Biobanks allow researchers to store and study human tissue for a variety of uses, including biomarker discovery, immunotherapy, and targeted treatment of diseases. To be useful, biobanks rely on the collection of samples, efficient isolation, production workflows, and sophisticated data mining tools to access the information and specimens once they’ve been collected and stored. Advances in all these areas are converging to propel biobanking into a new era.
After surgery to remove a tumor, tissue samples are frozen and saved in biobanks for research purposes. These samples usually come from primary tumors before drug treatment has begun. But molecular evolution may render them less than relevant when it comes to developing novel therapies.
Cancers contain a genetically diverse population of cells. Because of this heterogeneity, therapies may kill part of the cancer, but also provide a selection pressure, leaving behind a population enriched for resistant cells. As those cells proliferate unchecked, passing on their hardier genotype, a “new” (and potentially more dangerous) cancer arises. To develop drugs against these refractory cancers, researchers need to ascertain the new, drug-resistant molecular profile, which is often different from the one derived from the original tumor sample in the biobank.
“When we’re trying to match up specific molecular changes and mutations to select patients for a specific treatment to block that particular mutation, we may be way off if we’re using banked tissue,” says Gerald Batist, MD, CM, director of the McGill Centre for Translational Research in Cancer at McGill University in Montreal.
To trace these molecular changes, it’s necessary to sample the tumor at different stages of treatment. Batist has helped launch a large, multi-institution program to collect serial biopsies from metastatic cancers after they’ve undergone one or two different therapies. “We call this next-generation biobanking,” explains Batist, “because it’s biobanking of metastatic tumors in the context of specific treatments.”
Biobanking of serial biopsies allows for analysis of protein expression as it changes under the selection pressure of the first-line treatment. Samples from different timepoints can inform research on biomarker development, tracking changes between primary and metastatic disease, and identifying features of the metastatic process that could become drug targets.
“Once we know the mechanisms of resistance, we can predict resistance earlier,” asserts Batist, “or we can find mechanisms to delay or overcome resistance, and therefore improve the benefit of therapy.”
Batist and colleagues recently conducted a study of metastatic colorectal cancer in which they analyzed exome sequencing data from biopsies collected at diagnosis compared with samples taken after resistance developed. “[We began] defining specific changes in metastatic cancer that are not predictable from looking at banked primary tumors,” Batist recalls.
The tumors that acquired resistance to first-line treatments, they found, had the highest levels of variability over time, suggesting that the current model of banking primary tumor samples may need an overhaul.
“To follow the evolution of technology, we should be willing to abandon some of our long-held beliefs,” Batist declares. “I personally enjoy being very disruptive. I’m happy that’s now considered a positive attribute.”
Bridging the gap between patients and researchers
When Kaley Zeitouni was diagnosed with multiple sclerosis (MS) in 1998, she was just 12 years old. One of the youngest patients diagnosed with MS, Zeitouni embarked on a quest to learn more about her disease and treatment options. MS occurs rarely in kids and teens, and few resources were available to her at that time. Energized by a sense of purpose, Zeitouni rallied classmates and family members and founded a patient advocacy nonprofit, Youth Against Multiple Sclerosis, which has since grown to over 10,000 members. Meanwhile, she researched treatment options and participated in clinical trials.
“She did that with no Google and no cell phone back in 1998—and all from a wheelchair,” says Brian Neman, PhD, a former classmate of Kaley’s. Inspired by her drive, Neman partnered with Zeitouni to found Sanguine Biosciences as a resource for patients and researchers alike. From his experience working on a clinical trial for diabetes, he had seen firsthand the struggle researchers face trying to gather enough patients to populate a study. He thought he could ease that burden by partnering with patients and providing a concierge service to curate their data and connect them with researchers.
“We are helping accelerate medical research by working directly with patients to organize their data and specimens and provide it in an efficient way to researchers that are doing clinical trials,” asserts Neman. “That’s the novelty of our business—that we’re working directly with the individual.”
Sanguine officials say they decrease the burden on patients by accepting the task of collecting and organizing patients’ medical records, as well as providing in-home specimen collection and same-day delivery of samples to clients. For researchers, Sanguine provides a search platform to identify eligible patients. Currently, the company is developing software to further improve the search process, allowing researchers to narrow their searches by defining various inclusion and exclusion criteria, such as co-morbidities and treatment history.
In partnership with dozens of pharmaceutical companies large and small, Sanguine has contributed to over 500 completed studies looking at diseases such as lupus, rheumatoid arthritis, and leukemia. More than 30,000 patients have opted in to share their medical records with Sanguine, Neman points out. Most of them come to Sanguine by way of patient advocacy groups, such as Youth Against Multiple Sclerosis, or by social media, and most are pleased with the process. “The responses are quite strong,” reports Neman. “Once the individuals actually participate [in a research study], we have over a 90% retention rate.”
Nothing beats robots
Peripheral blood mononuclear cells (PBMCs) are an essential tool for studying disease at the cellular and molecular levels. Isolating these cells can be tricky, and success depends quite a bit on the skill of the technician. Different laboratories may obtain products with different yields and quality levels, making it hard to standardize results for clinical trials.
“The process of isolating PBMCs is a poorly reproducible process,” says Carlos L. Aparicio, PhD, CEO and president of ImmunoSite Technologies. “It’s labor intensive, and the skill set of each operator can impact the results.” Because different types of blood cells have different densities, density gradient centrifugation will separate the fractions and create a slim band of PBMCs between the plasma layer and the density gradient medium. That fraction can then be removed from the rest of the blood.
ImmunoSite, which specializes in assay automation, turned its expertise toward the process of PBMC isolation. “We got very good results, but it was a single tube at a time,” recalls Aparicio. At various stages, operator intervention would still be needed to move tubes to a centrifuge, say, or monitor the liquid handler. Full automation of all those steps, the company realized, would be so expensive it would be inaccessible to many laboratories. “The footprint, the complexity, and the cost of the system kept increasing,” Aparicio continues. Consequently, it appeared that the system would be limited to laboratories running large numbers of samples daily.
To create automation solutions flexible enough to help smaller labs, ImmunoSite explored modifications that would allow clients to keep their existing equipment. “Our strategy now is more of a modular approach to independent operations in the process,” Aparicio explains. “Yes, there will be interventions by the users, but they will involve relatively low-skilled activities by the user.”
The modular approach means that laboratories can incorporate automated steps into their setups in a way that makes sense for them. In some cases, ImmunoSite works with the end user to modify or reprogram its existing equipment to create an efficient automated process for consistent, reproducible results. Automating the most labor-intensive steps in the isolation workflow not only saves on labor costs, it also helps eliminate operator variability and errors.
Instead of recommending that end users replace laboratory machines, such as centrifuges, with costly automated versions, ImmunoSite focused on providing optimized consumables. ImmunoSite started out using off-the-shelf tubes and pipettes, but soon realized these weren’t ideal for the robotic systems the company was implementing. “Now we are creating our own centrifuge tubes, which are more amenable to an automated process,” asserts Aparicio. “Let’s say that our robot has to move in the x–y plane and that it also needs a z movement to collect all the cells. With the new consumables, we will have the robot move only in the z direction. That will save time and reagents and give us a better quality of the sample.”
Still, differences in machinery leave some wiggle room for variability in results between laboratories. That’s why ImmunoSite’s next goal is to standardize the equipment it offers.
“Now, we’re going to go a little further and either design our own system or really capture one system that we like,” declares Aparicio. “If it’s one single platform, one single consumable type, everything is more standardized, and that makes it more scalable for our clients.”
If you can’t find it, you can’t use it
Sometimes biological samples are collected with a specific purpose in mind, but other times they may be banked so they may serve purposes to be determined in future studies. Successful biobanking depends not only on collecting high-quality specimens, but also on having complete and easily accessible data for the available specimens. Without sophisticated management tools, the samples will languish in storage instead of being readily accessible for research.
Amgen has developed an informatics platform that compiles data from a variety of sources, making it readily available to researchers and amenable to sifting and sorting according to any number of criteria. “The scientist can drill down by disease, adverse events, additional lab values, medical history, or demographics,” indicates Lynn Wetherwax, senior manager, translational sciences operations, biobank, Amgen. “From there, the scientist can identify specimens that are available and appropriately consented for the proposed research and then order them for analysis to support a research hypothesis.”
Amgen’s Translational Research Information Platform (TRIP) integrates data from sources including clinical trial data, consent data, and specimen inventory data, enabling researchers to search, sort, and analyze that data through an easy-to-use portal. “The systems are designed to harmonize the data so it can be easily be queried for research purposes,” Wetherwax adds.
TRIP houses over 400 clinical datasets, more than 1600 nonclinical studies, almost 3 million biospecimens, plus genomics data. To make the best use of the large investment that a biobank requires, transparency of the inventory is critical. By providing easy access to comprehensive biobank data in a standardized format, Amgen’s platform is already facilitating research collaborations and accelerating the pace of discovery.
