Skip to main content
Premium Trial:

Request an Annual Quote

Arul Jayaraman Discusses Cell Arrays for Gene Expression Studies

Premium

At A Glance

Name: Arul Jayaraman

Position: Assistant professor, chemical engineering, Texas A&M University

Background: Scientific staff, Shriners Burn Hospital — 2001-2003; Instructor in surgery (bioengineering), Harvard Medical School — 2000-2003; Research fellow, Harvard Med/Massachusetts General Hospital — 1998-2000; Postdoc, biomedical engineering, MGH — 2000; PhD, biochemical engineering, University of California, Irvine — 1998

The corresponding author of a recently published Analytical Chemistry paper describing a microfabricated living cell array, Arul Jayaraman joins the growing number of researchers using live cells in gene expression studies to see the bigger picture. Jayaraman, who just completed a cross-country move from Harvard Medical School to Texas A&M University, took a few moments to discuss his work with Inside Bioassays.

How did you first become interested in a live-cell approach to gene expression studies?

In some of our initial work, when I was a postdoc at Massachusetts General Hospital in the late 90’s, we started studying gene expression with an eye towards using it to develop therapeutics. The lab that I was in was associated with the Shriners Burn Hospital, which is a pediatric burn hospital geared towards post-burn care. The lab had a very long and deep history of studying post-burn physiology — things like development of articifical skin, wound healing, metabolism, and so on. What my group was working on was using therapeutics for controlling liver gene expression. The liver is very important in the body and how it responds to burn injury, so I wanted to see if you could control expression of certain genes in the liver. As we started doing this, we began to realize that it was pretty much like a black box approach — we had no idea what was happening inside the cell. Things were happening with different dynamics, and were coming on and going off at different times. So that got us into microarrays. But the classic problem came, which was we would look at microarrays and we were slicing and dicing it as close as we could, given our financial constraints. But things change — expression changes — even on the order of a few minutes, to a few hours on the other end. And depending on how you took your samples, you could actually miss out on certain events. So if you’re doing a study over 24 hours, and you just monitor at 1, 2, 4, 8, and 24, which people normally do, you would miss something happening between hours 4 and 8, or 8 and 24. So that was the biggest motivating factor for getting in to this live-cell approach. We know dynamics are important; therefore we need to be in a position to monitor the dynamics correctly and continuously. The question is: How do we do it without sitting in front of a microscope all the time? So that’s what brought us into the live cell approach.

Touching on that comparison between live-cell analysis and microarrays, what are the advantages and disadvantages of each when considering gene-expression studies?

I would say that the live-cell approach complements the microarray approach. The high-throughput [pattern?] and processing capabilities of the microarray approach just cannot be met by the other method. The logistics of there being cell reporters for every gene and every process is going to be tremendous. So in that sense, they are not really competing technologies, but they complement each other. The way I look at it, you can use microarrays to get a broader picture of what’s happening and a coarser picture of what’s happening initially. Let’s say during your microarray studies you come up with 25 genes that are relevant to the process you are studying, and you believe those genes are really important to your problem, now you can go over the next year or year-and-a-half and develop these supportive live-cell reporters, and then study the dynamics of those genes in the context of your process more accurately. And with the information from these live-cell experiments, one could go back and try to design further microarray experiments. Now that you know a little bit about a subset very vigorously, can that information be applied to drive experimental design for a bigger set of genes in a more rigorous and accurate manner?

With the technology that is described in the Analytical Chemistry paper [Anal Chem. 2004 Jul 15; 76(14): 4098-103], you’re hoping to monitor the expression of several genes at the same time, right?

Yes, that’s correct.

Tell me a little about the technology and how you would go about monitoring this expression?

The basic approach for studying the dynamics continuously is that you want it to be a non-invasive measurement, so you can use a cell population and continuously monitor it. Green fluorescent protein, which has been there for some time now, offers you that capability. The other motivation for this comes from the heterogeneity in gene expression. One common drawback that people talk about when they are doing different microarray experiments, with different samples, at different times, is the fact that there is sample-to-sample heterogeneity. And that can introduce changes and variations in the type of measurements you make. One of the advantages of using [the live-cell] approach is that you can actually make these measurements using the same cell population. If you want to push it to the limit, you can actually do it based on a single cell. So if you have a cell population and you stimulate it, you can go back and continuously keep monitoring how the expression of a particular gene changes in that cell over ‘X’ number of hours. That is the basis of the living cell array technology — the living cell part.

The other part of it comes from trying to complement, or to approach the microarray threshold, in a sense. You can [assay] the GFP reporter cell lines in traditional tissue culture plates. If you’re studying four genes or eight genes, you could do it in a 12-well tissue culture plate and monitor the expression events. What this doesn’t let you do is control the stimuli. That is the part where the microfabrication comes in. If you think about this, in the body, when you have these different mediators — cytokines, and so on — they come in waves, almost. Different things come at different times, and they interact with each other. So in order to understand the whole effect, you need to be able to recreate those interactions and those dynamics as accurately as possible. We still have a long ways to go for that. In a classical tissue culture format, you add something, then after some time you take it out, you wash it out, and so on. That is not accurate control. There is always going to be some residual effect. Do I wash it out four times, or eight times, or what?

But with the microrfabricated platform, you can actually [have] control of how you deliver something to a cell. Not only can you monitor the expression of several genes at the same time, you can monitor the effect of several mediators on several genes at the same time, in a single experiment. So you can think of it as a matrix — one side of it is eight genes that you put in, and the other side would be, as a simple [example], eight different concentrations of a compound. Or it could be the same compound appearing at eight different times. Or it could be two different compounds interacting with each other at eight different concentrations. So in something like drug screening, where you’re interested in studying the effect of two drugs or something like that, this approach would be very useful.

Just to clarify, the microfluidics portion is what is delivering different media to the cells, and the cells are in an array format?

Yes, that’s right.

Are the cells on a microscope slide or something similar?

What we do is make a pattern or a mask of a microfluidic network, and there is a [polydimethylsiloxane] block that has been bonded onto a glass slide, so the sensor is actually on the glass, and the network is in the PDMS. Since they’re seeded through the device, they’re actually on a glass slide, and the slide is on a stage in a microscope so that you can visualize it from underneath. Anything that you add to stimulate the cells with goes through the fluidic network.

Have any patents been sought on this?

Well, the microfabrication and seeding of cells — it’s been there for some time. The conventional way of seeding cells in microfabricated devices is using micropatterning. Actually, some of the authors on the paper — Mehmet Toner and my postdoc mentor Martin Yarmush — they published work sometime in 1996 or 1997 on patterning two different cell types in a microformat. People have also used cells in microdevices in a non-arrayed format, in cytometry-type assays, and so on. But very little work has been done on actually seeding cells in a microdevice and studying the biology of it. We have, of course, a plate for this, but there’s a non-disclosure thing. But we have some related patents, but not this exact one.

The reason I ask is that it seems like it would be a very useful tool in drug discovery. Do you see it having a future for such an application?

Well, it could. At this point, we are still in the process of using it for fundamental studies, and mechanistic studies. But it definitely has potential to be applied to drug discovery studies. And especially with the might of industry in terms of manpower and material power, it can easily be applied. The key beneficial feature is that it lets you monitor things in parallel inside the cell without having to do destructive assays. The fluidics brings in the combinatorial aspect. For example, if you’re studying toxicology of the liver and you want to know how these 10 or 100 compounds affect these four or 40 genes in the liver, you could do the whole thing, conceivably, in a single experiment. Of course, I’m projecting into the future, there, but that is the power of the approach.

Do you think this type of approach can ever reach the level of throughput that microarrays, or even biochemical assays, have?

Probably. I would like to say I want it to, but the limitations are as follows: The ground work required is tremendous. You basically need to generate your reporter cell lines before you do even a single, meaningful experiment. That, in my mind, is the single biggest limitation. Designing fluidic networks and tackling the engineering problems involved in delivering substances, mixing — those are also an issue, but are more feasible. The biggest impediment to something like this is the work that has to be done up front. It’s almost like doing cDNA microarrays. When people started printing their own cDNA microarrays, the labs would take a year-and-a-half to get all the cDNA clones, and [do] PCR, and purify them, before they even printed their first array. It’s the same kind of lag time that would be there. That is a non-technical, but genuine, limitation.

The other issue that I think is important is that we really need to get a better understanding of how cells behave in these environments. You are, after all, putting them in a very, very small environment, and you are making measurements of a very small number of cells. So you have to have a very good handle on how cells behave in these environments, and what the technical issues are with making measurements of small numbers versus the large numbers of cells you would use in a biochemical assay or flow cytometry assay.

Getting back to your work, what type of research are you going to be applying this to in your new lab at Texas A&M?

Obviously I have an interest in the process of inflammation, but my research interests also focus around using the dynamics of gene expression for understanding and studying different problems. One aspect is definitely the inflammation, which carries over from my postdoc training. But another application that we are interested in is using this approach as a biosensor device. Bacteria form structures called biofilms. These are very coordinated and organized structures. They communicate with each other extracellularly secreted molecules called quorum sensing molecules. Basically, the bacteria keeps spitting out sensory molecules, and when the level of sensory molecules reaches a particular threshold, it knows it has enough members around it and it turns on a certain set of genes. Biofilms have been associated with antibiotic resistance for a long time now. And one of the key issues in controlling an infection is early detection. We believe that if you can detect these quorum-sensing molecules early on, then you have a better chance of actually preventing the onset of an infection. So we are using the microfluidics approach in a sensor format. Instead of using mammalian cells, you’re actually growing bacteria in it, but with the bacteria engineered to pick up a certain sensory molecule. The advantages are that you could detect multiple bacteria simultaneously in a very short period of time.

Another area we are beginning to work with, but haven’t really done too much yet, is to combine this with RNA interference. Because what this gives us is the ability to temporally regulate gene expression at different times. In conventional RNAi, you either knock out a gene before or at a certain time, but with this, you could actually flow in a knockout gene for a particular period of time, and then turn it back on again, or do all sorts of expression profiles. That would be useful in understanding genetic networks and how they control response to stimuli.

The Scan

Driving Malaria-Carrying Mosquitoes Down

Researchers from the UK and Italy have tested a gene drive for mosquitoes to limit the spread of malaria, NPR reports.

Office Space to Lab Space

The New York Times writes that some empty office spaces are transforming into lab spaces.

Prion Pause to Investigate

Science reports that a moratorium on prion research has been imposed at French public research institutions.

Genome Research Papers on Gut Microbe Antibiotic Response, Single-Cell RNA-Seq Clues to Metabolism, More

In Genome Research this week: gut microbial response to antibiotic treatment, approach to gauge metabolic features from single-cell RNA sequencing, and more.