Nine years ago on a Monday afternoon, Sandra Smith, a pastor’s wife and mother of three in DeWitt, Mich., learned she had an aggressive form of breast cancer. The real bad news, however, would hit the family later that week.
At first they thought their youngest, six-year-old Andrew, was just battling the flu. Then he started vomiting. He’d also developed a facial droop, and his gait seemed off. Smith remembers wondering if they were “making a big deal out of nothing,” even as they rushed to the emergency room.
Nothing could be further from the truth. An MRI scan revealed a large area of swelling in Andrew’s brain stem—clear evidence of a fatal childhood cancer that typically strikes between the ages of four and 10 and kills most within a year of diagnosis. Unlike the cells dividing uncontrollably in Smith’s breast, her son’s cancer, called diffuse intrinsic pontine glioma (DIPG), could not be fought with surgery or conventional chemotherapy. In DIPG the malignant cells entwine with normal brain tissue in a region that controls critical functions such as breathing and heart rate, making it impossible for a surgeon to remove. In more than 200 drug trials nothing has worked better than radiation therapy, which itself can only extend life a few short months in kids with DIPG. Andrew outlived the “typical” DIPG patient by surviving just over two years after his diagnosis, passing away at the end of 2009.
DIPG accounts for about 10 percent of childhood brain and spinal cord tumors. It is the second-most common pediatric brain tumor and the leading cause of cancer death in kids. Treatment options and survival rate for DIPG have not changed in 40 years—a predicament that likely helped nudge brain cancer past leukemia as the deadliest childhood malignancy in the nation, according to a recent report from the U.S. Centers for Disease Control and Prevention.
Today, however, the outlook for DIPG and other childhood brain cancers looks more promising, thanks to a surge of new research made possible by advances in gene-sequencing methods and tumor tissue donations from families who have lost children, such as Andrew, to these diseases. In recent years researchers around the world have used patients’ tumor tissue to generate dozens of cell lines and mouse models to study the basic biology of pediatric brain cancers. The time is ripe. In the dawn of precision medicine, which aims to customize disease treatment to the individual patient, genetics and basic science findings suggest why past trials may have failed and are guiding future and ongoing efforts to identify effective therapeutics for these devastating diseases.
Michelle Monje, an assistant professor of neurology at Stanford University, first encountered DIPG around 2002 as an MD/PhD student there. Working with her clinical mentor to care for a nine-year-old girl dying of DIPG, it was “the first time I’d come upon a disease we had no idea how to treat,” Monje says. “I felt so close to this patient and was devastated by my inability to help her.”
Back then there was little molecular data on DIPG. No animal models. No cell cultures. Generating such research tools requires tumor samples from patients. Yet since MRI scans can reliably diagnose typical DIPG and getting brain stem tissue is not trivial, biopsies were rarely done. With precious little tumor tissue to study in the lab, Monje says, research progress on DIPG had stalled for decades.
The tide started turning by 2007, when a team of surgeons in France reported safely obtaining biopsy samples from 24 children with DIPG using stereotactic techniques that use computer imaging to guide needle placement. That study invigorated longstanding efforts by a pediatric neuro-oncologist Mark Kieran, clinical director of the Brain Tumor Center at Dana-Farber/Boston Children’s Cancer and Blood Disorders Center, who had spent years pushing for DIPG biopsies in the U.S., initially without success.
By that time technological advances had made it possible to read DNA sequences from tiny bits of tissue, giving further impetus for a tricky surgical procedure now shown to be safe in trained hands. The Boston team began offering patients genomic sequencing of their tumors biopsied at diagnosis and relapse, to “see how the tumor is evolving and redirect the appropriate drugs to it,” Kieran says. Tumor profiles could help determine which patients might benefit from a newer class of drugs called targeted therapies, which hit specific proteins in the tumor rather than just kill off any dividing cell. Targeted therapies are a cornerstone of precision medicine.
Since 2009 researchers at Dana–Farber have sequenced brain tumors in nearly 1,000 children. Among kids with tumors classified as a low-grade glioma, up to 10 percent have a mutation in a gene called BRAF that is seen in some adult skin tumors. A few years ago, 32 children from Europe and North America with BRAF-positive gliomas entered a clinical trial of dabrafenib, a targeted therapy approved for melanoma patients with this mutation. At a conference in Copenhagen earlier this month, Kieran reported that 23 of the 32 kids improved on the BRAF-inhibiting drug—a response rate high enough that his team is offering continued therapy to trial participants with the mutation.
In 2012 Kieran and collaborators launched a clinical trial to biopsy tumors of children with DIPG, test them for several molecular markers and, based on the results, assign one of four treatment strategies. Two years ago a team led by Sabine Mueller, a pediatric neuro-oncologist at the University of California, San Francisco, initiated another DIPG trial. This study probes patient tumors using a more sophisticated technique, whole exome sequencing, which scours the entire protein-encoding portion of the genome rather than just checking for pre-specified markers. Based on each patient’s tumor profile the U.C. San Francisco team proposed up to four drugs that seem appropriate. It will take another one to two years to see if the drugs help.
Although helpful for some individuals, precision medicine is expensive, and some scientists suspect it may only modestly improve the lives of cancer patients in general. Cells within a single tumor can acquire different mutations, such that “even if there is an effective agent, it is likely to have limited benefit because molecular pathways that are active in other parts of the tumor will lead to tumor growth from different clones of tumor cells,” researchers wrote in a New England Journal of Medicine commentary published in September. And without specific drugs approved for DIPG, there is ongoing debate about whether biopsies offer real benefit to these patients.
Some labs have taken a less controversial route to DIPG samples—obtaining them as legacy gifts from families who agree to donate tumor tissue once their child passes. Smith, the Michigan mom, learned about legacy gifts through an online DIPG support group in spring of 2008, a half year after her son Andrew was diagnosed. Reading that post about removing the brain after death and donating the tissue “was horrifying to me,” Smith recalls. “But I understood [that without patient samples] there was no way for researchers to look at this tumor.” She shared the idea on a Yahoo group for DIPG families that she and her friend moderate. In November 2008, a mother from the Yahoo list called Smith in a panic. Her daughter was in her final hours, and the family wanted her tumor tissue donated but hadn’t made arrangements.
Autopsy tissue donations are logistically challenging. Once a child passes, the brain needs to be removed, with tumor tissue placed into sterile tubes, within six hours. Yet patients tend to die at home, far from the medical center, sometimes in the middle of the night or during a snowstorm on a holiday weekend. Some labs book on-call service from a tissue recovery team close to where the child lives. To have the biggest impact, samples should go to labs that can receive and process them the same day.
As Smith helped other families arrange tumor donations, she got to know some of the leading DIPG researchers, including Monje, who had just worked out a way to culture cells from autopsy tissue and use those cells to create mouse models of DIPG. In July 2011 Smith learned of a seven-year-old girl named McKenna, who was battling DIPG. Smith and Monje worked with the family and “made sure we had the required documents when the time came,” says McKenna’s mother, Kristine Wetzel, a high school teacher in Huntington Beach, Calif.
McKenna faded suddenly, and the family decided to donate her tumor tissue within an hour of her death. Although painful, the decision was “surprisingly comforting,” Wetzel says. “It was a way to fight back against the monster that had stolen our daughter.” The Wetzels have since helped other DIPG families with tumor donations and created a foundation to raise awareness and fund research in pediatric brain cancer. The foundation covers the cost of tissue donations to Monje’s lab and pays for a technician who maintains the lab’s DIPG cultures and has shipped samples to some 80 labs around the world. Support varies but usually amounts to about $100,000 per year, Monje says.
Autopsy tissue donations have “transformed the research landscape from an unapproachable problem, due to lack of material for research, to an unprecedented analysis of the DIPG genome,” says neurobiologist Suzanne Baker, who helps lead the Neurobiology and Brain Tumor Program at St. Jude Children’s Research Hospital.
A wave of papers by Baker, Kiernan and others at Washington University School of Medicine in St. Louis and elsewhere, revealed a surprise. Although DIPG has a gene signature distinct from other brain cancers, the rare childhood tumors share one striking feature: Nearly 80 percent of them have mutations in a gene that codes for a protein called histone H3. Histone proteins are like spools around which DNA wraps. A key player in epigenetics—the study of biological mechanisms that turn genes on or off—histones influence how easily DNA is accessed by enzymes that translate genetic code into working proteins. “Histone H3 is so fundamental…I would think lots of cancers would have these mutations,” Baker says. Yet they seem to be unique to DIPG and about a third of non–brain stem tumors in kids.
The genetic insights could not have come at a better time. Whereas some labs were busy plumbing whole genome sequencing data from DIPG tumor cells, others were testing potential drugs in cell cultures and mouse models generated from patient brain tumor samples. The idea was to carefully vet compounds in the lab before choosing which ones to test in a longer, costlier clinical trial. The approach is not revolutionary. Generally it is “the way you do medical research,” Monje says. But for DIPG, “we’d been unable to do this” for decades because there had been no cell cultures or experimental mice modeling the disease.
The situation improved in 2010 when Charles Keller, then at Oregon Health & Science University, organized a global screening effort. By then many labs were creating cell cultures from DIPG tumor tissue. As co-chair of a committee proposing drugs for DIPG clinical trials, Keller rallied the Monje lab and 12 other groups to pool resources for a collaborative study. His group dispensed 83 potential drugs onto well plates and sent them to the other labs for testing. At these far-flung labs researchers loaded the plates with DIPG cell cultures and looked for wells that turned blue—a chemical indication that the drug was killing tumor cells. Top drug candidates also improved survival in mice with implanted DIPG tumors.
In addition, the labs purified genetic material from their DIPG cell lines and sent DNA and RNA samples to Oregon for sequencing to establish a clear tie between the cells’ gene glitch and drug response. Checking for these connections is key, Keller says, because many drugs that looked promising based on mutations in the DIPG cells showed no effect in the cellular assays.
But there emerged a winner—a drug called panobinostat, which inhibits enzymes that chemically modify histone proteins. Coincidentally, the U.S. Food and Drug Administration approved panobinostat as a treatment for another cancer, multiple myeloma, as the global screening manuscript went to press in Nature Medicine. The results helped launch a clinical trial of panobinostat that opened for enrollment in May. Led by Monje, this trial will measure side effects and determine the best doses of the drug for treating children with DIPG. Panobinostat is not going to be a silver bullet, however. The lab data showed some DIPG cells develop resistance to the drug, suggesting it will need to be combined with other therapies to achieve a survival benefit in patients, Monje says.
One challenge with panobinostat is shared by many brain cancer therapies—delivering them effectively into the brain. “Many drugs don’t cross the blood–brain barrier so they are not getting to the tumor,” says U.C. San Francisco’s Mueller. Some researchers are using a procedure called convection-enhanced delivery to place drugs through small catheters directly into the brain tumor. Others are using nanotechnology to reformulate drugs so they can be more specific and durable—for example, by inserting molecular tags that direct the drug to molecules found uniquely on the tumor. It is possible that good drugs for DIPG already exist, Mueller says, but “we just don’t know how to deliver them correctly.” She is planning a future trial using convection-enhanced delivery of panobinostat in kids with DIPG.
In the meantime Monje’s lab and other groups are doing additional drug screens with epigenetic agents and combination regimens, and Keller founded a nonprofit cancer biotech to speed the movement of candidate drugs from basic science research to clinical testing. Each year his team organizes a weeklong crash course to teach families about pediatric brain cancers and explain how tumor tissue donations are driving research.
Some families make regular visits to the lab to see their child’s cells under the microscope. Wetzel’s family visited Monje’s lab eight to nine months after their daughter died of DIPG. “The first time I saw McKenna’s cells, I started crying uncontrollably,” Wetzel says. “I wanted to take every petri dish…and throw it against the wall, to the destroy the cell line like it had destroyed my daughter.”
Now, after trips to the lab about once a year, Wetzel feels differently. “I see it as McKenna’s last stand…her gift to the world and to children who will follow her. If McKenna can’t be here, let her make the world a better place in the one way she now can. It helps us to think there was some purpose to her death.”