Not so long ago, protein microarrays were the domain of gearheads. Conversations about them revolved not around how they could be used but around how many proteins could be spotted down, or what kind of labels worked best.
Sure, those conversations are still being had. But the remarkable thing about protein arrays is how rapidly they have gained traction in applied science — they're virtually commonplace now in disease screening studies, turning up in studies of asthma, Lyme disease, plague, and various forms of cancer, to name just a handful. Emanuel Petricoin, one of the pioneers in developing and using protein arrays, says that recent findings in genomic research "have basically placed protein arrays at the center of understanding cancer."
If that turns out to be the case, protein microarrays are certainly here to stay. One of the reasons so many people have adopted them, says Mike Snyder at Yale, is that "the mass spectrometry approaches have not really panned out the way people thought." With their superior sensitivity, protein chips have secured a foothold in the field and are now seeing advances both from continued technological improvements and by application in research settings.
A better mousetrap
Just as DNA arrays continue to see technical improvements, their protein microarray cousins are still a work in progress — and will continue to be, largely thanks to the complexity of protein molecules.
Johan Roeraade at the Royal Institute of Technology in Sweden is part of a European Union project aimed at making protein microarrays more robust and affordable. When the project launched, he and his collaborators identified a number of technical hurdles that would have to be cleared for these arrays to become a truly mainstream tool: throughput, speed of fabrication, challenges in the range of protein concentration, reproducibility, and more.
His team has worked extensively on a new approach to array fabrication that uses far less biomaterial and appears to be amenable to multiplexing. The major advance, he says, is that "we can make arrays with ultra small spots — down to a few picoliters today." The trick is covering the microslide with a liquid that doesn't mix with aqueous solutions or interact with biomolecules. Under cover of this liquid, his team has found it possible to work with extremely small volumes of material that would otherwise "evaporate in a split second" if exposed to air, Roeraade says. "Now we also have time to deposit things and we can look at the results in a microscope." It's a non-contact approach, meaning that proteins are spotted on the slides by pressurized capillaries that never touch the chips. That's an advantage when working with living cells, for example, Roeraade says.
At this point, Roeraade notes, the work has been demonstrated only in proof-of-concept studies. Next up is running the approach in parallel — "We have already made a 16-capillary array," he says — and ramping up to several hundred capillaries so that just a few blasts could produce an entire microarray. "Then we'd start talking about mass production," he adds. "That's our final goal."
Over at the Lawrence Berkeley National Laboratory, Frank Chen's team is studying how quantum dots can be used to improve protein micro-arrays. Used in place of fluorescent probes, quantum dots are less prone to photobleaching, more amenable to multiplexing, and can be seen even in very small numbers. A recent effort successfully demonstrated that quantum dots work well in reverse phase protein arrays, and Chen says that other work in the lab has managed to measure enzyme activities on a peptide array "without any purification of the enzyme at high-throughput levels." One of his goals is to produce a tool that would allow researchers to watch phosphorylation and other proteomic events in real time without stopping the reaction, he says. That could significantly enhance sensitivity and decrease the amount of material needed in each experiment, he adds.
One area of development that has come a long way is the availability of good antibodies. "We are seeing an increase in the number of well-performing antibodies with high affinity in the public domain," says Petricoin, who helped develop the reverse phase protein microarray. But while that's been a "tremendous boon" to scientists, it's no shortcut for the time-consuming validation step that's still required. "You can't just go into it thinking that every antibody is a good one," he adds. "You have to do your own internal validation." Berkeley's Chen says this represents a significant cost burden for researchers, and that more inroads must be made. Petricoin notes that large-scale projects like the Human Protein Atlas are expected to be a major source of publicly available antibodies to give protein arrayers a kick-start.
Beyond that, Petricoin says, much effort is going into building more sensitive and accurate detection systems. "We're also seeing — and going to see — a real fantastic boon in increased sensitivity in the immuno-assays themselves which you can then multiplex on a microarray," he says. That allows researchers to use less sample — clearing the way for clinical utility — and opens the doors for a variety of new types of proteomic profiling studies.
Yale's Snyder says there is still plenty of work to be done, and that many technology developers are focusing on making functional arrays with membrane proteins and antibody arrays.
Applied arrays
Even a quick search on PubMed will tell you that the real action in protein arrays is in applying them to new studies, particularly in disease screening. These arrays have already proven to be "very accurate and reproducible," says Snyder, noting that "it's for good reason these things are moving into the more [applied] phase."
Roeraade points out the versatility of the arrays, which can be used for any kind of project that previously might have been done with an -ELISA. "What is interesting is that such systems, if they're really low cost … [they] could be used for screening of populations for certain diseases [and] as a complement to genomic screenings," he adds.
Snyder's lab was among the first to show a clinical use for these chips with a study of ovarian cancer. His team went on a hunt for differentially reactive proteins in sera from cancer patients and normal controls, taking the antibodies to those for follow-up study and emerging with "quite a few antigens that are differentially expressed in ovarian cancer," he says. While the project was encouraging, Snyder says it remains to be seen whether the results will be useful as biomarkers in patient care.
The possibility that these arrays could be used in a clinical setting stems from the significant work done to improve their sensitivity so they can work with very small numbers of cells, Petricoin says. "Most of the proteomic platforms that exist now don't have adequate sensitivity" for that kind of application, he adds.
Petricoin and Lance Liotta performed early validation of their reverse phase array in prostate cancer, looking at cell signaling in pre-malignant lesions compared to advanced tumors. Today, Petricoin says, chips have improved to the point that they can now be used to profile "200 to 300 phosphoproteins at once, quantitatively, from a few hundred cells," he says. That tool has been licensed to Petricoin and Liotta's startup, Theranostics Health, which has been using it for signaling studies performed in collaboration with pharma and biotech companies.
Petricoin says the importance of the work that will be done with protein microarrays was highlighted last year when initial papers were published by the Cancer Genome Atlas. While the work those scientists performed was "stunning," Petricoin says, "what was more stunning to me was the conclusions that they drew … [that] cancer is a protein pathway disease." In the follow-up research that he feels will be necessary, "the protein array is really the only way to do it at this time," he says.