Center for Human Immune Monitoring, Stanford University Medical School
Name: David Hirschberg
Title: Director, Center for Human Immune Monitoring, Stanford University Medical School & Institute of Medicine.
Professional background: 2006-present, director, Center for Human Immune Monitoring, Stanford University Medical School & Institute of Medicine; 2004-2006, senior scientist and project leader, Agilent Technologies; 2000-2004, research scientist, Agilent Technologies.
Education: Doctorate and masters degree in neurobiology from the Weizmann Institute of Science, Rehovot, Israel.
A paper in this month’s issue of the Centers for Disease Control and Prevention’s Emerging Infectious Diseases journal describes the development of a panmicrobial microarray called the GreeneChip that the authors claim can be used to “facilitate rapid, unbiased, differential diagnosis of infectious diseases.”
Scientists at Columbia University’s Greene Infectious Diseases Laboratory led development of the 29,455 60-mer oligonucleotide probe array, modeled on Agilent’s platform, for vertebrate viruses, bacteria, fungi, and parasites. According to the authors, the chip can confirm the presence of certain viruses using samples from nasopharyngeal aspirates, blood, urine, and tissue.
In the paper, the researchers describe the first successful tests of the technology, which include detecting a previously undiagnosed fatal case of malaria that occurred during an outbreak of Marburg hemorrhagic fever in Angola in 2004-2005. In order to learn more about the array, BioArray News spoke with David Hirschberg, the director of Stanford University’s new Center for Human Immune Monitoring and a co-author on the paper, last month.
What started the process that resulted in this chip?
I used to be an employee at Agilent Technologies and I had a group there that was basically developing applications. So I am a biologist and I am trained as an immunologist. The reason that I came to Agilent was to leverage their technologies to do better biology. Our group was working on these cutting-edge things and we thought of using the [array technology] for other applications beyond gene expression and [comparative genomic hybridization].
So I went to visit Ian Lipkin, the director of Columbia’s Greene Lab, to talk to him about something completely different, and through Ian I met Gustavo Palacios — a Greene Lab scientist who eventually became the lead author on this paper. They were already working on building these kinds of chips and spotting the arrays themselves, and I offered that our technology would be useful in helping them build more consistent arrays.
We’d originally decided to make the arrays for looking at just cultures of viruses. I thought the technology was only sensitive enough to do that, but I felt we could give them an order of magnitude of greater sensitivity. So we began working on this sort of on the side using the Agilent platform and a couple of my supervisors let me do it. I think we built an array in about three months and within that time we had this beautifully working array that was more sensitive than any of us had thought. We started looking at clinical samples with it and could identify viruses with it. So we have revved it a couple times, but basically the data that was shown in the paper was done on these first couple of designs. The newer designs that have been developed since then are much more sensitive.
What features from the Agilent platform are evident in the GreeneChip?
There are several features to the array that make it different than standard protocols, but one use of the Agilent platform is the use of in situ synthesized oligos. Basically, Agilent can spot down As, Cs, Ts, and Gs, and it uses 60-mer oligos, rather than a smaller oligo — so it gives better sensitivity. And because we are very quick, we can just print out a few arrays, test everything, and then make changes to the probes and print out more. So we were able to rev it very fast.
There are two other Agilent features that we use that make the array very successful. One is that the database that we used to design the probes was very well thought out. It’s not a standard database. If you go to Genbank, for example, most of the viral sequences in there are biased towards what people study the most. Generally, things have to be completely sequenced before they are put in there. It is also not a curated database, so there are limitations with that. You get a lot of junk and you don’t get viral sequences that are more exotic and emerging.
So we chose to use the ICTV [International Committee on Taxonomy of Viruses, http://www.ncbi.nlm.nih.gov/ICTVdb/] database, which is curated, and the Lipkin lab at Columbia has made probe designs from that database and maintained it. As a result, it is a very highly curated database. Incompletely sequenced viruses are in there, and so we chose to use viral sequences that are expressed in, say, infection. So we are more likely finding the sequences that would be associated with an infection. And it is not only humans that are in there, it is all the intermediate hosts.
The third part of the array that made it function well was the labeling. We used random priming in a two-step process where we introduced a linker in the first PCR reaction and we hybridized that to the array, so it essentially does not contain any dye. So you are amplifying both the virus and the host, and so you hybridize it to the array and hopefully your pathogens stick to the array, and then you wash off the excess. Because of this, your signal to noise is very nice.
How did you decide which content to put onto the array?
Well that decision was made at Ian and Gustavo’s team at Columbia. But the way we arrived at the content was that we took advantage of the 60-mer oligo technology. The one nice thing about 60-mers is that you can have several oligo mismatches and still get a very good hybridization to the array. The mismatches actually correlate to how intense a signal you get. And so we design primers that would tolerate up to four mismatches.
The idea is that the way the probes are designed is by family, genus, and species. And so if you look at the ICTV database, it is like a family tree of viruses and how they are related. And so you can design a number of probes, like you can look at influenza at the family level, and then drill down for genus and species, and you can look at flaviviruses and all the different virus families there are.
The advantage of this is, as you are looking at an unknown virus that is, say, mutated from another influenza virus, you have enough probes that are actually pointing you in that direction, so the way the array works is that it guides you down a certain pathway. You’ll see certain family and genus probes but not species probes light up, or perhaps the species probes wouldn’t be as intense. That will guide you and let you know that the virus is probably from that family.
Many of the probes that Ian’s group eventually designed had been tested with PCR, so they had an idea of what they wanted to put on the chip already. So we started with viral cultures and then we started getting samples from the New York State Department of Public Health. Obviously, we haven’t been able to test everything on the array, but we’ve been able to test quite a few things.
There are a number of researchers developing viral identification chips. What is the end purpose for the GreeneChip and how does it fit in with arrays that are being developed with similar intention?
From my understanding, what the Greene Lab is trying to do is to build very broad detection systems. You know, we are putting all this effort into detecting influenza, but the Greene Lab feels that there is a huge need for broad detection systems that really identify large numbers; so that one sample can be tested and we can answer many different questions. So that was the idea behind this chip — it’s a broad surveillance tool.
Plus, you have the ability to rev the platform very quickly. So with the platform you can actually capture a virus and if it is unidentified, take it, sequence it, and develop new primers. You can then quickly set up and create a new array. But the overriding idea is broad detection.
You have some government authors on the paper. Would they be the ones that would contact you to use this tool in the event of an outbreak of some kind?
Yes, I think the US Centers for Disease Control and Prevention and the World Health Organization would do that. I think also that the lab at Columbia has been training representatives of other labs from around the world in how to use these tools. And I think what we would love to do now is to produce a lot of these arrays and to put them in the hands of these people, so we can disseminate this technology everywhere.
Right now at Stanford University I am also involved in starting up the immune monitoring center, and we are using assays to look at the immune system to measure people’s health. This tool fits in really nicely because the three main areas we are working with are autoimmune disease, transplantation, and infectious disease. I can certainly see us putting the chip to use in our lab.