By Meredith W. Salisbury
Still doing gene expression on a microarray? That’s so 2001.
For scientists on the bleeding edge, microarrays have become like any other ordinary lab tool — a technology platform that acts as constant challenge for what else to put on it. From microRNAs to proteins to chromatin immunoprecipitation, new uses for what was once a gene-expression-only technology have just begun to take off.
“Expression isn’t really the best thing to do” on an array — it just happened to be the first thing people did says Emile Nuwaysir, vice president of business development at NimbleGen. “It isn’t necessarily a more important type of genomic variation than the histone code or genotype or copy number variation,” he says. “What’s becoming obvious is that there’s all these other ways that the cell encodes information in the genome.”
And scientists want nothing more than to get at that information. Rich Fisler, a consultant who has extensively studied the microarray market, says he now sees demand from both customers and chip manufacturers to get beyond RNA expression profiling. Scientists are using the technology to look at “transcription factor analysis in a massively parallel approach,” he says as one example. Others are looking at DNA methylation — genome-wide studies that show which regions of DNA are silenced and which aren’t — or microRNAs, where high-density chips can help scientists track down candidate genes for therapeutic targets.
“Gene expression has of course been very useful, but that’s not the end of the story,” says Kevin Meldrum, head of genomics marketing at Agilent Technologies. “Now [scientists] can start to build a more pathway-oriented view” by compiling information from these different kinds of array experiments.
With the expansion of microarray uses comes an expansion of microarray users — and that may be a big step to getting this technology into a more clinical setting. Meldrum believes the more patient-oriented work that can be done with arrays will translate into diagnostic opportunities. “The biggest hurdle that we have is that these folks aren’t used to using microarrays,” he says. “The educational aspect of that is going to be one of the biggest challenges.”
On the pages that follow, GT offers an overview of the latest chip applications gaining traction in labs today.
Premise: Chromatin immunoprecipitation on a micro-array enables researchers to perform high-throughput studies of DNA-binding proteins across the genome. Scientists are using ChIP-on-chip technology to map transcription factors, for example. This technique is also known as genome-wide location analysis.
Bing Ren was a postdoc working on transcription regulatory networks in Rick Young’s Whitehead lab some six years ago when he began wondering how to scale up and improve the information he was getting from his experiments. “A key question that bothered me was lack of direct information about targeting of a transcription factor,” he says, noting that at the time, the way to find that target was a fairly rudimentary approach of comparing expression profiles in mutant and wild type cells. Having worked with chromatin immunoprecipitation in the past, Ren knew that there were more accurate ways to study transcription factors — it was just a matter of ramping up the throughput of ChIP research.
Microarrays seemed the obvious answer, and in Young’s lab, Ren had the resources to make the attempt. “The idea was pretty simple,” he says. “However, to get it to work took several months.” Early obstacles included increasing the nanogram amount of material yielded by a ChIP experiment to the micrograms needed for the DNA microarray; producing the highly specific antibodies required for chromatin immunoprecipitation; adding controls to deal with background noise; and developing software to analyze the new breed of data these chips would generate.
As Ren and the team overcame these obstacles, they faced another one in trying to build a chip to perform this work on a yeast genome. In typical gene expression studies, scientists can look just at the ORF regions, Ren says. “But if you are looking at transcription factor interaction sites you actually have to consider every base pair of the genome.” That meant squeezing a lot of additional information on what were at the time not the high-density chips we take for granted today. Now, ChIP-on-chip work is rapidly gaining popularity, and Ren credits a lot of that to vendors’ release of truly high-density arrays. “Now you can interrogate the genome in an efficient and high-throughput process,” he says. “That was a key in recent years that made ChIP-chip become a widely useable tool.”
‘Widely useable’ is something of an understatement. In the years since the members of Young’s lab bootstrapped their own ChIP chip, vendors have jumped right in and released their own chromatin immunoprecipitation array products. Emile Nuwaysir, vice president of business development at array service provider NimbleGen, says customer demand for these arrays “is growing tremendously.” He uses PubMed citations as one way to gauge growth in these markets. “In terms of the number of times it’s mentioned in the literature, it looks a lot like expression microarrays” when those first came out, he says. “It’s something that’s exploding. It’s exponential.”
As one of the pioneers of the field, Ren, now a faculty member at the University of California, San Diego, says he still gets e-mails daily from scientists asking how to perform ChIP-on-chip experiments. Among the more common uses of these arrays is mapping transcription factors. Others use it to identify regulatory sequences in the human genome, Ren says. “Those sequences are not so easy to spot by computational means because the sequences are highly divergent,” he adds. “So we developed this tool so that we can apply ChIP-chip to map transcription factor binding sites and use that binding event as a mark for regulatory function in the underlying sequence.” The technique, which allows researchers to mark promoters, enhancers, silencers, and insulators throughout the genome, “is now gaining a lot of acceptance in the field,” according to Ren.
Technological hurdles remain for ChIP chips, though. The most important one is the need for antibodies specific enough for chromatin immunoprecipitation experiments. These antibodies are responsible for pulling down proteins so scientists can track where the proteins bound on the DNA, so there’s no room for confusion between them. Ren says public-sector organizations such as NIH as well as private-sector companies “are devoting resources to producing ChIP-grade antibodies.” Data analysis, predictably, remains another challenge.
“We are only at the beginning,” Ren says. “I’m hoping that in the next couple of years that we’ll see rapid growth in [this] area.”
Premise: Using microRNAs identified throughout a genome, these arrays can help scientists home in on function as well as on candidate genes with varying disease states for therapeutic studies.
It wasn’t all that long ago that scientists were realizing there were lots of microRNAs scattered throughout genomes, so in some ways it’s remarkable that the studies of them have already become so high-throughput that they’ve been chip-ified.
Scott Hammond, now at the University of North Carolina at Chapel Hill, cut his teeth on microRNAs in Greg Hannon’s lab at Cold Spring Harbor Laboratory. In the early days, he says, there were “quite a few microRNA genes that had been identified, but no one had any idea what they were doing. The first step to identify candidate microRNA genes was to build a microarray.”
So he did. Since there were no vendors in the space at the time, Hammond and his team had to develop their own system. About three and a half years ago, he worked with the microarray core facility at UNC to custom-spot arrays with the microRNAs in question. There was also some technology development necessary to get the chips up and running, Hammond says. “We decided to develop a labeling method because the conventional labeling method would not work with microRNAs.”
“MicroRNAs are interesting pharmaceutical targets as well as markers of disease,” says Kevin Meldrum, who heads up the microarray marketing effort at Agilent Technologies. That’s just one reason the field is “gaining momentum,” he says. His company, like many chip vendors, has come out with a microRNA chip offering to help scientists study these genetic snippets. “We’re also working on an application basically utilizing microarrays to profile microRNA patterns in various samples,” he adds.
Indeed, cancer is a primary focus of microRNA arrays for Hammond’s lab, which continues to use homebrew chips because of the cost of commercial options. The rapid uptake of this technology comes as little surprise to Hammond. “It makes sense that people would want to look at microRNAs … once they were connected with disease states,” he says.
Proteins and Peptides
Premise: Why not use the massively parallel concept of microarrays for protein studies? Scientists are taking different approaches to tame finicky proteins for life on a chip, and the results are proteome-wide experiments.
Josh LaBaer joined the scientific field in the days when a researcher could spend an entire career focusing on a single protein. But that didn’t really fit with his goals of “understanding protein function and doing so in a high-throughput setting,” says LaBaer, now the director of the Harvard Institute of Proteomics. “We’re never going to get through the proteome at one career per protein.”
Microarrays have long been an obvious next step for protein studies, but the stability and resilience of DNA that has always made it so amenable to being spotted on a chip and stored away is sadly lacking in its protein counterpart. Spotting proteins on arrays while maintaining their structure and integrity has been the bane of more than one researcher. Michael Snyder made tremendous strides in this field when he printed most of the yeast proteome on a microarray. That work launched a company called Protometrix that was snapped up by Invitrogen, which continues to sell its protein array technology.
But protein spotting remains temperamental, as does storing the easily denatured products — and that’s why LaBaer and his team took a completely different approach to their protein-based array. Seeking an alternative to printing actual proteins, LaBaer hit on the idea of printing the gene that encodes for the protein and synthesizing the protein in situ only when it’s actually needed. “It took us a couple of years really to get that chemistry working,” he says. They finally succeeded, and the first version that was published in 2004 contained on the order of dozens of proteins. Since then, LaBaer has focused on increasing throughput, and the chips — known as NAPPAs, or nucleic acid programmable protein arrays — can now hold thousands of proteins. “We’ve preliminarily done the entire proteome of cholera on an array,” LaBaer says. Biological applications of these chips “can range from doing immunoassays and biomarker discovery to … even small molecule binding,” he adds. So far, immune response data from the NAPPA chips compared side by side with ELISA tests matches up nicely, according to LaBaer.
Protein interaction studies are a simple application for these arrays, and one that’s already up and running quite smoothly, LaBaer says. In addition, “we actually think that the immune response is doable now.” Using the cholera proteome chip, for instance, his team can take serum samples from people who have or are getting over the illness “and look at all the proteins in the organism and see which ones people are making antibodies to,” LaBaer says. It’s the first time people are really able to document that kind of immune response in a global way, he adds.
Being able to work with proteins in an array format is an “elegant solution,” according to LaBaer — so it’s no wonder that vendors are doing their homework for this field as well. Houston-based LC Sciences, which relies on an in situ synthesis microarray platform developed by sister company Atactic Technologies, has an offering in the works for protein and peptide arrays. While the product is just now getting close to market, Hebel says customers who have had early access to the technology “have been really floored at what they can do.” The chip has applications for kinase profiling and epitope mapping, among others, according to Hebel. In epitope mapping, for instance, “you want to really zero in on which parts of that peptide sequence are really important to that … binding event,” he says. The chips are essentially “sequence walks,” so users can “pick out of that sequence which amino acids are really important to be there for that antibody to bind,” he adds.
Premise: Microarrays provide a high-throughput alternative to sequencer-based detection of methylation, which has been linked to cancer and other diseases related to changes in DNA and cell regulation.
Rich Fisler has an enviable vantage point for the latest applications in microarrays. A former manager of the array business at PerkinElmer, Fisler is now a consultant for Beachhead Consulting, through which he gets to sit down with manufacturers and customers in the chip space to see what’s coming down the pike as well as where demand is greatest. From his survey of the field, “my own feeling is that methylation has the most interest,” he says.
Using microarrays to study methylation is not new — for example, the concept was the foundation of the technology platform for Orion Genomics, founded back in 1998 by Rob Martienssen, Dick McCombie, Rick Wilson, and John McPherson — but until recently it’s been an endeavor attempted in just a handful of labs. Over the years, papers have hit journals showing that DNA methylation state may have significant implications in cancer and other diseases where DNA silencing is an issue.
That, of course, helped spark the new boom in interest in methylation microarrays. Kevin Meldrum at Agilent says the company has just launched its methylation array offering. “I think that one’s actually going to be a big product for us,” he says. “Specifically in the cancer area, people are looking at methylation signatures” in oncogenes or tumor suppressors. “They think that some of these signatures might be useful as diagnostic tests,” he adds.
Those are the magic words to pique pharma’s interest — and that’s precisely where Fisler sees strong demand for these arrays. “Looking at methylation is a hot area,” he says. “I think that has a lot of growth potential.” He points out that even though methylation studies may be growing more common in basic research, it’s an application that is still fairly new to the drug discovery side. That’s a sizable market opportunity for these arrays, and one that could turn them into a mainstay for the field.
Emile Nuwaysir at NimbleGen also sees tremendous demand for global methylation arrays, which are in general just starting to become available to scientists. He says methylation, like other more structure-focused array applications, will help drive the growth of microarrays in the next few years. “Big government initiatives like ENCODE and the cancer genome project absolutely crystallized the need to measure this higher-level information structure,” he says.
Premise: Structural and other differences between genomes may be key to understanding evolution, disease, and a host of other biological questions. Microarrays can be used to compare entire genomes in a process known as comparative genomic hybridization, or CGH.
There’s simply no denying the popularity of array CGH. A PubMed search for the term “CGH” shows just a handful of results between 2001 and 2004, and then a veritable explosion: from 2005 to late this summer, some 450 papers were added to the scientific literature about this concept of using microarrays to compare genomes.
CGH is perhaps most often used to track copy number variation, a genetic trait that has been linked to various diseases — such as autism and Parkinson’s — in a number of studies. Indeed, cancer studies have been the beneficiary of much of this research, as the disease has long been studied in chromosome-scale comparisons. One drawback of the technique, according to experts in the field, is that array-based CGH cannot detect insertions, deletions, or rearrangements of stretches of DNA. (Other strategies, such as sequencer-based genome comparisons, can find those changes.)
Still, that hasn’t slowed the field down. Kevin Meldrum at Agilent says his company has “seen a tremendous uptake” in array CGH, calling it the “hottest new area” in microarrays. “We’re looking for gene copy number microdeletions or amplifications on the genome,” he says.
In fact, studies of copy number variation have become so trendy they’ve all but eclipsed SNPs, those genetic variations of yesteryear. The main meeting for SNP Consortium folks and other students of genetic variation has this year been renamed HGV2006 for “Human Genome Variation,” and its website assures attendees that copy number variation (and therefore array CGH) will be a main focus of the conference.