Biologists Tyler Jacks and Jonathan Weissman’s mice may look like any other lab mice — plain, pale, and furry — but cancer geneticists see them as a “technical tour de force.”
“It’s a pretty heavily engineered mouse,” said Weissman, a biologist at the Whitehead Institute.
Engineered into that mouse is intricate molecular machinery that allowed Weissman and his team to trace the ancestries of single tumor cells, placing them on a tumor’s family tree with an unprecedented level of detail. With those findings, scientists say that Weissman, in a new paper published this month, has begun to uncover new insights into the fundamental biology of tumors and how they evolve.
“It’s like a molecular flight recorder like the black box on a plane,” Weissman said. With this technology, developed by his lab and others, “you can reconstruct the history of how a tumor evolved and became aggressive, even try to understand the vulnerabilities that a tumor has at a very early stage.”
The detailed tumor evolution in the study is something that’s completely new to cancer research, said Kamila Naxerova, a systems biologist at Massachusetts General Hospital who studies tumor evolution and was not involved with the work. “There’s no comparison,” said Naxerova, one of several outside experts who called the work a tour de force. “This is something now uniquely enabled by the methods they used here — how different cancer cell states change over time as the tumor lineage evolves? There’s no prior data.”
Malignancies begin with a normal cell quietly accumulating cancer-causing mutations and passing them on to its pre-cancerous children. This lesion can fester for years, but over time these mutations can twist the shapes and structures of healthy tissues into a jungle of deformities. Often once a tumor is found, it’s an array of warped cells, many with vastly different appearances and features, all cobbled together like a cursed mosaic.
The diversity of tumor cells is part of cancer’s threat. Some of those cells can — and often do — pick up distinct abilities like growing aggressively, spreading across the body, or resisting certain therapies. The technology inside Weissman’s mice helps to disentangle some of the details of that evolution by creating a molecular ancestry record in every cancer cell.
This works by using a protein called Cas9, which scientists use to engineer mutations into genetic material, and engineered rolls of DNA inside the mouse’s cells that serve as a molecular record book. The scientists engineered their mice to carry these components, and together they make a molecular machine. The protein progressively scribes mutations into this DNA record, which then get passed down to all the cell’s descendants. “The Cas9 is constantly online, randomly making marks at different locations in that sort of scratch pad,” Weissman said. “So, when two cells share the same mark in the same place, that’s because it happened in an ancestor.”
Once the researchers engineered these mice, they forced the mice to inhale a virus that activated the molecular record book as well as common cancer-causing mutations in the mouse’s lung cells. “Then you wait months and over time, a subset of them grow and form a mass,” Weissman said. “Some of those cells get additional changes and become cancerous, and we’re able to follow how they evolved.”
At the end of the study, the investigators took the tumors out of the mice and analyzed them one cell at a time, looking for two main things. First, they looked at which genes were turned on in each cancer cell. That gave the scientists an idea of physical traits the cell had and whether it bore any resemblance to other healthy cell types — like other lung cells, gastric cells, or cells from other tissues.
Then, the researchers looked at the genetic record book to trace the cell’s lineage and place it on the tumor’s genetic family tree — or phylogenetic tree. That allowed them to reconstruct the evolutionary history of the tumor and pinpoint certain changes in time. For instance, if a mutation is shared across all branches in the tree, it must have occurred in the very beginning for it to have been passed down to every cell. Mutations present in only a few cells must have happened very recently — and the cells sharing them should also be closely related cousins or siblings.
“They were really able to analyze not only the cell state — the different flavors of cells that develop in tumors — but also which are related to each other,” said Monte Winslow, an associate professor of genetics at Stanford University who did not work on the study. “That’s really important. You know who is related to who. It really allowed them to understand the states that the cancer cells progress through — and some are quite interesting.”
In particular, Weissman and his colleagues found that cancer cells tend to move through a few main evolutionary pathways, starting with a malignant form that appears to be relatively slow-growing. In this study, that early stage resembled a type of lung cell — likely because this particular cancer model started in the lung. Then, at some point, the tumor cells begin to change more starkly.
“There is this period where the tumor is very plastic and changing back and forth between multiple cell types and states,” Weissman said. “Some look like lung cells or have a signature that gastric cells have.”
Then, he said, through all that shapeshifting, some of the tumor cells luck upon a state that makes them more aggressive and more likely to spread. Those cells then quickly proliferate and take over large swaths of the cancer’s family tree. In Weissman’s model, these cancer cells resembled mesenchymal cells, which play a normal role in wound healing. Picking up mesenchymal-like traits may be key to cancer’s ability to metastasize, since these cells have an ability to creep around the body.
“Normally, cells don’t migrate from where they started, but during wound healing, these mesenchymal cells become activated to fill in wounds,” Weissman said. “The tumor takes advantage of this natural transition to allow it to grow and leave its original spot and ultimately metastasize.”
When Weissman mutated additional cancer-forming genes in the mice, that pathway to aggressive growth and metastasis looked notably quicker, he said. “It’s like a detour – like a shortcut to get to the final, aggressive state,” he said.
The process fits with some hypotheses that scientists have made about tumor evolution, Mass. General’s Naxerova said. For one, while cancer cells may be flipping between cell states, there are a limited number of pathways they can take to becoming aggressive and metastatic — and those pathways may differ based on the initial cancer-causing mutations.
“It’s a huge problem in cancer research or treatment, that you have a colon, lung, or breast tumor and the therapies that work for one won’t work for another. The biology is not the same,” she said. “To see it’s not just the overall cell state but the trajectory is different because of that baseline mutation. It’s expected, but kind of cool to see it really happens like this.”
With further study, this kind of work may help further explain why some extremely high-risk mutations — like BRCA 1 or 2 for breast and gynecological cancers — can lead to such aggressive tumors, added Stanford’s Winslow. That question, like many others in basic cancer science, is still unanswered, but Winslow said the methods demonstrated in Weissman’s paper offer an alluring new way to discover that knowledge.
“This is a great example where an amazing new technology and technical tour de force can lead to new insights,” he said. “There’s a lot of future potential to answer questions about other cancer types, different cancer genotypes, and how different therapies or drugs might affect cancer evolution.”
That, Winslow said, might ultimately lead from this foundational science to information that could make a difference in the clinic. “We don’t know how things work,” he said. “When we make these discoveries, you find something really new in this or that, you have an opportunity that it will ultimately help patients.”