Moving cancer research out of the Petri dish and into the third dimension
Galen of Pergamos was a physician with a practice in second-century Rome. Among his clientele was the Roman Emperor Marcus Aurelius. Galen was also a prolific scribbler, so we know a lot about his ideas regarding disease, including his theory of cancer. Cancer, he wrote, was caused by "black bile" that flowed through the body; when it became trapped somewhere, it formed a malignant tumor.
He was wrong about black bile, though it is one hell of a good metaphor. But he was strikingly close to the mark with the flow theory. There are cancers, such as glioblastomas in the brain, in which the primary tumor can be deadly. But for most cancers, the original tumor does not pose the mortal peril. In more than 90 percent of cancers, what kills is metastasis. Cancer cells have a terrifying ability to move through the body and form new tumors in the bones, in the lymph nodes, in the lungs, in the liver and other internal organs. If a physician finds your tumor before the cancer has spread, you may survive. If the tumor has metastasized, cancer will probably kill you. Medicine still cannot do much to counter the flow of black bile.
What if that is, in part, because a substantial portion of cancer cell biology and cancer drug testing has been reliant on a ubiquitous piece of lab equipment?
German bacteriologist and military physician Julius Richard Petri invented the Petri dish in 1887. Unless it was invented two years before that by a Slovene, Emanuel Klein, or by a pair of Romanian microbiologists, André Cornil and Victor Babes. Unless it was invented a year before that—we're back to 1884 now—by English researcher Percy Faraday Frankland. Whatever its provenance, the two-piece flat cylindrical glassware (now frequently polystyreneware) has been used by scientists for decades to culture and study cells of all kinds, including cancer cells.
Denis Wirtz is a professor of chemical and biomolecular engineering in the Whiting School and director of the Wirtz Lab in the Physical Sciences–Oncology Center. He has dedicated the past few years to developing methods of studying cancer cells in three-dimensional environments. In a Petri dish, cells are cultured on a substrate so thin as to be two-dimensional. Wirtz believes that this 2-D microworld-in-a-dish so distorts the cells and cell behavior as to cast doubt on a significant portion of critical cancer biology. He and his research team have been working with 3-D matrices that more resemble human tissue, growing cancer cells in them and observing how those cells move about. The difference has been so dramatic to Wirtz that when he talks about it, he becomes an evangelist for cell biology in three dimensions. To figure out metastasis, he says, scientists must work in 3-D. And it would be a good idea to take hundreds of drugs that were deemed failures after testing them on cells in a dish and test them again in three dimensions. Wirtz thinks pharmaceutical companies may have missed medicines that will work because of their reliance on Herr Petri's invention.
In journal articles, cancer researchers refer to "the metastatic cascade." It is a remarkable process. At the disease's point of origin, cancer cells proliferate and clump, creating the primary tumor and forming blood vessels to nourish themselves in a process called vascularization. Before long, malignant cells begin to detach from the original tumor and move through the surrounding tissue until they run into a blood vessel. Able to deform themselves, they squeeze between the endothelial cells that form the blood vessel's wall and enter the bloodstream (this is known as intravasation). As the heart pumps the blood, it pumps the cancer cells as well to other parts of the body. The cells tumble and bump into the blood vessel's walls, and when that happens some of them adhere and push between the endothelial cells once again and exit the bloodstream (extravasation). Now they are inside a lung, or the liver, or some other organ. Malignancies constantly shed cells by the millions, and almost all of them die before they can do more harm. But enough survive the rough ride through the blood to lodge somewhere new and begin the explosive proliferation characteristic of cancer.
Crucial to understanding this process is figuring out cell motility—how cells move. For many years now, researchers have studied motility usually by growing cancer cells in a Petri dish and observing how they move about. When cultured on a substrate in a dish, cancer cells typically flatten and move in a slow, seemingly random fashion by pulling themselves along by their leading edges. They also form strong attachments, called focal adhesions, to the floor of the dish. These adhesions consist of proteins that aggregate on the bottom of the cell. Up to a few years ago, that was the picture scientists had of how cancer cells move.
In 2010, Stephanie Fraley, then a doctoral student in Wirtz's lab, was studying motility, but not in a Petri dish. She had become curious about the possibly distorting effects of the planar environment of the dish. What would happen if she placed cancer cells in something that more resembled human tissue? From a fibrosarcoma cell line called HT-1080, she took cells and inserted them in a gel prepared from collagen I, the protein that forms most connective tissue in the human body. The gel, formed in a cylindrical well, was only a few millimeters thick, but that was enough to constitute a three-dimensional environment for something as tiny as a cancer cell. Then Fraley watched what happened.
What she saw was startling. For one thing, the cells were no longer flat. They were more spherical, with long protrusions at each end that had not been observed in a dish. The proteins that correlate with the metastatic potential of cancer cells and, in a dish, were located mostly on the bottom of the cell, now were diffuse throughout the cell. Focal adhesions barely existed. The cells did not crawl along in the laborious, erratic fashion observed in two dimensions, but moved rapidly through the 3-D environment by first extending protrusions fore and aft and anchoring them in the collagen matrix, then contracting like springs and releasing one protrusion to snap the cell in the opposite direction.
Fraley's work suggested that much of how cells behaved in a Petri dish was an artifact of the 2-D environment. The cells moved as they did not because that's how motility works in cancer cells but because the cells were in a dish. Put them in a 3-D environment and everything changed. To Wirtz, the implications were staggering. If cells in a 3-D matrix similar to cancer's actual environment behave so differently from cells grown in a dish, then much of what scientists thought they knew about motility, which is central to metastasis, had to be reconsidered.
For decades, various cell biologists had been thinking about experiments using various extracellular matrices, or ECMs, because cancer cells move through the human body's connective tissue, which is an ECM. As far back as 1980, a paper by three UCSF Medical Center researchers in the journal Cell noted that tumor cells "are more likely to resemble their in vivo counterparts when maintained on an extracellular matrix than on plastic." A National Institutes of Health investigator, Kenneth M. Yamada, published an influential paper 21 years later in Science titled "Taking Cell-Matrix Adhesions to the Third Dimension." It became the most-cited paper to date by Yamada, whom Google Scholar ranks second among all cell biologists in research citations. Wirtz, too, had been pondering the implications of working in three dimensions. Fraley's results convinced him that 3-D was going to be transformative for cancer research. "Stephanie's paper was the trick," he says. "There was no turning back. You forget that in research sometimes the reason you are doing things [a certain way] is for convenience, not because it faithfully reproduces what we already know from studying cancer in vivo."
Scientists know how to work in a dish. Electron microscopy and other important research tools work well only in a dish. Vital sources of research dollars, like the National Institutes of Health, fund studies of cells in a dish. Pharmaceutical companies have developed sophisticated automated processes that screen cancer drugs by testing them on malignant cells—in a dish.
The Petri dish has influenced the fundamental direction of much cancer research, in Wirtz's view. Because the two-dimensional environment of the dish lends itself much more to studying cancer cells' explosive proliferation than motility, that's what scientists have studied. Wirtz argues that the result has been a diversion from the vital study of how cancer cells migrate. Roger Kamm, another oft-cited researcher and a professor of biological and mechanical engineering at MIT, says, "I agree, and would go a little further. It's not only migration that matters, but all steps in metastasis: epithelial-mesenchymal transition [vital to enabling cancer cell invasion], intravasation, extravasation. And all of these involve studies that cannot be done by standard cell culture techniques. The bottom line is that we now have excellent methods for controlling primary tumors, but have precious little knowledge about how to prevent cancer cells from spreading to remote sites."
"Ninety-five percent of funding from the National Cancer Institute is about tumor shrinkage," Wirtz says. "Why? Because we can see it! Because then everyone is happy! Pharmaceutical companies are happy, they're selling stuff. The scientists are happy, they're publishing, they're getting funding. The patients are happy, for a while, until it doesn't work."
Wirtz adds, "We've obsessed about proliferation. The first statement in any textbook about cancer is that cancer is a disease of high proliferation. I say this is completely wrong! Completely wrong! We are discovering that often the very cells that successfully metastasize are those that proliferate the least. We're not trained to think about metastasis because it's harder. We have blinders to the point where we don't even think about blinders."
Wirtz is slender, boyish in appearance, bespectacled. When he gets excited, his accented English becomes emphatic and he uses his hands a lot. He has formidable instant command of detail when talking about his work, but says he is entirely reliant on his assistant, Tracy Smith, to keep him on schedule and tell him where he needs to go. His office and lab are in Croft Hall on the Homewood campus, and one day as he walked to a meeting for which he was already late, he had to ask for directions to Gilman Hall. Teased about this, he held up his hands and said, "I've only been here 18 years! How am I to know?"
He works in cancer cell biology but is not a biologist. "I've yet to take my first course in biology," he says. His background is in physics, which he studied as an undergraduate at the Free University of Brussels. He says he picked that course of study because he thought physics afforded the best opportunity for graduate study in California, where he wanted to live for a while. "I figured I'd do that for a couple of years and then go back to Belgium and be a nuclear physicist, or something like that. I had no intent to stay in the U.S. or apply physics to biology. None of it." He did indeed make it to California when he went to Stanford for a doctorate in classical physics. His PhD adviser, Gerald Fuller, steered him toward study of the long molecules known as polymers. Looking for polymers to investigate, he veered toward biology. "DNA, to me, was a polymer," he says. "The laminate that makes up the nucleus was a bunch of polymers. I thought, 'Wow, there are polymers everywhere! I'm going to have a fantastic time!'"
He studied physics at Stanford within the chemical engineering department because the university's physics department was more oriented toward classical physics. When he found a job at Johns Hopkins, it was in chemical and biomolecular engineering. This provided a path into cancer biophysics and connections to Johns Hopkins oncologists, pathologists, and cancer cell biologists who, he says, put him on to the right questions to ask. "There's a beautiful back-and-forth between biology and physics," Wirtz says, though 10 years ago he could not have predicted that because physicists had been reluctant to dive into the complexity of cells. "They think there's beauty in simplicity and bringing phenomena down to fundamental interactions and universal behavior." Cells vary in such complex ways, universal behavior is hard to discern. To a biologist, there's beauty in complexity; physicists took a different view, Wirtz says. "We hate the alphabet soup of proteins. We hate this diversity of cells. I don't, I love it, but most engineers and physicists tend to be kind of pushed away from it. It's too bad because physicists and engineers are the best trained to handle complexity and extract what really matters."
Wirtz looks at cancer metastasis as a mechanistic process involving forces that engineers and physicists are well-equipped to study. For example, the extracellular matrix that cancer cells inhabit subjects them to confinement forces, especially as the cells proliferate and become more densely packed in the tumor. When cancer cells force their way through the walls of blood vessels to enter the bloodstream, that subjects them to more compression forces, as does the reverse process of the cells pushing out of the blood vessels into other organs or tissues. During their migration through the circulatory and lymph systems, the cells are subject to shear forces. All of that can be studied as physics and engineering. "Ask biologists what are the units of force, and they haven't a clue," Wirtz says. "They know what a force is of course, but they don't know how to measure it." Enter the physicists. Wirtz recalls attending meetings of the American Society for Cell Biology 10 or 15 years ago and finding two or three physicists like himself. Now when he attends he finds several hundred, he says, and more and more biologists who are learning physics.
Wirtz Lab currently has a dozen doctoral students and seven postdocs; among them are electrical engineers, biophysicists, biologists, and a physician. Wirtz has turned the focus of the lab entirely toward the study of cancer cells in 3-D, with what he describes as dramatic results. For example, when a tumor grows in a human body, the tissue around it tends to stiffen. Often this stiffening is what the fingers detect when someone first finds a lump that turns out to be a malignancy. "Cells in 2-D dishes tend to migrate toward stiffer surfaces," Wirtz observes. "The cells have a sense of touch called mechanosensing, and cells on flat surfaces love to go from soft, pliant surfaces to a stiff surface. That's called durotaxis." But that poses a paradox: If cancer cells prefer stiffer surfaces, and tumors stiffen the tissue around them, why do the cells migrate to softer tissue in the body to form new tumors? Why don't they stay in the stiff collagen capsule that is home? When Wirtz observed migration in three dimensions, cancer cells did the opposite of what they do in a dish: "Cells will tend to go to the softer part of a 3-D gel from the stiffer part. So what we had learned in 2-D does not translate in the 3-D case. Differences in stiffness might indeed promote metastasis by promoting cell migration, but a 2-D experiment will never show that."
Much of the Wirtz Lab's research in the last few years has been figuring out how to do its research. "Life is hard in 3-D," Wirtz says. "Every measurement that we take for granted in 2-D, like how to measure the protein content of a cell, or protein activity, or protein localization, the shape of cells, cell motility—all of those become so much harder, if not impossible." Conventional electron microscopy becomes impossible in three dimensions. To achieve high magnification and high resolution, you have to work at very short distances of under 1 millimeter. Comparable three-dimensional microscopy will require a lens that does not yet exist. With current technology, researchers can't even locate the nucleus or mitochondria in 3-D. "It's just a moving blob of light, basically," Wirtz says. "We don't have the subcellular resolution that we have come to take for granted in 2-D."
He is especially interested in how malignant cells create and exploit pathways out of the original tumor site and through the body to secondary sites. The tissue matrix in which a tumor begins to form may feel soft and pliant to the touch, but at the cellular level it's dense, and this density inhibits the proliferation and motility of cancer cells. But those cells somehow effect structural changes in the collagen. Ordinarily, within collagen are fibers that are evenly distributed with no favored orientation—they are every which way. Cancer cells aggregate the fibers and orient them into thick strands that Wirtz calls "freeways from the tumor to the blood vessels." To travel these freeways, the malignant cells must still get through the dense collagen matrix. So they secrete enzymes that have no effect on the fibers but digest the surrounding collagen, clearing the way. It's as if the cancer cells were hiking through a jungle, and to follow the path at their feet, they hack through the dense jungle growth. All of this needs to be studied in three dimensions, Wirtz says, because none of it happens when cells are grown in a Petri dish.
When the topic moves to drug screening in 2-D environments, he says, "Before drugs are tested in clinical trials, the way pharmaceutical companies identify new compounds is to develop thousands, tens of thousands, hundreds of thousands of compounds, and then subject cancer cells to them in dishes to see how proliferation is affected. But that's presuming the 2-D case is somewhat relevant to cell proliferation in 3-D." This brings him to the example of paclitaxel, a chemotherapy drug marketed as Taxol. Tested on cells in dishes, it was found to have little effect on motility, but some ability to inhibit cell division when applied in high doses. So it has been used to treat ovarian, breast, and lung cancers for 50 years. Test paclitaxel on cells in a 3-D matrix, Wirtz says, and you get much-different results. In 3-D, the drug has little effect on tumor growth but does inhibit migration. These results, Wirtz argues, suggest that the drug should be rethought. Stop using high dosages in an often futile attempt to stop growth of the primary tumor, and start using more targeted dosages to inhibit metastasis while the primary tumor is attacked by better weapons.
Wirtz says, "When you do a conventional drug screen, you end up with maybe 10 candidate drugs that are really the big killers of cancer cells in 2-D. All right. Then you move on into years of development not only to produce this in large quantities but do clinical trials in mice, blah blah blah blah." Next are hugely expensive human trials. "Then eventually most of the drugs fail. So there is all that lost development cost, based on that very first promising 2-D screen." A 2-D screen that, perhaps, could not produce meaningful results because the dish imposes so many changes on cell morphology and behavior. Conversely, a bad outcome in a 2-D experiment could lead to rejection of a drug that actually might be found to halt metastasis when tested on cancer in three dimensions. Wirtz believes many drugs abandoned after failed Petri dish tests should be revived and tested again in 3-D matrices. There might already be something out there that can stop a prostate or breast tumor from exploding throughout the body.Wirtz is moving full speed ahead on his conviction that three-dimensional matrices are essential to cancer cell biology. He is not alone in this conviction. Donald E. Ingber of the Wyss Institute at Harvard wrote in the journal Trends in Cell Biology that 3-D microenvironments still require validation but "could have profound effects on drug discovery and environmental toxicology testing." (Ingber has been working on a marriage of microbiology and microtechnology to produce more sophisticated matrices he calls organs-on-chips.) Kenneth Yamada continues to work on and extol the value of 3-D matrices. At Johns Hopkins, a number of researchers are engaged with Wirtz on various research collaborations. Winston Timp, an assistant professor of biomedical engineering in the Whiting School, says, "It is vastly underestimated how important this is, especially when it comes to motility. A 3-D microenvironment lets us bridge this gap between information we have from 2-D and the animal or human model. The more we can do in an in vitro environment faster and more effectively, the more we'll be able to get to better drugs to help human health, better screening that identifies diseases, and more information for plumbing the unknown in biology." Various researchers point out how much more complicated the science becomes when you add the third dimension. "The beauty of the Petri dish is that it is a reductionist approach," Timp says. "It's easy to do and you get results which make sense. When you move stuff to three dimensions, everything becomes a lot more complicated." Technology and protocols relied on for many years no longer work well. The number of variables that need to be controlled for in 3-D scales up fast. And while some commercially available 3-D matrices now in use greatly improve on the dish, they remain a long way from replicating the remarkable complexity of actual human tissue (though some research groups are getting closer, Timp says). Stephanie Fraley, now a postdoctoral research fellow in the School of Medicine, further cautions that collagen gels and other experimental matrices may create artifacts of their own that distort observations and research results. And a model will always just be a model. As John Isaacs, a professor of oncology and cancer biology in the School of Medicine, says, "I have always lived by the belief that 'a model is a lie which can tell you a lot about the truth.'"
Timp, for one, isn't ready to toss out every Petri dish in the lab. "It makes sense to do the easiest experiment you can that will give you the information you want," he says. "So sometimes there's value in doing things in 2-D first. That said, validation in a 3-D environment is still a vital confirmation that the 2-D experiment is real." Then he laughs and adds, "You have to avoid the tendency to do the shiniest experiment. Because 3-D is cool, there's no question about it. It's incredibly cool."
Scientists commonly face a fundamental dilemma that Wirtz cites in talking about cancer cell biology: There is what can be measured, but then there is what should be measured. Andrew J. Ewald, an assistant professor of cell biology and oncology in the School of Medicine, says, "I think Dr. Wirtz is exactly right when he says that scientists do what they can get funding to do, and that enough biology is turning out differently in 3-D tissues than in 2-D Petri dishes [to indicate] we have to revisit a large fraction of our existing conclusions. The challenge for funding is that this often means proposing to retest existing models and hypotheses, something study sections [that review grant applications] are loath to do."
Wirtz says, "I am not in the business of wanting to prove people wrong. We are not in the business of dismantling knowledge. But here we are, years into it, discovering again and again how many things we took for granted."