Unraveling Metastasis, the Lethal Agent of Cancer


Cancer is not just one disease. Were it, we could imagine a single cure. Instead, we find that cancer is a mix of clinically separable diseases. In common parlance, we are accustomed to naming each type based on the body part from which it originates: breast cancer, prostate cancer, brain cancer, cancers of the blood.

Still, these are roughshod descriptions. Within each group, we find subdivisions and branches. Genotyping promises a more precise distinction, but even this falls short. The same genetic mutation can drive aggressive cell proliferation in one human while remaining quiet in another. Cancers are unique, hence the current drive towards “personalized medicine.”

Despite this multiple and amorphous identity of cancer, there is a shocking commonality to the way that cancer kills people. Cancer is a disease that spreads, and while we label cancers based on their origin, it is cancer's ability to colonially invade distant tissues that almost always kills. These colonies are called metastases, and the phenomenon of metastasis causes >90% of cancer-related deaths.

Dr. Devon Lawson, a post-doctoral researcher in the lab of Dr. Zena Werb at UC San Francisco, is committed to researching the problem of metastasis. Lawson, Werb, and company have made important initial strides with a study published in the October 1st edition of Nature. In this study, they have discovered differences in the cellular profiles of small versus large metastases, and have shown that these differences may provide important clues for how metastasis starts.

As Lawson frames the question, “What I really wanted to know is: what’s the cell that starts metastasis?”

You might be wondering: if metastasis is so important, why is so little known about its biology? Lawson explains: “It’s really hard to study and very expensive, because you can’t do it in any system but a mouse… it’s a whole organism problem.”

Lawson emphasizes that cancer is a systemic disease. The problem of studying metastasis in the lab is the problem of balancing the necessary complexity of the system with the abilities to probe, measure, and quantify.

To recapitulate systemic complexity, this study used PDX models. PDX stands for Patient-Derived-Xenograft, and this procedure involves taking cancerous cells from a human patient and injecting them into a benign mouse. The cells form a tumor, but more importantly, they turn the mouse into a “cancer system,” and can cause metastasis in sites distant from the injection.

To quantify the biology within this complex system, the authors of this study used new tools for single-cell analysis to probe gene expression with previously unseen detail. By looking at one cell at a time, researchers can discern subpopulations within a tissue that might be missed were one to look at the entire tissue in bulk.

Consider an example gene X. If half the cells in a tissue express high levels of gene X, but the other half barely express it at all, bulk analysis, which is easier to perform than single-cell analysis, would suggest that all the cells express an intermediate amount of gene X. But the reality is there are two subpopulations.

Subpopulations are important in the mammary gland. Breast cancer arises from cells in the breast epithelium, and this epithelium contains at least two cell types. In the context of milk production, the luminal cells secrete milk, while the basal cells provide the contractile force to pump this milk to the nipple.

The mammary gland also has stem cells capable of generating both luminal and basal daughters. The stem cells have traditionally been grouped with basal cells since the proteins on their surface are more similar to basal surface proteins. Current research is doing much to purify and analyze the stem cells on their own, but for the purpose of this study, basal cells and stem cells were indistinguishable.

Lawson and company created PDX models with cells from triple-negative breast cancer patients and saw metastases colonizing the entire mouse: brain, lymph node, lung, bone marrow. They also observed significant differences in the size of these metastases. The big ones are classified as “high-burden” and the smaller ones as “lower-burden,” but there was also a spectrum of sizes between these.

By analyzing the gene expression in the single cells of different metastases, Lawson and company saw that some metastatic cells expressed genes more common in normal luminal cells, while other metastatic cells expressed genes more common in normal basal/stem cells.

Furthermore, the ratio of these two populations was quite different in high versus low burden tissues. High burden tissues looked more luminal and low burden tissues more basal/stem-like, but each metastasis contained cells with each signature.

The basal/stem-like cells in the metastases also had important differences from basal/stem-like cells in normal breast tissue. One example is that they showed extremely high levels of a protein called Sox2, usually found in embryonic stem cells and associated with the ability to produce many types of cells in great abundance.

As Lawson explains, “This is a pluripotency marker, not just a tissue stem cell marker… tissue stem cells don’t really express it.”

As often is the case, the data only generates more questions. Specifically, Lawson and company want to know if low-burden tissues advance to become high-burden over time (this is known as the hierarchical model), or if low-burden and high-burden tissues are seeded by different types of colonizing cells.

While Lawson emphasizes that this study did not prove that the hierarchical model is the dominant one, they saw some results that argued in its favor. Specifically, they saw that the basal/stem-like cells from low-burden tissues were capable of producing high-burden tumors when transplanted to new mice.

Proving or disproving the hierarchical model will be important for advancing our understanding of the origins of metastasis, and perhaps for preventing metastasis in patients. If the source of metastasis is a single type of colonizing cell, we can attempt to target these cells specifically, or prevent them from advancing to later metastatic stages. Conversely, if the source of metastasis is a variety of cells, this will be important to know as well.

Lawson’s pursuit of these questions is not ending with this paper. She is moving to a tenure-track faculty position at UC Irvine, where her lab will focus almost exclusively on metastasis, looking at not only breast cancer, but also prostate and other cancers to understand how metastasis operates in different systems.

While cancer is not just one disease, metastasis is the one problem that makes most cancers lethal. Personalized medicine looks to decode each cancer individually, but a treatment for metastasis could in theory provide a skeleton key.