When pathologist Adrienne Flanagan joined the UK's Royal National Orthopaedic Hospital in 2002, the first version of the human genome was brand new. At the time, she received a piece of advice: Bank as many tumor samples as you can; they'll be needed for future genomic studies. "So, I biobanked almost any tumor that I could lay my hands on since 2002," Flanagan says. "And the person who gave me that advice is a chap called Professor Michael Stratton, who is now the director of the Wellcome Trust Sanger Institute."
Flanagan is now working with researchers there, using the bone tumor samples that she has saved over the years. As part of the International Cancer Genome Consortium, researchers at the Sanger Institute were interested in analyzing many different tumor types, including bone tumors. As a whole, the ICGC aims to study the genomics, transcriptomics, and epigenomics of 50 different tumor types, from breast to renal cancer. A local UK charity called the Skeletal Cancer Action Trust raised money to kick-start the international project. Now, with Flanagan's team, the group aims to search for mutations in 500 bone tumors: 250 osteosarcomas, 200 chondrosarcomas, and 50 other bone tumors.
In parallel, Flanagan and Sanger researchers, including Peter Campbell, are working to develop a way to determine the specific characteristics of a person's bone tumor and use that to track disease progression and the patient's response to treatment.
For both arms of the project, the researchers are beginning to see results.
Very little is known about why some people develop bone tumors, particularly osteosarcomas, which account for about 35 percent of all bone cancers. With sequencing and further analysis of tumors, Flanagan, Campbell, and their colleagues hope to uncover a genetic cause for bone tumor development. "The aim is to discover the complete library of genetic alterations found in osteosarcoma," Flanagan says. "This information will unlock the secret as to why this tumor type grows, and will enable the identification of crucial genetic abnormalities that drive the development, growth, and spread of this cancer."
For that cohort, the researchers turned to the biobanked tumor and matched normal samples that Flanagan has been collecting as well as others from their collaborators at the ICGC, including samples from Norway and possibly Austria, Egypt, and the Netherlands.
Obtaining enough samples of sufficient quality is a challenge, Flanagan says. Osteosarcomas are rare — there are about 800 new cases diagnosed in the US each year, according to the American Cancer Society. In addition, how the cancer is treated makes it difficult to get a sizable amount of tissue to work with. Bone tumors are diagnosed using needle biopsies that take a tiny sliver of tumor, about 10 millimeters long and about 2 millimeters thick, Flanagan says. Then the patient is given chemotherapy for a few months before the tumor is removed. "Patients with osteosarcoma are generally treated with neo-adjuvant chemotherapy. In many cases, their tumors have been killed by the therapy, leaving very little tissue for analysis," she says. Further, she adds, making sure there are matched normal samples for comparison is also an issue.
Despite these challenges, the team has completed the first phase of this part of the project. For that, they used the Illumina HiSeqs and the bioinformatics pipeline in place at Sanger to examine a portion of their samples. "We looked at 70 tumors and to do a sweep to see if there were any common mutations that would pop out," Flanagan says, adding that they are now expanding to study a larger cohort to home in on the mutations behind bone cancer.
To track response
Treatment for bone cancer has remained largely the same since the 1970s, Flanagan says, and the 60 percent five-year survival rate has been fairly steady since the 1980s. "The treatments that exist at this stage are really pretty crude," the Sanger Institute's Campbell adds. Furthermore, clinicians currently use costly imaging techniques to determine how patients are responding to treatment and whether they are staying in remission after treatment ends. Instead, the team is working to develop a blood test-based way to track how tumors respond to chemotherapy.
As bone tumors grow, some of their cells get into the patient's bloodstream, where they circulate. Campbell and his colleagues at Sanger are taking samples of those circulating tumor cells from a patient to determine the genetic abnormalities of that specific tumor. That set of abnormalities serves as a fingerprint that can be tracked over time in that patient's plasma to determine the burden of disease. "This study involves measuring tumor--specific DNA in the plasma of patients with osteosarcoma ... to measure the response to chemotherapy. Having a test like this would inform the oncologist if a patient was responding to chemotherapy, and therefore allow different treatment options to be considered. We anticipate that this research will completely change the way patients with cancer are monitored," Flanagan says. "This is a step towards a personalized delivery of cancer treatment."
Campbell adds that they have piloted this approach in a patient with osteosarcoma. As he and his colleagues reported in Genes, Chromosomes, and Cancer last year, they sequenced DNA from fresh-frozen tumor samples to identify rearrangements in that patient's tumor. Then the researchers developed nested, real-time PCR assays for several of those rearrangements, and took multiple blood samples from the patient during the course of her treatment to test. "The level of tumor DNA broadly correlated with clinical relapse," Campbell says, adding that the researchers could see the tumor DNA levels increase as the patient failed to respond to salvage chemotherapy.
For the next phase of this part of the project, the researchers plan to follow about 100 patients with soft tissue sarcoma and bone tumors to gauge their responses to chemotherapy and see whether their cancers might recur. "We think that we should be able to get good results from this study within a couple of years. We've been collecting tumor samples and following patients now for about a year," Flanagan says.
Eventually, the researchers hope the test could be used to detect tumors earlier, using a less expensive technique. "We should be able to get useful results from this study within the next three to five years," Flanagan adds.
Participants: The UK's Royal National Orthopaedic Hospital and the Wellcome Trust Sanger Institute, which are using samples from international collaborators at Leiden University Medical Center in the Netherlands, Norway's Oslo University Hospital, and elsewhere
Funding: The Skeletal Cancer Action Trust and the Wellcome Trust