Swedish Institute for Infectious Disease Control
At a Glance
Name: Henrik Nordström
Title: Investigator, Swedish Institute for Infectious Disease Control
Professional Background: 2001 present, Investigator, Center for Microbiological Preparedness, Swedish Institute for Infectious Disease Control.
Education: 2001 MS, engineering biology, Umeå University, Sweden.
Two years ago Henrik Nordström, a scientist at the Swedish Institute for Infectious Disease Control (SMI), gave the keynote lecture at the winter meeting of the European Society for Clinical Virology in Copenhagen, Denmark, entitled "Do DNA microarrays have a place in diagnostic virology?"
In this month's issue of the Journal of Medical Virology, Nordström and his colleagues set about proving that microarrays do. Working together with scientists from the Karolinska Institute, the Royal Institute of Technology, the University of Umeå, and Sweden's Defense Research Agency, Nordström debuted a microarray customized to detect seven different flaviviruses.
The chip detects yellow fever, West Nile virus, Japanese encephalitis, and Dengue viruses 1-4, and Nordström contends that it can pinpoint what virus a patient is suffering from before developing antibodies to the disease, giving physicians a significant head start.
BioArray News first spoke with Nordström about the flavivirus chip at Cambridge Healthtech Institute's Microarray Data Analysis meeting in Washington, DC, last August, and caught up with him last week to talk about the new paper.
You and your co-authors come from a variety of different institutes. So how did your group coalesce to work on this particular microarray?
Since we are using RT-PCR normally and they only look for one virus, this looked like a good way to see if we could use one test for several different viruses. That was the basic idea. Five years ago we started thinking about that. I started the work at the Pathology Research Institute in Umeå and I have moved down to Stockholm and started the collaboration with the [Royal Institute of Technology] because they are very good in working with microarrays. With regards to the Karolinska most people here at the Swedish Institute for Infectious Disease Control have an affiliation with the Karolinska Institute. The Karolinska is [right next] to us and we are sharing microarray equipment with them.
The infectious disease institute together with the Microbiology and Tumor Biology Center at Karolsinka is the biggest institution in Sweden with the most researchers, I think. We have a joint core facility for sequencing and microarrays, proteomics, mass spectrometry, and this kind of stuff.
When you decided to put the array together, why did you specifically pick seven different flaviviruses?
They are among the most interesting flaviviruses and I think the ones that cause most disease in the world, especially here in Europe and Asia. There is the greatest chance of encountering these viruses [although] there are some other flaviviruses like St. Louis encephalitis. But that's mostly in the US, and there are also some Australian ones. But these are the flaviviruses that are most important for Europe, Asia, and Africa.
If you look at how many cases of disease per year the four dengue viruses cause 400 million cases of disease all over the world every year. I think yellow fever is still several hundred thousand cases per year. And then you have this Japanese encephalitis in Southeast Asia and that's already a bad disease.
All of these seven are mosquito-borne also, and for the people that are sick, the first days they don't have any antibodies developed. It takes about three or four days. And during those first days it's the nucleic acid you look for determining which virus is causing the infection, rather than the antibodies.
Would you say there is a strong need for better testing?
Yes, because there is an indirect need for a test that can verify which virus it is. You have to conclude that it is an infection with these viruses because you don't have any real treatment with these viruses because you can't use antibiotics. There are some antiviral drugs, but to conclude that [the patient] has a viral disease and not a bacterial disease is important. And you don't know what disease it is sometimes. People come and they have hemorrhagic fever and [they] really want to know [what] viruses [they have]. You can't directly see which virus infection it is just by looking at where the patient came from. Like if you have a person that comes from somewhere in Africa and come to Sweden and gets sick you don't know really know where they have been, then you have a lot of different diseases from which to choose. So this is a way to determine what it could be rapidly just by one test.
So it is a very useful first-time screening tool, I would say. Of course, it is difficult to use in poor countries because it is expensive, but it is not that much more expensive than running a RT-PCR plate. And there's a lot of RT-PCR use. But there are disadvantages with that because it's very easy that you get new strains that have mutations in the primer sites, and then you get false negatives. So you'll actually have the virus but you can't tell which virus it is. I tried to design our method to be more tolerant of mutations by looking at several different regions on the genome and I think it has worked, and compared to RT-PCR in some cases.
You told me that you actually had a patient that had come to Sweden and you used the chip to identify the virus …
There were some Indian sailors working on a ship that came to Gothenburg, a port on the west coast of Sweden, and two of them fell ill with internal bleeding in their stomach and intestines. The [people that were treating them] wanted to know as quickly as possible what it could be because they thought it could be some kind of hemorrhagic virus.
So they were looking at two different viruses, this Congo-Crimean hemorrhagic fever and these four Dengue viruses. But then it turns out that both were negative for both of these patients. Then they didn't know. So then I tried with a microarray virus a day after and found that this was Dengue 2. And then they took samples later and saw that it was a Dengue infection, they could now look at antibodies. The results were later confirmed with RT-PCR.
It is very useful to know early on what kind of infection you have. If it had been a Congo-Crimean infection, it would have been much more difficult to take care of the patients. You have to isolate them and treat them. Dengue infections are not as severe. The thing is that the multiplex RT-PCR for these four viruses was negative, but what I found with the microarray was positive.
Can you describe the actual array itself? What tools are you using in the experiment?
To begin with, for making the microarray, we have amplified five different regions of these seven viruses. So it's 500 nucleotide fragments approximately. We attach those on an aminosilane microarray slide Corning UltraGAPS. So there are non-PCR-modified nucleotide fragments with non-covalent binding to the microarray slide surface. I spotted that in collaboration with the Royal Institute of Technology and they have Genetix Q-Array spotters that work very well.
For the hybridization and so on, you can track the viral RNA with the Qiagen extraction kit, then we make cDNA with Invitrogen's superscript using these primers for these five regions. We [also] use an [GE Healthcare] ASP hybridization station to hybridize it because you want to have fixed conditions for these hybridizations.
We read the slides in an Axon scanner, which is very robust and easy to use.
Did you communicate with any of the companies whose tools you used?
Not really. The problem has been with optimizing the [GE Healthcare] hybridization system so we have had to do it ourselves.
Do you think it is possible to add more viruses to the chip in time?
I am doing it now. I am adding hemorrhagic fever viruses to the array. The problem is to amplify the sample because you get such little RNA in a sample and you have to get more of it. It's hard to do the amplification when you are looking for a number of viruses at the same time and you have little RNA.
Random amplification amplifying everything in the sample is not very efficient. Even in the easiest samples you find the sensitivity is one-tenth, or you have to have 10 to a hundred times more material than when you use a more specific amplification. So you won't find anything using random amplification if you get too little RNA in a sample.
The real problem with the flavivirus microarrays is to come up with a good amplification strategy when you have a lot of viruses. That's the main thing.
How do you plan to publicize the array and what are your goals for the invention?
Well I hope people will read the Journal of Medical Virology publication and I am traveling to a lot of conferences. I think people will get to know it that way.
It's probably difficult to start a company based on this because you have these patents, like Affymetrix patents, so its difficult to start selling Affymetrix microarray slides for example, but I could start selling plates for people containing the PCR products and then allow them to spot the arrays themselves. That is possible. But at the moment I think it would be hard to start selling microarray slides, due to these patents, like [those that claim methods for] how dense you can attach DNA fragments on a glass slide. That makes it difficult. But compared to other techniques, I think that spotting is the cheapest way to do this and you really have to start running samples and you have to cut down on the cost otherwise it is not useful in the clinical lab. I think that institutes like our infectious disease institute nowadays have microarray equipment, and it is possible that once they have the protocol to do these types of things they can adopt these methods. I think it's doable for all the institutes like ours that have access to this type of equipment.