Paraffin-embedded tissue samples like these contain DNA and RNA that could harbor cancer’s secrets.Credit: MaXPdia/Getty Images
Patricia Basta does not want to disappoint her colleagues. Cancer researchers approach her regularly for help obtaining nucleic acids from preserved tissue. One might want DNA from breast tumours for a genetic study of treatment failures. Another might want to analyze DNA methylation profiles in brain tumours to understand epigenetic changes as tumours progress and are treated.
But the tissue samples received by Basta, who directs a biospecimen processing core and repository at the University of North Carolina at Chapel Hill, are almost invariably FFPE tissue samples —preserved in formalin and embedded in paraffin wax. This ubiquitous method of preserving tissue is considered the gold standard for tissue preservation, and has been used by biobanks for more than a century. However, it can be difficult to extract enough high-quality DNA and RNA to analyze with modern quantitative PCR or next-generation sequencing methods, and current extraction methods that lose or degrade material exacerbate the problem.
“We do the best we can,” Basta says. “But when you don’t get what you need, it’s a disappointment for everyone.”
Stored in pathology departments and biobanks all over the world, FFPE tumour samples provide a unique opportunity for researchers to link molecular profiles and clinical outcomes without having to obtain fresh or fresh-frozen samples from patients.
In the past, scientists had to rely mostly on fresh tissue samples as material for molecular analyses. But more recently, scientists have shown that molecular signals teased from archived samples can differentiate cancer subtypes, and reveal what drives tumours to spread and become more aggressive over time.
“If we’re able to look at large numbers of patients — which is possible with archival tissue — then we can assess known variations in populations by age, sex, or race, and other things which we ordinarily can’t do because we simply can’t collect enough fresh-frozen material,” says Robert West, a pathologist and cancer researcher at the Stanford University Medical Center. “An improved ability to pull genetic signatures from FFPE samples would broaden our ability to study and predict treatment responses, and connect research findings with clinical outcomes.”
But wider efforts to mine FFPE samples for insights into cancer have run into a persistent stumbling block: Although FFPE preservation retains a tissue’s cellular details, it alters its DNA and RNA, making extraction and purification more difficult. A new extraction technology could change that.
Laborious Extractions
Next-generation sequencing helps researchers analyze huge numbers of fragments by determining where they overlap and then knitting them together into a virtual sequence. The ability to work with fragmented RNA or DNA has raised hopes for genomic or genetic research using FFPE tissue samples. “But traditional extraction methods work well only if you have about a hundred or more cells worth of DNA or RNA,” says Juan Santiago, a professor of mechanical engineering at Stanford University.
FFPE samples make that difficult because formalin, which is formaldehyde in solution, covalently cross-links DNA, RNA, and proteins to each other. To free the nucleic acid for extraction, these crosslinks must be reversed. When the crosslinks are reversed during lysis, a step just prior to extraction, the nucleic acids get degraded into small fragments that cause poor analytical results.
Commonly used solid-phase extraction methods require removing paraffin wax by physical and chemical means, treating samples with chemicals that remove covalent crosslinks, letting RNA and DNA bind to silica columns, washing away extraneous material or impurities, then washing off the nucleic acids into an elution buffer. The process is laborious, often requires toxic chemicals, and doesn't work well when there’s not much DNA or RNA to extract. This often causes labs to burn sample and time trying repeatedly to extract enough material.
“We've had all this wonderful progress in sequencing and bioinformatics, but the methods for purifying DNA and RNA are 40 years old,” Santiago says.
To better extract nucleic acids from FFPE samples, scientists have tried nonchemical methods for removing paraffin, as well as solid-phase extraction methods that bind RNA and DNA to magnetic beads instead of columns, but with only limited improvement. To improve extraction efficiency, Santiago turned to a microfluidic approach called isotachophoresis (ITP).
Instead of separating molecules by having them physically bind to a surface, ITP separates them based on how fast they travel through an electric field. ITP has long been used to separate ions in analytical chemistry. Santiago adapted it for nucleic acid purification, and later co-founded Purigen Biosystems, a genomic solutions company in Pleasanton, California. The company now sells an instrument and kits that use ITP to automate the extraction and purification of DNA and RNA from FFPE samples with far less manual effort than earlier methods. (Santiago is an advisor to the company.)
Our instrument “really simplifies the workflow,” says Klint Rose, Purigen Biosystem’s other founder and its chief scientific officer. After preparing a sample, “you can load it into our instrument, press go, come back in an hour, and you'll have purified DNA or RNA for your downstream analysis.”
Searching the Archives
Easier extraction of FFPE samples could eventually pay off in the clinic, generating better results from existing therapies. That’s the goal of University of North Carolina investigators who asked Basta to extract RNA in samples isolated from a patient before and after lymphoma treatment. Only a few tumour samples were available, and Basta worried that yields from a standard column-based extraction would be too low. Instead, she sent the samples to Purigen Biosystems, which used its Ionic Purification System to extract them.
The experiment yielded enough RNA to generate an in-depth read of the transcriptome, Basta says. The results were “remarkable considering how little starting material we had to work with.” The analysis could reveal biological reasons why chimeric antigen receptor T-cell (CAR-T) immunotherapy sometimes fails.
Archived samples could also yield fundamental insights into cancer biology. In his laboratory at Stanford, West studies pre-invasive neoplasias to spot the biological changes that let small precursor breast lesions called ductal cell carcinoma in-situ (DCIS) grow into invasive breast cancer. Few fresh-frozen DCIS samples are available for research, and to trace the biological transformation, “we need to look at large numbers, and for that we need archival material,” West says.
Studies on archived samples have begun identifying genetic risk factors and biomarkers. Researchers from Charité–Universitätsmedizin Berlin have shown that gene signatures in FFPE breast-tumour samples correlate with HER2 status, which is valuable information for predicting disease severity and potential response to targeted treatments.1 And University of Manchester researchers classified diffuse large B-cell lymphoma into genetically defined subtypes — an important step for discovering new drug targets and developing better therapeutic options.2
“Such retrospective studies will ultimately benefit people with cancer,” Santiago says. Now that the promise of more efficient isolation is being realized, “we can begin to exploit this amazingly large resource of FFPE tissues sitting in cabinets and warehouses all over the world.”