June 1, 2018 (Vol. 38, No. 11)
Biomarkers, the Molecular Signatures Characteristic of Disease, are Seldom Hidden in Plain Sight
Despite his prodigious powers of deduction, Sherlock Holmes would have failed to solve many a whodunnit had he lacked his trusty magnifying glass. Today’s biomedical sleuths face a similar predicament.
They’re terrifically skilled in figuring out disease mechanisms, but they rely on investigative tools. The outcome of many a biological whodunnit, for example, depends on researchers and clinicians having the right biomarkers, measurable indicators of biological states or conditions. Biomarkers can reveal clues—molecular signatures and disease-related patterns—that would otherwise remain hidden.
If one finds that the right biomarkers are not ready to hand, the solution is elementary. One must develop them. Less obvious, however, is how the average biomedical researcher should go about solving the case of the missing biomarker. Since help is no longer available at 221B Baker Street, biomedical researchers seek guidance at conferences instead, such as the 11th Biomarker Summit, which recently took place in San Diego.
At this event, would-be Sherlocks and latter-day Watsons alike learned how biomarker and diagnostic development can be taken from discovery to translation to commercialization. Prominent themes included the importance of collaborating across multiple scientific disciplines, accelerating the discovery of tumor neoantigens, employing whole-slide multiplexing to retain spatial context of the tumor microenvironment, and deploying ultrasensitive digital ELISA, to detect brain injury biomarkers in blood.
With the growing availability of patient-specific information derived from proteomics, metabolomics, genomics, etc., a paradigm shift is occurring from one-size-fits-all healthcare to personalized medicine. “The ultimate goal is to provide the right treatment to the right patient at the right time,” remarked Bruce W.M. Jordan, Ph.D., vice president, international business leader, Personalized Healthcare Solutions, Centralized and Point-of-Care Solutions, Roche Diagnostics. Dr. Jordan gave the keynote address at the conference.
“We need to better understand disease biology by looking across scientific disciplines,” he explained. “In oncology, one can look for overexpression or genetic changes, but for other diseases, identifying specific biomarkers of a drug target requires a more sophisticated analysis.”
Dr. Jordan elaborated while bringing up the example of Alzheimer’s disease (AD): “Diseases are rarely monogenic. A biomarker signature involving several pertinent biomarkers would better define disease subtypes and stratify patients in clinical trials. Protein folding is probably more critical than genetic profiling in AD.”
To obtain such relevant biological signatures, Dr. Jordan advised, investigators should engage in a more holistic approach to biomarker identification. “This process will be strongly enhanced as people take up a cross-disease approach,” he maintained. “For example, people are considering how inflammation may underlie molecular pathologies across central nervous system diseases. I applaud the encouragement of discourse across disciplines because innovation often happens at the space between disciplines.”
Although oncology has been a main focus, future targets include central nervous system diseases as well as cardiovascular diseases and inflammatory diseases. Dr. Jordan said that he believes the shift has already begun: “Within the next five years, we are likely to see some game-changing advances in understanding biochemical and disease drivers that will hopefully lead to improved earlier diagnosis and to meaningful treatment options for people with dementia.”
Tumors often accumulate somatic mutations that can alter downstream protein expression, creating unique amino acid sequences that occur in neoantigens. New therapies aim to turn the tables on such malignant transformations by utilizing these changes for targeting or assessing therapeutics.
“In the last three years, there’s been an explosion in drug therapies seeking to direct the immune system to recognize tumor antigenicity,” remarked Steve Siferd, senior director of biopharma business development at Cancer Genetics. “Historically, identifying tumor-specific mutant antigens was a long and protracted process involving extensive molecular cloning and cumbersome screening methods to identify which neoantigens correlate to response to immunotherapy.”
To expedite neoantigen identification, the company developed an integrated discovery pipeline. “While others check primarily at the DNA level, our AntigenID™ workflow also profiles RNA expression to prioritize mutations likely to be neoantigens,” Siferd explained. “The extensive computational analysis of DNA, RNA, and the immune repertoire [can help investigators] understand the total mutational burden in the context of the tumor microenvironment.”
The collected data is used to rank-sort the results. “Often companies are trying to determine which patients respond to the therapy and which do not,” Siferd noted. “We can generate a list of neoantigen candidates listed by significance and how best to validate those. At the end of the day, we work with clients in an interactive and collaborative way to help them optimize their current therapeutic and identify biomarkers that better stratify patient responses to immunotherapy.”
Although multiplexing can be used to detect and analyze many biomarkers from one sample, much of the technique’s potential is unrealized because in many analyses, samples undergo homogenization, which destroys all spatial information about the tumor microenvironment.
“There are multiple cell subpopulations in the tumor microenvironment, including tumor-infiltrating lymphocytes and stroma cells,” observed Louis Levy, director of corporate and business development at Ultivue. “These heterogeneous cell populations define the tumor microenvironment whose analysis has become the new game-changing challenge of immuno-oncology.”
According to Levy, histopathology must not only detect multiple cell populations, but also determine how they interact spatially. To facilitate spatial analyses, Ultivue has developed the InSituPlex™ technology platform, which images and analyzes marker protein targets in formalin-fixed paraffin-embedded tissue samples. The immunofluorescence technology uses stable DNA–DNA interactions.
Biomarker antibodies are tagged with a unique DNA barcode. Then, a complementary DNA fragment coupled to a fluorophore binds to the barcode, enabling protein marker detection and quantification. “In a fiveplex whole-slide assay format, InSituPlex relies on a single antigen retrieval step, a single staining, a single amplification step, and a single imaging step,” Levy detailed. “Our kits are compatible with existing workflow and instrumentation and provide the performance and flexibility demanded in a wide range of use cases from clinical research and translational research to discovery applications.”
The company offers access to its platform as a service for clients and will soon release validated fiveplex kits for whole-slide spatial multiplex biomarker analysis. “Such analysis,” Levy predicted, “will be instrumental to designing quantitative scoring and objectively demonstrating the clinical relevance of specific markers in the context of companion diagnostic test development.”
Central Nervous System Biomarkers
Sensitive detection of biomarkers of brain injury and disease has chiefly relied on invasively obtaining samples of cerebrospinal fluid. “Detecting biomarkers in blood is an even bigger challenge because of the ultralow concentrations of such brain-derived proteins,” noted Andreas Jeromin, Ph.D., a scientific and medical advisor at Quanterix.
Trace quantities of brain-derived proteins can be detected by the company’s single molecule array (SIMOA™) technology—and not just in cerebrospinal fluid, but also in serum, plasma, urine, and cell extracts. “This digitally based ELISA format,” Dr. Jeromin asserts, “provides about 1,000 times more sensitivity, on average, than conventional ELISA.”
The technology is based on capture agents such as antibodies that are attached to the surface of paramagnetic beads. Following incubation with the sample, a secondary, enzyme-linked antibody is added, and beads are individually distributed into arrays consisting of 215,000 microwells per array, which can hold only a single bead per well.
The subsequent addition of substrate produces signals: “on” wells (fluorescence) and “off” wells (no fluorescence). The assays not only allow the quantification of low-abundance biomarkers down to single-molecule level, but also span a dynamic range of 4–5 logs. In addition, the technology can be multiplexed and is an open design technology, allowing for custom assay development.
“The SIMOA technology,” Dr. Jeromin emphasized, “can detect proteins, DNA, RNA, and even microRNA, which can be quantified with great sensitivity, similar to that obtainable only at the limits of polymerase chain reaction (PCR) technology, but without the bias and issues of PCR.”
The company’s first-generation instrument, the HD-1 Analyzer, was developed to have a “huggable” size and deliver high-throughput performance. More recently, Quanterix released a benchtop reader (SR-X) for integration with existing plate-washers and readers. Thus far, the sample-conserving, ultra-sensitive technology has been utilized in more than 700 phase trials and cited in more than 100 publications. It continues to expand central nervous system and other applications including infectious diseases, inflammatory diseases, and cancer.
Expecting the Unexpected
Of the many topics that were covered at the Biomarker Summit, those pertaining to biomarker identification, validation, and translation have been emphasized by this article, which encourages biomarker investigators to emulate, in their own domain, the mystery-solving ways of Conan Doyle’s fictional character, who demonstrated how a combination of careful technique and sharp reasoning may expose subtle clues and produce broad and possibly surprising conclusions.
“You know my methods,” Holmes said. “There was not one of them which I did not apply to the inquiry. And it ended by my discovering traces, but very different ones from those which I had expected.”