A lot of good points raised but most exhibit a proclivity to the primary tumor "doing things" that subsequently result in metastases - we should also consider the roles played by the pre-cancerous niche and how the microenvironment where the  metastasis occurs facilitates the "seeding" of the cancer- so how would one weight the role of a traveling cancer cell against the role of the environment where the metastases ends up occurring?

No. New research indicates that cancers metastasis via resident cancer stem cells. Most cancer cells that proliferate are not responsible for the spreading of the tumors; this creates barriers for designing therapeutic.

Agree with Micheal. Every deadly cancer evolve to a metastatic state, but only a short amount of cells aquiere such phenotype.

Only a small number of primary tumor cells are capable of successfully forming a metastasis, and these have come to be known as METASTASIS-INITIATING CELLS (MICs) [15] in contrast to TUMOR-INITIATING CELLS (TICs) which are essentially tumorigenic but not necessarily metastasis-capable, with all MICs being a subset of TICs, but not all TICs, or even many of them, being MICs. Thus, data shows that while close to 90% of cells that have escaped the primary tumor are capable of completing all the steps of metastasis up to and including extravasation, only 2% of these cells can develop micro-metastases and even fewer (0.02%) can then, from micrometastatic status, develop into macrometastases (but see fuller discussion below) [13], complete with angiogenesis-supported tumor vasculature to finally, at the tail end of the metastatic cascade, become well-vascularized and clinically-detectable in distal sites [14]. The key steps of the metastatic cascade are briefly summarized below:

THE METASTATIC PROCESS: BRIEF SUMMARY

The multi-step process of metastasis is a complex and coordinated choreography encompassing:

(1) local infiltration of tumor cells into the surrounding/adjacent tissue (tumor cell penetration through the ECM / the basement membrane),

(2) intravasation (endothelial transmigration of tumor cells into vessels),

(3) hematogenous survival and translocation, that is, the tumor cell survival in the circulatory system and its translocation through the bloodstream to microvessels of distant tissues

(4) extravasation (exit from bloodstream, and

(5) adaption to the foreign microenvironment of distant site tissue and subsequent colonization (cell proliferation and the formation of a macroscopic secondary tumor) in competent organs.

Note: Subprocesses (2) and (3) together, that is, the combination of intravasation + hematogenous survival constitute collectively what is known as hematogenous dissemination.

SO WHAT ABOUT CANCER STEM CELLS (CSCs) IN THE METASTATIC PROCESS

One fascinating - and critical - question that this raises is the role of cancer stems cells (CSCs), and whether they are necessary and/or sufficient for metastatic development, which devolves to the question of whether tumoral metastasis-formation is restricted to, and co-extensive with, cancer stem cells (CSC)?

The answer, as I will show below, is No:

CSCs are neither necessary, nor sufficient (as to their insufficiency, this is a point my colleague Fatih Uckun has made elsewhere), for metastasis formation, but I will go further to show that more broadly, neither are they necessary or sufficient for cancer in general, neither for carcinogenesis nor tumorigenesis, and my goal is to put forward a more nuanced appreciation of the difference between metastasis-initiating cells (MICs), and tumor-initiating cells (TICs), on the one hand, and cancer stem cells (CSCs) on the other.

THE BOUNDARIES OF CANCER STEM CELLS

With regard to solid tumors, CSCs for human colon cancer have also been found to consist of functionally distinct clones [1]. Using a molecular "barcode" marking/tracking strategy, Sebastian Dieter and colleagues at the German Cancer research Center (DKFZ) demonstrated that at least three major clones were found in CSCs - and, we note, genetic heterogeneity was not a significant contributor to the functional differences between the three types of CSCs:

(1) a clone that maintained self-renewal activity over the long-term ((LT-TICs) in serial transplants, with extensive self-renewal and metastatic activity,

(2) a clone that lack detectable self-renewal and metastasis-forming potential [called tumor transient amplifying cells or T-TACs] , and

(3) a clone with delayed tumorigenic activity that became apparent in secondary or tertiary recipients [called delayed contributing TICs (DC-TICs)].

Of note, the clone with long-term self-renewal activity disseminated to bone marrow and was alone responsible for metastasis, and indeed only LT-TICs drove tumor initiation, self-renewal, and metastasis formation, and would therefore be the only CSC clone to constitute a quantifiable target for the eradication of self-renewing tumorigenic and metastatic colon cancer cells. These findings thus suggest the existence of clonal diversity that conforms to a hierarchical organization originated from CSCs. Most critically, it strongly suggests that the processes of tumor formation, self-renewal, and metastasis formation are decoupled and disjunctive from each other, indicative of the fact that they must be regulated independently of each other.

These long-term tumor-initiating cells (LT-TICs) with their unique properties of joint self-renewal, and metastasis formation are essentially migrating CSCs, or mCSCs, a concept first introduced by Thomas Brabletz and colleagues [2] at the University of Erlangen (Germany). Migrating CSCs possess both an element of stemness (inclusive therefore of true self-renewal) and mobility as required for metastatic formation. These cells appear to undergo epithelial-mesenchymal transition (EMT) at the invasive front of the primary cancer and then migrate to colonize new tissues, with the production of the cancer heterogeneity commonly observed in metastatic colonies being consequent to its acquired stemness, recognizing that CSCs that have undergone EMT can disseminate, and more critically that such disseminating cancer cells that retain stem-cell functionality can form metastatic colonies. The authors in fact characterized potential migrating CSCs in colorectal cancer as developing from stationary cancer stem cells by combining the two core features of stemness and EMT.

Furthermore precisely those cells that express high levels of nuclear beta-catenin characterize migrating CSCs [3], with high nuclear beta-catenin expression being strongly correlated with metastasis and thus also compromised survival, in line with the understanding that Wnt/beta-catenin signalling, allied with Notch and Hedgehog contribution, is centrally involved in maintaining a stem cell phenotype. Indeed, beta-catenin signalling has been identified as a critical factor in the maintenance of a CSC phenotype in epidermal tumours in general [4], and recent evidence establishes that beta-catenin serves as a marker for cancer stemness and tumour cell clonogenicity [5].

SUMMARY OF CONCLUSIONS

So this leaves us with these conclusions:

(1) The set of all tumor-initiating cells (TICs) contains, but only as a subset, the set of cancer stem cells (CSCs).

(2) Thus not all CSCs are metastasis-formative: as witness, in colorectal cancer (and other solid malignancies), the existence of CSCs which are tumor transient amplifying cells (T-TACs) that fundamentally lack detectable self-renewal and metastasis-forming potential.

(3) That "stemness" in a tumor cell - the essential qualities of a cancer stem cell (CSC) - is an insufficient condition for metastasis-formation, which in addition requires mobility that not all CSCs have; (4) co-existent stemness and mobility define the set of migrating CSCs and are also known and have been extensively studied in the literature as metastasis-initiating cells (MICs), logically a subset of tumor-initiating cells (TICs).

(4) Migrating CSCs (mCSCs), aka metastasis-initiating cells (MICs) should be viewed as not "metastatic CSCs, but rather as metastatic-capable CSCs. Thus Sofia Merajver and colleagues [6] at the University of Michigan Comprehensive Cancer Center recently found that for highly metastatic and normal breast cancer cell lines, RhoC expression - a subgroup of the Rho-GTPase (guanosine triphosphatases) family of proteins that is itself a subfamily of the Ras superfamily - is necessary to induce metastasis and, critically, that RhoC overexpression alone was sufficient to cause metastasis (even inducing metastasis in a non-metastatic mammary epithelial cell line (MCF-10A)), an in vitro finding they also confirmed in vivo in mice.

(5) This has several broad implications: in the landmark research of Anne Hakem and colleagues [7] in Ontario, RhoC was decisively established, in vitro and in vivo, as crucial for and an undeletable component of metastasis, and was fundamental for mobility, since loss of RhoC in the tumors impaired both their motility and invasiveness, true also for pancreatic cancer [8], lung cancer [9], and head and neck cancer [10], among others. (And note of course, as always, that colonization and metastasis formation by CSCs is highly niche-dependent).

(6) In addition, collectively, this suggests that not all metastasis are CSC-induced, that is, that cancer stem cells are not necessary - nor are they sufficient - for metastasis formation and that there are metastasis-initiating cells (MICs) that are not cancer stem cells / CSCs (and this is also suggested powerfully by another line of evidence, namely that not all cancers harbor cancer stem cells (CSCs) [11]).

(7) Finally, we can go one step further: not only is it the case that not all metastases are cancer stem cell-induced and that there are MICs (metastasis-initiating cells) that are not CSCs, but we also have evidence that stem cells are not requisite for carcinogenesis and tumorigenesis: Dinorah Friedmann-Morvinski and colleagues [12] at The Salk Institute for Biological Studies demonstrated that CNS differentiated brain cells which include corticol neurons), can be reprogrammed by oncogenic factors, as for instance by transduction by oncogenic lentiviral vectors of neural stem cells (NSCs), astrocytes, or even mature brain neurons, to become progenitor-like cells that then develop into brain tumors in the form of malignant gliomas. Indeed, most differentiated cells in the CNS can undergo dedifferentiation upon defined genetic alterations, generating an NSC or progenitor state that enables the initiation and maintenance of tumor progression, and this directly challenges the widespread but unrefined conviction that only stem cells can induce cancers of the brain, and that clearly we have here a case in which malignant gliomas - which are in addition highly aggressive - are produced by these terminally differentiated neurons such as mature neurons and astrocytes, both these non-stem cell types being shown to induce gliomas, and that is moreover the case that not all aggressive tumors are associated with CSSs.

So we now know through this breakthrough research that multiple different CNS cell types - not just neural stem cells as originally thought - can be transformed into glioblastoma multiforme (GBM) in an animal model robustly recapitulating key aspects of the human CNS disease. As the lead investigator, Inder Verma, put it starkly in a recent interveiew: "any cell in the brain that gets an oncogenic insult has the ability to dedifferentiate [and form tumors]".

LESSONS LEARNED

It is clear therefore, that for both metastasis-formation and for carcinogenesis / tumorigenesis, cancer stem cells (CSCs) need not (always) apply. This brings us away from the miasma of the cancer stem cell (CSC) hypothesis as somehow an "Ur-process" of all cancer, to a more mature perspective that I have been advocating against the CSC "hyper"-media, that

(1) not all that is significant that needs to be said about metastasis is said by the CSC hypothesis, and

(2) not all that is significant that needs to be said about cancer is said by the CSC hypothesis.

With, to date after decades (40+ years) of CSC research and hundreds and hundreds of published papers, the absence of any methodologically robust human clinical - not just preclinical animal-model level - evidence of CSC as critical to carcinogenesis, tumorigenesis, malignant transformation and metastasis formation - indeed after four decades we even lack any evidence of dedifferentiation as occurring in primary tumors in vivo - the business of unraveling oncogenesis and the abiding mysteries of cancer and metastasis goes on, and remains wider, and significantly beyond, just the scope and boundaries of the cancer stem cell (CSC) hypothesis.

References

1. Dieter SM, Ball CR, Hoffmann CM, et al. Distinct types of tumor-initiating cells form human colon cancer tumors and metastases. Cell Stem Cell 2011;9:357–65.

2. Brabletz T, Jung A, Spaderna S, Hlubek F, Kirchner T. Opinion: migrating cancer stem cells-an integrated concept of malignant tumour progression. Nature Reviews Cancer 2005;5:744–9.

3. Malanchi I, Peinado H, Kassen D, et al. Cutaneous cancer stem cell maintenance is dependent on beta-catenin signalling. Nature 2008; 452: 650–53.

4. Braun KM. Cutaneous cancer stem cells: beta-catenin strikes again. Cell Stem Cell 2008; 2: 406–08.

5. Vermeulen L, De Sousa E, Melo F, et al. Wnt activity defines colon cancer stem cells and is regulated by the microenvironment. Nat Cell Biol 2010; 12: 468–76.

6. Rosenthal DT, Zhang J, Bao L, et al. RhoC impacts the metastatic potential and abundance of breast cancer stem cells. PLoS One 2012; 7(7):e40979.

7. Hakem A, Sanchez-Sweatman O, You-Ten A, Duncan G, Wakeham A, et al. (2005) RhoC is dispensable for embryogenesis and tumor initiation but essential for metastasis. Genes Dev 19: 1974–1979.

8. Suwa H, Ohshio G, Imamura T, et al. Overexpression of the rhoC gene correlates with progression of ductal adenocarcinoma of the pancreas. Br J Cancer 1998; 77(1):147-52.

9. Ikoma T, Takahashi T, Nagano S, Li YM, Ohno Y, et al. (2004) A definitive role of RhoC in metastasis of orthotopic lung cancer in mice. Clin Cancer Res 10: 1192–1200.

10. Islam M, Lin G, Brenner JC, et al. RhoC expression and head and neck cancer metastasis. Mol Cancer Res. 2009 Nov;7(11):1771-80.

11. Visvader JE, Lindeman GJ. Cancer stem cells: current status and evolving complexities. Cell Stem Cell 2012 Jun 14; 10(6):717-28.

12. Friedmann-Morvinski D, Bushong EA, Ke E, et al. Dedifferentiation of neurons and astrocytes by oncogenes can induce gliomas in mice. Science 2012 Nov 23; 338(6110):1080-4.

13. Luzzi KJ et al. Multistep nature of metastatic inefficiency: dormancy of solitary cells after successful extravasation and limited survival of early micrometastases. Am J Pathol 1998; 153(3):865–873.

14. Nolte SM, Singh SK. Origins of Metastasis-Initiating Cells (pp 229-246). In R.K. Srivastava and S. Shankar (eds.), Stem Cells and Human Diseases. Springer Science+Business Media B.V. 2012.

15. Mimeault M, Batra SK. New advances on critical implications of tumor- and metastasis-initiating cells in cancer progression, treatment resistance and disease recurrence. Histol Histopathol 2010; 25(8):1057-73

Just one short comment, if I may:

Critical thinking may be of higher value than overtaking opinions although such mutated to a dogma, by meaning:

“An apple found in a car (correct observed Mutation / genetic changes) is not synonymous of prove apples grow in cars (mislead interpretation of correct findings)”

In this regard someone may read in detail - and I mean or suggest, not in a receptive reading manner - the following as it may be of relevance:

(1)   Epistemology of the origin of cancer: a new cancer paradigm. BMC Cancer 2014.

http://bmccancer.biomedcentral.com/articles/10.1186/1471-2407-14-331

(2)   Cell-cell communication in the tumor microenvironment, carcinogenesis, and anticancer treatment. Cell Physiol Biochem 2014.

http://www.karger.com/Article/FullText/362978

(3)   Imagine a World without Cancer. BMC Cancer 2014.

http://bmccancer.biomedcentral.com/articles/10.1186/1471-2407-14-186

(4)   Genomics, microRNA, epigenetics, and proteomics for future diagnosis, treatment and monitoring response in upper GI cancers

http://clintransmed.springeropen.com/articles/10.1186/s40169-016-0093-6

(5)   Somatic Mutation Theory - Why it's Wrong for Most Cancers:

http://www.karger.com/Article/FullText/443106

thanks for the interesting question

best

Björn

Article Somatic Mutation Theory - Why it's Wrong for Most Cancers

Article Epistemology of the origin of cancer: A new paradigm

Article Cell-Cell Communication in the Tumor Microenvironment, Carci...

Article Genomics, microRNA, epigenetics, and proteomics for future d...

Article Imagine a world without cancer.

A perhaps provocative thought: Could it be that we have got the wrong end of the stick when thinking that metastasis per definition is derived from the primary tumor (just like we are wrong about the Somatic Mutation Theory (see Bjorn's contribution)? Based on Konstantine's elegant deliberations, it appears that all assumptions from decades ago are out of the window, so we need to go back to the drawing board. The dogma that metastasis originate from the primary tumor is based on common biomarkers. However, this commonality can also explained by having common ancestry. I think of my friend who suffers from reoccurring lesions in his liver identified as non-endocrine pancreatic tumor cells, but they keep reoccurring even after total and successful removal of the primary tumor in the pancreas! If I am not mistaken, it might be time to start considering that primary tumors are not the source of metastasis at all.

A lot of good points raised but most exhibit a proclivity to the primary tumor "doing things" that subsequently result in metastases - we should also consider the roles played by the pre-cancerous niche and how the microenvironment where the  metastasis occurs facilitates the "seeding" of the cancer- so how would one weight the role of a traveling cancer cell against the role of the environment where the metastases ends up occurring?

Over 25 years ago, metastatic tumor progression was considered to late events, i.e. accumulating mutations and, for example, p53 mutations in colon carcinoma. At that time I should analyze only GBM grade IV, some highly aggressive glioblastomas for p53 mutations. As a young scientist I used a "free" week when my collaborator was on holiday to add some low-grade gliomas as my extra work... I thought there was no real data to assume p53 to be just a late event... and I found p53 mutations even at high frequencies in grade II and grade III astrocytomas (but no mutations in a very different biologic tumor entity, grade I pilocytic astrocytoma.

There may be many different types of tumors with different metastatic mechanisms; sometimes the metastases were found in patients without being able to detect the primary tumor. The theory of cancer stem cells may fit well with this observation, since tumor stem cells should have many tools to migrate, at least better than more differntiated cells.

With the growing field of epigenetics, another view may add some missing links in fully understanding cancer development. Perhaps, Björn and Ijaz critical views on the current mainstream point to epigenetics (although they both may see it differently).

At a recent meeting in the field of cancer research it was new to me that infections can influence epigenetics. I didn't ask the philosophical question but in theory one may assume that infections can induce by several tools changes in epigenetics which may change a cell's behaviour without even any classical (other) mutation...