Send to printer »

Feature Articles : Apr 1, 2011 (Vol. 31, No. 7)

Serum Profiling Making Mark on Predictive Medicine

Treasure Trove of Information Being Uncovered Is Advancing Field Rapidly
  • Vicki Glaser

Blood is a readily available and rich source of biomarkers that have the potential for use in diagnostics and companion diagnostics development, disease characterization and monitoring, and predictive drug response and toxicity analysis. Serum profiling using a variety of techniques to identify disease- or treatment-related proteins, antibodies, or oligonucleotides is leading to the identification of well-defined biosignatures and patterns of biomarker expression that are playing an important role in research, preclinical testing, patient stratification for clinical trials, therapeutic assessment, and increasingly in commercial diagnostic and prognostic applications and in drug selection.

At the American Association for Cancer Research annual meeting, Exiqon will present data from its diagnostic program aimed at early detection of colorectal cancer (CRC). The company applied its miRCURY LNA™ Universal RT microRNA PCR system to detect a miRNA signature in blood plasma indicative of CRC. The signature was derived from comparisons of profiles of several hundred miRNAs in about 200 patients with CRC and a comparably sized control group.

By the end of 2011, Exiqon expects to complete a validation study that will include about 5,000 patient samples. “It is hoped that this will eventually lead to a blood sample miRNA-based test that will identify patients at high risk of having early-stage (0-II) CRC with significantly greater precision than the currently used—but not very accurate—fecal occult blood test, which also suffers from poor patient adherence,” says Niels Montano Frandsen, Ph.D., product manager at Exiqon.

Patients identified as high-risk would be referred for follow-up testing such as colonoscopy.

Dr. Frandsen believes that despite the challenge of detecting miRNAs due to their small size, they “have all the hallmarks of a new class of powerful diagnostic and prognostic biomarkers of disease, drug efficacy, and toxicity.”

He notes their stability in clinical sample material such as FFPE tissue samples, blood serum and plasma, and urine. “We have found that miRNA levels in plasma are virtually unaffected by sample handling; samples can be left at room temperature for extended periods and thaw/frozen multiple times with no effect.”

Exiqon’s miRCURY system incorporates a universal cDNA synthesis reaction followed by PCR amplification with two miRNA-specific primers, which is possible through the use of LNA technology. The use of SYBR Green enables melt-curve analysis, which allows for verification of the amplification products.

The company currently offers two human miRNA panels and one mouse/rat panel, each containing 370 miRNAs on a 386-well plate, and will soon launch a second mouse/rat panel. In addition, toward mid-year, Exiqon plans to introduce its new Pick-a-Mix customized panels for which customers can specify a set of desired miRNAs and design a panel for a specific disease indication.

VeriStrat from Biodesix is an example of a serum-based test used in the clinic that can predict response to a second-line biotherapeutic treatment option. The test identifies patients who are likely or unlikely to benefit from treatment with erlotinib, an EGFR-TKI in patients with advanced non-small-cell lung cancer (NSCLC). Supporting data was generated following analysis of the BR.21 pivotal trial for erlotinib.

The VeriStrat test is available to clinicians through Biodesix’ CLIA-accredited laboratory. The test classifies serum samples from patients with NSCLC as either “good” (survival following erlotinib) or “poor” based on the results of mass spectrometry and analysis of an eight-peak protein signature. The result is available within 24–72 hours.

In addition to its in-house development and commercialization program, Biodesix collaborates on projects aimed at identifying patterns of serum protein expression that can be developed into predictive tests for use as companion diagnostics for targeted therapies, to predict and monitor treatment response, or to stratify patients for clinical drug testing. The company applies its ProTS® technology platform, based on MALDI-MS, which separates the serum proteome into 300–500 protein peaks.

The accompanying software normalizes and aligns the peaks and identifies signatures that can be used to distinguish between clinical subsets of patients. The algorithms are trained across large numbers of clinical samples and positive and negative controls; the software then applies that knowledge to define protein patterns that can differentiate good from poor responders to a drug against a background of genetic variability and variability in sample preparation and test conditions.

Paul Beresford, Ph.D., vp of business development and strategic marketing at Biodesix, anticipates broad clinical utility for VeriStrat across multiple types of cancer—including colorectal, head/neck, pancreatic, and breast—in which the test appears to detect a tumor/host interaction. The current test is designed to work with a dried serum spot that can be stored and transported at ambient temperature.

The next-generation test will require a blood sample obtained with a pinprick and spotted directly onto a collection card. Via active diffusion the serum component would separate from the red blood cells, yielding a dried serum spot that can then be punched out of the card at the laboratory and the serum extracted.

In a recent Phase II trial of first-line combination therapy with erlotinib and the oral multiple kinase inhibitor sorafenib in patients with advanced NSCLC, VeriStrat analysis was able to classify 48 of 49 serum samples with only one indeterminate result. Median overall survival among the patients identified as “VeriStrat good” (13.7 months) was significantly greater than for the patients classified as poor responders (5.6 months).

Targeting Antibodies

Sherry Dunbar, Ph.D., director of scientific marketing at Luminex, observes that in recent years serum-based immunological assays designed to measure an antibody response to an analyte have been moving away from classical capture-based ELISA sandwich assay formats and toward a more direct, higher-throughput, and more easily automated assay design.

In addition to its two earlier flow cytometry-based analyzers—the Luminex 100/200™ and the Flexmap 3D systems—Luminex introduced the Magpix analyzer and MagPlex® superparamagnetic microspheres, to which an antigen of interest can be directly coupled. Assay processing, including analyte and reagent addition, washing, and detection, are performed via magnetic separation.

Whereas the Luminex 200 can process up to 100 different reactions per well and the Flexmap 3D analyzer has a multiplex capacity of up to 500 analytes per well, the Magpix was designed as a lower-cost, more compact option capable of multiplexed analysis of 1–50 analytes.

Once serum biomarker studies have moved beyond the discovery stage and have identified a subset of analytes relevant to a particular disease process or drug response, “for most applications people are looking at 30 or fewer analytes,” says Dr. Dunbar.

She describes the development of protein-based assays to detect antibodies indicative of exposure to an infectious disease or as a measure of vaccine efficacy as a growing area of interest in the clinical application of serum biomarker assays. Dr. Dunbar notes that the Luminex 200 has already been cleared by the FDA, with expectations of future FDA submissions for the Flexmap 3D and the Magpix systems.

Serametrix partners with various cancer research institutes to identify novel tumor antigen panels that can then serve as the basis for developing assays targeting serum antibodies to tumor antigens such as the NY-ESO-1 cancer/testis antigen. Serametrix’ Seromic™ Profiling Assay is a multiplexed system that can simultaneously detect serum antibody levels against a panel of antigens associated with a particular tumor type, including lung, melanoma, breast, prostate, pancreatic, and colorectal cancer.

The company offers the Seromic antibody-profiling platform on a fee-for-service basis to evaluate serum antibodies in clinical trial samples and measure drug responses to cancer treatment strategies. Revenue from this business unit supports the company’s in-house biomarker assay-development program aimed at commercializing companion diagnostics.

“Vital to the successful development of novel anticancer drugs is a good understanding of the patient’s response to those drugs,” says Henry Hepburne-Scott, Ph.D., CEO of Serametrix. Targeting antibodies as tumor biomarkers offers certain advantages, he adds. Immunoglobulins overall have similar rates of decay, so even if the level of antibodies in a sample declines, the level of a particular antibody to other antibodies in the sample will be constant. “Although you may get low signals, you will not get confounding signals.”

Serametrix’ flagship assay targets the NY-ESO-1 antigen in melanoma and lung cancer. Different types of cancer tend to provoke a host immune response to varying degrees. Melanoma, for example, is a relatively immunogenic disease. Similarly, in lung cancer, circulating antibodies against tumor-associated antigens may reach detectable levels even before the tumor itself is large enough to detect on a CT scan. In addition to using serum antibody profiling to predict and monitor drug response to immunotherapeutic agents, it also holds promise for early detection of cancer.

From Discovery to the Clinic

ProFACT Proteomics has developed a set of tools for biomarker and drug target discovery that generates a profile of serum proteins based on their functional characteristics. The company’s SeraFILE™ platform comprises a multidimensional surface library and sequential separation protocols that reduce the complexity of the serum proteome and creates subproteome molecular profiles based on functional activity—typically enzymatic activity. SeraFILE protocols utilize mild binding and elution conditions to maintain protein function throughout the separation process. They incorporate a variety of surfaces for proteome subfractionation.

Matthew Kuruc, co-founder and president of ProFACT, describes the company’s Functional Proteome Prospecting™ approach as being complementary to conventional mass spectrometry-based proteome analysis strategies that emphasize protein identification.

“You can often detect functional features even before you can identify a protein,” says Kuruc. By monitoring the level of protein function relative to the total protein content in the various subproteomes, “you can monitor where a function goes in the different subfractions and select the fractions that have the most enrichment for that function.”

Each separation step provides further enrichment. Throughout this iterative process, each subproteome is scored for a desired activity using a variety of techniques including enzyme bioassay, post-translational analysis, and mass spectrometry.

ProFACTS’ partner in research consumables, Biotech Support Group develops products for proteomic sample preparation and enrichment, such as AlbuVoid™ for albumin depletion. The chemistry underlying AlbuVoid selectively inhibits binding of albumin, allowing the albumin-containing fraction to flow across the separation surface and into a collection vessel. The advantage of this strategy, rather than selectively binding the albumin, says Kuruc, is the ability simultaneously “to capture and concentrate the underlying proteins on the surface.”

Alphylase, a CRO that provides protein-analysis services to companies involved in research and clinical development of protein pharmaceuticals, received GLP accreditation by the Danish Medicines Agency earlier this year for use of its bioanalytical assays in human and animal serum samples for preclinical and clinical pharmacokinetic/toxicokinetic applications. The company’s Alpha-Quant™ assays are based on stable isotope dilution LC-MS/MS and multiple reaction monitoring to perform absolute multiplex quantification of specific proteins and protein variants in complex samples without the need for antibodies.

Ejvind Mørtz, Ph.D., COO at Alphalyse, reports “increasing interest from our clients in targeted protein biomarker quantification using LC-MS/MS methods. These methods can specifically quantify low nanogram protein amounts in a milligram protein serum sample within 15 percent accuracy.”