“When she was good, she was very, very good,” wrote poet Henry Wadsworth Longfellow of his girl with the curl in the middle of her forehead. “But when she was bad, she was horrid.” Much the same could be said of glial cells, the “glue” of the human brain.

Glial cells handle the behind-the-scenes chores that keep the neural system humming. A living scaffold that envelops the neurons, they insulate the synapses. Glial cells also regulate metabolic activity in the brain, provide quick bursts of fuel for hard-firing neurons, regulate neurotransmitter and ion levels, and maintain the blood-brain barrier, which protects against pathogens.

Fully half of the volume of the human brain consists of glial cells (the other half is made up of neurons) and they grow fast, supported by a rich network of blood vessels. Gone haywire, as in the malignant tumors known as glioblastomas, they can wreak havoc. The most common and aggressive form of brain cancer, glioblastoma has a dismal prognosis. Doctors measure a patient’s prospects in months rather the years and offer surgery—if possible—to buy a little time. Median survival for people with glioblastoma diagnosis is just 14 months; few survive more than two years, and the five-year survival rate is only 4 percent (compared with 97 percent for some forms of prostate cancer).

Stark brain tumor survival statistics were a given when Anna Lasorella, MD, was a medical student and later a pediatric oncologist at the Catholic University School of Medicine in Rome. It was terrible,” says Dr. Lasorella, now associate professor of pathology & cell biology and of pediatrics at P&S. “All of the children were dying. The treatment was having no effect.”

“We knew the type of drugs we gave to these children probably wouldn’t work,” says Antonio Iavarone, MD, Dr. Lasorella’s husband and collaborator and now professor of neurology and of pathology at P&S. His investigations of the cellular biology of brain tumors date back to the late ’80s when he was a trainee in pediatric oncology at the Catholic University School of Medicine. “That was the main reason I chose to do research on brain tumors. If you don’t do research at a high level, it’s basically impossible to expect a change in the likelihood of a cure.”

Today, Drs. Lasorella and Iavarone are optimistic that prospects will improve for people diagnosed with glioblastoma. In September 2012, the journal Science published their finding that a small subset—3 percent—of glioblastoma cases are caused by a unique genetic fusion, essentially a splicing error introduced during stem cell division. When the fused gene was transplanted into the brains of mice, 90 percent developed glioblastoma. Within eight months, the mice had died from the tumor. Testing a variety of pharmaceutical compounds, the team found one that targets the protein produced by the rogue gene. Mice treated with the compound survived twice as long as their counterparts, whose untreated tumors grew exponentially. “From the moment the paper was published, we could easily have found patients to test a drug that targets this gene,” says Dr. Iavarone, who regularly receives calls and e-mails from desperate patients and their families. He has begun negotiations with drug companies to develop treatments based on the findings. “We have the immediate possibility of having a real impact in the clinical setting.”


Scientists were just beginning to delve into the genetic underpinnings of breast and colon cancer in the late ’80s, when Dr. Iavarone’s zeal for the promise of basic research caught Dr. Lasorella’s imagination. At the time, most medical investigators considered brain tumors a lost cause. “The brain is very difficult to study and it’s difficult to get tissue,” she says. “It was kind of a neglected area and Antonio passionately believed that research was the only way we could change the fate of these patients.”

Their first papers—a series investigating the mechanisms by which a radiopharmaceutical compound used to treat neuroblastoma disrupts the cancer—were published in 1991. They married in 1999, soon after they accepted research posts in the United States. “At some point,” says Dr. Lasorella, “our personal and scientific lives kind of merged.”

As with any good marriage of the minds, the intellectual partnership the two have forged— much of it focused on the role of ID2, a protein implicated in brain cancer that inhibits stem cell differentiation—relies heavily on the synergy of their scientific sensibilities. Dr. Lasorella takes a developmental approach, toggling between the deranged trajectory of cancer growth and the arc of a healthy cell’s growth and differentiation. “I try to translate what we learn from normal development of the brain to this process that has been perturbed,” she says, “and look at whether there’s a way to re-establish a normal situation, reverse the tumor.” Dr. Iavarone’s approach is infused by an interest in genetic aberrations. “He tries to understand whether those cells can be fixed genetically,” she says. “In that way we complement each other.”

In Italy, Dr. Iavarone ran the lab associated with the pediatric oncology clinic overseen by Dr. Lasorella, who still keeps in touch with the families of some of those early patients. “I have very wonderful memories,” she says, “even of children who weren’t doing well. They teach you a lot.”

At Columbia, Drs. Iavarone and Lasorella head independent lab groups that frequently join forces. Neither physician pursued clinical licensure in the United States. “I felt that when you treat patients with a very terrible disease, they deserve 100 percent of your attention,” says Dr. Lasorella. “When you work on the science that might transform their prognosis, you have to devote yourself 100 percent. For me, 50-50 wouldn’t work. I knew that I get involved very heavily and the patients are my priority. If I were working for them, I wouldn’t be able to carry my scientific responsibility.”

Despite Dr. Lasorella’s full-time research appointment, her strong clinical perspective persists and infuses their partnership, Dr. Lavarone says. “I am convinced that research and clinical work should always go together,” he says. “To find a cure, you need people who have different visions.”

Among the visionaries with whom they collaborate is Andrea Califano, PhD, chair of the new Columbia Department of Systems Biology and associate director of the Herbert Irving Comprehensive Cancer Center, who trained first as a physicist and later became a systems biologist. “Anna and Antonio combine rigor and scientific insight of the highest caliber with a sincere desire to complement their approach with computational methods to elucidate mechanisms that drive brain tumorigenesis,” says Dr. Califano.

The trio’s first effort revealed the role of C/ EBPß and STAT3 as synergistic master regulators of the mesenchymal subtype of human high-grade glioma, a particularly aggressive form of brain cancer. “That work was seminal in developing the master regulator analysis approach, which allows us to first assemble regulatory models for a cancer of interest and then use them to identify genes that represent the elusive masterminds of cancer,” says Dr. Califano. The trick, he explains, is using regulatory networks to discern the drivers from the passengers in the cancer genome. “This technique goes from the full genetic signature of the tumor to the handful of genes that regulate a particular tumor, allowing it to survive and progress.”

Each cancer is different, says Dr. Califano, making regulatory models all the more important for scientists intent on developing personalized treatments. “If I give you a box with all of the parts of a 747, you’ll have a box with 6 million pieces,” he says. “But you still won’t be able to assemble the plane, because you don’t have an assembly manual, a blueprint. Every kind of cancer has its own assembly manual—each is very different in the way it’s built. In that original glioblastoma study and now with our current collaboration on neuroblastoma, we have been working with Anna and Antonio to write and analyze the assembly manuals of these tumors.”

In the case of their 2012 Science paper on the glioblastoma fusion gene, Drs. Iavarone and Lasorella assembled a dream team of two dozen scientists, a mix of MDs and PhDs with training in neuropathology, neuro-Oncology, biomedical computation, algorithm optimization, and computer science working at academic medical centers throughout the United States and in Canada, China, Italy, and Taiwan. “Science is a global enterprise,” says co-author Raul Rabadan, PhD, assistant professor of biomedical informatics, who has collaborated with Drs. Iavarone and Lasorella to craft multiple algorithms—including those used in the Science paper—to interrogate the cancer genome.

Beyond giving hope to people with glioblastoma, Drs. Rabadan, Iavarone, and Lasorella created a new technique to help scientists analyze the deluge of data associated with the cancer genome. Known as TX-Fuse, the algorithm hunts for unique fusion genes within the RNA of a tumor sample. In addition to revealing the glioblastoma fusion gene detailed i their Science paper, TX-Fuse has been used to interrogate more than 200 tumor samples. “It’s unlikely that we will find a single gene fusion responsible for most glioblastomas,” says Dr. Lasorella, “but we may be able to discover a number of other gene fusions, each accounting for a small percentage of tumors and each with its own specific therapy.”

Dr. Iavarone traces the project’s intellectual legacy to the development of Gleevec, no the standard of care for chronic myelogenous leukemia. In 1960, a pair of scientists discovered that people with CML share a common anomaly, dubbed the Philadelphia chromosome. As technology improved over the next three decades, CML researchers used emerging sequencing techniques to identify the anomaly as the unique fusion of two genes, which produced a novel protein, BCR-Abl tyrosine kinase enzyme. In 1990, a Los Angeles-based team showed how the enzyme affects white blood cell production. Over the next 20 years, scientists honed in on the design of a tailored compound to derail that novel protein. “Fusion genes result in the production of proteins don’t exist in normal cells,” says Dr. Iavarone. A drug that targets the fusion protein has the chance of being incredibly selective.” In the case of CML, it took four decades for the relevant drug to reach patients. In 2001, the DA fast-tracked approval of imatinib, marketed as Gleevec, which nearly tripled the five-year survival rate for CML, from 30 percent to 89 percent. “Gleevec has resulted in a major beneficial effect in terms of prognosis and survival for these patients,” says Dr. Iavarone, who notes that the identification of gene fusions in other cancers has led to the development of additional therapeutic compounds. “In the case of glioblastoma, there had never been evidence of gene fusion. That’s why we started to look for it.”

The science behind Gleevec spanned 40 years, lurching forward as the tools for genetic and computational analysis became available. The groundwork laid by Drs. Iavarone and Lasorella with their collaborators—hastened by supercomputers and advanced sequencing— took 20 months. After identifying promising RNA sequences from the glioma stem cells of nine tumors and refining their approach, the team sequenced another 97 samples from the NIH Cancer Genome Atlas. Then the algorithm TX-Fuse churned through the resulting flood of sequencing data seeking novel replication errors among adjacent genes.

What they found was FGFR-TACC, a fusion of fibroblast growth factor receptor and transforming acidic coiled-coil, genes vital to late-stage cell division. “It’s clear that the development of sequencing and computational techniques is becoming more and more important to teams trying to understand the complexity of data,” says Dr. Rabadan, a theoretical physicist. “I see my work as a translator—taking a problem in biology and making it intelligible to a computer, something that can be coded and solved. Anna and Antonio move the information from a line in the computer code to something that can make a drug, can make a difference.”

When cell division stays on course, identical daughter cells result. When the process goes awry, as when FGFR-TACC interrupts late stage mitosis, chromosomally unstable aberrants result. To reveal the clinical implications of that particular mitotic meltdown, the team conducted experiments in mice to confirm the gene’s effect and identified a compound that inhibits FGFR kinase, an enzyme central to the fusion gene’s action. “If you block the function of the gene fusion, you can have a very strong anti-tumor effect,” says Dr. Iavarone. “That’s why we’re very excited about the therapeutic opportunities.”

After Science published the team’s findings, Drs. Iavarone and Lasorella started hearing from patients and their families hoping to enroll in a clinical trial. Sadly, for patients who already have been diagnosed with the rapidly growing cancer, science will not be fast enough to change their prognosis. “We feel intense pressure,” says Dr. Iavarone. “Strong, basic research should be immediately followed by therapeutic intervention—especially for patients like these, who have a particularly dismal prognosis.”

Dr. Lavarone and Dr. Lasorella have begun developing a screen compatible with clinical pathology protocols to identify the FGFR-TACC gene in glioblastoma biopsies. “Ninety percent of patients have surgery,” he says. “After the tumor has been removed, it can be analyzed for the presence of fusion genes and any other genetic alterations. Identifying this fusion is great for this particular subgroup of patients. That doesn’t mean that the remaining 97 percent should receive no treatment. They should receive a targeted treatment for their genetic situation. The goal is personalized therapy.

Source: the 2013 annual report of the Columbia University College of Physicians & Surgeons