Inside Darwin’s Tumor

Cancer evolves. Those two words may sound strange together. Sure, birds evolve. Bacteria evolve. But cancer? The trouble arises from the fact that cancers, unlike birds and bacteria, are not free-living organisms. They start out as cells inside a person’s body and stay there, until they’re either wiped out or the person dies.*

Yet the same forces that drive the evolution of free-living organisms can also drive cancer cells to become more aggressive and dangerous. Evolution becomes our inner foe if mutations disable a cell’s self-restraint. The cell multiplies. Sometimes a new mutation arises in its descendants. If the mutations allow the cancer to grow faster, the cells carrying it will take over the population of cancerous cells. Natural selection and other processes that drive evolution on the outside start driving it on the inside.

Like so many other scientists, researchers who study cancer evolution have jumped on new technology for sequencing genomes on the cheap. They’re now starting to publish fine-grained histories of the disease, tracking individual mutations as they arise and spread. Nature has just published a fine example of this new research. I particularly appreciated the informative pictures they came up with to accompany the paper, one of which I’ve included here. You can click on the picture for a bigger version. And below the picture, I’ll explain what it means.

In the new paper, Li Ding and colleagues at Washington University describe a study they carried out on eight people suffering from acute myeloid leukemia (AML), a disease of the immune system. In people with AML, stem cells in the bone marrow that would normally turn into white blood cells instead become cancerous. Treatments include bone marrow transplants and chemotherapy. Unfortunately, AML has a nasty way of bouncing back from chemotherapy, and the drugs become useless to stop it. As a result, a lot of people who seem at first to be in remission eventually die of the cancer.

The Washington University scientists reconstructed the history of the cancer in each patients by sequencing genomes from a number of cells. To determine the normal, original genome, they sequenced DNA from a healthy skin cell. They then sequenced genomes from cancer cells taken from the patients when they were first diagnosed. And then they looked at genomes of cancer cells that emerged after the patients relapsed. From this survey, they came up with a catalog of new mutations that emerged over the course of the cancer. They could then go back into the blood samples and estimate what fraction of the cancer cells had a given mutation at a given point in time.

This figure illustrates the sad chronicle of one particular woman they studied. When she was in her late 50s, she suddenly came down with a sore throat and began to bruise easily. A bone marrow biopsy confirmed she has AML. She got chemotherapy, and then a stem cell transplant. Although she seemed to go into complete remission, the cancer returned 11 months after her diagnosis. The chemotherapy drugs that had previously been so effective now could not stop the cancer. Other drugs failed, too. Two years after her diagnosis, she died.

On the left of the figure, the cancer begins. A single stem cell mutated and became the founder of the cancerous lineage. we start with normal cells. (The cell is dark, and the grey dot marks its original mutation. HSC stands for hematopoietic stem cells).

The cancer cells grew in number, and as they did, they accumulated a lot of mutations, some of which are listed in the figure next to the star. All of these mutations, one after the other, took over the entire population of cells–a signature of natural selection. When the woman went to her doctor, however, the cancer had diversified into a number of different lineage, each carrying additional, distinctive mutations. Over half of the cells belonged to a lineage marked here in purple, known as cluster 2. Cluster 3, marked in yellow, was made up cells with a separate set mutations. And from within Cluster 3 emerged yet another lineage–Cluster 4, marked in orange. The dots in each circle show the sets of mutations that accumulated in each cluster.

The chemotherapy knocked down all the clusters of cancer cells to such low numbers that doctors couldn’t find them any more. But they were still there. And when exposed to chemotherapy drugs, the most successful cluster was not the one that had been most successful back when the cancer was diagnosed. It was the relatively rare Cluster 4. Apparently, it had mutations that made it better able to withstand the chemotherapy drugs. Some its descendants later picked up new mutations, which enabled them to reproduce quickly and take over the cancer population, as they resisted new chemotherapy drugs as well.

“The AML genome in an individual patient is clearly a ‘moving target,'” the scientists right conclude. “Eradication of the founding clone and all of its subclones will be required to achieve cures.” Easier said than done, of course. The parallels between this research and studies on antibiotic resistance in bacteria are sobering. But at least now we’re starting to see what kind of evolutionary challenge we’re really up against.

(*For one very cool exception to this rule, consider the case of Tasmanian devil facial tumors, which travel from devil to devil. They evolve too, though.)

12 thoughts on “Inside Darwin’s Tumor

  1. Very interesting. I wonder if this fact about the evoluton of cancer cells, could refer to the origin of life. If our cells could turn over to be cancer cells which resemble bacterial cells in its michanisms of evolution, this could be a proof to the suscebtible fact that our body was constructed of many bacterial cells that assemble and gone through transformation any how to get eventually the eucariotic cells of today. Then when cancer occures this means that the cell comes back to it’s origin.

  2. Very interesting. I wonder if this fact about evolution of cancer cells, could refer to the origin of life. If cancer cells resemble bacterial cells in the evolution, this could mean that the origin of our cells was assembling of bacterial cells. Then when cancer take place this means that the cell come turn over to it’s origin.

  3. Catching evolution in action – fascinating!

    Evolution becomes our inner foe if mutations disable a cell’s self-restraint.

    Though I know very little about biology, when I learned to think of cancer in this way, it was an exciting new way of looking at it. We think of ourselves as one organism, but we’re made up of billions of cells, and no one says they have to toe the party line. Most of the time, our cells cannot run away with our bodies because they have no potential for differential reproduction – thus no way to compete. Cancer is (to put it perhaps over-simply) what happens when some of our cells gain the ability to reproduce on their own, thus paving the way for competition within our bodies.

  4. Nihaya: The pattern of small isolates taking over under heavy selection is a general numerical pattern of any highly variable, rapidly reproducing system. So, something analogous may have happened with regards to early lifeforms, but I would not apply it too literally to cancer – they do turn on some ancient genes but they’re still highly derived human cells.

    This was an excellent summary of the research, and I salute whoever built the figure, as well – it’s one of the best summary figures for cancer line change over time that I’ve seen. A big thumbs up to the artist responsible.

    What I would be curious to know next (and I imagine the researchers are looking at this one) is how timing of initial treatment builds into the overall picture of rare isolate takeoff. Early treatment would be expected to limit the time that is available for rare, potentially resistant variants to enter the cancer cell population, so “early is better” would seem to hold true, as usual. Beyond that, however, given that resistant forms seem to do poorly at first in many cases, it could be that many of them lose out to the common strains in some patients before the onset of treatment. As a result, the early diversity of isolates may be fluctuating heavily early on, and this might produce “sweet spots” for treatment onset, where the cancer population happens to be particularly low on diversity as a result of local, competition-related extinction events. If so, oddly enough, it might be time for cancer researchers to start poking around in the extinction literature.

  5. There’s something a little bit misleading about this diagram, a confusion between the percentages of the various cancer sub-lines and the absolute number of cells. If you look at the initial diagnosis date, it’s clear they’re representing fractions. But to the left of that they have a swooping curve that seems to indicate growth in the absolute number of cancer cells. If it really were the absolute number of cells, it would imply that the number of cells of the initial mutation (the steel-blue) increased greatly, but then decreased as the other cancer sub-lines originated and grew. The percentage declines, but not the absolute number of cells. This is a minor quibble about some exciting work.

  6. The vast number of mutations show that much of the genome is conserved, but portions are poorly conserved. The polar opposite of our normal cells, which conserve all portions of the genome.
    One wonders if THAT might be a target to attack, the poorly conserved portions, rather than the usual chemo targets. Perhaps a target in those poorly conserved regions could be triggered with disasterous results for the cancer cells. Perhaps giving those cells an advantage over their peers, with a massive vulnerability, leaving that population as the survivor, then administering a secondary drug that targets the implanted vulnerability.
    The trick is, as is typical in chemotherapy, killing the cancer whilst not killing the host.
    Since natural selection is in full operation with cancers, why shouldn’t WE be the ones making the selection? Permit only certain cancerous cells, with known vulnerabilities to survive to the final round of treatment, then utilizing the vulnerability to destroy the surviving cancerous cells?

Leave a Reply

Your email address will not be published. Required fields are marked *