With Biomarkers, Don’t Trust, Just Verify

Everyone Agrees That Biomarkers Can’t Be Taken at Face Value

Biomarker development is hardly coextensive with drug development, but in every developmental stage—from exploratory research to clinical trial to therapeutic application—there is significant overlap.

As medicine becomes more personalized, more dependent on molecular insights, the overlap between drug development and biomarker development is bound to grow, as genomic and proteomic investigations have demonstrated.

Molecular biologic methods, chiefly genomic and proteomic techniques, are increasingly capable of tracking multiple biomolecules in parallel. Consequently, it is becoming easier to identify potential biomarkers—characteristics that may be objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to a therapeutic intervention.

This definition for biomarkers, it should be noted, is not limited to genomic and proteomic features. It also encompasses other kinds of biomarkers—physiological measures (such as blood pressure), imaging results (such as available via radiographs), and basic blood analytes (such as glucose levels). All these measures may be used in combination to strengthen an assessment, that is, to increase a developer’s (or a regulator’s) confidence that a candidate biomarker means what it is supposed to mean.

It is here that we come to the crux of the matter: How do we determine whether a particular biomarker is a reliable guide for a particular purpose? As it happens, there are degrees of trust, each appropriate for a different purpose. For example, the “fit for purpose” standard suffices for a developer’s internal deliberations, or for publishing research results. More stringent requirements must be satisfied if biomarkers are to achieve validation and/or qualification. (For more information about these requirements, see this article’s sidebar, “Biomarker Validation and Qualification—the FDA Way”).

Example of a biomarker assay that uses hybrid technology. A purified monoclonal antibody is linked to magnetic beads, then added to the sample. A magnetic field separation is used to isolate the protein or peptide of interest. Further analysis is done via liquid chromatography-tandem mass spectrometry (LC/MS-MS). [Pfizer]

Biomarker Applications

Applications of biomarkers range from setting up diagnostic and prognostic screens, to predicting drug efficacy, to serving as surrogate endpoint in a clinical study.

“Clinical biomarkers must be simple and robust enough to be available in a hospital laboratory setting,” states Mahmoud Loghman-Adham, M.D., medical director for Shire. “For example, a mass spectroscopy assay that works well in a research setting is far too specialized and expensive to be employed in a clinical situation.

“Until about 10 years ago, biomarkers were an afterthought, Now, biomarkers are developed in the preclinical stage, beginning in animal models, continuing through the phases of clinical development.

“There is a great deal of interest in finding biomarkers that predict the risk for developing a disease, long before the symptoms become clinically evident. Given the efficacy of early interventions for cancer and cardiac problems, well as kidney diease, this is most welcome.”

According to Dr. Loghman-Adham, clinical trials could be accelerated or treatments could be started more expeditiously if appropriate biomarkers were available. For example, if a biomarker were available that let clinicians catch acute kidney injury early, they could act before the condition progressed beyond treatment.

At present, the detection of kidney disease depends on “traditional” laboratory tests such as the determination of BUN (blood urea nitrogen) and creatinine levels. Unfortunately, these tests tend to indicate a problem long after damage has occurred.

Despite such limitations, Dr. Loghman-Adham remains optimistic. “Biobanks of blood and other bodily fluid samples are a rich source that can be mined to discover and validate new potential biomarkers for various maladies,” he notes. “The patient information must be de-identified, and the patients must consent to having samples stored for several years to be used for further studies.”

Comparison of the change in the level of an ideal biomarker with the changes in the levels of traditional markers of kidney function such as glomerular filtration rate (GFR) or serum creatinine. Changes in traditional markers often occur late in the course of disease, when irreversible damage may have already occurred. [Mahmoud Loghman-Adham, M.D., Shire]
Public-Private Partnerships

“In Europe, there are hundreds of publicly funded biomarker projects, designed to support medical decisions with the final goal of tailoring treatments for individual patients,” states Thomas Joos, Ph.D., deputy managing director, Natural and Medical Sciences Institute, University of Tübingen. “New biomarkers will better predict drug response and allow the best treatment regimen selection for each patient.

“You need to combine robust assay technologies, solid assays, and high-quality samples for biomarker discovery and validation projects. Then data analysis is simple and fun. Specifically, we look for biomarkers that indicate an early stage of a disease. Ideally such a biomarker should not be present or be increased under healthy conditions.

“Biomarker qualification processes are resource and time-consuming, requiring a careful selection of representative sample sets. Hence, the success rate of such projects is more likely to be achieved in a combination of private and public partnership programs.”

Dr. Joos notes that in Europe, there is a public-private partnership between the European Union and the European pharmaceutical industry. It is called the Innovative Medicines Initiative (IMI), and it serves to speed up the development of better and safer medications for patients.

“The IMI accelerates drug development by supporting collaborative research projects between industrial and academic partners,” explains Dr. Joos. “The IMI also supports biomarker efforts such as the IMI SAFE-T project, which qualifies biomarkers to be used in assessing drug-induced organ injuries.”

In the United States, similar work is carreid out by the Predictive Safety Testing Consortium (PSTC). This group brings together pharmaceutical companies to share and validate innovative safety testing methods under advisement of the FDA, the European Medicines Agency (EMA), and the Japanese Pharmaceutical and Medical Devices Agency (PMDA). Thus far, the PSTC has agreed upon a set of proteins that can indicate liver damage.

“We cannot see the dynamics of a living cell when looking at genomic data,” points out Dr. Joos. “Proteins studied with immunoassay techniques provide an excellence source for biomarker candidates and provide a much closer look into cellular activities.”

Biomarkers in Clinical Trials

 “Developing and validating biomarker assays to support clinical trials presents many challenges,” states Jenny Y. Zhang, M.D., senior manager and biomarker assay specialist, Clinical Assay Group, Pfizer. “Compliance with FDA guidelines is key to successfully launching a bioanalytical assay. Keeping an open channel of communication with regulatory entities is critical for successful biomarker development.

“Our group works mostly with pharmacodynamic (PD) biomarkers, which demonstrate the connection between a drug and its desired biological effect.”

In recent years, a higher degree of emphasis has been placed on researching biomarkers, as they play a key role in the decision to continue with or abandon a project (go/no-go decisions), maintains Dr. Zhang.

“Pfizer develops and validates the PD biomarker assays for clinical trials. Data generated by a validated biomarker assay is often used to select the appropriate dose of a new drug in clinical trials. The proof of mechanism is linked to the proof of concept via the biomarker data,” says Dr. Zhang.

“Variability in biomarker data comes from many sources: the intrinsic nature of the biomarker; variability within a patient and between patients; whether the medication is taken with or without food; and the circadian rhythm of the biomarker. Hence, one of the goals before starting a clinical trial is to produce a validated biomarker assay. The analytical method needs to differentiate between normal and disease samples and evaluate the progression from one state to another.’’

Getting accurate measurements of a biomarker can be challenging for a number of technical reasons, according to Dr. Zhang. Since no blank matrix sample exists, free of endogenous analytes, the choice of a substitute matrix is challenging.

Other hurdles include the fact that several isoforms of the same protein can be found in the body, which can confound the biomarker assay. One approach to address this problem is to combine several techniques.

“Using mass-spectroscopy alone to quantitate the level of oxytocin in serum did not achieve the desired sensitivity, whereas the employing the antibody quantification was not specific enough,” informs Dr. Zhang. “But by combining the techniques of antibody recognition and magnetic beads to enrich and isolate oxytocin from the serum, we were then able to selectively and precisely measure serum oxytocin using a hybrid mass spectroscopy approach.”

Sample integrity is another issue. Attention must be paid to the details of sample handling. How quickly will the samples be frozen after collection? Does the protein or peptide used as a biomarker undergo any further biological degradation after sample collection? Can a preclinical biomarker be translatable to a clinical application? Can we measure biomarkers in an accessible and useful target matrix?

All these questions must be investigated and resolved before initiating a clinical trial, especially when the biomarker data are used for primary and secondary endpoints. Exploratory biomarker assays utilized for internal decision making do not require the stringent validation of a clinical biomarker assay.

Assay Development Services

Support for those hoping to build biomarker assays that have multiplexing capabilities is available from Meso Scale Discovery, a division of Meso Scale Diagnostics (MSD). The company also offers a service to develop assays for biomarkers, says Fiona Coats, Ph.D., MSD’s vice president of marketing.

“Normally, our customers in this area are large pharmaceutical/biotechnology companies,” she notes. “We work with clients from the initial stages of assay development through to providing them with a validated assay for one or multiple biomarkers.

“Our expertise is in the area of developing a robust assay, using our knowledge in electrochemiluminescence (ECL). Scientists use our assays to get precise and reproducible measurements of proteins in complex sample matrices.”

Most people are using proteins as biomarkers, but there is also a trend toward using DNA or RNA as biomarkers. The custom-coated plates offered by Meso Scale Discovery can be modified to detect nucleic acids as well.

Meso Scale Discovery recently launched U-Plex®, a platform that allows people to design their own multiplexed assay. Crucially, this allows researchers to choose their own panel to build a biomarker assay, when a commercial assay may not be available. The U-Plex platform is designed to help customers to create their own biomarkers panels while allowing customization by species, antibodies, and number of plates.

U-Plex is used along with MSD’s Multi-Spot® immunoassay plates, which support multiplex assays in 96- and 384-well formats.

“Each well of the MSD Multi-Spot® plates has the capacity for 10 unique spots to measure 10 different analytes,” details Dr. Coats. “MSD’s platform assists with the challenges of multiplexing and reproducibility in biomarker development.

“A positional camera is used in the plate reader to determine the signal from each of the 10 spots in the well. ECL technology has very low background and cross-signaling for great sensitivity. Many of the commercially available ELISAs are colorimetric in nature and do not allow detection at the picogram level of protein concentration.

“Technology from Meso Scale Discovery permits measurement of a very large dynamic range of multiple targets in the same well. This is beneficial for our customers who have a very limited sample size, say from a biobank.”

Stringency is a keystone of biomarker validation. Despite these challenges, much progress is being made in characterizing and qualification of biomarkers for a variety of applications.

The concentration of the protein of interest within a sample is a critical aspect of target validation. In blood alone, there about 100,000 proteins spread over a large range of concentrations—normally 6 to 10 orders of magnitude. Most novel biomarkers are in the lower end of the range; hence immuno-based assays are needed to achieve the required sensitivity. [Thomas Joos, Ph.D., University of Tübingen]

Biomarker Validation and Qualification—the FDA Way

“One must be careful to distinguish between validation and qualification,” states Mahmoud Loghman-Adham, Ph.D., medicaldirector, Shire. “There are carefully defined regulatory pathways for biomarker validation and qualification.”

According to one review (Hunter et al., Current Drug Targets), validation is the process of assessing the biomarker and its measurement performance characteristics, and determining the range of conditions under which the biomarker will give reproducible and accurate data, whereas qualification is the evidentiary process of linking a biomarker with biological processes and clinical end points. In short, validation is about a biomarker’s performance metrics, whereas qualification is about a biomarker’s approved meaning.

“Validation is the process of showing that the biomarker accurately predicts the diagnosis, prognosis, or clinical endpoint in question,” explains Dr. Loghman-Adham. “Qualification requires interaction with regulatory health authorities such as the FDA. It is a laborious process that normally requires a consortium to be formed between academia and industry to bring a biomarker to fruition.”

The FDA has established detailed guidelines for the validation and qualification of biomarkers. For example, in 2013, the FDA updated its Bioanalytical Method Validation document by issuing a draft guidance that included biomarkers and diagnostics. The FDA has also established the Biomarker Qualification Program, which handles requests for regulatory qualification of a biomarker for a particular “context of use” in drug development. Once a biomarker has been approved through this process, it can then be employed in clinical trials, as a companion diagnostic, and for other health-based situations.