Send to printer »

Feature Articles : May 1, 2012 (Vol. 32, No. 9)

Biomarker Discovery Methods Evolving

  • Greg Crowther, Ph.D.

Biomarker research has made substantial progress recently beyond the proverbial low-hanging fruit. Emerging technologies and techniques are helping make biomarker detection and application faster, cheaper, and more diverse. This was the focus of talks at GTC’s recent “Biomarker Summit” conference.

Autoantibodies, which recognize the body’s own antigens, can function as biomarkers not only for autoimmune disorders but also for diseases like cancer, in which atypically high levels of proteins can trigger autoantibody production.

Both Life Technologies’ ProtoArray® microarray and its ProtoPlex™ Luminex® assay have roots in the company’s successful recombinant expression of over 9,000 full-length human proteins. These proteins are printed on the ProtoArray microarray, which can be used to identify possible differences in autoantibody prevalence among samples from different populations (patients versus healthy controls, for example).

According to Lisa Freeman-Cook, Ph.D., manager of R&D, researchers consider the ProtoArray microarray a good tool for discovering autoantigen/autoantibody candidates, but have had trouble following up on these candidates with new samples and independent methods. “They were using many different low-throughput techniques—ELISAs, Westerns, immunohistochemistry—and were getting stuck in the workflow,” she said.

The new ProtoPlex assay, which Dr. Freeman-Cook described as “a multiplexed ELISA on beads,” addresses this bottleneck. A client may select up to 80 candidate protein antigens (chosen from Life Technologies’ library and/or provided by the client) to be included in every well of a 96-well plate; each antigen is distinguishable from the others because it is conjugated to microsphere beads with a unique dye-based “bar code.” Thanks to the multiplex design, as little as 3 microliters of serum is required for testing in triplicate.

Validation of the ProtoPlex assay has come in part from a study of systemic lupus erythematosus (SLE). Forty-seven possible autoantigens originally identified with the ProtoArray microarray, along with known autoantigens such as Ro52, were tested against new serum samples from 174 women who had SLE, another autoimmune disease, or were healthy. Fifteen of the antigens remained strongly associated with SLE in the new study; each reacted with antibodies from at least 20% of the SLE samples but did not react with control samples.

“ProtoPlex is an ideal custom validation platform because each customer will be interested in a different subset of proteins from their discovery study,” Dr. Freeman-Cook noted. Life Technologies has anticipated these interests to some extent by scanning customers’ published reports, making a list of potential cancer autoantigens, and conjugating 116 of them to beads for use in the ProtoPlex assay. Increasing use of this assay to discover novel cancer biomarkers thus appears likely, she remarked.

Ezose Sciences grew out of a Japanese consortium between scientists at Hokkaido University and industry partners such as Shionogi & Co., a mid-sized pharmaceutical company that launched Ezose as a U.S.-based subsidiary in 2009. Its specialty is high-throughput glycomics—the analysis of sugars in biological systems, analogous to genomics and proteomics.

Ezose considers glycomics a promising platform for biomarker discovery because glycosylation influences the stability, activity, binding, and immunogenicity of many proteins, and because altered glycosylation patterns are seen in numerous diverse diseases.

Two glycosylation-based biomarkers are already clinically informative, noted Scott Siegel, Ph.D., vp of business development at Ezose. Measurement of CA19-9 (Lewis “a” antigen) sialylation is used to guide pancreatic cancer management, and detection of fucosylated alpha-fetoprotein (AFP-L3) aids in risk assessment and early recognition of hepatocellular carcinoma.

“So there are actually already glycan biomarkers out there in clinical use,” he summarized. “Think about how many more could be discovered if you actually had a tool to get in there and do the research.”

Ezose aims to provide such a tool via its GlycanMap® technology. Dr. Siegel indicated that GlycanMap has three key components: (1) bead-based chemoselective glycan enrichment known as Glycoblotting, (2) robotics for automated processing of samples in 96-well plate format, and (3) MALDI-TOF mass spectrometry coupled to bioinformatic analysis.

In a 2010 collaboration between Ezose and Genentech, GlycanMap’s results were comparable to those of a traditional low-throughput method in which glycans were fluorescently labeled and detected via HPLC. With this validation in hand, Ezose is now applying the GlycanMap platform to biomarker discovery projects such as a diabetes study recently launched in cooperation with Merck.

Protein Phosphorylation States

The clinical significance of certain proteins’ phosphorylation state is well-established; for instance, persistent tyrosine phosphorylation in the transcription factor STAT3 is observed in a variety of tumors. As for autoantibodies and glycomics, though, tools for large-scale investigation of phosphorylation biomarkers are just starting to emerge.

One firm working in this area is Evotec. The company’s PhosphoScout® platform permits identification and quantification of thousands of cellular phosphorylation events.

“In contrast to other technologies, PhosphoScout does not require phosphate-specific antibodies,” explained CSO Cord Dohrmann, Ph.D. “The detection of phosphate groups is based on masses in the mass spec; we can therefore comprehensively detect phosphorylation on tyrosine, serine, and threonine residues.”

For a given sample, “The phosphorylation status of up to 10,000 to 15,000 individual phosphorylation sites can be measured separately with high accuracy and reproducibility,” Dr. Dohrmann noted.

Evotec currently offers its phosphoproteome analysis as a service to clients, but it is also pursuing discovery research in which phosphoproteome signatures may be identified as possible biomarkers. Along these lines, a PhosphoScout analysis of the effects of the cancer drug sorafenib, a protein kinase inhibitor, suggests that this drug may inhibit the mTOR pathway in prostate cancer cells.

Hypothesis-Free Sequence Discovery

These days, looking for biomarkers among genetic mutations and mRNA transcripts is not especially novel in and of itself. But Genomic Expression is developing technology to make this analysis faster, cheaper, and less biased.

The heart of Genomic Expression’s technology is the creation of a single-stranded library of nucleic acid tags of defined length from double-stranded DNA or cDNA, followed by decoding of all possible sequences in the library on a GEx Digital Chip. Since all possible sequences are decoded, no assumptions need to be made about which ones will be informative—i.e., biomarker discovery can proceed in a hypothesis-free fashion—and even samples of unknown sequence (e.g., from microbiome samples) can be analyzed.

As new cancer drugs are developed, companies must come up with a biomarker strategy that will guide usage of the drugs, explained CEO Gitte Pedersen. “We can assist them in doing that without having to hypothesize around the target or the mechanisms—or the mechanisms for nonresponders,” she said.

The current version of the GEx Digital Chip can be read by existing array-scanning equipment to which many labs have access. It therefore generates data quickly and in a compact, accessible format, whereas methods such as RNA-Seq require both next-generation sequencing and extensive bioinformatics analysis to obtain interpretable data, Pedersen noted.

Biomarker Grants

Bayer Healthcare’s Grants4Targets program was launched in 2009 as a way of bringing new participants and ideas into the drug discovery pipeline, especially in areas like cardiology, oncology, and gynecology, reported Khusru Asadullah, M.D., who serves as head of global biomarkers at Bayer. About a year ago, the program was broadened to include biomarkers explicitly.

This reflected the program’s initial success as a catalyst for drug discovery research, Bayer’s interest in the field of biomarkers, and the fact that some early applications concerned molecules that could conceivably represent both drug targets and biomarkers. Applications written purely from a biomarker perspective are now receiving funding, and Dr. Asadullah anticipates seeing many more of these proposals in the future.

While applications from all sources are permitted, the program is designed especially “to provide something of a door-opener for academic groups that may not have experience collaborating with big pharma,” Dr. Asadullah said.

Unusual features of the Grants4Targets program include the simple application, fast review process (about eight weeks from submission to decision), lack of intellectual property restrictions, and partnering of grant recipients with specific Bayer scientists.

According to Dr. Asadullah, many grant recipients consider the access to industry resources at least as valuable as the grant money itself. “We provide to every project an internal Bayer scientist as a kind of caretaker,” he said, “and we are trying to support the projects by giving advice, and sometimes even by providing data and running some experiments and applying technologies that might not be available at the academic site.”

Overall, the Grants4Targets has received about 600 applications to date and made 77 awards. While it is too early to expect many success stories from the biomarker side, one drug target project has already progressed to lead optimization at Bayer, and another is at the lead-generation stage.