Deputy head, Department of Immunotechnology
At A Glance
Name: Christer Wingren
Position: Deputy head, Department of Immunotechnology, Lund University, Sweden, since 2005. Researcher and teacher at Lund since 2003. Guest teacher, 2000 — 2003.
Background: Postdoc, Ian Wilson's laboratory in the Department of Molecular Biology, Scripps Research Institute, 1997-1999.
PhD in biochemistry, Lund University, Sweden, 1997.
Christer Wingren is scheduled to give a talk on using antibody-based microarrays to do high-throughput cancer proteomics at this month's OncoProteomics World Congress in South San Francisco, Calif. ProteoMonitor spoke with Wingren to find out more about his array technology and how it was developed.
How did you get into developing antibody arrays?
I did a PhD in biochemistry. My PhD work was focused on antibodies, but at that time I was more looking into relationships between surface properties and function.
Then after that, I did a postdoc in the United States. I was at the Scripps Research Institute in San Diego. I worked with Professor Ian Wilson. I actually spent two years doing X-ray crystallography, which I really enjoyed. And then after that, I went home to Sweden, and I joined the department that I'm at now. That is how I got back to working with antibodies, and then antibody microarrays.
What did you work on when you first got back to Sweden?
The first project I got involved with had a slightly different focus. I looked more into structural work on antibodies, trying to find a high-throughput manner to crystallize antibodies.
I worked on that for not that long, and after that we wanted to focus and get into the field of protein microarrays. And since I joined a department headed by Professor Carl Borrebaeck, who has a strong background in antibodies and immunology, it was kind of natural to combine our backgrounds and to get into the field of using not just using protein arrays in general, but using antibody microarrays.
So it was your department head who got you interested in antibody microarrays?
Yes. We started up this project in around 2000. Then, protein microarrays were just really starting to emerge. They were really in their infancy. So basically, we started from scratch to set up the technology and get the whole platform working.
Initially, it was a lot of small, focused antibody arrays just demonstrating proof of principle, sensitivity, specificity, and stuff like that.
How did you go about making the arrays? Did you spot them yourself?
The very first arrays were done manually, and they were pretty basic. But then fairly early on we realized that if we were to do this in a good manner, we would need an automated way of fabricating them. So we quite early on purchased an instrument called a Biochip Arrayer, which is a non-contact printer which arrays antibodies in the picoliter scale. That is what we still actually use to fabricate our antibody arrays.
The instrument, at that time, was from Packard BioScience. But Packard later on merged with PerkinElmer.
What was the content that you were putting on the early microarrays?
A lot of people are using monoclonal and polyclonal antibodies, but we took on a slightly different approach. That was based on work done by Professor Carl Borrebaeck, who had been involved in designing a recombinant antibody library array.
So we have recombinant antibodies that we put down on our chip. And this library is not based on intact antibodies. It's just the variable domains of the antibodies. They are called single-chain Fv fragments.
Basically, we have access to a library with around 210 different clones. So basically, if we were to take every antibody we had access to, we could make a chip of that size. From a practical standpoint, you can't really do that, but in theory you could.
Why did you decide to use the variable domains of the antibodies?
Well, the variable domains are the parts of the antibodies that bind to antigens. So those are the parts that carry the antigen specificity. They are called ScFv, or single chain Fv.
If you wonder why we went to that fragment, instead of using the full, monoclonal antibodies, the answer to that is that we want to create an antibody microarray platform that you can easily scale up. And if you rely on monoclonal and polyclonal antibodies, and you want to make an array with, let's say, 10,000 antibodies, that would mean that you have to produce 10,000 monoclonal antibodies, and that is a tremendous amount of work and resources, if you are able to do it. I think that's really challenging.
We, on the other hand, [we] can take the library and quite easily select any number you would like from the library. So the source of the content put down on the chip is not limiting.
And have you scaled up chip production?
Well, we are in the process of doing that. We're working with a couple hundred of our fragments right now. But before you can actually scale it up, there's a lot of basic, practical stuff that you have to optimize. You have to have a really good, high-performing platform with simple stuff optimized, like what substrate to print them on, and do you want to purify all your antibodies, or would you be able to make an array using crude expression supernatant? How should you label your samples to have a good readout? There are basic technological setups that haven't really been developed and optimized yet. So we've been focused a lot on doing that, and we haven't really gotten to the point of scaling up production yet, but that is something that we will address in the near future.
I think today we have basically designed, I would say, a state-of-the-art recombinant antibody microarray technology where we can screen clinical samples with picomolar to fentomolar sensitivity, meaning that you can screen serum samples and find really low-abundance analytes. That is what we have today, and we are using a few hundred antibodies while doing that.
Basically, the next phase is that we would like to scale it up. We will address that shortly.
What was the most challenging part of developing your recombinant antibody microarrays?
I think it's like setting up any assay or new technology. It requires that you not only have one part — so not only access to antibodies — but you have to take a really broad view to it. You need to be good at surface chemistry, you need to be good at working with antibodies. If you deposit them onto a chip, are you sure that they're still active? Can you design the antibodies to be more stable on the chip? And there are other things on how to handle a really complex samples.
All of our setup, and our competitors' setups, are all based on fluorescent readouts. And one major challenge is how to label a proteome and be sure that you actually label it in a manner that you can see all the analytes in it. And we have spent some time in optimizing the label protocol for the samples that we are running.
It's difficult to pinpoint one challenge. I would say that there [is] a combination of challenges. If you succeed in addressing them all adequately, you will have a really good platform, but if one of them fails, it will impair your whole platform.
Does fluorescent detection pose a limit on sensitivity?
Well, not really. We're down in the picomolar to fentomolar range, and that is actually without any signal amplification. I think we can even push it a little further if we could like to. So for the moment, sensitivity is not really a limitation for our platform at least.
Are you applying the technology you have developed to cancer studies?
Since the platform was so new when we started, the first couple of years was basically dedicated just to developing the platform. But now, during the last couple of years, as we have seen the platform develop into really good technology, we are more and more addressing various applications. And that will be, at least initially, focusing within cancer research.
During the last years, we have actually started the first major projects where we are trying to do protein expression profiling of various serum or tissue samples from various cancers, using our technology.
Are you planning on commercializing these arrays?
Our main focus, since we are academically based, is more of an academic perspective, but we always have an eye open for something to come up so that it might be commercialized. We have no direct plans, but in the future, it might come to be.
Are there any companies that make recombinant antibody arrays?
As far as I'm aware, the antibody arrays that you can buy today are based mainly on monoclonal antibodies.
Have you filed any patents for this technology?
No, not yet.
How many antibodies are there on the arrays?
The arrays that we're doing, you can run them as small, focused arrays with about 10 antibodies, and then you can also have a scale of maybe a few hundred. And then if you have maybe 10 replicates of each one, that means about 1,000 datapoints per array. That's basically the scale we are in right now.
What kind of cancers are you focusing on?
So far, we have looked into breast cancer, pancreatic cancer, and we have run one project in gastric adenoma carcinoma. But we're surely not limited to those three. We have [a] broad, open mind on that, and we have a number of projects starting up where we will look at three or four additional kinds of cancer.
Our knowledge is the technology and the antibody side. And then within these projects, we are collaborating with different clinicians who can provide the clinical samples and clinical know-how. We rely on quite a lot of collaborations when we want to study the clinical samples.
Do you think this technology is something that could be used in a clinic?
In the end, I surely hope that that could be the case. If the technology develops and performs the way we hope, our long-term aim is to find something that could be used either for biomarker discovery, or for diagnostics, and then you can see a clear and direct application within the clinics. That's our ultimate goal — to find something you can actually use in real life.
How expensive is it to produce one of these arrays?
It's difficult to put a number on it. I know for sure, that if you go out on the market, and you want to buy an antibody array, it's fairly expensive. But as always with developing technologies, arrays would be very expensive in the beginning like, for example, DNA microarrays were. But then the more it's used, and the more standardized the technology is, you can surely see that the price is going to go down.
I'm not really into business side, but if I were to guess about cost, it depends on what kind of circumstances you have when you design your array. If you have to first buy the antibodies from one company to generate your chip, it's going to be more expensive than if you have it all in house, and can produce it on your own, and then sell it. I think you need to have some of the basic stuff in house. Then you could, in the end, produce a product that's an affordable price.
And there are small, dedicated antibody arrays today that are basically used as point-of-care applications. So it's fully doable to get a product out there at an affordable price.
Aside from scaling up production, are there other projects you are working on for the future?
Working again closely with Professor Carl Borrebaeck, we have one other exciting track. We're also trying to scale down the arrays, going from the microarray format to the nanoarray format. That means instead of having a few thousand spots per square centimeter, you might have a few thousand spots per square millimeter.
That's still in a very early stage, but there are a lot of efforts going on, if you look in the world of academics, where they try to scale it down even further. I can foresee in the long-term future, nanoarrays might be the next step.
What's the advantage of having so many spots per square millimeter?
I guess the big advantage is that you will miniaturize everything even further, and that means that the amount of sample and reagents that will be consumed is even less. And you can surely foresee with clinical samples that if you have small biopsies and you want to analyze them, it really requires miniaturized assays.