It was the challenge that attracted Yiannis Ioannou to study Niemann-Pick type C — a neurodegenerative lysosomal storage disease for which there is no treatment — more than 15 years ago. At that time, researchers did not know the cause of the fatal disease, which presents its first symptoms at around age seven or eight. While geneticists have since determined that variants in the NPC1 and NPC2 genes are behind this presentation of the disease, there is still much work to be done to better understand the pathophysiology of NPC and, ultimately, to develop treatments that target the disease pathway.
"We can't test a treatment and say 'OK, we're going to put this number of children on this and then we're going to wait five years to see what happens,' because it's just not feasible," Ioannou says. "And that's where biomarkers come in."
He and his colleagues at the Mount Sinai School of Medicine in New York are working to validate novel biomarkers for NPC in order to track disease progression. He has teamed up with Rong Wang, a proteomics expert at Mount Sinai, to do just that. Using LC-MS/MS, they have detected more than 50 protein and lipid signatures that differ between NPC patients and healthy volunteers. A $423,750 grant from the National Institute of Diabetes and Digestive and Kidney Diseases has put them on track to validate the most promising of these markers.
"The problem with most biomarkers ... is that usually the validation is really difficult. And that means that we have hardly any good biomarkers at present," says Emile Voest, chair of the oncology department at University Medical Center Utrecht in the Netherlands.
Once biomarkers have been validated — that is, they have been tested and measured using "well-established performance characteristics" and are supported by evidence showing that they have physiological, toxicological, pharmacological, or clinical significance, according to US Food and Drug Administration standards — they have reached the point at which they can be put up for "qualification" consideration at the regulatory level. FDA considers biomarkers to be qualified if there is an "established scientific framework" or "body of evidence" that shows they are clinically significant, according to a Biomarkers in Medicine paper published earlier this year.
To meet these regulatory requirements, researchers from industry and academia are teaming up to share expertise and precompetitive data, and these collaborations are just beginning to show some successes.
But perhaps the largest problem facing biomarkers — even qualified ones — for diagnostic use, determining disease progression, and other applications geared toward the personalization of medicine is that, even when taken collectively, individual biomarker investigations cannot possibly capture all of the biological intricacies that cause healthy systems to function properly, and diseased ones to go awry.
"Right now, people spend a lot of time and effort making separate measurements at the DNA level, at the RNA level, and then at the protein level, and then try to stitch the data all back together," says the University of Illinois, Urbana-Champaign's Ryan Bailey. "But, of course, every technique has systematic biases in the way it obtains data, so therefore, just dealing with the heterogeneity of the data-analysis process from these different techniques can make it challenging to track differences across the levels of biological complexity."
[ pagebreak ]
Biomarkers have "always been a part of the conversation" at FDA, says Federico Goodsaid, the biomarker qualification process co-ordinator at the agency. In the last few years, the biomarker conversation at FDA and the European Medicines Agency has hinged on efforts aimed at the regulation or qualification of these biological signatures.
Since the early 2000s, Goodsaid and others at FDA have been crafting and revising a process to qualify individual biomarkers on the basis of their specific use. They have finally tailored the process to their liking, he says, and expect to release a draft guidance for qualifying drug development tools — including biomarkers — after a final round of vetting by year's end.
"The development of novel biomarkers is an extremely important part of the development of personalized health care," Goodsaid says. "We really would like to receive as many qualification requests as are supported by data."
In addition, FDA has its eye on reducing redundancies in efforts to find biomarkers to aid drug development, Goodsaid says. Indeed, through the finalized biomarker qualification process, the agency "would like to be able to qualify biomarkers across multiple drugs so that we get a consensus in the agency of what those biomarkers are useful for ... so that you don't have to keep justifying the use of the same biomarker of the same context of use," he says. To do this, FDA requires that the proposed 'context of use' — such as screening for CYP2C19 variants to determine clopidogrel metabolizer status — for all biomarker submissions be -strictly defined. In this way, the agency hopes that a successful biomarker qualification could benefit R&D as well as trials across pharma.
For biomarker qualification, "the burden of proof is high — as it should be — so it can be difficult to get appropriate, high-quality data that really meets that need," says Sonia Pearson-White, a scientific program manager of the Foundation for the National Institutes of Health Biomarkers Consortium. Data requirements can vary from submission to submission based on trial design, technology platform, and proposed use, among other factors.
No academic lab, pharmaceutical company, or funding body is equipped to bear the costs associated with biomarker discovery, validation, and assay development to the point of regulatory qualification and subsequent laboratory adoption on its own. To that end, collaboration-savvy scientists have spearheaded public-private partnerships devoted to accelerating the development of personalized medicine, and are working to address the biomarker qualification needs common to industry, academia, regulatory bodies, and the patient populations they serve.
"One of the things that we're trying to do is to diminish the risk and cost to individual companies by really pooling efforts; making sure — by working with FDA from the beginning — that we're doing the right thing, collecting the right data, and spreading the risk among the different participants," Pearson-White says of her group. The Foundation for the NIH — along with partners at NIH, FDA, and the Pharmaceutical Research and Manufacturers of America — provided support to establish the Biomarkers Consortium in late 2006 in an effort to facilitate the "development and validation of biomarkers using new and existing technologies," Pearson-White adds.
[ pagebreak ]
The European Union's Innovative Medicines Initiative also supports public-private partnerships in the context of "open innovation," says its director Michel Goldman; qualifying biomarkers for oncology drug development are among the program's top priorities.
In addition, the Critical Path Institute was established in 2004 in response to an FDA report on challenges to developing new medical products — drug products, biological products, and medical devices. The C-Path Institute's Predictive Safety Testing Consortium was subsequently developed in 2006 to allow its corporate partners to "share internally developed, pre-clinical safety biomarkers," and now includes 16 pharmaceutical firms as members, according to Elizabeth Gribble -Walker, director of the PSTC.
These groups — the Biomarkers Consortium, PSTC, and IMI — all function as neutral liaisons between pharma and its regulators. As such, consortia leaders are constantly striving to maintain the delicate balance of cost-sharing in a precompetitive environment. They are also pushing to inform regulatory decision making — with some substantial successes to date.
According to Walker, PSTC operates under the principle of "broad participation, broad input." The Biomarkers Consortium, too, aims to distribute biomarker qualification costs — time, intellectual property, and financial commitments — across 61 contributing members from NIH, pharma, academia, nonprofits, and FDA. "We can't be in a space where we're just going to validate one company or one marker," Pearson-White says. In an effort to maintain a fair balance, the Biomarkers Consortium aims to qualify myriad biomarker types across diverse technological platforms.
FDA's Goodsaid says that one of the largest successes of public-private consortia is that they are able to bring together pharmaceutical companies that likely will see no tangible benefit from their collaborations — at least right away.
One small success came in June 2009 when a team led by Merck Research Laboratories' John Wagner, along with the Biomarkers Consortium's Metabolic Disorders Steering Committee, published their demonstration of the utility of adiponectin as a biomarker to predict glycemic efficacy in Clinical Pharmacology & Therapeutics. In a December 2009 editorial in the same journal, Eli Lilly's Steven Paul and Steven Eck note that the "Biomarkers Consortium provided the basic vehicle that enabled sharing of the primary data and analysis by agreed-upon standards."
Indeed, Pearson-White says, the team "got some really ground-breaking data-sharing agreements together ... and developed ways of maintaining confidentiality and addressing antitrust issues" among the participating sponsors.
As yet, adiponectin has not made it "to the level where it [is] ready for regulatory qualification," Pearson-White says, adding that the team is "in the early stages of putting together a follow-on project" that includes designing a prospective clinical trial to really "pull it to that level."
PSTC's Walker says she has witnessed an "enormous reluctance" in pharma to use novel toxicity biomarkers that companies have discovered and validated in-house "because of the fear that — FDA, EMA, whomever — the regulators are not going to interpret the data the same way that the pharmaceutical company has."
To move beyond this reluctance, PSTC had to ensure that the data interpretation for a set of seven novel biomarkers for nephrotoxicity was uniform across the researchers and the regulators. The team compiled validation data from industrial sponsors, independent labs, and FDA itself before submitting the markers for qualification. In June 2008, both FDA and EMA issued statements outlining their joint qualification of the kidney biomarkers as "acceptable in the context of non-clinical drug development for detection of acute drug-induced renal toxicity." Both agencies acknowledged that the seven PSTC markers "provide additional and complementary information to the currently available standards," and said that they would consider the clinical use of these biomarkers on a case-by-case basis.
[ pagebreak ]
According to Walker, 12 of the 16 firms that participated in the nephro-toxicity submissions are now actively incorporating the seven markers in their pre-clinical renal toxicity testing.
Still, adiponectin serves as a reminder that extensive collaborations and additional validation data alone may not be enough to push biomarkers through the qualification process and into the clinic.
Individual biomarkers, even some that have made their way through the regulatory system to be qualified, may not be enough to capture the complexities of health and disease. To get a better picture of how healthy systems can misfire, researchers are combining tools and developing new technologies to assess multiple biomarkers simultaneously.
Over at Dartmouth's Norris Cotton Cancer Center, Murray Korc, Lorenzo Sempere, and their colleagues are using genetic and protein biomarkers in combination to more accurately characterize pancreatic masses. Korc, chair of Dartmouth's department of medicine, says that it can be difficult to distinguish malignant pancreatic cancer from benign masses due to the confounding effect of chronic pancreatitis. As such, he and Sempere are quantifying both cancer-specific microRNA signatures and protein biomarkers; while they investigate miRNAs as indicators of disease status, they can simultaneously detect the cell type of origin based on the co-expressed protein markers.
"These microRNAs have been reported to be over-expressed in cancer, and the implication is that they're expressed in the cancer cells," Korc says. "Just because the microRNA is elevated in cancer doesn't necessarily mean that it's in the cancer cells," Sempere adds. "By combining the microRNA signal with protein markers [that] are expressed in different cell types, you can better interpret where the changes of those microRNAs occur."
Korc says that, down the line, the co-detection of miRNA and protein biomarkers "may tell us which patients will respond to which treatments," allowing clinicians to personalize cancer therapies.
It is this intra- and even inter-personal variability that has driven UIUC's Bailey to develop a multiparameter miRNA-protein biomarker detection and quantification platform that he has shown to be highly sensitive and specific. He also says it provides extensive potential for multiplexing and mass production.
In 2007, Bailey was granted an NIH Director's New Innovator Award to develop and validate his arrays of silicon photonic microring resonators, which allow for the concurrent, label-free detection of clinically relevant proteins and miRNAs.
Photonic microring resonators are optical waveguides that, when coupled with light, support resonant optical nodes that are sensitive to the refractive index at the surface of the capture structure. For biological applications, the micro-rings can be functionalized via a capture agent — antibodies for protein capture, cDNA for miRNAs — so that binding events "displace a small amount of water," Bailey says. "Since organic molecules like proteins and DNA have a higher refractive index than water, this changes the refractive-index environment near the microring and ends up changing the resonance wavelength of the cavity."
[ pagebreak ]
In a series of three papers published between November 2009 and June 2010, Bailey and his team demonstrated the applicability of their technique to capture and quantify carcinoembryonic antigens, a panel of five protein markers — CEA, PSA, EFP, TNF-a, and IL-8 — and several miRNAs implicated as "valuable disease markers." At present, Bailey and his team are working to tease out capture agent issues that "limit sensor performance" in order to bring his chips to the masses.
"Our thought is that you can learn a lot more about a system if you can measure more of the proteins and nucleic acids from that system. The information is really powerful; information is the key driver here," Bailey says. "We're trying to integrate multiple levels of biomarkers onto the same chip, because they all contain a slightly different piece of complementary information about the relative health or disease in a system."
Bailey emphasizes the importance of co-developing technologies to capture biological complexity while investigators work to qualify biomarkers for clinical applications. "You can discover all of the biomarkers that you want, but if you don't have cost-effective solutions for detecting those in practical populations, the downstream applicability of those biomarkers is going to be significantly limited," Bailey says. "I think there's a lot to be said for coupling discovery and validation ... [while] developing practical downstream diagnostic solutions."
At the intersection of biomarker qualification and clinical quantification, PSTC's Walker expects that as regulators solidify their guidelines and as labs submit markers for qualification, more and more technological platforms — like Bailey's — will become fit to measure increasingly relevant endpoints.
"I see all agencies as being actively interested in reviewing novel biomarkers," she says. "Once the agencies have their processes in place and have gone through a number of reviews ... and disseminated the results of those, I really see that anything that has practical applications and utility can then use that [established] pathway" to qualify markers using innovative, multiplex-capable technologies going forward.
SIDEBAR: Biomarker Qualification in a Nutshell
The first draft of the Biomarker Qualification Pilot Process, issued in March 2007, included six steps from submitting a biomarker for qualification to the decision. First, sponsors submitted their request to the Interdisciplinary Pharmacogenomic Review Group to qualify a biomarker for a specific use; the IPRG then assembled a Biomarker Qualification Review Team, whose primary function was to "evaluate study protocols and review study results for the qualification of novel biomarkers ... using appropriate pre-clinical, clinical, and statistical considerations." The review team then assessed the biomarker context and available data from any Voluntary Exploratory Data Submissions in an effort to determine a qualification study strategy upon which the sponsor and the agency could agree. Finally, FDA reviewed the qualification study results and either accepted or rejected the biomarker for its suggested use.
In a Biomarkers in Medicine paper published this April, Goodsaid and his colleagues describe a refined, dual-phase approach to the biomarker qualification process at FDA. The first phase — or the "evaluation process" — involves a Biomarker Qualification Coordinator, who receives and evaluates the initial sponsor request. From there, a review team "requests a briefing document and data from the sponsor to evaluate the exploratory biomarker." The team and sponsor then meet in person to request additional data and to initiate the second phase of the process or to reject the request at this stage. During phase two, the review team evaluates the full data package and drafts recommendations for directors at various offices within the Center for Drug Evaluation and Research. These directors make the final decision and send their rejection or acceptance of the biomarker to the sponsor. Accepted biomarkers are then subjected to the "appropriate actions," such as an FDA-issued label update, or inclusion in an agency biomarker database.
"We strongly believe that this is the right process for qualifying biomarkers," Goodsaid says, adding that "we want this process to serve not only to qualify biomarkers, but also to promote the development of new and novel biomarkers. ... We want this to really pull through the whole pipeline in biomarker development."