SEARCH

SEARCH BY CITATION

Keywords:

  • cancer stem cells;
  • extracellular matrix;
  • metastasis;
  • microenvironment;
  • niche

Abstract

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Metastasis is an inefficient process and most cancer cells fail to colonize secondary sites. There are several possible reasons for this. First, the nature of the infiltrating cells is important as a small population of cancer stem cells has been shown to have exclusive metastasis-initiating potential. Secondly, supportive niches are required to promote the outgrowth of disseminated tumour cells. Such niches are either produced prior to the arrival of cancer cells in the target organ or are induced ad hoc upon cell infiltration. Components of the extracellular matrix (ECM) have been found to play a role in establishing these niches. This has highlighted the importance of the ECM for metastatic progression, and suggests that such components may provide alternative targets for treatment of metastatic disease.


Hierarchical organization in cancer

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Homeostasis is achieved in many tissues in the adult organism through a hierarchical organization with a small population of tissue-specific stem cells at the apex. Stem cells possess the ability for long-term self-renewal giving rise to cycling (transit-amplifying) cells, which in turn generate terminally differentiated cells [1]. Whilst these cells fulfil the essential functions of the tissue, they are often short-lived and need to be replaced by a constant and finely adjusted supply from the multipotent stem cell population. The identification of cancer stem cells (CSCs), i.e. the discovery of organizational hierarchy in tumours, has fundamentally changed our understanding of cancer biology. Similar to their normal tissue stem cell counterparts, this small subset of cells has the unique ability to initiate and perpetuate tumour growth and heterogeneity in serial transplantation experiments. For some tissues, it has been shown that CSCs can be derived directly from normal tissue-specific stem cells or from early progeny which regain essential stem cell properties [2, 3]. It is now well accepted that future therapy will have to target this essential CSC subpopulation and that we must start to elucidate the detailed biological programme that maintains CSCs and ensures their function. It is noteworthy that tissue stem cells are associated with specialized niches, which provide a defined cellular and molecular microenvironment to control stem cell self-renewal and differentiation [4, 5]. This has led to the idea that there are similar interactions between CSCs and their stromal niche. As discussed below, such interactions are indeed observed and participate not only in primary tumour formation but also have an important role in metastatic colonization.

Programmes defining ‘stemness’ of stem cells

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

The unique transcriptional networks and signalling pathways of tissue-specific stem cells and their embryonic counterparts, embryonic stem cells (ESCs), which enable them to self-renew and to give rise to multiple cell lineages, provide the foundation of ‘stemness’. It is expected, although not yet proven, that similar programmes are also active in CSCs. Several studies have been conducted to define a core transcriptional programme of stemness, but only a few individual genes were found that were shared amongst different stem cell populations. The inability to describe a unifying consensus signature suggests that, on a molecular or per-gene level, distinct mechanisms allow the self-renewal state to be perpetuated. However, a higher-order systems level analysis, which took into account groups of genes identified by multiple independent observations, has revealed that ESCs and adult stem cells can be organized into two main groups [6, 7]. Some tissue-specific stem cells, such as retinal and a subset of neuronal stem cells, share a core transcriptional programme with ESCs. By contrast, the majority of adult tissue-specific stem cells, including neural crest, skin, bone marrow haematopoietic stem cells (HSCs) and mammary stem cells share a distinct transcriptional programme. It is interesting that many epithelial CSCs demonstrate gene expression patterns similar to those of ESCs, indicating reactivation of embryonic programmes. By contrast, some acute myeloid leukaemia (AML) stem cells utilize a core programme which shows remarkable overlap with that of normal HSCc [8]. As leukaemia stem cells showed a much more pronounced overlap with ESC signatures in a different tumour model [9], it appears that separate mechanisms can maintain stemness depending on the cellular origin and oncogenic driver mutations in a particular tumour.

The understanding of these basic mechanisms has been further advanced by the recent discovery that the differentiated phenotype of cells in the adult organism can be manipulated to revert to an undifferentiated state. Such reprogramming of somatic cells into induced pluripotent stem cells is possible by overexpression of a combination of four transcription factors, Oct4, Sox2, Klf4 and c-Myc, or a subset thereof [10]. However, the low efficiency of this process provides a clear indication for the existence of epigenetic barriers, i.e. histone modification, DNA methylation and chromatin packaging patterns, which preserve the identity of differentiated cells. In somatic cells, alternative differentiation programmes and pluripotency factors are characterized by an inaccessible repressive chromatin environment of their regulatory regions that prevents transcriptional activation [11]. Therefore, even after reactivation of a pluripotency network, the epigenetic memory of the differentiated state of the cell still needs to be erased. This requires silencing of tissue-specific genes that were expressed previously and the re-establishment of a permissive chromatin structure at differentiation genes of alternative cell fates. In this respect it is interesting to note that cancer cells typically show a low level of CpG methylation compared with their normal tissue counterparts [12, 13].

From the core pathways involved in reprogramming and ESC maintenance, several key factors have been implicated in controlling CSC biology. In addition to c-Myc, which is a well-described proto-oncogene, Oct4 was recently shown to control the stem cell phenotype in a variety of solid tumours which spontaneously develop upon p53 ablation [14]. Sox2 is another example of a reprogramming factor that has been implicated in the CSC programme [15]. Similarly, a key transcription factor for ESC stemness, Nanog, has been functionally implicated in the maintenance of CSCs in several epithelial cancers [16, 17]. It appears that switching back to this embryonic programme plays a major role in the CSC phenotype and may enable to specifically target this population in the future.

Determinants of metastatic spread

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Whilst it is known that metastatic spread and impairment of multiple organ functions is the main cause of death by cancer, the exact mechanisms that enable cancer cells to grow at a distant site are still only partially understood. It is clear that the organ distribution of metastasis follows characteristic patterns for a given primary tumour, a phenomenon termed organ tropism [18]. Whereas for example breast cancer preferentially metastasizes to the lymph nodes, bone, liver, lung and brain (in order of decreasing preference), other primary cancers show different patterns of metastasis (e.g. non small-cell lung and kidney cancer preferentially spread to brain and lung respectively). Two main theories have been proposed to account for the organ tropism of cancer metastasis: the ‘seed and soil’ theory by Paget in 1889,[19] and Ewing's ‘mechanical arrest’ theory in 1928 [20]. Paget compared the spreading of distant metastases from a primary tumour with the distribution of seeds by a plant; seeds are carried in all directions by the wind, but can only germinate if they fall on receptive soil. The mechanical hypothesis of metastatic spread suggests that patterns of metastasis are determined exclusively by lymphatic and venous drainage and that cells are mechanically arrested in the first capillary bed they encounter.

The dominant pattern of gastrointestinal tumours metastasizing to the liver can be explained by the direct connection between these organs and the liver via the portal vein. However, prostate primary tumours metastasizing to the spine and kidney primary tumours preferentially metastasizing to lung tissues are difficult to explain by the mechanical theory of drainage patterns, and suggest differential tissue compatibility and active homing mechanisms as major drivers of the metastatic process. Detailed analyses of autopsy studies provide some estimation of the relevance of these two theories not only by investigating the incidence of metastatic lesions but also by taking into account haemodynamic patterns between primary and secondary sites [21]. Findings from such analyses have revealed that on average about two thirds of the secondary sites from all types of primary tumours can be explained by drainage pattern. Of interest, in support of Paget's theory, a considerable number of ‘hostile’ sites were identified that should contain metastases based on haemodynamic patterns, but were clearly under-represented. For example, metastases to the skin, which are usually clinically easy to detect, are comparatively rare for cervical or prostate cancers. Overall, it seems that both mechanisms can apply, but are not mutually exclusive. Cancer cells arrive in secondary target sites either by passive transport based on the drainage pattern of the primary site or by specific attractive signals. Once extravasated, they must either encounter or induce a suitable microenvironment for initiation and maintenance of secondary tumour growth.

The situation is further complicated by the fact that cascades of metastases have been frequently observed for certain cancers [22]. For example, for primary carcinomas of the colon or pancreas, first metastases often occur in the liver. From there a second generation of metastases spreads to the lung and subsequently via the arterial system to further sites. The involvement of such cascades in the formation of metastatic patterns has been recently confirmed by detailed DNA sequencing data to define genomic rearrangements in primary and secondary tumours of individual patients [23]. These data revealed clonal relationships amongst metastases, which confirm metachronous seeding of target organs as a major mechanism. This process of metastasis from metastases typically involves parallel evolution of cancer cells at the secondary site independent of the genetic events at the primary site.

The requirement of a supportive microenvironment becomes evident when the overall efficiency of the metastatic process is evaluated [24]. Cancer cells that begin to metastasize must overcome several obstacles before colonizing a new tissue: they have to leave the primary tumour, enter the circulation, extravasate, establish themselves as a new colony of growing cells to initiate a micrometastasis and finally induce neoangiogenesis to progress to a clinically detectable macrometastasis. In experimental mouse models, less than 0.1% of initially seeded tumour cells can successfully complete the entire metastatic process and form macroscopic metastases. Of importance, not all stages are equally inefficient, but rather the main selection occurs during the colonization of the secondary target site when tumour cells need to initiate growth. By contrast, most tumour cells appear to be capable of survival in the circulation and to extravasate into the secondary site. Similar observations have been made in patients [25], which further emphasizes the role of the stromal microenvironment at the secondary site in controlling metastatic spread.

Local microenvironments and stem cell niches

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Local microenvironments play an important role in regulating cell behaviour during embryonic development by allowing stable compartments with groups of interacting cells to be formed. Within such cellular communities, the exchange of signals either via direct cell–cell communication or via the deposited extracellular matrix (ECM) promotes the control of various aspects of the behaviour of cells including their proliferation, phenotypic differentiation and migration. It has become increasingly evident that similar principles also apply to cancer biology where a large variety of cell types are known to interact [26]. In addition to the cellular components of such a microenvironment, noncellular constituents and especially the ECM are now known to be important contributors to cancer progression.

The ECM and its structural and supportive role in maintaining tissue morphology have been well characterized. The composition of the ECM is surprisingly dynamic and can adapt to modulate cellular biology. Multiple mechanisms affect the regulation of the ECM and control its production, degradation and three-dimensional remodelling. The complex composition of the ECM with many biochemically distinct elements including proteins, glycoproteins, proteoglycans and polysaccharides is suitable for such highly dynamic behaviour. In addition, the arrangement of ECM components can be dynamically adapted with individual components cross-linked together via covalent and non covalent modification. However, this complexity cannot be fully determined with the currently available experimental tools.

Structurally, ECM components constitute both the basement membrane and the interstitial matrix. The basement membrane is produced cooperatively by epithelial and stromal cells to separate the epithelium from the underlying stroma, whereas the interstitial matrix is mainly generated by stromal cells. These two matrices differ principally in their porosity, rigidity and topography (spatial arrangement and orientation), with the basement membrane being more compact and less porous than the interstitial matrix. This is due to the specialized composition of type IV collagen, laminins, fibronectin and linker proteins, such as nidogen and entactin, that connect collagens with other protein components in the basement membrane. By contrast, the interstitial matrix is rich in fibrillar collagens, proteoglycans and various glycoproteins such as tenascin C (TNC) and fibronectin [27].

Furthermore, the ECM represents a highly glycosylated and charged protein network that can bind to many secreted growth factors, including the transforming growth factor (TGF) β family, bone morphogenetic proteins (BMPs), Wnts, hedgehogs, epidermal growth factors and fibroblast growth factors (FGFs) [26]. Binding to the ECM restricts the distribution of these factors in the tissue and can either act as sink to prevent growth factor action until release upon remodelling of the matrix or promote their signalling activity by creating highly enriched and localized sources. The ECM is also a source of biologically active signalling fragments, which are generated upon matrix degradation by proteolytic cleavage and modulate tumour cell apoptosis and angiogenesis [28]. Some ECM components have been shown to contribute to adult stem cell niches, for example osteopontin, TNC and biglycan, and to control the size of the stem cell pool in different systems [29-31].

ECM-mediated anchorage to a local niche allows stem cells to stay in contact with specialized niche cells, which produce paracrine signalling molecules that are essential for maintaining stem cell properties. Furthermore, ECM components can directly regulate stem cell fate by modulating signalling pathways that play an important role in stem cell biology; for example, TNC can affect Notch, Wnt, FGF and BMP signalling [30, 32]. Anchorage also enables stem cells to establish cell polarity, which is a prerequisite for undergoing asymmetric cell division. This division mode can control the fate of two daughter cells by selective inheritance of cytoplasmic components or distinct localization of the newly generated daughter cells relative to the niche [33]. Changes in cell polarity or in the machinery that controls asymmetric division can lead to the failure of stem cells to differentiate. For example, loss of function of Musashi, Numb or Lgl1, components involved in regulating cell polarity, has been implicated in excessive stem cell expansion and can direct brain tumour and leukaemia formation [34, 35]. Therefore, because of the essential roles of the ECM in the normal stem cell niche, it is likely to participate in creating a CSC niche during cellular transformation.

The metastatic niche

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Organ tropism of metastasis involves interactions between infiltrating cancer cells and the local microenvironment. It has been hypothesized that the metastatic cells rely on extrinsic signals from a supportive microenvironment to establish themselves as new colonies at a distant site. Three independent groups recently provided evidence that support this concept [32, 36, 37]. To elucidate the niche signals that support cancer cells to successfully colonize distant organs, we studied the genetic profiles of niche regions adjacent to normal tissue stem cells and found that periostin (POSTN) was highly expressed not only in normal stem cell niches but also in the stroma of the primary tumour and in newly forming metastases [36]. Periostin is a secreted matricellular glycoprotein of the ECM which is highly conserved in vertebrates and belongs to a fasciclin domain-containing family of molecules. Whilst matricellular proteins are highly expressed during development [38], periostin expression is downregulated in the adult, except in mesenchymal niches in close contact with tissue-specific stem cells. Of note, in the stromal compartment of breast tumours (both human and mouse), periostin is widely expressed by αSMA+VIM+ fibroblasts [36]. It is also deposited in secondary lesions in the lungs and in many, but not all, human lymph node metastasis from breast cancer patients. These results are of particular interest because the presence of cancer cells in lymph nodes is currently used in the clinic to determine metastatic spread in breast cancer. Therefore, the detection of periostin in lymph nodes could be used as a potential indicator of distant metastasis, whilst its absence may denote the presence of transient cells without much metastatic power. However, this hypothesis remains to be rigorously tested.

To determine the effect of periostin on tumour development and metastasis, we generated a knock-out mouse for periostin and crossed it with the MMTV-PyMT breast cancer model. We found that although the primary tumour was not affected, these mice demonstrated a dramatic reduction in the number of metastases. Moreover, when knock-out tumour cells were injected into wild-type or knock-out hosts in an experimental lung metastasis assay, we again observed a reduction in metastases in the knock-out recipients. These data indicate that periostin secreted by activated lung fibroblasts in response to metastatic cancer cells plays a crucial role in the colonization phase, creating the right environment for progression to secondary tumours. We further showed that the mechanism by which periostin acts is through recruitment and presentation of Wnt ligands to the tumour cells, thus enhancing Wnt signalling in CSCs, which are the only cells in these tumours that are responsive to these ligands (Fig. 1). The Wnt pathway, known to have a major role in development and cancer, is particularly important for the maintenance of stem cells in both normal and cancer tissues [39, 40].

image

Figure 1. The secondary site provides a supportive niche that stimulates Wnt and Notch signalling pathways in infiltrating cancer stem cells (CSCs) to increase viability and stem cell expansion. Two ECM components play a major role in that niche: periostin (POSTN), which accumulates and presents Wnt ligands to the Wnt receptors Lrp and Frz; and tenascin C (TNC), which promotes Wnt and Notch signalling intracellularly.

Download figure to PowerPoint

The role in metastatic colonization of another ECM molecule, TNC, was recently demonstrated by Oskarsson et al. [32]. The authors showed that breast cancer cells infiltrating the lungs produced and secreted TNC, and that this protein was required for successful metastasization. Tumour cells produce TNC when they first reach the lungs and form micrometastases but, as they grow, the production of TNC switches to activated fibroblasts (myofibroblasts) in the stromal compartment. Hence, it is the autocrine production of TNC that initially drives the outgrowth of micrometastases. TNC acts through enhancement of Wnt and Notch signalling in tumour cells, which eventually strengthens the fitness of these cells giving them a survival advantage. O'Connell et al. showed that S100A4+ lung fibroblasts, which proliferate in response to the presence of infiltrating tumour cells, secrete both TNC and vascular endothelial growth factor A (VEGF-A) to support metastatic colonization [37]. Depletion of this fibroblast population reduces TNC levels and impairs metastatic growth. Together, these data underscore the importance of the role that the microenvironment plays during the metastasization, in which some tumour cells are able to instruct the host tissue fibroblasts to secrete ECM components required for their establishment.

Both periostin and TNC have been identified in patients as markers of cancer progression and found to correlate with a higher risk of metastasis or to predict poor prognosis in a variety of cancers (Table 1). However, it is noteworthy that depletion of either of these ECM components in mouse models does not affect tumour growth, but only establishment of metastases. One of the various possible reasons for this is that the microenvironment at the site of origin is better adapted to support tissue-specific stem cells and CSCs of the autochthonous tissue (mammary gland) and is therefore richer and more complex. This complexity, which comprises different supportive proteins may more easily compensate for the loss of a single component. By contrast, it is likely that the secondary site can be optimized to support local (lung) stem cell populations, which may involve different molecular cues. Thus, inhibition of a single component that is required to mimic the original niche with which the disseminated cancer cells were familiar can have dramatic effects on the ability of infiltrating cancer cells to survive.

Table 1. Cancer types in which the ECM proteins periostin and tenascin C have been associated with poor prognosis
ProteinType of CancerReferences
PeriostinNon small-cell lung, breast, gastric, colorectal, prostate, endometrioid ovarian, epithelial ovarian, nasopharyngeal and papillary thyroid cancers, malignant pleural mesothelioma and cholangiocarcinoma [61-72]
Tenascin CBreast, colorectal, gastric and bladder cancers, malignant pleural mesothelioma, astrocytoma, adipocytic tumours and clear cell renal cell carcinoma [73-83]

CSCs as the source of metastases

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

As discussed previously, intrinsic characteristics of individual tumour cells and their ability to adapt to new environments partially explain their potential to metastasize with respect to other cancer cells in the tumour. Revival of the concept of CSCs [41, 42] as the driving force of tumorigenesis has led to the hypothesis that CSCs could be responsible not only for tumour growth and recurrence but also for metastasis. We recently demonstrated that a population of CSCs – identified as LinCD24+CD90+ – exclusively retains the ability to metastasize in in vivo experimental settings, thus providing proof of concept evidence for this hypothesis [36]. Similar results have been obtained for other epithelial cancers [43].

We also monitored the relative amounts of CSCs during the process of metastatic colonization, and observed a 10-fold increase in this population during the first 2 weeks after the cells had infiltrated the new target site followed by a subsequent decrease over the next 2 weeks [36]. This indicates that whilst CSCs are essential to initiate colonization, local mechanisms must be involved in established metastases which restricts the size of the CSC population. At the secondary site, periostin produced by lung fibroblasts serves to maintain the stemness of these cells through the activation of Wnt signalling. It is possible that localization of such niche signals within the supportive stroma becomes rate limiting for the maintenance and expansion of CSCs (Fig. 2). As cells lose direct contact with the stroma in emerging metastases, CSCs fail to maintain their stem cell state and start to (aberrantly) differentiate. Also, TNC was found to be upregulated in breast cancer tumourspheres, which are enriched with CSCs compared to classical attached cell cultures, and this correlated with the expression of stemness genes [32].

image

Figure 2. The proposed presence of a supportive metastatic niche may explain the transient changes in the levels of cancer stem cells (CSCs) during metastatic colonization. By enriching and restricting stem cell-promoting signals to the stromal niche, direct contact between stem cells and the niche becomes rate limiting for the expansion of CSCs. As cells leave the niche, the stem cell state can no longer be maintained. With increasing expansion of the metastatic colony, the relative amounts of CSCs will therefore decline.

Download figure to PowerPoint

The cancer cell niche accounts for the summary of the tumour's cell tolerances and it is conceptually defined by the boundaries within which the tumour can live, grow and spread/seed. Given the requirements of cancer cells in terms of nutrients and O2 to divide and ensure secondary tumour growth, the niche for disseminated cancer cells involves their capacity to activate the angiogenic switch. The difference between a micrometastasis and a clinically detectable macrometastasis is related to the capacity of the cancer cells to modulate the environment in such a way that new blood vessels sprout from preexisting vessels [44, 45]. Periostin, TNC and osteopontin have been previously linked to the induction of angiogenesis in different systems [46-49], which indicates that these ECM proteins – and probably others – are fundamental components of the niche. These ECM proteins are able to regulate VEGF and its receptors and induce angiogenesis in endothelial cells. Thus, ECM molecules supply the necessary resources for successful metastatic colonization and secondary tumour growth by: (i) allowing cancer cells to recreate a similar environment to that of the primary tumour in a distant site, (ii) providing CSCs with factors required for their maintenance and fitness and (iii) providing a nutrient supply through the induction of the angiogenic switch.

The premetastatic niche

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

To evaluate the seed and soil hypothesis, whether or not the correct microenvironment in distant organs is prepared even before the arrival of cancer cells has been investigated. Kaplan et al. described the presence of a population of VEGFR1+VLA-4+ bone marrow-derived cells (BMDCs) in future metastatic sites before they were colonized by melanoma and Lewis lung carcinoma cells [50]. The authors termed these areas the premetastatic niche. Homing of these VEGFR1+VLA4+ cells was mediated via the induction of the ECM protein fibronectin, which is a ligand of VLA-4. Fibronectin therefore supports the assembly of these premetastatic BMDC clusters. In fact, it was later reported that tumour cells secrete lysyl oxidase (LOX), an amine oxidase that plays a role in crosslinking collagens and elastins in the ECM, to provoke systemic alterations and induce the formation of the premetastatic niche [51]. Under hypoxic conditions, breast cancer tumours secrete LOX, which accumulates in premetastatic sites. This favours the recruitment of CD11b+ myeloid cells that adhere to cross-linked collagen IV and produce matrix metalloproteinase 2, which in turn cleaves collagen and makes it easier for BMDCs and tumour cells to invade the area. In addition, other factors may play a role in the formation of the premetastatic niche. For instance, cancer cells secrete factors such as interleukins 6 and 10 that activate the S1PR1–STAT3 pathway in myeloid cells. This in turn promotes activation of fibroblasts and the upregulation of premetastatic niche molecules including fibronectin [52]. Targeting the pro-invasive S1PR1–STAT3 pathway in Cd11b+ myeloid cells eliminates de novo formation of premetastatic niches and metastasis, and reduces preformed metastatic niches. In addition, the expression of tissue factor (also known as coagulation factor III or CD142) by tumour cells induces the formation in the future metastatic sites of platelet clots, which recruit myeloid cells [53]. This myeloid population, predominantly composed of CD11b+ cells differentiated along the macrophage lineage, can enhance the survival of metastatic melanoma cells through the formation of platelet clots. In models of impaired function of monocytes/macrophages, tumour cells still induce the formation of clots, but tumour cell survival is decreased, indicating that myeloid cells are required for tumour cell survival in the premetastatic niche. Moreover, the increase in CD11b+ cells associated with the formation of the premetastatic niche could be responsible for the reduced antitumour activity of natural killer cells at this site, and therefore would indirectly contribute to enhance the survival of metastatic cells [54].

Other factors secreted by tumour cells that affect the formation of the premetastatic niche include VEGF-A, TFGβ and tumour necrosis factor α. These factors induce the expression of chemoattractants such as S100A8 and S100A9 by myeloid and lung endothelial cells, and promote the homing of tumour cells to the premetastatic sites as well as the invasion of circulating tumour cells through a p38-mediated activation of invadopodia [55]. To deliver factors to both local and distant microenvironments, the cancer cell may use exosomes, which are small secreted vesicles derived from the endocytic pathway. It has recently been shown that melanoma cells use exosomes to deliver signals that prime the future metastatic sites [56, 57]. These exosomes, which are regulated by Rab27a, instruct BMDCs to contribute to tumour growth and metastatic colonization through the transfer of different molecules such as the Met oncoprotein [57].

Granot et al. recently showed that, in addition to these largely supportive events, the premetastatic niche in the lungs of different breast cancer models also demonstrates considerable accumulation of Cd11b+Ly-6G+ tumour-entrained neutrophils that inhibit metastatic colonization through the production of H2O2 [58]. Regardless of the fact that the reported differences could in part be attributable to the use of different models, these data demonstrate that immunosurveillance of innate immune cells contributes to the inefficiency of the metastatic process; however, this protection eventually collapses resulting in metastatic growth.

Conclusions

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References

Why is targeting the microenvironment emerging as a promising therapy for metastatic disease? One of the major barriers to effective cancer therapy is the heterogeneity of the disease itself. New single-cell sequencing techniques and genome-wide analyses have confirmed the presence of clonal diversity and the heterogeneity of cancers [59, 60]. Furthermore, tumours are not static, but mixed entities in constant evolution. Because of (i) the presence of drug- and radiation-resistant CSCs, (ii) the outgrowth of different – often more aggressive – tumour populations resulting from the selection pressure by different chemotherapeutic treatments and (iii) the relative lack of specificity of current agents, cancer treatment remains a challenge. Thus, targeting the host microenvironment, a less dynamic and probably more predictable system, may provide better options for interfering with cancer progression. In particular, modulating the establishment of secondary tumours, which appear to rely on stromal support for their growth, holds great promise for future therapies.

References

  1. Top of page
  2. Abstract
  3. Hierarchical organization in cancer
  4. Programmes defining ‘stemness’ of stem cells
  5. Determinants of metastatic spread
  6. Local microenvironments and stem cell niches
  7. The metastatic niche
  8. CSCs as the source of metastases
  9. The premetastatic niche
  10. Conclusions
  11. Conflict of interest statement
  12. References