considering the most frequent locations of metastases, it becomes evident that they have one feature in common: they have a large pool of organ-specific macrophages (Kupffer cells in the liver, alveolar macrophages in lungs and microglia in the brain, peritoneal and bone marrow macrophages, lymph node macrophages). This feature is not characteristic of the heart, gastrointestinal tract organs, skeletal muscles, kidneys, or gonads – and in these organs solitary metastases occur relatively seldom. Obviously, the endothelium of these organs has to be adapted to an active immigration into them of macrophage precursors under both normal and inflammation conditions when the need for restitution of the physiological macrophage pool is especially strong.