Protein Microarrays at the Cusp

Courtesy of Ciphergen Biosystems Wildlife biologist Marissa Irwin had no idea that cells in her body had gone haywire. In response to scrambled protein signaling pathways, certain cells turned cancerous and grew into tumors that took over the 37-year-old woman's ovaries. Though her doctors settled on a diagnosis of ovarian cancer, they could not determine what kind of neoplasia was causing the tumor growth--an important factor in determining the proper treatment. Daniel Chan is one of a growing number of researchers who are trying to change all that. As director of the Biomarker Discovery Center at Johns Hopkins School of Medicine, Chan works with a relatively new and still developmental tool in his quest to make such diagnoses more straightforward and to provide an early screening method. He works with protein-detecting microarrays. The latest developments with these microarrays make it possible for researchers such as Chan to survey tens to thousands of proteins in biological samples simultaneously. If the technology works as advertised, it could illuminate the protein makeup, or proteome, of cells and body fluids, as well as the interplay between these proteins. That, in turn, could provide scientists with a better understanding of disease pathology and development, cell signaling, and other complex cellular processes. Researchers have a number of kinks to work out, however, if protein chips are to move from the research lab into the clinic. MICROARRAYS EXPLAINED Protein arrays, like their precursor DNA microarrays, enable a massively parallel experimental approach. In a DNA microarray or biochip thousands of distinct DNA sequences, each representing a single gene, are printed on a solid substrate such as a glass microscope slide. Incubation with a labeled sample, such as mRNA from tissue, reveals which of the genes represented on the array are expressed in the sample, and their relative levels, thus producing a snapshot of the tissue's global gene expression pattern. Panoramic though such pictures are, they reveal little about the actual biology of the cell. "It's the proteins that really carry out the function of cells. It's not the DNA or the RNA," says Mark Schena, visiting scholar at Sunnyvale, Calif.-based ArrayIt (a division of TeleChem International), which develops protein microarray technology. "Protein microarrays are direct assays that measure the molecules of interest, rather than inferring function from levels of mRNA," says Schena. That's an advantage, because mRNA levels do not always correlate with protein abundance.1 And, DNA arrays offer no insight into posttranslational modifications, such as phosphorylation or glycosylation, which can affect a protein's biological function. Protein-detecting microarrays can fill in these gaps. "There's a lot of information that protein chips provide that DNA chips intrinsically can't," says Larry Cohen, president and CEO of Zyomyx, Hayward, Calif., which develops protein arrays. Protein array technology uses antibodies, cell or phage lysates, or recombinant proteins or peptides.2 Researchers incubate the microarray with a biological sample such as blood serum to determine which proteins in the sample adhere to which features on the chip. Two basic forms of arrays exist: antibody arrays, such as the Ab Microarray 500™ from BD Biosciences' Clontech division, Palo Alto, Calif.; and general protein arrays, such as the Yeast ProtoArray™ from Protometrix of Branford, Conn., containing nearly 5,000 polypeptides (the majority of the proteome) from Saccharomyces cerevisiae. The former chips are used to quantify a given set of proteins in a cell lysate or other biological sample; the latter help scientists monitor protein-protein and protein-small molecule (e.g., drug) interactions. A QUESTION OF ORIENTATION Array manufacturers can print molecules on either two- or three-dimensional surfaces. Two-dimensional substrates are typically coated with aldehydes that bind to the primary amines of the proteins' lysine residues, allowing the proteins to adhere to the chip in various orientations and exposing different faces for binding, says Gavin MacBeath, a microarray researcher and assistant professor of chemistry at Harvard University. Three-dimensional surfaces, in contrast, such as Irvine, Calif.-based Miragene's Z-grip and Boston-based PerkinElmer Life Sciences' HydroGel™, employ thin membranes that provide a matrix for protein attachment. The resulting solution-like environment, says Amy McCann, PerkinElmer product manager, allows the sample to interact with all surfaces of the arrayed proteins. "It's still being hotly debated whether 2-D or 3-D surfaces are better," Schena says. "At the end of the day they may have equivalent performance." Generally, less material loads on a 2-D surface than a 3-D surface, he says, but after exposing the array to a biological sample, unbound molecules tend to get trapped in the 3-D matrix, increasing background noise in some cases. "The all-important signal-to-noise ratio may be the same for 2-D and 3-D surfaces," Schena concludes. Some applications require capture proteins to be attached to the surface in a specific orientation, and several companies have adopted strategies to deal with this. Zyomyx and UK-based Procognia enforce a uniform orientation by modifying capture proteins with biotin and coupling them to streptavidin-coated chips. Beckman Coulter, Fullerton, Calif., meanwhile, uses DNA hybridization: Each antibody is chemically coupled to a DNA oligonucleotide whose complement is arrayed on the chip. Hybridization of the chip and antibody pool creates the working array, with antibodies uniformly positioned, explains Beckman Coulter's strategic business manager, Brendan Yee. Courtesy of Molecular Staging ROLLING, ROLLING, ROLLING: Molecular Staging's rolling circle amplification method is one way to amplify signals on protein arrays. FAREWELL, 2-D GEL? Protein arrays are certainly not the first technology to offer a glimpse of the proteome; two-dimensional gels have been offering such peeks for 30 years. But while protein arrays can detect and quantify specific proteins in a rapid, high-throughput, and automated process that requires minimal sample and reagents, 2-D gel electrophoresis is laborious, lengthy, difficult to reproduce, low-throughput, and reagent-intensive.3 "Two-D gels will be dinosaurs in five years," predicts Thomas Kodadek, professor of internal medicine and molecular biology at the University of Texas Southwestern in Dallas, whose research is developing an innovative approach to microarrays. "They're too irreproducible and too cumbersome." Indeed, protein microarrays may be the method of choice for proteomics researchers in the future. But, like any technology in its infancy, protein microarrays have a way to go before researchers can rely on them for consistent and accurate results. First, proteins present unique arraying challenges. They are less stable than nucleic acids, must maintain a correctly folded structure in order to function, and have varying viscosities that make spot-to-spot uniformity difficult.1 Also, many suppliers provide antibodies against denatured, rather than native, proteins, and when these reagents encounter natively folded proteins, they sometimes don't have the required level of specificity.4 "Typically less than 5% of commercial antibodies, which have not previously been validated for use in a sandwich-style ELISA [enzyme-linked immunosorbent assay], work in a microarray format," says MacBeath, who recently surveyed more than 100 commercial antibodies from various vendors. Antibodies directed against intracellular proteins, such as those involved in signal transduction, are especially difficult to obtain. Most antibodies being used, he says, recognize cytokines. In response to this problem, some companies, such as BD Biosciences' San Diego-based Pharmingen division, are working on higher-quality antibodies. Even high-quality antibodies have drawbacks, though. Each antibody has a characteristic binding affinity for its antigen, explains Lance Liotta, codirector of the Clinical Proteomics Initiative at the National Cancer Institute's Center for Cancer Research. Quantifying the amount of antigen becomes difficult, because such a measurement depends both on the concentration of the analyte and the binding affinity of the antibody.2 On a DNA array, each element has a roughly identical affinity for its target, and spot-to-spot comparisons provide a fairly accurate assessment of target concentration. But if the antibodies on a protein array have differing binding affinities, then correlating signal intensity with the amount of antigen present in the sample produces inaccurate results. "There needs to be a relationship of the amount of analyte and the intensity of the spot," Liotta explains. "Otherwise, you can't measure any analyte with precision." (Liotta provides a complete list of antibodies and commercial suppliers available at clinicalproteomics.steem.com.) One company that is circumventing this problem is Boulder, Colo.-based SomaLogic, which uses photoaptamers instead of antibodies as capture agents. Photoaptamers are single-stranded DNA molecules that fold into 3-D shapes with high affinity and specificity for specific target molecules. According to Helen Petach, senior director of scientific affairs, the company is currently assembling a library of photoaptamers with distinct specificities. SIGNAL AMPLIFICATION Protein microarrays face another problem, one that also plagues other proteomics technologies: dynamic range. The proteome is dynamic and fluid, and protein concentrations can vary by 10 orders of magnitude.2 Proteins present in much higher concentrations can drown out weak signals or cross-react with other array elements, reducing experimental accuracy. What is needed, therefore, is signal amplification to improve the chips' sensitivity. "With appropriate signal amplification technology, it isn't unreasonable that we can get to the point that we can detect single molecules," MacBeath says. Though researchers cannot amplify proteins in a PCR-like fashion, several companies are developing workarounds. Molecular Staging, New Haven, Conn., uses an approach called rolling-circle amplification (RCA).5 In RCA, antigen-detecting antibodies are coupled to an oligonucleotide. Addition of a circular DNA complementary attached to the oligo, nucleotides, and a specialized DNA polymerase, initiates a polymerization reaction that produces multiple copies of the plasmid, which remain tethered to the array. As the nascent sequence grows longer, more fluorescent "signal detectors" can bind to it, amplifying the signal, explains Stephen Kingsmore, Molecular Staging's chief operating officer. San Diego-based Genicon Sciences, meanwhile, is foregoing fluorescent dyes in favor of nanometer-sized particles of silver or gold. Shining a direct light on the particles causes resonance light scattering, producing a signal that is much more intense than fluorescence. Genicon product manager Gillian Wotherspoon says the technology improves the sensitivity of the array by 10-fold on average. Miragene likewise has turned from fluorescent dyes to enzymes, says Stewart Lebrun, director of research and development. Alkaline phosphatase, for instance, catalyzes a reaction that produces a colored precipitate, creating an amplified signal that is more intense than fluorescence, he says. Protein arrays sometimes have high background noise when fluorescently labeled proteins adhere to the spaces in between the printed spots. Procognia coats its arrays with a "protein resistance layer" so that protein binding takes place only on the printed spots. "We have the highest level of sensitivity," says Johanna Griffin, chief commercial officer of Procognia's US division, explaining that the company's arrays can detect femtogram quantities of proteins. Miragene also coats its arrays with a substance, called Z-block, to prevent nonspecific binding. While companies use creativity and innovation to surmount technical hurdles, the real test lies in the hands of researchers such as Chan, who remains optimistic despite the current shortcomings of protein microarray technology. "We're quite interested and believe that we will be successful in identifying cancer biomarkers using these approaches," says Chan, who uses the ProteinChip® system from Fremont, Calif.-based Ciphergen Biosystems in his work. But he realizes that using protein microarrays for cancer screening or routine diagnostic procedures is still years away. For cancer survivor Irwin, who lost her ovaries, uterus, and cervix before having children, the technology may be coming of age too late. Yet she remains hopeful that technologies such as protein microarrays will empower physicians to prevent and better treat cancer and other diseases, so other women can be spared her fate. Laura Lane (laura_lane@yahoo.com) is a freelance writer in San Francisco. References 1. M. Templin et al., "Protein microarray technology," Trends Biotechnol, 20:160-6, 2002. 2. L. Liotta et al., "Protein microarrays: Meeting analytical challenges for clinical applications," Cancer Cell, 3:1-9, April 2003. 3. P. Mitchell, "In the pursuit of industrial proteomics," Nat Biotechnol, 21:233-7, March 2003. 4. G. MacBeath, "Protein microarrays and proteomics," Nat Genet, 32:526-32, 2002. 5. B. Schweitzer et al., "Multiplexed protein profiling on microarrays by rolling-circle amplification," Nat Biotechnol, 20:359-65, 2002. function sendData() { document.frm.pathName.value = location.pathname; result = false if (document.frm.score[0].checked) result = true; if (document.frm.score[1].checked) result = true; if (document.frm.score[2].checked) result = true; if (document.frm.score[3].checked) result = true; if (document.frm.score[4].checked) result = true; if (!result) alert("Please select") return result; } .myradio{background-color : #94bee5} Please indicate on a 1 - 5 scale how strongly you would recommend this article to your colleagues? Not recommended 1 2 3 4 5 Highly recommended Please register your vote

Protein Microarrays at the Cusp

Interested in reading more?

Become a Member of