Immune cells in breast tumors that support cancer — and can be eliminated

It always amazes, impresses and delights me when something basic and simple is discovered, even now, after so many years of intensive research.

That was my reaction on reading about the type of immune cells that are found in a breast cancer lump in mice.

It was known that throughout mammary gland tissue one could find macrophages, large phagocytic white blood cells. That’s true for any tissue of the body; macrophages are there. Macrophages act as scavenger cells mopping up bacteria, viruses, and debris from tissues all over the body. They remind me of the brush-whirling trucks that clean the streets of our cities, removing trash that builds up near the sidewalk curbs, when most residents are asleep.

Macrophages also play an important role in the immune response because they are able to transfer some of the proteins they engulf to their cell surface, rather than digesting them to their component amino acids, and “present” them to lymphocytes. Lymphocytes that recognize the protein (the protein now serves as an antigen) are stimulated to divide, expand their population, and then target residual antigen in the body.

Researchers in NY (1) asked if macrophages found in mammary tumors are the same type of macrophages found in the rest of the mammary tissue. Or are they different? What a simple question! And if they are different, how are they different, and how does that difference matter? But to begin with, how can you tell one macrophage from another?

That was the hard part. The investigators had to examine macrophages from normal mammary gland tissue and from tumors within that mammary tissue, and then compare them with respect to a variety of different properties.

But once that process was done, their data led to the simple conclusion that the two populations of macrophages were indeed different. There were regular Mammary Tissue Macrophages (MTM’s), and there were altogether different Tumor Associated Macrophages (TAM’s). The two types of cells have different cell surface markers, differentiate from different precursor cells, and express different sets of genes.

This is shown for mice, and only for breast cancer, but presumably this is a general phenomenon. Tumors have their own type of macrophages. We didn’t know that before this study.

Why does this matter? For one thing it means that the micro-environment of the tumor must be different from that of normal mammary tissue. That follows from the fact that the macrophages are able to detect that difference. It could be useful to know what difference is detected, at the molecular level, by the macrophages, because that might be exploitable for development of new therapies. Secondly, the tumor associated macrophages may play a functional role in the support of tumor growth.

The latter possibility was tested by inhibiting the differentiation of the TAM’s in mice. This was possible once it was known that they differentiated from a different set of precursor cells than MTM’s, and the molecular regulator of that differentiation was determined.

The researchers found – hold onto your seats – that if TAM’s were blocked from differentiating, if these special macrophages did not therefore take up residence in the tumor, then the tumors did not continue to grow. Further investigation showed that TAMs block the body’s own cytotoxic T cells from attacking cancer cells. By eliminating TAMs, the body’s immune system can succeed against cancer cells.

This is not only a very satisfying piece of research, it opens up entirely new avenues for cancer therapy. Instead of (or in addition to) killing cancer cells (as with radiation and chemotherapy) it may be possible to limit cancer by blocking TAM cells in tumors and letting the body do the work of killing cancer cells.

The big question now is whether tumors of other tissues also have special Tumor Associated Macrophages, and whether each tumor has its own type of TAM or whether there is a universal property to TAM’s. If the latter, then it could be possible for one type of drug to help suppress any type of cancer.

This may take a number of years to work out, but it is wonderfully promising.

1. Franklin RA, Liao W, Sarkar A, Kim MV, Bivona MR, Liu K, Pamer EG, Li MO. 2014. The cellular and molecular origin of tumor-associated macrophages. Science 344:921-925.

Cancer cells that drink themselves to death!

I really like this paper, partly because of the fabulous results presented — this could lead to a real “cure” for brain tumors (glioblastomas), one of the most deadly cancers — and partly because its approach flies in the face of the new dogma of personalized medicine through genomics (1).

 It starts with the premise that mutations in the many different genes that cause cancer may be having affects on cells in more ways than just in the control of cell division and migration, ways that we perhaps don’t much about right now. If so, they reasoned, then perhaps by interfering with one of those “side effects” of cancer genes, we might be able to halt the cancer itself.

They started with a grab bag of over a thousand different chemicals, treated samples different strains of glioblastoma cells and normal cells with each one, and assayed for something different in the appearance of the tumor cells that did not occur in the normal cells.  That’s all. Using a microscope they looked to see (with the help of a computer program) if there were size changes, shape changes or some other kind of change such as cell death.

And, wouldn’t you know it, they found that several compounds, with one particular one standing out, caused such changes. The compound killed glioblastoma cells in culture (and, they soon found, also in living mice), but had no affect on normal cells. 

What is particularly interesting is that before the cells die, they filled up with clear vacuoles. And before that the cell surface became turbulent. Filling in the gaps, they showed quite definitively that the active compound (which they call Vacquinol) was causing cells to form channels that brought fluid into the cells, a drinking process called macropinocytois. The fluid entered the vacuoles and the vacuoles increased in number and size until the cells literally burst – and died.

No one has ever seen this before. In fact, this represents a new form of cell death, a new way that cells can die that has not been reported before. Something about the physiology of the glioblastoma cells (we still don’t know what that is) makes them sensitive to very low concentrations of Vacquinol which causes them to drink themselves to death!

Extraordinarily, the few other cancer cell types tested were, like normal cells, unresponsive to Vacquinol. In this study only glioblastomas responded.

The researchers were lucky. They did a marvelous job of analyzing what was happening to the cells in response to Vacquinol, and showed that because of its effectiveness in living mice (6 of 8 vacquinol treated tumor-induced mice were still alive at 80 days, while all untreated mice with tumors were dead by 60 days), it is ready for clinical trials, but they were lucky that some compound in their random bag of chemicals actually works, especially since it works only on the cancer type they used, and not others.

One reason glioblastoma tumors have been so intractable is that even after treatment, even after many cells are killed by the chemotherapy drug used, disease recurs as the tumor continues to grow. There seems to be a reservoir of cells that are not affected by the cytotoxic drug. They may represent a stem cell-like population that is generally quiescent, which means that that they do not divide for long periods of time and therefore would be unaffected by any drug that targets the proliferative process.  But Vacquinol is different. It doesn’t target proliferation; because it affects a fundamental property, it can impact on quiescent cells as well. The kill rate is therefore substantially higher than for a traditional chemotherapeutic drug.

Let’s hope that we will see other cancers tested with whatever grab bags of small molecule chemicals that are available.

1.         Kitambi SS, Toledo EM, Usoskin D, Wee S, Harisankar A, Svensson R, Sigmundsson K, Kalderen C, Niklasson M, Kundu S, Aranda S, Westermark B, Uhrbom L, Andang M, Damberg P, Nelander S, Arenas E, Artursson P, Walfridsson J, Forsberg Nilsson K, Hammarstrom LG, Ernfors P. 2014. Vulnerability of glioblastoma cells to catastrophic vacuolization and death induced by a small molecule. Cell 157:313-328.



Coumadin at very low dose may slow metastasis

   Natural Killer cells are part of the immune system, and unlike their T-lymphocyte cousins, they seem to have an innate ability to kill virally infected cells and cancer cells. This is not something they learn from experience, as do other T cells. They act as if by instinct.

   Strangely, even though we have this natural defense system in our bloodstream, we still develop cancers that spread, via the bloodstream, to other parts of the body. And it is this spread of a cancer, metastasis, which makes cancer such a lethal disease.

   This has been a troubling puzzle for some time now: if we have Natural Killer cells, why don’t they kill cancer cells? Or at least, why not those that leave the primary tumor and enter the blood stream, where NK cells are found?

    So it is big news when a study comes along that identifies a way that Natural Killer (NK) cells can be influenced to do a better job killing cancer cells. Such a study has recently been published (1).

    It had been known that the deletion of a particular gene from mice provided them with greater resistance to the spread of cancer, and an increased lifetime of mice with cancer, compared to their normal counterparts. That gene encodes a protein with a tongue-twister name called “E3 cbl-b ubiquitin ligase”.  Let’s just call it “E3” here.

    The question that was addressed is what is the function of that protein, and what cells are affected by the elimination of that protein (via its gene deletion or inhibition of gene expression)?

    Along the way they made an unexpected discovery, that low doses of Coumadin (also known as warfarin), the inexpensive and widely used drug that retards the coagulation of blood, is effective at unleashing the cancer-killing ability of NK cells.  It reduced colonization of several difference cancer types, and increases the lifespan of mice with cancer. Its impact was just about as good as a specially designed drug that “awakens” NK cells. Coumadin was effective even at concentration so low that blood-clotting is not at all affected, which means that negative side effects are virtually non-existent. It should be emphasized that Coumadin had no effect on the growth of the primary tumor, and it did not totally prevent metastasis, but it slowed the process of metastasis significantly, and reduced the spread of the cancer.

    To repeat, the study was limited to mice, to two different strains of mice, so it may be possible that there will be no positive benefits of low dose Coumadin in human beings. But, since Coumadin is so inexpensive, and has been used for so long with so many people, and at the low concentrations used has no negative side effects, and it ie possible, perhaps likely, that it will work in humans as well as in mice, why not prescribe it as soon as a person is diagnosed with any type of solid tumor. It just may slow the spread of cancer enough to buy time, and to allow drugs that prevent the growth of tumors to have their maximal effect.

     If I had a solid tumor I would ask this of my doctor. It is not a cure, but together with a drug that limits growth of a tumor, it could offer increased hope.

    If you are a doctor reading this (to the end), and who has read the paper referenced below, can you think of a reason for NOT prescribing low dose Coumadin?

     Now let’s get back to the study, for those who would like to follow the investigative process.

     The researchers started their study by reproducing the published observation that deletion of the E3 gene in mice slows metastasis and extends life. They also observed increased infiltration of NK cells into the tumor mass, indicating that NK cells were somehow influenced by the E3 deletion. That influence could have been indirect, for example, tumor cells with missing E3 might secrete something that attracts NK cells. Or E3 deletion could have directly altered the properties of the NK cells themselves.

    They next isolated NK cells from E3–deleted mice and tested them in normal mice that still had the E3 gene. Once again metastasis was slowed. It follows that E3 deletion affects the NK cells directly. Please note that this does not rule out the possibility that other cells are also affected in some way by the  elimination of E3. 

    What does the E3 enzyme do? It adds a chemical group called ubiquitin to existing proteins thereby changing their functional properties. This is referred to a post-translational modification, because it alters a protein after it has already been synthesized (translated from the DNA code).

    The researchers then asked, which proteins in the NK cells are modified? They tested 9000 different proteins for ubiquitination by E3, and found that one protein stood out. That protein was one of a three-member of family of proteins (each of which could be ubiquitinated) collectively known as TAM.

     There are things known about TAM. It appears in cells other than NK cells, it is part of a trans-membrane protein found at the surface of cells. Its exterior part serves as a signal receptor (known signaling molecules, or ligands, include Gas6 and Protein S) and its interior, cytoplasmic, side has tyrosine kinase activity. When a TAM molecule binds to its ligand, it is ubiquitinated by E3, and this induces endocytosis, or the formation of a phagocytic vesicle.

    Incidentally, TAM also function in the phagocytosis of apoptotic cells, and mutations in TAM are associated with some types of autoimmune diseases.

    It seems that TAM negatively regulates the NK cells; Gas6, which activates TAM and induces endocytosis, prevents killing activity, while E3 deletion which blocks endocytosis awakens NK ability to kill cancer cells.

    The investigators reasoned that by inhibiting the tyrosine kinase activity of the receptor they could suppress the negative regulation of NK cells. So they set out to design a small molecule inhibitor of TAM based on the known structure of the active site of the tyrosine kinase portion of the molecule. In this they succeeded, and they called their inhibitor LDC1267.

    By the way, this approach stands in contrast to the development of most other drugs, which are discovered by screening thousands of known compounds, many extracted from living plants, fungi and other organisms.

    LDC1267 worked. It worked just as effectively as if the E3 gene had been deleted. So it, or a similar drug that may work even better in humans, will likely soon be tested in clinical trials.

    Indeed, Foretinib, a drug presently being tested in clinical trials, that inhibits some tyrosine kinase receptors, was recently tested with glioblastoma, a highly invasive and high morbidity tumor (2). The authors write: “Foretinib treatment in vivo abolished MerTK [the M in TAM] phosphorylation and reduced tumor growth 3-4 fold in a subcutaneous mouse model”. When M was completely inhibited, it completely stopped the growth of the tumor. (In this case, interestingly, Foretinib had a direct affect on the tumor cells themselves, in vitro, slowing growth and reducing migration.  So TAM activity may regulate both cancer cell behavior as well as NK cell behavior).

    It turns out that Gas6 and Protein S, the proteins that activate TAM, are Vitamin K dependent proteins. They need Vitamin K in order to be able to bind to TAM.  Vitamin K is a molecule of known importance in bone construction, and in blood coagulation.

    Blood coagulation? That brings us back to Coumadin, because Coumadin is an antagonist to Vitamin K. That is why Coumadin is useful as a so-called “blood thinner”. (It doesn’t really thin the blood, but by antagonizing Vitamin K it does decrease the probability of blood clot formation).

     Now we have a molecular explanation of just why Coumadin (warfarin) is such a good activator of NK cells. Because Coumadin antagonizes Vitamin K it reduces the functionality of Gas6 and Protein S. In turn that means that their ability to bind to TAM is reduced. Since Gas6 activates TAM and negatively regulates NK activity, Coumadin prevents activation of TAM and positively regulates NK cells In effect, TAM is inhibited, similar to its inhibition with LDC1267.

    In their Figure 4, a section of which I copied for this blog, one can readily see the splotches of melanoma cancer that had metastasized to the lung tissue in control (vehicle) mice. By contrast the number of metastatic colonies is significantly reduced when mice are treated with their inhibitor LDC1267. It is striking that common Warfarin (Coumadin) is just about as effective.


   It seems like before too long we will see some powerful anti-metastasis drugs made available to cancer patients.

   But meanwhile, I hope that doctors will start prescribing very low dose Coumadin to their cancer patients.


 1.         Paolino M, Choidas A, Wallner S, Pranjic B, Uribesalgo I, Loeser S, Jamieson AM, Langdon WY, Ikeda F, Fededa JP, Cronin SJ, Nitsch R, Schultz-Fademrecht C, Eickhoff J, Menninger S, Unger A, Torka R, Gruber T, Hinterleitner R, Baier G, Wolf D, Ullrich A, Klebl BM, Penninger JM. 2014. The E3 ligase Cbl-b and TAM receptors regulate cancer metastasis via natural killer cells. Nature 507:508-512.

2.         Knubel KH, Pernu BM, Sufit A, Nelson S, Pierce AM, Keating AK. 2014. MerTK inhibition is a novel therapeutic approach for glioblastoma multiforme. Oncotarget 5:1338-1351.



New insight into the most deadly form of breast cancervv

This could be important, or at least it could become important. In the journal Science, in an issue whose cover story is breast cancer, a short note appeared describing a gene that could be involved in causing breast cancer (1).

As the editorial in Science points out, a great deal has been learned about breast cancer in the last 20 years, and new treatment options have been developed Still, it is the leading cause of cancer death among women.

Some 5-10 % of breast cancers are caused by hereditary mutations in BRCA 1 or 2, which normally aid in the repair of damaged DNA (2)and for which there is no cure.

Most cases of breast cancer however, are associated with changes in the genes for hormone receptors, like estrogen receptors, progesterone receptors and epidermal growth factor receptors. Treatments exist for these kinds of cancers. For example, Herceptin is used for treatment of breast cancers with epidermal growth factor receptor mutations, and in the case of estrogen receptor mutations, tamoxifen is used. In both cases the drugs alter the behavior of the receptors without eliminating some of their important positive functions. Interestingly, tamoxifen seems to work best on women whose mammary tissue also expresses high levels of progesterone receptor as well as estrogen receptor (3). (To the best of my knowledge, there are no drugs working to modulate progesterone receptor protein in breast cancer treatment).

But – and here is the point relevant to this discussion — about 15-20% of cancers are classified as “triple negative”, meaning they show no signs of elevation of any of the three hormone receptors (and presumably are BRCA negative). These cancers are generally more life threatening, because they are more metastatic, and there is no treatment for them.

Now, a group of investigators from Cold Spring Harbor and the University of Toronto, report that triple negative cancers display high levels of a possible receptor for G protein activation (4). G proteins are important mediators of cell signaling. They are located in the cytoplasm and can be coupled to a receptor in the membrane when that receptor is bound by a signaling molecule, or ligand. When coupled, the G proteins are activated and can transmit the signal to other molecules in the cell that then migrate into the nucleus and activate a varied set of genes, including genes in growth control.

There are hundreds of different, but related, G proteins, and there are hundred of different, but related coupled receptors. For many of those receptors, the ligand is not known. Such is the case with the receptor reported in this study. It is likely a signal receptor at the cell surface of mammary cells, but its signal at this time is not known. It is a receptor without a known signal, an “orphan” receptor. Each orphan receptor discovered is given a number, in sequence. In this case it is G-Protein coupled Receptor 161, or GPR161.

But the important thing is that GPR161 levels are elevated in triple-negative cancers. This is more cause than effect, because in cell cultures, experimentally elevated expression of the gene for GPR161 induced proliferation and increased the cells’ invasive properties.

The hunt must now be on for the signaling molecule for GPR61, because if the activity of the “receptor” could be modulated pharmacologically, as others are with tamoxifen and Herceptin, another category of breast cancer could be controlled.


  1. Kiberstis PA. 2014. Negative Reinforcement. Science 343:6178.
  2. Campeau PM, Foulkes WD, Tischkowitz MD. 2008. Hereditary breast cancer: new genetic developments, new therapeutic avenues. Human genetics 124:31-42.
  3. Stendahl M, Ryden L, Nordenskjold B, Jonsson PE, Landberg G, Jirstrom K. 2006. High progesterone receptor expression correlates to the effect of adjuvant tamoxifen in premenopausal breast cancer patients. Clinical cancer research : an official journal of the American Association for Cancer Research 12:4614-4618.
  4. Feigin ME, Xue B, Hammell MC, Muthuswamy SK. 2014. G-protein-coupled receptor GPR161 is overexpressed in breast cancer and is a promoter of cell proliferation and invasion. Proceedings of the National Academy of Sciences of the United States of America 111:4191-4196.

A promising new drug to treat Acute Myeloid Leukemia

    The financial news was abuzz today about a report by Agios, a biotech company, that in a small Phase 1 study, a drug they are developing was remarkably successful with a subset of Acute Myeloid Leukemia (AML) patients. The patients tested had been through at least one round of chemotherapy, and up to four rounds, without success. The Agios drug, AG-221, was tested for how well it would be tolerated by AML patients. The results were very good, in that adverse side effects were few and limited. The big news though was that 6 of the 7 evaluable patients showed complete remission of their leukemia.
    It is rare to see such impressive results in a preliminary study that deals with dose tolerance. In response to the news, the stock price of Agios rose significantly today.
    Some 10% of AML patients might benefit from AG-221, those with a mutation in a metabolic enzyme, isocitrate dehydrogenase-2. Cells with that mutation produce a higher than normal concentration of 2-hydroxyglutarate, which leads to brain cancers (gliomas and blastomas) as well as AML. How that oncometabolite causes cancer is still being examined. In any event, additional clinical trials with AG-221 are on their way.

Tumor, parasite, and ancient dog — all at once


This is a story about an extraordinary tumor.

You know that most cancers arise from mutations in a person’s gene complex, though in some cases a viral infection can cause cancer. An example of the latter is the human papillomavirus, or HPV, a virus is transmitted sexually and that can cause cervical cancer in women.

Now comes word that there is a tumor in dogs that is also sexually transmitted, but it is not caused by a virus. Instead infection occurs by the passage, during sex, of some of the cancer cells themselves. By this means they take up residence in a new dog. It is referred to as the Canine Transmissible Venereal Tumor, or CTVT.

After it is passaged it can cause the growth of a tumor in the recipient organism.  Tumor growth commonly occurs in the genital orifice, but can also appear in other places, including the skin. It generally does not metastasize, and quite inexplicably it regresses after a period of a few months. After that the affected dog is immune from further infection. But meanwhile it can have passed the tumor cells on to other dogs through sexual activity.

For all practical purposes, it acts like a parasitic organism, growing in a host and transmitted by a particular behavior; it is not generally harmful to the host organism, and death from the tumor is rare.

This cancer cell has been passed from dog to dog for many generations. In fact, according to genomic analysis, it may have originated some 11,000 years ago from a breed of dog that no longer exists.

It has survived in dogs as if it were a parasitic single celled organism. But it is not a protozoan. It is a cell that originated from a dog, but no longer belongs to a dog.

This is like a HeLa cell that uses host dogs instead of Petri dishes to provide its growth conditions.

I don’t know of anything comparable.

Meanwhile, according to the investigators from England and Australia who did the genomic analysis[1], it is extraordinary that the cell has survived all this time. Its genome has undergone 1.9 million base substitutions, undergone thousands of rearrangements, and has lost 646 genes. The cell has changed dramatically over millennia, while allowing it – or enabling it – to grow and reproduce and survive passage from one dog to another — and produce tumors that do not kill their host organism.

As they write, “Our results provide a genetic identikit of an ancient dog, and demonstrate the robustness of mammalian somatic cells to survive for millennia despite a massive mutation burden”.

These are two “firsts” for me; a tumor that is transmitted from one organism to another by tumor cells themselves, and a mammalian cell that has survived away from its original organism for over 10,000 years.  Extraordinary.

1.         Murchison EP, Wedge DC, Alexandrov LB, et al. Transmissible dog cancer genome reveals the origin and history of an ancient cell lineage. Science 2014;343:437-40.

Cells that lead the metastasis migration — can they be stopped?

A tumor is an abnormal growth of cells, but cancer occurs when some of those cells leave the tumor mass, enter the circulatory system and colonize new parts of the body.  We usually think about single cells breaking away from the pack and setting off on their own to establish these new colonies. But, according to recent research findings, when cells break away from the tumor mass they do so usually as groups of cells.

It now turns out that the leader cells have a unique identity.

This was the result of a study conducted at Johns Hopkins and the UCSF Medical School in San Francisco.[1] The goal now is to selectively target those cells to block the migration, and thereby inhibit the development or extent of cancer.

To investigate the characteristics of the leader cells the researchers made use of a method they recently developed for culturing breast tumors as three-dimensional structures, rather than the more established method of culturing cancer cells as single cells. A three-dimensional structure more closely reflects what occurs in an organism. The method involves embedding small chunks of highly metastatic tumors from mice, containing 200-1000 cells, in a collagen gel.

They observed that within 2-3 days, small strands of cells projected from the tumor mass, like a finger pointing to a new diection. “Because the cells leading these invasive strands were highly protrusive and migratory we refer to them as ‘invasive leader cells’”. Since these cells were so morphologically distinct, they were able to determine, at a cellular level, if their molecular properties were distinct from the tumor mass.

Indeed they were. They contain proteins that distinguish them from the tumor mass, while sharing some molecular features as well. To simplify, I will focus on one of the markers, perhaps the most significant.

That protein is cytokeratin-14 (K14), an intermediate filament protein which contributes to the skeletal structure of cells. 93% of leader cells were positive for K14, while the rest of the tumor cells were K14 negative. So K14 is associated with invasive leader cells.

They then asked whether K14 positive cells could be seen in intact tumors in mice. For this they examined sections of tumors derived from mice. They found that K14+ cells were concentrated at the boundary of tumors with surrounding tissue (stroma), and that those cells led invasive strands into adjacent muscle. The same was true for several different types of breast cancers.

These experiments in culture and in several types of intact tumors were repeated for human tumors, and again the same results were obtained. From the observations on cells in 3D culture, insight into the molecular features of invasive cells in a variety of breast cancers, both in mice human, had emerged.

Now for the clincher. The investigators genetically engineered mice breast cancer cells such that the emergence of  K14 was blocked. These cells were grown in mice and formed tumors. Those tumors were then cultured in the 3D models, side by side with K14+ tumor 3D models. In contrast to the K14+ models, in the K14- models “both K14 and collective invasion were markedly reduced”. Further, among K14- tumors that formed in vivo, there was also a striking reduction in collective invasion at the tumor-stromal border.

These data indicate that K14 does not merely appear in the leading invasive cells as a consequence of differentiation, but that K14 is required for strands of collective invasion to occur.  The authors write, “targeting a basal invasive program expressed in a small minority of tumor cells is sufficient to disrupt the invasive process in an advanced carcinoma”. The prediction is that tumors in which K14 activity is blocked would be less metastatic than normal. Perhaps we will see that data in their next paper.

This is such a wonderful approach because any therapeutic development would not only be effective in a variety of breast cancers, but would also likely be effective for other metastatic cancers as well.

[1] Cheung KJ Gabrielson E, Werb Z, and Ewald AJ (2013) Cell 155:1639-1651.  Collective invasion in breast cancer requires a conserved basal epithelial program.

A new FAKt about metastasis

Most cancer deaths are due to metastasis, where tumor cells detach from their original site of tissue growth and spread to other parts of the body.

Metastasis requires tumor cells to break free from the tissue in which it is growing, crawl through a “basement membrane” (a collagen-based structure that envelops the tissue), crawl across the extracellular matrix, attach to a capillary, and crawl through openings made in the capillary wall to enter the blood stream. Unless tumor cells can enter the blood stream, and depart from it, they cannot migrate to other body sites, and metastasis would be thwarted, just as it would be impossible to travel home from work if the doors to the subway or train station refused to open.

Researchers have just clarified the way in which tumor cells open the “doors “of the capillary and enter the blood stream.  This finding can lead to ways to at least diminish metastasis if not eliminate it.

When tumor cells attach to capillaries, they secrete VEGF, a compound that causes endothelial cells that make up the wall of the capillary, and that are normally very tightly sealed to one another, to separate apart just enough for the tumor cell to crawl through and enter (or leave) the blood stream. VEGF binds to a receptor on the cell surface of endothelial cells; the receptor’s cytoplasmic side has enzymatic activity. Like other growth factor receptors it can add phosphate groups to tyrosine amino acids of selective proteins. 

In this case, phosphorylation of a molecule, cadherin, is promoted. Cadherin is responsible for the endothelial cells being able to adhere tightly to one another. When cadherin is phosphorylated, however, they lose their grip, and the endothelial cells can form opening between them.

However, an important step on the pathway was missing. While VEGF binding to its receptor PROMOTES phosphorylation of cadherin, it does not actually do the phosphorylation of cadherin. Some intermediate is needed.

Researchers from San Diego (Jean et al., 2014) recently identified that intermediate, an enzyme called focal adhesion kinase, or FAK. Focal adhesion kinase plays an important role in supporting attachment between cells, as well as attachment of cells to a substrate. When tumor cells secrete VEGF it causes FAK to accumulate at the endothelial cell junctions. (Another protein, src, also accumulates at the junction, but fails to do so if FAK activity is blocked). Now it is known that FAK does not merely accumulate at the cell junction but actually modifies the properties of the junction through its phosphorylation of cadherin.

 To summarize, when tumor cells secrete VEGF it binds to its receptor on endothelial cells, which then stimulates FAK phosphorylation, and FAK in turn phosphorylates cadherin resulting in a weakening of the adhesion of endothelial cells, allowing tumor cells to pass between them.

This is clearly and important signaling pathway because it allows tumors cells to enter into and exit from the circulatory system. Limiting access to the circulatory system could decrease metastasis and therefore limit that aspect of cancer that most affects mortality.

In mice, that was the case. By genetically eliminating FAK activity in adult endothelial cells the investigators prevented melanoma cell metastasis in mice.

Interestingly, this did not prevent the growth of the tumor, only its metastasis.

Therefore, the VEGF-Receptor-FAK-src-cadherin pathway has emerged as an important set of therapeutic targets, creating new opportunities for preventing cancer deaths. Inhibitors of FAK are already being tested in clinical trials. As the pathway continues to be clarified, new pharmacological agents will surely be sought.

Jean, C., Chen, X. L., Nam, J. O., Tancioni, I., Uryu, S., Lawson, C., Ward, K. K., Walsh, C. T., Miller, N. L., Ghassemian, M. et al. (2014) ‘Inhibition of endothelial FAK activity prevents tumor metastasis by enhancing barrier function’, J Cell Biol 204(2): 247-63.



A New Explanation for Metastasis

Understanding how metastasis occurs may allow the most important approach to preventing cancer deaths; that is because most cancer deaths occur due to metastasis of epithelial tumor cells. The key question is how tumor cells of epithelial tissue, whose general characteristic is one of constant shape and immobility, acquire the ability to change to a crawling shape and move out of the tumor mass, travel into the blood stream, and arrive at a location they can colonize.

It is generally thought that epithelial tumor cells undergo transition to mesenchyme-like cells that can move and crawl to other sites in the body. The transition  is thought to be triggered by products of other cells, like the senescent stroma cells that are neighbors to epithelial tissues, perhaps together with changes at the level of the genome. The ability of transitioned cells to move away from the tumor itself may also depend on the composition of the extracellular matrix the cell encounters.

I call your attention to a new possibility for metastasis suggested in a paper (Lazova at al, 2013) from the Pawelek lab at the Yale School of Medicine, with the assistance of the Denver Police Department Crime Lab. They argue that metastatic cancer cells may be derived from a fusion of mobile cells such as leukocytes together with epithelial cancer cells. The fusion product would have characteristics of the cancer cell, but also characteristics of leukocytes, namely the ability to move, to crawl into and out of blood vessels, and to squeeze into tissues throughout the body. It would also have a novel genetic makeup.

This is an interesting possibility, but a very difficult one to demonstrate. It was first proposed over a hundred years ago. But till now there has been no experimental evidence of metastatic  cancer cells in humans with hybrid properties.

The researchers used an innovative approach. They examined a metastatic brain melanoma taken from a person who had received a bone marrow transplant.  Some time after the transplant the patient developed a melanoma that had then spread to other parts of the body, including the brain.

Here is the interesting part. Since the person had received a bone marrow transplant, his or her own bone marrow cells were all destroyed, meaning that all new white blood cells were derived from the donor. The donor and host have different genes. Therefore if host melanoma cells fused to donor bone marrow-derived cells then it might be possible to show that metastasized melanoma cells also have properties of leukocytes, and contain donor genes as well as host genes.  That would constitute evidence that that the metastasized cells were derived from a fusion of cells. And if that were the case, that could account for the mobile, colonizing properties of the metastatic cells.

This is indeed what the investigators found, giving rise to a new way to view the epithelia-to-mesenchyme transition, and to the properties of metastasized cells. This is important, because once it is understood HOW metastasis occurs, it then becomes possible to develop drugs that can PREVENT metastasis from occurring.

The authors are quick to point out that this is only a single instance, only one type of cancer from only one patient. As they say, “The extent to which this mechanism is operative in other tumors remains to be determined”.

I look forward to the results of future studies; preventing metastasis is the key to limiting cancer deaths. A method to prevent metastasis would have general applicability for a variety of tumors, and would not depend on treating each cancer with medication to counter the specific mutation(s) that caused the tumor.


Lazova R, LaBerge GS, Duvall E, Spoelstra N, Klump V, Sznol M, Cooper D, Spritz RA, Chang JT, Pawelek JM (2013) PLOS one 8:1-7.        A melanoma brain metastasis with a donor-patient hybrid genome following bone marrow transplantation: First evidence for fusion in human cancer.

Impact of tumor heterogeneity on genomic approach to cancer therapy

            I noted earlier (see The Personalized Cancer Treatment Bandwagon Continues to Grow, June13, 2013) that both in the US and Great Britain, policy decisions have been made to conduct a full-out campaign to develop a personalized medicine approach to cancer treatment. The idea is to assay by genomic analysis, a person’s cancer for the underlying genetic mutation (or mutations) associated with the cancer, and to provide a specific drug therapy directed against the altered protein.  Treatment would be highly specific.

            The first stage of this approach is to sample as many different types of cancers as possible in order to identify as many mutations as possible that may be causative (some may simple be associative but not causative). The second stage is to develop drugs that interfere with the altered protein product of the mutated gene. The third stage is to perform clinical trial to see which of the drugs are most effective with fewest side effects. This could be a lengthy process.

            A review article appeared in Nature recently that identify significant obstacles to the development of personalized cancer medicine (Burrell et al 2013). The main point of the paper is that most cancers are genetically heterogeneous, which means that there are different mutations in different cells of a tumor, and that different metastatic foci may be different with respect to the mutations they carry. That is because the mutation rate in tumors is often higher than in normal cells due to what is called “genomic instability”. “Most solid tumors, and haematopoietic malignancies display at least one form of genomic instability”. Consequently, future genomic analyses will have to search for more than just the dominant mutation, because even if cells carrying that mutation are eliminated therapeutically, that could leave other cancer cells still capable of growing. More than one specific drug would be needed for therapy.

            In addition, the authors warn that even identifying all the mutations may still be insufficient, because among genetically identical cells there is a range of different phenotypic characteristics. “Phenotypic heterogeneity…is determined …also through stochastic events in gene expression and protein stability, epigenetic divergence and micro-environmental fluctuations”.  The particular microenvironment in which a tumor is found can have a significant influence on the gene expression of a tumor.

            Finally, as a tumor evolves with time the genetic pattern may change. Therefore, for personalized medicine to be successful, genomic assays will have to be repeated throughout the course of therapy.

            Personalized cancer medicine may be a lot more complicated than first thought. Universal therapies for cancer (which I have written about here) are looking better and better, and may be closer at hand.  Perhaps that is where the research money and effort should be directed.

Burrell et al, Nature 501:338.  The causes and consequences of genetic heterogeneity in cancer evolution.