Dr. Harvey has noted a growing interest in reverse-phase arrays over the past 12–18 months. One of the advantages of spotting whole-cell lysates is the ability to capture all the proteins in a cell and interrogate the protein mixture with any available antibody. “The high protein-binding capacity of FAST slides makes them advantageous for protein-expression profiling,” said Dr. Harvey.
Antibody Breakthrough Needed
A panel discussion entitled “Fly It or Capture It? Deciding Between Mass Spectrometry or Protein Arrays” was led by Karin Rodland, Ph.D., science lead for NIH programs in the biological sciences division at the Pacific Northwest National Laboratory (PNNL;). In the microarray area, key issues include the specificity, selectivity, and availability of affinity reagents; strategies for working with poorly immunogenic antigens; and maximizing the tiling density of arrays.
Proteomics strategies using plasma for therapeutic-target discovery as well as diagnostic and prognostic applications face several challenges including the large dynamic range of protein expression, the need to detect and measure low-abundance proteins that are most likely to yield the most specific and sensitive biomarkers and drug targets, the difficulties inherent in working with complex protein mixtures, the demand for high-throughput processes and quantitation of protein expression, and the ability to distinguish between posttranslationally modified proteins.
PNNL takes a bottom-up approach to plasma proteomics based on global protein profiling using fourier transform ion cyclotron resonance mass spec to separate and detect tryptic peptides in plasma samples and identify potential biomarkers. The researchers then produce recombinant antibodies to proteins of interest and convert those into a protein microarray format for clinical validation.
Using mass spectrometry, “you avoid focusing on the usual suspects” and can take an unbiased approach to discovery, said Dr. Rodland. Despite continuous improvements in MS technology aimed at expanding its dynamic range and sensitivity (with a goal of protein detection at ng/mL concentration levels), “where cytokines and shedded proteases exist, the technological leap needed to make MS a robust, quantitative, clinical, point-of-care technique is 5–10 years off,” reported Dr. Rodland.
She identified the three main advantages of microarray technology: it is quantitative, it requires small sample sizes compared to traditional ELISAs, and it enables multiplexing of antibody/antigen pairs, although cross-reactivity is a concern with multiplexed arrays.
The main challenge at present is having access to the necessary “antibody pairs that are sufficiently specific, do not cross-react, and maintain their biochemical stability when immobilized on glass slides or beads,” said Dr. Rodland. Strategies in development to overcome this obstacle include novel surface chemistries, recombinant antibodies, and the use of autoantibodies—IgGs the body makes against a tumor or other foreign cell.
The cost and time needed to generate suitable affinity reagents as well as questions about their stability and shelf-life continue to limit the applicability of protein microarrays for clinical diagnostics. “We need a breakthrough in how to generate affinity reagents quickly,” emphasized Dr. Rodland. This would enable researchers to take differentially expressed proteins, candidate biomarkers, identified from profiling studies and quickly and cheaply produce antibodies to make large, multiplexed arrays of candidates to test against patient sera. The results could yield panels of multiple biomarkers that might offer the specificity needed for diagnostic and prognostic applications.
Autoantibody profiling using reverse-capture microarrays is the technique Brian Liu, Ph.D., director of translational research and assistant professor at Brigham and Women’s Hospital, and colleagues are using for biomarker discovery in sera, with prostate cancer as a model system. First, they capture native proteins from diseased (prostate cancer) tissues or cells using highly specific mAbs spotted on a glass slide. Next, they purify antibodies (IgGs) from patients with either prostate cancer or benign prostate hyperplasia (BPH), a noncancerous enlargement of the prostate, and label them with fluorescent dyes.
Using the previously captured antigens as bait, the labeled antibodies can then be washed over the slides. The labeled antibody that reacts with the captured antigens can be detected to yield autoantibody profiles that recognize proteins that are somehow altered in the disease state. They chose to work with autoantibodies because of their ability to identify disease-related proteins and their enhanced stability in serum.
Toward Clinical Biomarkers
Dr. Liu’s group published initial work demonstrating their ability to use a reverse-capture microarray platform to distinguish, with good accuracy, between sera from patients with prostate cancer or with BPH. They identified 28 unique antigen-autoantibody reactivities that were able to discriminate prostate cancer from BPH with P-values <0.01. After a one year or longer follow-up of postsurgical prostate cancer patients, only 1 of the 28 autoantigens used in the assay remained differentially targeted by autoantibodies.
Ongoing work in Dr. Liu’s lab involves expanding the number of patient samples tested to establish a statistical classification system and to use that to identify a panel of biomarkers that can be incorporated into a multiplexed assay for differential profiling. The assay they are developing requires only 50 µL of serum, comparable to a drop of blood.
“When you are dealing with such a large amount of low-abundance proteins, you realize that it is not always the case that individual autoantibodies of interest are present at the same levels in individual patients,” reported Oliver Tassinari, Ph.D., a research associate in Dr. Liu’s laboratory. By studying large numbers of patient samples against a large group of candidate markers, “you can filter the variables down to a specific group of proteins that consistently react differently and may potentially serve as a fingerprint for a certain disease or stage of disease,” he said.
“It is not very often that one target is considered a good biomarker,” added Dr. Liu. What appear to be the most significant antigens in one or a few individuals might not be consistent across patients. Using combinations of biomarkers, however, it may be possible to differentiate between cancer and benign disease. Dr. Liu pointed to prostate specific antigen as an example of a protein that, although prostate specific, only accurately distinguishes between cancer and benign disease about 50% of the time and has a high rate of false positives.
Dr. Liu envisions the transfer of his lab’s technology to the clinical laboratory taking the form of multiplexed protein arrays on glass slides, beads, or nanoparticles. Ideally, label-free detection systems could be developed for directly assaying autoantibodies against protein targets; these systems might include surface plasmon resonance or electroconductivity assays.
Schott is a technical glass company that entered the biochip market in 2002, producing borosilicate glass slides with or without coatings for a range of microarray applications. The company has seen a significant increase in interest in three distinct protein microarraying market sectors: academic researchers, producers of commercial kits, and companies developing diagnostic assays.
At “PepTalk,” Dietmar Knoll, Ph.D., senior chemist at Schott, discussed the factors to consider in selecting the best surface chemistry for a particular application. Some proteins such as antibodies are inherently quite stable and do not tend to denature. For these proteins, traditional 2-D coatings may be sufficient. For spotting more challenging proteins such as enzymes, however, a 3-D surface coating can help enhance protein stability. Both Schott’s thin- and thick-film 3-D coatings—for example, nitrocellulose—are optimized to maintain protein stability and activity, reported Dr. Knoll.
The main considerations in selecting a surface and binding chemistry are the type of proteins being arrayed, the spotting buffer to be used, and the method for blocking the surface area surrounding the protein spots to avoid nonspecific binding, according to Alistair Rees, microarray solutions product manager at Schott. For example, when arraying cell lysates, the surface chemistry selected must tolerate the high concentrations of detergent in the print buffer needed to protect against protein precipitation, noted Rüdiger Dietrich, Ph.D., director of R&D and technical support at Schott’s Nexterion product group.
Dr. Dietrich also identified several trends in protein microarray technology and applications: use of assays designed to analyze protein function (versus binding); label-free detection methods to eliminate the effects labeling might have on 3-D protein structure and thereby, function; strong interest in multiplexing of samples on a single slide; and miniaturization using smaller sample and reagent amounts and creating smaller spots to preserve resources, reduce costs, and enable more samples to be tested in smaller arrays.
“Fabrication challenges relate mainly to the diversity of the material being arrayed and the wide range of fluid properties,” reported John Austin, Ph.D., president of Aushon BioSystems.
One advantage of the deposition technology used in Aushon’s 2470 Arrayer, according to Dr. Austin, is its ability to create high-quality arrays with low coefficients of variation in a wide range of fluid viscosities. He noted a growing demand for arrays printed onto nitrocellulose membranes on glass slides, in particular for spotting cell lysates for reverse-phase microarray applications in biomarker discovery and protein-expression studies.
In December, Genomic Solutions, which offers the MicroGrid II arraying platform (originally developed by BioRobotics), was purchased by Digilab. In his presentation, Jim Galt, Ph.D., manager of technical support and applications at Digilab Genomic Solutions, described the various types of substrates onto which the Microgrid II can create an array of protein samples, including amino-silane coated glass, gold-plated glass, membrane-coated glass, microfluidic substrates, and membrane-coated plate wells. The MicroGrid can accommodate up to 24 source plates and create arrays on as many as 16 plates or 120 slides in one programmed run, he said.
As demand for miniaturization and high-throughput array applications increases, companies are developing instruments capable of printing microarrays on the well bottoms of microtiter plates. The MicroGrid can print more than 1,200 distinct spots in each well of a 96-well plate and at least 450 spots in the wells of a 384-well plate, Dr. Galt added.