Run, Spot, Run

---

By Patricia Nicholson

Spotted DNA microarrays — glass slides dotted with high-density arrays of DNA that provide high throughput analysis of up to 30,000 genes at a time — have revolutionized genetic research. New techniques, new technologies and new ideas are now making it possible to bring the speed and efficiency of microarrays to new applications.

James Woodgett, PhD, scientific director of the University Health Network (UHN) Microarray Centre (Toronto, ON), says that about 95 per cent of microarray applications remain nucleic acid-based, but new approaches are extending their scope. The UHN Microarray Centre — Canada’s largest manufacturer of microarrays for academic use — is capable of producing 1,000 arrays in a week, and services 260 labs worldwide. Some of the new technologies the centre is working with include CpG island arrays (promoter arrays), protein arrays, comparative genomic hybridization (CGH) arrays and cell-based arrays.

Colleen Nelson, PhD, head of the Prostate Centre at the Vancouver General Hospital’s Gene Array Facility (Vancouver, BC), is also experimenting with new types of arrays. Nelson’s lab is exploring protein-based arrays — printing both proteins and antibodies — and also cell-based arrays with the goal of performing high throughput functional characterization of genes. The lab has just started to print CGH-based arrays as well, and has also printed promoter arrays for functional binding assays as well as for chromatin immunoprecipitation and hybridization.

“I think it’s twofold,” Nelson says of the forces driving new microarray technologies. “It’s the high throughput nature as well as the miniaturization. So you can analyse more with less of the sample. So that gives it a great deal of efficiency. It’s also very rapid. It seems to be quite robust for many of these applications, and it allows you to multiplex a lot of questions into a single experiment or set of experiments.”

On the Spot

Spotted arrays are one of two main types of microarrays, the other being Affymetrix Inc.’s (Santa Clara, CA) GeneChips® — oligonucleotide arrays manufactured using a proprietary photolithographic process and available only through Affymetrix. Spotted arrays can be produced independently in any lab equipped with a robotic arrayer.

Daniel Tessier, PhD, project leader at the MicroArray Lab of the National Research Council of Canada’s (NRC) Biotechnology Research Institute (Montreal, QC), says his lab is interested in developing technologies that will be applicable to many types of spotted arrays.

“We have applied for Genome Canada money to actually develop different platforms, different surfaces, different chemistries, and are also going from micro to — I’ll use the word loosely — nanoarrays to try to maximize surface area,” Tessier says. “If you reduce the size of the spot, you’ll eventually be able to deposit more features on the same surface area.”

Tessier’s lab is working with NRC engineers to come up with new substrates for spotted arrays.

“We’re looking at developing different surfaces using nano-imprinting and embossing,” Tessier says. “Generating surfaces that would be different from what we currently know in terms of glass surfaces.”

These surfaces are intended to improve both price and design flexibility.

“Glass needs to be highly manufactured and there are costs associated with glass,” Tessier says. “By looking at other types of surfaces, more plastic-like . . . you could think that you would be able to manufacture much cheaper surfaces.”

The real advantage, though, could be in the texture of those surfaces. Instead of depositing spots onto a flat substrate, new slides could provide special topography that would make it possible to deposit spots of any material at different levels or heights.

“Using confocal microscopy you can perhaps think of avoiding increasing signal-to-noise by considerably reducing the background, because you’re working at a different plane of view,” Tessier says. He uses the example of depositing material on the tops of “mounts” on the slide surface. Anything below the level of those mounts would not be captured by the scanner.

In terms of the material that gets arrayed onto the glass substrate, Woodgett says CpG island arrays and CGH arrays both use nucleic acid spots. In the case of CpG island arrays — which derive their name from the high frequency of the CG sequence, connected by a phosphodiester bond, in these sections of DNA — it is pieces of DNA representing the promoters of genes that are spotted on the slide. CGH arrays use genomic DNA such as bacterial artificial chromosomes (BAC). The spots on protein-based arrays can vary, but are usually either BAC-expressed proteins or antibodies.

“With antibodies, you usually have to have two antibodies,” Woodgett adds. “You have a capture antibody and then you have a detector antibody, and that’s the one that is labelled.”

Of course, different types of spots are only the starting point. The important thing is the different types of data that they can produce.

“Gene expression arrays just tell you the amount of the relative concentration of the RNA for a particular gene — for thousands of genes,” Woodgett says. “DNA copy number arrays, which are CGH — comparative genomic hybridization — arrays, show you whether or not there are changes, either gain or loss, at the genetic level. So there may be gene amplification. That’s primarily of use in cancer biology.”

Cell Arrays

Cell-based arrays were pioneered by Dr. David Sabatini, PhD of the Whitehead Institute for Biomedical Research (Cambridge, MA). They work on the principle of reverse transfection: cells are added to DNA samples on a slide, rather than DNA added to cells in microwells or Petri dishes. In 2001, Sabatini and another former Whitehead Institute researcher founded Akceli Inc. (Medford, MA) to commercialize the technology.

“For cell arrays, what you spot are full-length cDNAs,” Woodgett says. “Then the cells that overlay the spot take up that DNA and express the gene of interest. So on a single microarray, you might have one or two thousand genes that you’re interested in, and then being expressed in these clusters of cells, and then you provide the various assays. For example, you can look to see whether or not a specific pathway is modulated by developing an assay to that pathway and looking to see if that’s modulated by expression of a particular gene. There are a variety of other tests you can do; it’s quite variable and very flexible.”

One of the factors slowing down the adoption of the technique is that it requires the generation of full-length cDNAs.

“There are a number of programs doing that — the (U.S. National Institutes of Health) Mammalian Gene Collection, for example, which has about 10,000 full-length genes,” Woodgett says, but adds that there are not yet any commercial sources of that type of technology.

“The other approach is to use RNAi or small interfering RNAs (siRNAs),” Woodgett says. “You spot those and the cells then take up the RNAi so you can look at the effect of selectively knocking down the expression of a particular type of gene. Again, doing it in a very high throughput manner.”

Cell-based arrays, he says, can be used to screen for molecules that would interfere with a particular pathway.

“You can screen, for example, the entire genome for all of the genes which influence a particular signalling pathway or change the rate of apoptosis or change the rate of cell proliferation,” Woodgett says.

Nelson’s lab is using cell arrays for functional characterization of genes that typically become activated during cancer progression.

“That becomes a more limited subset of genes that we’re interested in that we want to characterize more thoroughly, but we want to do it in a reasonably high throughput fashion,” Nelson says. “So that’s where we’re interested in expressing those genes in mammalian cells and then challenging them, for instance, with chemotherapeutics or other therapeutic agents to see how they modify therapeutic response.

“We’re also printing promoter-based arrays that are transfected into cells, and try to get a sense of what those promoters are responsive to,” Nelson adds. “So that would be promoter-reporter gene constructs.”

Protein Arrays

Applications for the protein and antibody arrays produced in her lab are quite specific because they have more limited subsets, Nelson says.

“Primarily the antibody-based arrays that we’re printing are largely targeted at signalling molecules,” she says. “Our applications have largely been aimed more toward characterization of which pathways might be on, turned on or modified in some way due to a particular stimulus.”

Woodgett describes protein and cell arrays as the next step up in terms of validation.

“There isn’t a direct correlation between the level of protein or the activity of a protein with the level of expression of a gene. You may have a gene that you identify (for which) there’s more of the RNA being produced, but its half-life, the half-life of the protein, may be also increased so that there’s no net increase in the amount of the protein,” Woodgett says. The advantage of using cell or protein arrays is, “It just provides the next step or next steps to identify and to hone down on particular genes which are implicated in various diseases. And it maintains the high throughput nature of arraying technology.”

Tessier says that although the spotting involves different chemistries and different buffers, protein arrays work in much the same way as DNA arrays.

“You would have your material spotted on a glass slide and you would do a hybridization or a protein-protein interaction study by overlaying your proteins on the chip with proteins that are fluorescently labelled, for example, and you would see where signals would be left behind. And essentially you would scan and look for residual fluorescence on different spots. And you could say well, my protein seems to have bound to proteins x, y and z — more or less in the same way as DNA arrays,” Tessier says. “This is what seemingly makes it easy and transferable.”

But, as Tessier points out, protein chemistry is very different from DNA chemistry.

“Proteins are not the same thing as DNA. They’re much more complex. There aren’t four bases as there are in DNA, there are 20 different amino acids. There’s structure, there’s stability, there’s proteolysis. Chemistries are completely different in terms of immobilization. So it’s not as simple as it seems,” Tessier says. “That’s why it’s been a bit slow in the development.”

Woodgett says he expects the Microarray Centre to make CpG island arrays available by the end of the summer, and cell arrays by the end of the year. Protein arrays, he says, are also on the horizon.

“We will be making those arrays available later — I’m not sure when, to be honest, because it really depends. There’s a lot more involved in it,” Woodgett explains. “The half-life of protein arrays is much shorter than cDNA arrays. You spot in a solution to try to prevent the proteins from denaturing. That can be tricky. These only last a week or so, so they’re a bit more difficult to work with.”

Despite the challenges, Tessier says he anticipates that protein detection will become a major application for microarrays.

“DNA and RNA are not the effector molecules in the cells, it’s proteins that are,” he says. “Obviously detecting proteins in either serum or any biological fluid is of great interest, and those methods of detection will obviously be very trendy or necessary in the next few years. So we are trying to foresee the need in our technology development to accommodate for protein chemistries in general.”

He also says he expects to see far more focus in protein arrays than is the case with DNA arrays.

“Contrary to gene discovery arrays like we know them for DNA — where you try to spot down every single gene on an array — trying to express and to spot down every protein of an organism, to me, is almost not feasible,” Tessier says. “I don’t think we’ll see protein discovery arrays. We’ll see very focused arrays on, say, kinases or cytokines or receptors.”

On the Horizon

Although the new array technologies have yet to be widely adopted, their advantages are compelling.

“I think a lot of people have caught on to various different ideas as to how to utilize the efficiency of an array format to create large amounts of data, but in quite a co-ordinated way that takes advantage of routine data analysis tools and software packages,” Nelson says. “I think that’s where you’re going to see a lot of people springboarding off the principles of microarrays and applying them to address their scientific questions.”

Nelson says the development of new applications for microarray technologies is probably still on the upward swing.

“You’re more limited by your imagination than what the actual technology can do. I think that you’re going to see more and more of this. I think people have been extremely pleased at how valid their information is in these different techniques, provided that you are prudent in terms of knowing what you’re spotting and what its limitations are, and how you’re putting it down. But I think people have been extremely amazed at how you can actually print down pieces of DNA and get proteins to bind to them,” she says. “That level of functionality is miles ahead of simple hybridization kinetics that drive the cDNA microarrays.”

Nelson adds that miniaturization, in combination with the efficiency of microarray technology, is driving down the costs of arrays.

“So a lot of people can perform these experiments that have thousands of data points for a few hundred dollars in a few days time,” she says. “And that’s just unparalleled compared to where we were before, trudging away with one gene over a year, or a protein over a lifetime.”