It could be nature’s cruelest joke—the molecules that give us shape at the beginning of our life also can lead to the end of it.
The genes and proteins that help sculpt a single cell, the fertilized egg, into a complex multicellular organism are also responsible for the birth of many cancers.
This link between embryogenesis and tumorigenesis has been suspected for more than a century, but only within the last 20 years have the striking molecular similarities between these two monumental events come into focus.
“These developmental molecules that cell biologists know and love—our favorite proteins that operate in the early embryo—are the same molecules that seem to go haywire in cancer,” says Jason Jessen, Ph.D., assistant professor of Medicine and Cancer Biology at Vanderbilt University Medical Center.
The rapid and exponential cell division, differentiation and cell movements that characterize embryonic development bear a close resemblance to those involved in tumor initiation and metastasis.
“It does make sense because in the developing embryo, so many things are happening: cell migration, cell specification, cells interacting with each other,” Jessen says. “So, if those proteins get activated in an adult cell, it’s no wonder it can have dire consequences.”
Definitive links between the two processes have been a long time coming, mostly because of the academic divide between the separate cultures of developmental biology and cancer research. Now that divide is beginning to close.
One of the first links between cancer and embryogenesis was made in the early 1980s.
Roel Nusse, Ph.D., and Harold Varmus, M.D., identified a cancer-causing “oncogene,” which they called int-1, in a mouse model of breast cancer. When int-1 is activated or turned on by the “integration” (thus its name) of a mouse mammary tumor virus into its DNA, a tumor forms.
Around the same time, Christiane Nusslein-Volhard, Ph.D., and Eric Weischaus, Ph.D., who were studying development of the fruit fly, Drosophila, found that a gene they called wingless was involved in setting up the polarity of the embryo. When the protein encoded by the gene is defective, the fly fails to develop proper body segment boundaries—and wings.
Nusse and colleagues soon determined that the two seemingly unrelated genes were homologs—genes similar in structure, function and evolutionary origin, and found throughout the animal kingdom. So the names (int-1 and wingless) were combined, and Wnt was born.
“I can imagine that they were thinking: ‘What in the world is this cancer protein doing in a Drosophila embryo?’” says Jessen. “It’s a classic example of the two worlds coming together.”
Wnt is probably best known for its involvement in one of the earliest aspects of development, the formation of the primary body axis.
“You have a ball of cells, and somehow this signal tells the ball of cells which parts form the head and which form the tail,” says Ethan Lee, M.D., Ph.D., assistant professor of Cell and Developmental Biology at Vanderbilt who studies the Wnt pathway in frog (Xenopus) embryos.
In 1989, Andrew McMahon, Ph.D., and Randall Moon, Ph.D., demonstrated that injection of the int-1 (Wnt-1) gene into Xenopus embryos induced the formation of a secondary axis, resulting in a two-headed tadpole.
“This was critical because it was the first biological assay for int/Wnt-1, and it linked a proto-oncogene to a developmental process,” he says.
This multipart pathway has a number of other developmental roles in patterning the brain, heart and limbs and possibly in stem cell differentiation. Additionally, mutations that activate the Wnt pathway have been linked to cancers of the colon, skin, blood, liver and several other tissues.
One of the strongest links between Wnt and cancer was revealed with the discovery that mutations in the APC (adenomatous polyposis coli) gene—a component of the Wnt pathway—were required for the initiation of colon tumors.
Researchers have already identified more than a dozen Wnt ligands (proteins that bind Wnt receptors and initiate signaling), and new components of the pathway are being cloned and added to the already complicated system at a rapid pace.
“It really looks like a mess,” Lee says. “I would say we are basically ‘stamp collecting’ right now, putting together a picture that is very complex.”
Finding the switch
To make sense of the overwhelming data on this pathway, Lee and colleagues developed a mathematical model to examine how the pathway is regulated. They found that a protein called axin may be the limiting factor.
“Based on the model, we can propose that controlling axin levels and its turnover is the major way by which the pathway can be regulated,” Lee says. “Perhaps that is the way the pathway can be turned on or off.
“Interestingly, the genes that this pathway turns on are classic examples of proto-oncogenes,” he says. Researchers believe that if they could find a way to selectively switch the pathway “off,” they could halt tumorigenesis.
Toward that goal, Lee and colleagues are taking advantage of high-throughput screening methods and an in vitro model based on extracts of Xenopus embryos that Lee developed as a postdoctoral fellow.
“We were able to recapitulate the pathway in a test tube,” says Lee. This method provides an efficient tool to screen for molecules that either inhibit or activate the pathway.
“We’re now using this assay to do drug screens,” Lee says. “The idea is that if any of these molecules work out, they could be used as tools to study the pathway—and further down the line, as potential drugs to inhibit the pathway.”
Lee is screening the large catalog of small molecules available through the Vanderbilt Institute of Chemical Biology, as well as extracts from medicinal plants and herbs through Harvard University’s high-throughput screening facility.
Although currently there are no chemotherapeutic drugs that specifically inhibit the Wnt pathway, pharmaceutical and biotech companies have taken an interest.
“In the future, I think you’ll see more companies targeting developmental pathways with the realization that they play a role in cancer,” Lee predicts.
Jessen is also attempting to unite the worlds of developmental and cancer biology with the help of a tiny tropical fish.
Zebrafish have a long history in developmental biology research. Their embryos are transparent and develop outside the mother. They also develop rapidly and are inexpensive to maintain, making zebrafish embryos an efficient model for studying genetic and environmental factors that influence early development.
While indispensable for development research, they haven’t been widely used in cancer research—yet.
As a postdoctoral fellow in the lab of Vanderbilt developmental biologist Lila Solnica-Krezel, Ph.D., Jessen realized that one of the key events he was studying in zebrafish—gastrulation—might offer some insights into the aspect of cancer that is the most deadly—metastasis.
Cell migration
Gastrulation is a time in early development when an initially amorphous ball of cells begins taking on its adult shape due to rapid and extensive cell movements. Metastasis also is characterized by cell movements—cancer cells break off the primary tumor and spread throughout the body.
“My main interest is trying to understand the fundamental migratory differences between metastatic (invasive) tumor cells and primary (non-invasive) tumors,” Jessen says. He is currently searching for molecular signals in the Wnt pathway that might underlie cell motility during both metastasis and zebrafish gastrulation.
A major goal of the Jessen lab is to develop a model of melanoma, a deadly form of skin cancer, by transplanting human melanoma cells into zebrafish embryos. The idea of using the zebrafish embryo to model a disease that afflicts humans is very exciting, Jessen explains.
“We can manipulate the embryonic environment to determine what kind of environmental cues (such as the Wnt pathway) might influence tumor cell migration.”
An added benefit of zebrafish cancer models is the ease and cost-effectiveness of doing in vivo drug screens.
“What’s interesting is that fish tumors look very similar to human tumors,” Jessen says. “And you can bathe (zebrafish embryos) in chemicals and look for molecules that inhibit or promote growth of the tumor.”
Jessen is also combining developmental biology and cancer research through a more traditional approach, using the zebrafish embryo to determine how proteins associated with cancer and metastasis regulate cell migration normally, such as during gastrulation.
“It is important to remember that for the majority of cancer proteins, we know very little about how these proteins function to control basic cellular activities such as motility,” Jessen says.
While both cancer and development are exceedingly complex processes, insights about developmental pathways like Wnt are slowly beginning to reveal the genetic underpinnings of tumorigenesis.
“It’s so complex that no one lab or company is going to come up with the answer to cancer,” says Jessen. “I see myself as trying to fill in some of the key gaps in our knowledge.”
Another pathway involved in both embryonic development and cancer is the “Hedgehog” (Hh) pathway. First identified in fruit flies, it is named for the “short and prickly” appearance of fly embryos that have an abnormal Hh protein due to a genetic mutation.
Like Wnt, Hh is a secreted signaling molecule involved in the patterning of the embryo. Also like Wnt, Hh appears linked to the formation of tumors in those tissues where it is required for development—the cerebellum, foregut, prostate, skin and lung, for example.
In some cases, the Hh protein “may tell some cells to proliferate,” says Michael Cooper, M.D., an assistant professor of Neurology at Vanderbilt. In other cases, “it may tell cells to differentiate along a certain lineage… to become a motor neuron,” for example.
The exact instruction imparted by the Hh signal depends on what type of cell is receiving the signal and the location of that cell.
“Hh regulates a number of cell types, but most importantly, stem cells or progenitor cells,” says Cooper, who is studying the role of Hh in primary brain tumors called gliomas. “We’ve learned that these pathways regulate not only stem cells in development, but also in tumorigenesis.”
Stem cell theory
Most tissues have stem or progenitor cells well into adulthood. When adult tissues like the epithelium require repair or renewal because of an injury or normal cellular turnover, these cells most likely respond by activating or reactivating the developmental pathways that led to that tissue’s formation in the first place—pathways like Hh and Wnt.
“Stem cells or progenitor cells can respond to Hh by self-renewing, that is by dividing to form more stem or progenitor cells,” Cooper says. “In a setting where mutations can occur and accumulate, the process becomes dysregulated, and the self-renewal process may become turned on in a way that it can’t be turned off.”
When the pathway can’t be turned off, the anomalous cell divisions can lead to tumor formation. This is known as the “stem cell” theory of tumorigenesis. Indeed, the Hh pathway appears to be activated in many cancer types, particularly within cells that have a stem- or progenitor-like appearance.
While a research fellow at Johns Hopkins University, Cooper was examining how compounds known to cause birth defects (teratogens) interfere with Hh signaling. The research unexpectedly pointed towards the Hh pathway as a possible chemotherapeutic target.
Working with Philip Beachy, Ph.D., professor of Molecular Biology and Genetics at Johns Hopkins, Cooper was looking at how the teratogenic compounds jervine and cyclopamine cause a range of birth defects of the face and brain, from mild holoprosencephaly, such as cleft lip, to the most severe and fatal form of holoprosencephaly, cyclopia (development of a single, centrally-positioned eye).
Beachy and colleagues had demonstrated earlier that cholesterol played a critical role in Hh signaling during development; to be active, the Hh protein must be cleaved and one end of the protein modified by cholesterol. If this modification was inhibited, birth defects such as holoprosencephaly resulted.
Cooper suspected that these teratogens, whose chemical structures are similar to the structure of cholesterol, were somehow interfering with cholesterol modification of Hh, thus inhibiting Hh signaling required for development.
“We thought that we had this mechanism all figured out before we’d done a single experiment,” Cooper says. “But we were wrong. And it was the most spectacular mistake ever!”
They eventually determined that the compounds inhibited Hh signaling not by interfering with cholesterol modification in the Hh-generating cell, but by inhibiting receiving cells from responding. And because the Hh pathway was known to be activated in a number of cancers, it immediately became clear that these chemicals, which can produce such horrible birth defects, might have some redeeming value in treating cancer.
Since this discovery, Beachy and colleagues have demonstrated the effectiveness of cyclopamine against several tumor types in animal models.
“So far, those tumors types that require Hh signaling for their growth shrink in animals treated with cyclopamine,” Cooper says.
Cyclopamine and related compounds are now being investigated as possible chemotherapeutic agents by pharmaceutical companies.
“This has become a spectacular molecule—not only is it interesting as a biological tool, but it may have therapeutic value in treating tumors,” says Cooper.
Whether or not the revelation that embryonic development and cancer share fundamental pathways results in new therapeutic treatments for cancer, the traditional “departmental” borders that defined research in years past have been broken down.
“Cancer biology and developmental biology have traditionally been two separate fields,” says Cooper, “but now they have intersected in a wonderful way.”
Researchers at Vanderbilt University Medical Center have found that attachment of cholesterol to Sonic hedgehog, a mammalian version of the Hedgehog protein, controls finger and toe development in mice. The paws of a normal mouse embryo are shown in panels A and D. Mice lacking cholesterol-modified Sonic hedgehog (panels B and E) have malformed digits, while those expressing half the amount of Sonic hedgehog without cholesterol (panels C and F) develop extra, ectopic digits. In addition to directing development, the Sonic hedgehog pathway—named for the video game character—is also involved in a number of human conditions, including cancer.
© 2006 National Academy of Sciences, U.S.A.