September 1, 2005 (Vol. 25, No. 15)
Obstacles Still to Be Overcome, But Promise Remains
Anyone looking for short-term returns on investments in biobanks is bound to be disappointed. In fact, such returns will not be seen for decades, according to Jan Carlstedt-Duke, Ph.D., dean of research at the Karolinska Institutet (Stockholm). He spoke at the IBM Healthcare and Life Sciences Biobank Summit III, which was held earlier this year at the institute.
However, setting up biobanks, such as the Karolinska Institutet Biobank, is essential because they could help identify disease sub-populations and disease markers, allow major follow up of the long-term effects of a therapy, as well as determine the effects environmental exposure have on disease states, says Dr. Carlstedt-Duke.
To show the potential of a biobank, he presented studies where a collection of samples has been used to gain an understanding of diseases or treatments. One study, headed by the institutes Lars Klareskog, Ph.D., used blood and serum from 10,000 twins to determine if there were any consistent biological markers for rheumatoid arthritis.
The researchers in Dr. Klareskogs group identified antibodies to citrullinated proteins, many of which were present in patients up to 10 years before the onset of the disease. Using this set of samples has meant we have been able to identify a potent predictor of rheumatoid arthritis, notes Dr. Carlstedt-Duke.
Another application he cited was the use of in-patient registry records to determine the side effects on patients taking anti-TNF therapies for rheumatoid arthritis. Comparing information on patients following standard treatments to those on anti-TNF-based therapies showed that the incidence of side effects, including tuberculosis and lymphomas, were statistically similar in both patient populations.
Researchers concluded that these side effects are related to rheumatoid arthritis rather than the anti-TNF therapy.
Since Sweden has a national healthcare system with many patient registries of diseases, such as cancer, dating back as far as the late 1950s, we have access to a great resource of detailed patient history. This can be mined for a range of purposes.
When this patient information is put together with data from biological samples it could provide great insights into what causes diseases along with a long-term healthcare and economic impact, notes Dr. Carlstedt-Duke.
Patrice Milos, Ph.D., senior director of pharmacogenomics, translational biomarkers, and DNA sequencing at Pfizer Global Research and Development (Groton, CT), discussed examples of how a collection of biological samples has been useful at Pfizer in examining clinical efficacy related to human gene variation, as well as finding biomarkers for potential application to outcome studies.
A subset of Pfizers biobank of 70,000 anonymized DNA samples with associated phenotype data was used to identify common human variants of the cholesteryl ester transfer protein (CETP) gene.
This protein, which increases high-density lipoprotein (thought to have a protective effect against heart disease), and the variants identified provide some insight into the relationship of the protein to human disease. These human variants were then assessed for efficacy against Pfizers CETP inhibitor and found to exhibit similar levels of compound inhibition.
Dr. Milos also presented an example of how a large collection of biological samples can be used to identify biomarkers, which might enable better prediction of clinical outcomes.
From 825 patients participating in a Norvasc clinical study with angiogenic evidence of coronary heart disease, over 3,000 serum samples were collected at 12, 24, and 36 months post receiving Norvasc for the treatment of hypertension and angina.
Individuals experiencing clinical events were studied to identify proteomic markers of these events. The initial samples (70 cases and 70 controls) had common proteins such as albumin and haptoglobulin removed and were then run on 2-D PAGE to identify any novel proteins associated with clinical events.
This study, performed in collaboration with Celltech (Oxford, U.K.), has identified seven candidate biomarkers, whose presence could be used in future to determine the efficacy of new therapies for heart disease.
The use of biobank samples will help us understand disease progression, which in conditions such as hypertension is poorly addressed. Additionally, using biobanks to identify biomarkers for specific conditions could give us the tools to determine the potential effectiveness of drugs as early as Phase I, notes Dr. Milos.
This in turn, will lead to more informed decisions on continuing with trials and potentially a lower attrition rate of drug candidates.
According to Brian Clark, CEO of a new national cancer biospecimen resource, OnCore UK (London), one of the major difficulties facing biobanks is funding.
Most institutes will spend more on animal research than on biobanks. Yet coupled with the correct sample information, many human biosamples have the potential to be a viable alternative for clinical research on living humans and could allow clinical trials to progress faster, he points out.
However, Clark, along with many other delegates at the summit, agreed there were many more challenges to overcome before achieving this situation. One area identified as important in collecting clinically relevant tissue is patient trust.
The concept of a biobank is still quite sensitive with the general public especially in the U.K., as the Alderhey Hospital incident is still present in the minds of the British public, says Clark.
He is referring to the unauthorized post-mortem collection during the 1990s of childrens hearts for research purposes at Alderhey Hospital in Bristol, U.K.
We must learn to communicate with the public so they be-come willing to provide biosamples with plenty of information to make them useful in research and not just the samples we can obtain easily, as has been the case in the past, says Clark.
Scientists are often unaware of the emotional journey that a sample has been on and especially what donating a sample means to a cancer patient. The purpose of a biobank is to minimize the need for researchers to be directly involved in the cycle of collection and handling of biological samples and act as an honest broker between donor and researcher.
Therefore, positive and proactive communication which could potentially cost as much as many enabling technologies is key to making this happen, but most biobanks assign very little budget to this activity.
Dr. Carlstedt-Duke echoed this sentiment of public cooperation leading to better quality samples. He emphasized that the Swedish government has been proactive about communicating the merits of biobanking.
Indeed, in January 2003, the government passed a biobank law stating that patients must give informed consent for any tissue or blood taken and kept for longer than two months to be saved in a biobank.
The law was passed so quickly that academics and hospitals had to work together to find a way to implement it, continues Dr. Carlstedt-Duke. This has resulted in Sweden now having one of the largest biobanks in the world.
Sweden, like the U.K., has a public health system and many patient registers have been set up since the late 1950s and now include registers for cancer patients, malformations, and twin births so that detailed patient history is also available with some samples.
Since our biobank law also controls the way samples are used and further regulates quality and security policies, patients in Sweden are not just willing but are keen to take part in life science studies, says Dr. Litton.
He confirmed the Swedes eagerness to participate in life science research by adding, In a recent clinical survey posted to 50,000 post-menopausal women, there was a 40 percent response to a questionnaire, which took one hour to complete.
Both Drs. Litton and Carlstedt-Duke believe this level of public support from the Swedes will contribute to producing more clinically relevant samples.
John Potter, M.D., Ph.D., senior vp and director of the division of public health sciences at the Fred Hutchinson Cancer Research Center (Seattle), also pointed out the importance of obtaining sufficient data on environmental exposures to carcinogens when collecting samples for a biorespository.
The massive increase of the incidence of colon cancers, for example, in Japan since the 1960s, can only be explained by environmental factors as the genetics of a human population cannot change this rapidly.
Therefore, when we are sampling for a biobank we need to gather information on people and their exposure history [to potential carcinogens] not just their tumors.
If we dont do this kind of research well just be collecting samples in biobanks for the sake of it and will never gain an understanding of the causes of diseases like cancer.
Sample handling and storage was identified at the summit as one of the key obstacles to providing the best clinical biosamples.
We are still assessing what format to store our biospecimens in the OnCore UK bank as there is no one size fits all because different researchers require tissues for a whole range of applications, says Clark.
Supplying frozen samples is the current benchmark but fresh tissue is much more useful for functional assays in translational medicine.
Denis Hochstrasser, M.D., vice dean in the faculty of medicine and a leading proteomics expert at Geneva University and University Hospitals (Geneva) added, Most blood or serum that is kept in biobanks is stored frozen so is might be worthless for several types of proteomics analysis.
Since proteins are like a Lego system in water they can break up easily or get modified when stored in water for a long time even in a frozen state, whereas DNA is generally more robust in solution.
Additionally, in serum there are around a million proteins; some of them in very low concentration so 2-D PAGE has too low a dynamic range to see many of these low copy number proteins in frozen serum.
Dr. Hochstrasser also presented an example of the effects that storing a blood sample on a bench for six hours has before biochemical testing. The results showed that even something as simple as storage conditions had a profound effect on the levels of metabolite within the sample.
Before setting up a biobank, scientists should define what it is they want to find out from those samples, he advised.
For example, if they want those cells to live again they have to freeze them slowly to freeze the fluid around the cells to prevent the cells from breaking open. Since there is currently no single method of storing samples that will fit all purposes, samples that are going to be used for a range of applications, such as genomics, proteomics, and metabolomics, need to be prepared in different ways, explained Dr. Hochstrasser.
Stem cells need to be frozen with DMSO and samples for proteomics studies need to be prepared by total tryptic digests followed by lyophilization, he noted.
If this type of preparation does not occur then many of the samples being put into biobanks today will be worthless for future proteomics or metabolomics research, he warned.