NEW YORK (GenomeWeb News) – When the biomedical research community was still abuzz over the completion of the Human Genome Project, three scholars at the Massachusetts Institute of Technology published a sobering article. Although biobanks could be great enablers of genome-based personalized medicine, they said, such repositories are fraught with problems.
"They are generating a lot of controversy — about ethical and scientific considerations," the trio of authors wrote in Pharmaceutical Discovery in 2004. Stan Finkelstein, Anthony Sinskey, and Scott Cooper were proponents of personalized medicine, but they highlighted criticisms others were promoting.
There were many issues, they wrote, and there was chatter among scientific journals and societies about whether biobanks could ensure privacy, whether they would be useful in studies beyond single-gene disorders, and whether genomic medicine would be feasible.
While nearly a decade has passed since that article was published — and although the promise of personalized medicine has already begun to be realized — biobanks still present a thorny set of problems for researchers wading through seas of genomic information for bits of meaning.
One struggle biobanks and their curators face is finding samples that have the broad consent that wide-ranging genome-based science sometimes demands and that is also ethically sturdy enough to protect donors, banks, and researchers. Another barrier is education. Patients, doctors, and nurses often aren't familiar with laws, regulations, and standards surrounding submitting samples to biobanks, and thus a bureaucratic paralysis can creep in that keeps them from storing and sharing research samples and data. Further complicating matters is that such standards and regulations also vary regionally and from bank to bank, creating a patchwork that scientists have to navigate.
A booming field, wide-ranging practices
In the past dozen or so years, the number of biobanks in the US has risen dramatically. A study published early this year by researchers at the University of North Carolina at Chapel Hill generated a list of nearly 800 US biobanks and found that 59 percent of those were established after 2001.
These biobanks may house more than 50 million samples, and they are playing "an increasingly important role in biomedical research," lead author Gail Henderson, professor and chair of the Department of Social Medicine at UNC, told GenomeWeb Daily News in January.
Despite their proliferation and their value, Henderson found, there is little empirical information about the kinds of policies or practices biobanks follow in handling, storing, and managing their samples. It is, she said, a "Wild West" out there.
While these collections may be invaluable to genomics researchers who seek disease-linked genetic variants or biomarkers to diagnose disease, stratify patients, or set their therapeutic dosage, it is hard for biobanks to be useful if scientists have not heard of them or do not know how they operate or whether they share specimens or information.
"There's no list or database of biobanks that a researcher can easily go to to find all of the biobanks that might have specimens of interest," Nicole Lockhart, a program director in NHGRI's Ethical, Legal, and Social Implications research program, told GWDN.
"That really poses a significant difficulty [for researchers] to try to figure out where you can even find specimens. And then, should you find a bank that might be appropriate, it's not always clear what their access policies are…" Lockhart said. "[And] it's not always clear whether their specimens would be appropriate for you. What do they actually hold? How are they preserved? How are they consented? Trying to kind of figure out all of that can really be a significant amount of work."
Henderson's study found a wide range of types of biobanks in various institutional settings, such as hospitals, universities, and institutes, and she said that they follow different practices, policies, and standards.
"Some of the differences between these banks make sense. A bank that was formed from clinical trial specimens might have a different access policy than a bank that was formed for more general research use," Lockhart said.
But what works best for the biobanks may also make life hard for scientists, particularly if they are trying to use specimens from multiple banks, which Lockhart said can create significant challenges.
"There are a lot of differences, even in the way people freeze specimens. Are they snap-frozen in liquid nitrogen? Are they frozen in isopentane? Are they formalin-fixed and paraffin-embedded? There can be kind of a huge range in how specimens are processed," she added.
Even while researchers struggle to find out all these details about specimen handling, there is a further complication: The banks themselves may not know the history and status of all their samples, or they may not be able to change their protocols to suit scientific demands.
"If you're a bank that has a lot of remnant samples [left over] from clinical research, you may not know precisely how every specimen in your bank was handled," Lockhart pointed out. "And you may not be able to control how they are handled. Because, obviously, you're collecting specimens from the department of pathology or something, and their first priority is, and needs to be, the patient. Once their diagnosis is complete they might give you remnant tissue, but they may not be willing to change their protocols based on what you think they should be."
Adrienne Flanagan from University College London has been wrestling with some of the challenges facing both researchers and biobanks for years. These problems can be particularly acute for scientists like Flanagan who seek to understand and develop ways to treat rare diseases.
In her efforts to identify biomarkers and other diagnostic signatures for sarcoma, Flanagan has encountered all of the above issues, and then some. One that is particularly bothersome for genomics researchers like her is that some collections do not include the matching normal tissues from the patients.
"Normally if you freeze a tumor, you need blood or some attached muscle. So a lot of people who sampled tumors over the years never collected the blood or the normal [matching] tissue. Therefore... it is of limited value," Flanagan said. Without matched normal tissue, researchers don't have anything to compare the tumor tissue to for an in-depth analysis.
Education and training are top priorities
When Flanagan decided to start a biobank to house bone and soft tissue samples that would address some of the problems she faces in finding samples for sarcoma research, she began to encounter other issues.
"Sometimes patients contact me to say, 'Well, I have a rare tumor like a cordoma or chondrosarcoma and we'd like some research done. How can we participate?' And they're having their surgery done the next morning, or maybe even a week [later]," Flanagan said. "And then I find that these hospitals may not have the liquid nitrogen to take the sample, or that nobody's ever taken a sample and frozen it before in their lives."
Scenarios like that highlight what Flanagan and others see as the most critical problem confronting genomic research that relies on biobanking and biobank data: that scientists and physicians do not know enough about how such repositories work and how they can use them.
"There's a huge lack of education about even simple processes — and they really are simple processes. … As a result of that, there's a shortage of samples, particularly for rare diseases," Flanagan told GWDN.
Alison Hubel, a professor of mechanical engineering at the University of Minnesota, couldn't agree more. She has done extensive research into biobanking technologies and practices and has launched the Biopreservation Core Resource, or BioCoR, at Minnesota.
"I think that biobankers need a sense of education, training, and empowerment. I think a lot of people think that somebody is going to tell them how to do things correctly — the magic biobank fairy or something. The reality is that it's going to require time, experience, training, et cetera," Hubel said.
"If there is one really global challenge, it is creating the intellectual foundations of biobanking so that people can rely on that intellectual foundation to create quality systems and ensure that the biospecimens are of high quality," she added.
Further, many institutions and physicians worry about providing samples to biobanks because they are concerned that they might breach certain regulations, laws, or ethical boundaries, and those worries, too, are caused by a dearth of good information, Flanagan said.
"A lot of that is lack of education. People don't know what is allowed to be done and what isn't allowed to be done. So, the default position if you ask people to do things is not to do them," she said.
She said that research units and professional bodies should step up to provide training to doctors and nurses who work directly with patients so that they can explain what the laws are and what the benefits are of using samples with clinical information in research, and so they can get their patients to agree to provide their samples for use in an array of research projects.
"Educating doctors and nurses on how to collect tissue and how to label it and how to consent patients would be a huge benefit, take the mystery out of it," Flanagan told GWDN.
The sticky problem of consent
If there is one overarching theme that cuts across the entire spectrum of biobanking worries, it is that there are a suite of sticky issues surrounding consent. Consent is particularly critical in genomic research projects like genome-wide association studies and other efforts in which researchers seek to access many samples from diverse populations or from very specific groups to connect genetic variants and clinical information to diseases and phenotypes. The donors often may not know who has access to their samples and information and what kind of studies these researchers are conducting.
Genomic information is uniquely problematic, too, because while it can be used to predict disease risk, it can also be used to identify patients in databases, and it can even affect patients' relatives.
Today, there are questions about whether the proper systems and practices are in place in hospitals and other institutions to help doctors and nurses to gain consent from patients so as to enable genome-based research.
"If you ever try to get a sample out of a non-teaching hospital, it's very difficult. One doctor has to try to get hold of another doctor who is doing the operation, and then speak to the pathologist who will try and coordinate with the individual patient," Flanagan said.
"I've done that myself around the world, and it can take me hours, phoning people in their homes at night before they do the operation in the morning to try tell the patient what will be done and that they want the sample flown to the UK," she added.
Teaching hospitals at universities, Flanagan said, are often better prepared to consent patients for research because it is a priority.
Mark Watson, an associate professor of pathology and immunity at Washington University in St. Louis, told GWDN that consent-related issues are "the biggest regulatory hurdle" confronting whole-genome sequencing researchers.
"Because various biospecimen collections have been obtained with various types of [institutional review board] approval, sometimes it is hard to negotiate which collections of biospecimens can be used for what particular studies, [particularly] if they involve things like whole-genome sequencing," Watson said.
Another consent-related problem is that genome technologies have allowed investigators to start using a wider range of samples, because these tools are now more tolerant of samples that were collected and preserved in paraffin five, 10, or 15 years ago, Watson said.
"That's great when you think of the number of samples you can potentially access for doing genomic studies now. Unfortunately, those things that were collected five, 10, or 15 years ago were not collected with the type of consent and regulatory approval that many studies demand now," he said.
Whether those samples are usable with whatever type of consent that they have will be up to the discretion of the IRB at the researcher's institute, and it will depend on how that information will be used or shared.
"It used to be if you were doing an immunohistochemical study for a single protein, looking at protein expression from a block that was five or 10 years old, there were less issues there. Now that folks are interested in taking tissue blocks from a decade ago and doing whole-genome sequencing, that has given a lot of people appropriate pause to say, 'Well, how much can you do without having patient consent?'" Watson noted.
For her biobank of rare cancers at London's Royal National Orthopaedic Hospital, Flanagan has created a consent system that works for her specific research purposes.
"I wrote a [consent] document and got approval at a national level to collect tissue form any patient coming to this hospital. That allows me to collect any tissue that is surplus to diagnosis — if the patient consents to it," she said.
The agreement also gives her access to look at clinical data and all patient imaging, enables her to access a patient's medical information from other hospitals where they may have been treated previously, and makes it possible for her to share — but not to sell — patient data with private companies.
"In that application, I said that I would use the tissue for genome sequencing, and that I also could use it to make cell lines, which would be permanently established, and that I could also put it into animals. So, I could really do almost any experiment that I wished to, except that I could not use the tissue to clone a human being," Flanagan told GWDN.
She said that in the global research culture of today, "the world is so small that you never actually know when you may want to collaborate with somebody else. So, it's better to get consent that is as broad-ranging as possible, and it shouldn't be time limited. You don't want to have to go back to patients."
While programs like Flanagan's in London may sound ideal, Lockhart suggested that there likely will not be a 'one-size-fits-all' model that will solve these consent problems.
"It's also important to keep in mind that not all people feel the same way. So, not all research participants are going to have the same level of comfort with data sharing, or the same level of comfort with genetic research," she said.
One proposal to address this is to offer consent agreements that include options for patients, she added.
"Letting some research participants say: 'I'm comfortable with open access. I'm comfortable with controlled access. But I don't want any data sharing,'" might be one such arrangement, Lockhart said.
Consent with options like those might "complicate researchers' lives … when only a portion of data can be shared, but it may allow you to recruit more diverse populations.
"It may seem frustrating at times, or like a patchwork — 'Why can't we just harmonize all this?' — but there are advantages in the way our human subjects protection system has been set up in this country. It is to give a lot of discretion to local institutions and local populations, [because] they may have considerations for their populations," she told GWDN.
What is being done? Quite a bit
What is certain about all of the wide-ranging issues surrounding biobanking is that there is a bustle of activity in this area. Institutions and people are pursuing a wide array of avenues for solving these problems — some are focused on education, while others are developing and sharing best practices.
The College of American Pathologists has recently launched the CAP Biorepository Accreditation Program, aimed at teaching methods for ensuring consistent and verifiable quality of biorepository practices and biospecimens.
NHGRI, meanwhile, has provided $5.7 million to create the Return of Results Consortium, and it has created the Genotype-Tissue Expression project. Additionally, the National Cancer Institute has developed a set of guidelines and best practices for biobanks to follow.
In Minnesota, Hubel's BioCoR effort has taken a three-pronged approach that includes education and training courses on biopreservation practices and cellular therapies, a biobanking support service, and a research program that aims to develop new technologies for improving biopreservation processes.
Internationally, the EU recently said it has provided €8 million ($10.5 million) to the fund the second phase of the Biobanking and Biomolecular Resources Research Initiative, and the International Society for Biological and Environmental Repositories has a forum to address the various technical, legal, ethical, and managerial issues.
These are just a few among many different initiatives that have been launched to improve biobanking practices, to improve sample and data quality, and to do all of this in an ethical manner.
Flanagan told GWDN that she has been working to help other people set up biobanks by showing them what practices have worked best in her program.
"Because we have learned from our mistakes. … One of the things that we can offer is to show other people how to do it, to share our experiences setting up biobanks and save them time and energy," she said. "I don't wish anybody to make all the mistakes that I made."